measurement disturbance effects on rasch fit …/67531/metadc279376/... · logit residual index...
TRANSCRIPT
![Page 1: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/1.jpg)
Vrv
MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT
STATISTICS AND THE LOGIT RESIDUAL INDEX
DISSERTATION
Presented to the Graduate Council of the
University of North Texas in Partial
Fulfillment of the Requirements
For the Degree of
DOCTOR OF PHILOSOPHY
By
Robert E. Mount, A.A., B.S., M.A., C.R.C.
Denton, Texas
August, 1997
![Page 2: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/2.jpg)
Mount, Robert E., Measurement disturbance effects on Rasch fit statistics and the
Logit Residual Index. Doctor of Philosophy (Educational Research), August, 1997, 194
pp., 15 tables, 2 illustrations, references, 32 titles.
The effects of random guessing as a measurement disturbance on Rasch fit
statistics (unweighted total, weighted total, and unweighted ability between) and the
Logit Residual Index (LRI) were examined through simulated data sets of varying sample
sizes, test lengths, and distribution types. Three test lengths (25, 50, and 100), three
sample sizes (25, 50, and 100), two item difficulty distributions (normal and uniform),
and three levels of guessing (no guessing [0%], 25%, and 50%) were used in the
simulations, resulting in 54 experimental conditions. The mean logit person ability for
each experiment was +1. Each experimental condition was simulated once in an effort to
approximate what could happen on the single administration of a four option per item
multiple choice test to a group of relatively high ability persons. Previous research has
shown that varying item and person parameters have no effect on Rasch fit statistics.
Consequently, these parameters were used in the present study to establish realistic test
conditions, but were not interpreted as effect factors in determining the results of this
study.
Rasch fit statistics were found to be robust to varying levels of guessing and to
distribution types. The unweighted total fit statistic was more sensitive to fit problems
far away from the average ability of the group in which the fit problems occurred (fit
problems away from the item difficulties). The weighted total fit statistic was more
![Page 3: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/3.jpg)
sensitive to fit problems centered on the item difficulties. It was also found that, as the
probability of guessing the correct answer increased, low-ability persons tended
consistently to guess the correct answer inducing item bias (item familiarity) into the
tests. These items were detected by the unweighted between fit statistic. In conditions
involving minor fit problems and misfitting items, the LRI was able to identify group
membership of the persons in which the fit problem occurred. Therefore, it is necessary
to use the unweighted total, weighted total, and between fit statistics in combination with
the LRI to diagnose fit problems for a more accurate assessment of individual differences.
![Page 4: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/4.jpg)
Vrv
MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT
STATISTICS AND THE LOGIT RESIDUAL INDEX
DISSERTATION
Presented to the Graduate Council of the
University of North Texas in Partial
Fulfillment of the Requirements
For the Degree of
DOCTOR OF PHILOSOPHY
By
Robert E. Mount, A.A., B.S., M.A., C.R.C.
Denton, Texas
August, 1997
![Page 5: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/5.jpg)
TABLE OF CONTENTS
Page
LIST OF TABLES v
LIST OF ILLUSTRATIONS vii
Chapter
1. INTRODUCTION 1
Overview Properties of Estimators Rasch Estimation Methods Rationale for the Study Research Question Definition of Terms Delimitations
2. REVIEW OF THE LITERATURE 15
Historical Perspective Measurement Disturbances Rasch Fit Statistics Logit Residual Index Purpose of Study
3. METHODS AND PROCEDURES 33
Data Set Construction Simulated Data Sets Rasch Analysis Statistical Analysis
4. RESULTS 39
Effect of Guessing on Rasch Fit Statistics Simulation Design Effects
![Page 6: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/6.jpg)
Table of Content (continued)
Chapter Page
Detection of Guessing by Rasch Fit Statistics Guessing and the Logit Residual Index
5. FINDINGS AND CONCLUSIONS 59
Effect of Guessing on Rasch Fit Statistics Detection of Guessing by Rasch Fit Statistics Guessing and the Logit Residual Index Summary Conclusions Further Study Recommendations
APPENDIX
A IP ARM Control File Parameters 72
B A Summary of Item Fit Information by Experiment 76
C A Summary of Misfitting Item Statistics by Experiment 83
REFERENCES 192
![Page 7: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/7.jpg)
L I S T O F T A B L E S
T a b l e P a g e
1. D e f i n i t i o n o f E x p e r i m e n t s 3 5
2 . M e a n S u m m a r y o f I t e m F i t I n f o r m a t i o n f o r E x p e r i m e n t s 1 - 9
( N o r m a l l y D i s t r i b u t e d I t e m D i f f i c u l t y D i s t r i b u t i o n s a n d N o
G u e s s i n g ) 4 0
3 . M e a n S u m m a r y o f I t e m F i t I n f o r m a t i o n f o r E x p e r i m e n t s 1 0 - 1 8
( N o r m a l l y D i s t r i b u t e d I t e m D i f f i c u l t y D i s t r i b u t i o n s a n d a 2 5 %
C h a n c e o f G u e s s i n g C o r r e c t l y ) 4 1
4 . S u m m a r y o f M e a n I t e m F i t I n f o r m a t i o n f o r E x p e r i m e n t s 1 9 - 2 7
( N o r m a l l y D i s t r i b u t e d I t e m D i f f i c u l t y D i s t r i b u t i o n s a n d a 5 0 %
C h a n c e o f G u e s s i n g C o r r e c t l y ) 4 2
5 . M e a n D i f f e r e n c e s f o r E x p e r i m e n t s W i t h N o r m a l l y D i s t r i b u t e d
I t e m D i f f i c u l t i e s W i t h N o G u e s s i n g , 2 5 % , a n d a 5 0 %
C h a n c e o f G u e s s i n g C o r r e c t l y 4 3
6 . S u m m a r y o f M e a n I t e m F i t I n f o r m a t i o n f o r E x p e r i m e n t s 2 8 - 3 6
( U n i f o r m l y D i s t r i b u t e d I t e m D i f f i c u l t y D i s t r i b u t i o n s a n d N o
G u e s s i n g ) 4 4
7 . S u m m a r y o f M e a n I t e m F i t I n f o r m a t i o n f o r E x p e r i m e n t s 3 7 - 4 5
( U n i f o r m l y D i s t r i b u t e d I t e m D i f f i c u l t y D i s t r i b u t i o n s a n d a 2 5 %
C h a n c e o f G u e s s i n g C o r r e c t l y ) 4 5
8. S u m m a r y o f M e a n I t e m F i t I n f o r m a t i o n f o r E x p e r i m e n t s 4 6 - 5 4
( U n i f o r m l y D i s t r i b u t e d I t e m D i f f i c u l t y D i s t r i b u t i o n s a n d a 5 0 %
C h a n c e o f G u e s s i n g C o r r e c t l y ) 4 6
9. M e a n D i f f e r e n c e s f o r E x p e r i m e n t s W i t h U n i f o r m l y D i s t r i b u t e d
I t e m D i f f i c u l t i e s W i t h N o G u e s s i n g , 2 5 % , a n d a 5 0 %
C h a n c e o f G u e s s i n g C o r r e c t l y 4 7
![Page 8: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/8.jpg)
Table
List of Tables (continued)
Page
10. Mean Differences for Experiments With Normal and Uniformly Distributed Item Difficulties and Varying Levels of Guessing (No Guessing, 25%, and 50%) 48
11. A Comparison of the Frequency of Misfitting Items Detected by Rasch Fit Statistics in a Normal Distribution of Item Difficulties at Varying Test Lengths and Levels of Guessing Using %2 50
12. A Comparison of the Frequency of Misfitting Items Detected by Rasch Fit Statistics in a Uniform Distribution of Item Difficulties at Varying Test Lengths and Levels of Guessing Using x2 52
13. A Comparison of the Frequency of Misfitting Items Detected by Rasch Fit Statistics in Normal and Uniformly Distributed Item Difficulties Using % 54
14. Number and Percent of Misfitting Items Detected by Rasch Fit Statistics Across Experiments Involving Normal and Uniformly Distributed Item Difficulties 55
15. Number of Misfitting Items and Number and Percent of LRI Values by Experimental Conditions 56
![Page 9: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/9.jpg)
LIST OF ILLUSTRATIONS
Figure Page
1. Rasch Model Notations 7
2. Table Format for the Display of Experimental Data 40
![Page 10: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/10.jpg)
CHAPTER 1
INTRODUCTION
Overview
Stevens (1946) defined measurement as the assignment of numerals to objects or
events according to rules. However, the measurement of individual differences is not as
straightforward as the definition implies; it is fraught with several measurement
problems: (a) No one approach is universally acceptable; (b) measurements are based on
limited samples of behaviors; (c) measurements are subject to error; and (d) the units of
measurement are not well defined (Crocker & Algina, 1986). Of the problems associated
with the measurement of individual differences, the accuracy of measurement is probably
the most important. A measure can only be as accurate as the ruler (scale) used to obtain
the measurement.
Traditionally, the measurement of individual differences was based on the
classical true score theory and its associated statistics and scales of measurement
(nominal, ordinal, interval, and ratio). The irony is that measurements obtained using this
approach were found to be sample specific and that generalizations beyond the reference
sample should be made with caution. In addition, the scales of measurement were found
to have arbitrary zero points and unequal measurement intervals. Therefore,
measurements obtained using these scales may be as arbitrary and unequal as the ruler
![Page 11: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/11.jpg)
used to make them. Given these characteristics, estimates (statistics) obtained in
identifying individual differences tended to be biased, inconsistent, insufficient, and
inefficient when used with nonnormal distributions.
What was needed was a true linear scale that had an absolute zero point and equal
interval units of measurement. In 1960, one such scale with associated statistics, was
developed by Rasch (Rasch, 1980). Known as Rasch Analysis, this approach allowed for
the independent estimation of person ability and item difficult parameters. In addition,
the statistics used in this approach were found to be consistent, efficient, sufficient, and
unbiased.
In the assessment of item function in the Rasch model, one observes the residual
trends among ability groups. The Logit Residual Index (LRI), a statistic introduced by
Smith (1991b), is a measure of how far an item deviates from the common slope that is
fitted for all items (model curve) and the residual trend among ability groups. Thus,
items with LRI values greater than zero will have an item characteristic curve (ICC) that
is steeper than the modeled, curve and items with values less than zero will have an ICC
that is flatter than the modeled curve. The residual trend is an indication of how well, or
less well, different ability groups performed on an individual item. The purpose of this
investigation is to test the effects of varying test parameters and levels of measurement
disturbance on Rasch fit statistics and the Logit Residual Index.
![Page 12: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/12.jpg)
Properties of Estimators
Rarely, if ever, are characteristics about a complete population known; therefore,
researchers make inferences about a population based on a representative random sample
taken from the population. A population refers to all members in the entire group having
some common characteristic. For example, a population may include all members in a
classroom, school, city, community, county, state, nation, or the world. As group
membership increases, it becomes increasingly more difficult to obtain measures on all
characteristics due to restrictions in time and costs, or the population size increases too
rapidly. A population may be finite, a known or countable number of members, or
infinite, a population that is so large that group membership is not known.
Characteristics about a population are called parameters and characteristics about a
sample are called statistics. The most commonly used estimates about populations are
measures of central tendency (mean, median, and mode). They identify the most typical
measures in a normal distribution.
The arithmetic mean is the most commonly used measure of central tendency. It
is a simple arithmetic average determined by summing all scores in a distribution of
scores and dividing by the total number of scores. The mean is an appropriate measure of
central tendency when the score distribution is normal and the level of measurement is on
the interval or ratio scale. The median is the midpoint in a distribution of scores when
arranged in order of magnitude. Stated differently, it is the point on the score scale below
which 50 % of the scores fall. It is also equivalent to the 50th percentile. The median is
![Page 13: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/13.jpg)
an appropriate measure of central tendency when the level of measurement is on the
ordinal scale and the score distributions are other than normal. The mode is the most
frequently occurring score in a distribution of scores. However, it is not a dependable
measure of central location. Depending on the shape of the distribution, it is possible to
have two (bimodal) or more modes (multimodal). The mode is an appropriate measure of
central tendency when the level of measurement is on the nominal scale.
In a normally distributed population of scores, the mean, median, and mode will
coincide or be the same value. However, in nonnormal distributions, these values differ,
and, therefore, certain estimators have more desirable properties than others. The
desirable properties of estimators are consistency, efficiency, sufficiency, and
unbiasness. These serve as criteria for determining preferences for one method of
estimation over another.
An estimator is considered to be unbiased when the mean of a sampling
distribution of means approaches that of the population parameter as the number of
samples of a given size increases. That is, a statistic is unbiased when it shows no
systematic tendency to be either greater than or less than the population parameter. For
example, it can be shown that the variance
S2 =E(X-X) 2 /n
is a biased estimate of the population variance (cr2) (Ferguson, 1981).
![Page 14: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/14.jpg)
Consistency implies that an estimator more closely approximates a population
parameter as sample size increases. The arithmetic mean is a prime example of a
consistent estimator. It more closely approximates the population parameter as the
sample size increases.
Efficiency is implied by sampling variance. It refers to the variability of estimates
from sample to sample, or the degree of sampling error associated with the estimator.
That is, if the sampling error is less than the sampling error associated with any other
method of estimation, the estimate is considered to be efficient (Ferguson, 1981). More
explicitly, the estimator that has the smallest standard error is more efficient (Walker &
Lev, 1953).
An estimator is sufficient for estimating a population parameter if it exhausts all
the information about the population parameter from sample data. For example, the mean
is a sufficient estimator of the population mean (JJ,), because, once the sample mean is
known, any other statistic computed from the sample data (such as the median or mode)
would provide no further information about the population mean (Neter, Wasserman, &
Whitmore, 1978). Given a normal population, the mean provides an estimate of (j, that
satisfies all the desirable properties of a good estimator (consistent, efficient, sufficient,
and unbiased) (Walker & Lev, 1953).
Rasch Estimation Methods
Measurement implies the determination of the quantity, quality, or some other
characteristic of an object or attribute. It answers the questions concerning how many,
![Page 15: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/15.jpg)
how often, and how much of a particular object or attribute exist. The process involves
the assignment of units or numbers in a logical fashion along a dimension or scale. When
an object or attribute is measured, it is assigned a specific position along a dimension or
numerical scale. Traditionally, we have used four scales or levels of measurement: (a)
nominal scale-when numbers, names, or words are used to identify or label individuals
or objects; (b) ordinal scale—when numbers or words reflect the order of things, (c)
interval scale-has equal units of measurement and an arbitrary zero point; and (d) ratio
scale—has equal units of measurement and a true or absolute zero point. Each scale of
measurement has its own rules and makes different assumptions about the measurement
process. These scales are prevalent in the traditional classical true score approach to the
measurement of human traits.
The Rasch measurement model, unlike the classical true score model, attempts to
explain the effect of a person's ability on item performance. The Rasch model frees the
estimation of a person's ability from the item difficulty, and the estimation of the item
difficulty is freed from the person's ability. In short, the more able the person, the better
the chances for success on an item, and the easier the item, the more likely a person is to
solve it (Wright & Stone, 1979). It has been shown that no other mathematical model
allows the estimation of person ability measures (Pv) and item difficulty calibrations^)
independent of each other (Anderson, 1973; Barndorff-Nielsen, 1978; Rasch, 1961;
Wright & Stone, 1979). The logistic function (probability of a correct response) in the
Rasch model,
![Page 16: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/16.jpg)
p { * v i - 11 Pv.8,! = exp (|3V - 8,)/[l + exp(fiv - 6,)],
P b°th linearity of scale and generality of measure (Wright & Stone, 1979).
Rasch called this particular characteristic "specific objectivity." The symbols and
associated definitions used in the Rasch model are presented in Figure 1.
1 r ~ ability level = = = = = = = = = = = = = = = = = = = = 8 difficulty level "
rv test score of person v " L the number of items in the test H the average difficulty level of the test M t
mean person ability © the variance in difficulties of the test items
i an item on the test Pi sample p-value of an item i V person
Pv ability level of person r score on the Test
individual item difficulty 8i item difficulty level
Xvi person response
Figure 1. Rasch model notations.
Several Rasch measurement models have been identified: (a) rating scale, (b)
poisson, (c) binominal, (d) dichotomous, (e) partial credit, and (f) many faceted.
However, for a measurement model to wo*, there must be some method of estimating its
parameters. Rasch identified six estimation methods: (a) the LOG method, (b) the PAIR
method, (c) the FCON method, (d) the UCON method, (e) the PROX method, and (f) the
UFORM method. Of the six estimation methods, PROX is the only estimation procedure
m which item and person parameters can be easily calculated by hand.
![Page 17: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/17.jpg)
PROX is a normal approximation estimation procedure that expresses item
difficulty calibrations and person ability measures on a common linear scale. This
measurement unit is called a logit. The procedure assumes that: (a) person abilities (Pv)
are more or less normally distributed [with mean (jj) and standard deviation (cr)], and (b)
item difficulties (8j) are assumed to be more or less normally distributed with average
difficulty (H) and standard deviation (co). Consequently, the effects of the sample on
item difficulty calibrations and that of test length on person ability measures can be
summarized by means and standard deviations on the variable being measured (Wright &
Masters, 1982). The PROX estimation procedure frees the scores from the effects of
sample size and test length, then transforms them into a logit measure (Wright & Stone,
1979). The PROX estimation of a person's ability can be found without iteration as
bv = H + (1 + (<o2/2.89)K In [rv/(L - rv)],
with a standard error of
SE(bv) = (1 + ( C O 2 / 2 . 8 9 ) k [L/rv(l - r v ) f ,
a test height of
L
H = J ] di / L ,
i
and a variance estimate of
L
fi>2 = ( J > 2 -LH 2 ) / (L-1) . i
Item difficulty dj can be found as
![Page 18: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/18.jpg)
d, = M + [1 + O2/2.89f In [(1 - Pi)/Pi]
with a standard error of
SE(dj) = (1 + O'2/2.89)'/2 [l/NPi/(l - Pi)]*
The Rasch model uses the logit function,
In [(1 - PiVPi],
to transform the item p-value into a linear equal interval scale. In the Rasch model, Pj is
calculated as:
Pi = Sj/N,
where S, is the number of satisfactory responses (correct answers) and N the number of
persons. The PROX estimation method is most appropriately used for calibrating new
items, because item difficulties among a sample of new items tend to approximate a
normal distribution, and a sample of persons tends to be normally distributed (Wright &
Stone, 1979).
It has been found that Rasch estimation methods are unbiased, consistent,
efficient, and sufficient (Anderson, 1973; Andrich, 1988; Haberman, 1977; Wright, 1977;
Wright & Stone, 1979). Therefore, the Rasch estimation methods are preferred over
those of the traditional classical true score model. In the traditional classical true score
approach, a person's ability is based on a total test score, usually expressed as total
correct, a percentage of 100, or a percentile rank. The total test score has been shown to
![Page 19: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/19.jpg)
10
reflect ordinal and curvilinear characteristics which are not conducive to meaningful
interpretation.
Rationale for Study
The measurement of individual differences based on the classical true score
approach and its associated statistics has been found to be biased, inefficient, inconsistent,
and insufficient, especially with nonnormal distributions. The most often used scales of
measurement (nominal, ordinal, and interval) have arbitrary zero points and unequal
measurement intervals, and the results are sample specific. In addition, a single error term
(standard error of measurement) is used for all examinees (Allen & Yen, 1979).
Consequently, there is no way to identify specific objectivity among items and persons
(independence of person ability and item difficulties) using the classical true score
approach.
Test data are useful only if there is some correspondence between the items on the
test and the latent trait being measured. In addition, the data should fit the measurement
model used in constructing the test. In the classical true score approach, chi-square (%2) and
the point biserial correlation are used as indexes of goodness of fit. Chi-square is used to
test the difference between observed and expected events, and the point biserial correlation
is sometimes used as an index of fit for items on a test. The problem with % as a fit index
is that there are different sampling distributions as the degrees of freedom change. With the
point biserial correlation, it is unclear what magnitude of value is needed to establish an
![Page 20: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/20.jpg)
11
acceptable item, mid it is sensitive to the score distribution of the sample. The Rasch model
has overcome many of the problems associated with the classical true score approach:
1. A standard error of measurement is provided for each examinee and item.
2. The standard error of measurement can be tested for significance.
3. The measurement scale has an absolute zero point and equal interval units.
4. The item and person parameters are independent.
5. The parameter estimators are unbiased, consistent, efficient, and sufficient.
Given these advantages, the Rasch model provides a more accurate estimate of a person's
ability than does the classical true score approach (Allen & Yen, 1979; Wright & Stone,
1979), and it allows for the independent diagnosis of measurement problems associated
with items, persons, and item by person interactions.
Although several research studies have shown the effects of guessing on Rasch fit
statistics, no studies have used the Logit Residual Index (LRI) in conjunction with item fit
statistics in helping to identify the effects of measurement disturbances. The purpose of this
Monte Carlo simulation is to investigate the effects of varying test parameters and levels of
measurement disturbance on Rasch fit statistics and the Logit Residual Index in the
detection of misfitting items in the single administration of a four option per item multiple
choice test as experienced in a classroom situation. Rasch fit statistics are applied to
simulated data sets of varying sample sizes, test lengths, item difficulty distributions, and
levels of measurement disturbance.
![Page 21: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/21.jpg)
12
Research Questions
Although Rasch estimation models possess the desirable characteristics or
properties of estimators, what effects do measurement disturbances have on Rasch item
fit statistics and the LRI when varying test lengths, sample sizes, item difficulty
distributions, and levels of guessing as a measurement disturbance? To test the effects of
guessing and varying test parameters on Rasch item fit statistics and the LRI, the following
research question is addressed:
1. What effect does medium and high levels of guessing have on Rasch item fit
statistics and the LRI when varying sample sizes, test lengths, and item difficulty
distributions?
Definition of Terms
Chi square (x )~a descriptive measure of the magnitude of the discrepancies between the observed and expected frequencies.
Consistency-an estimate more closely approximates the population parameter as the sample size increases.
Estimator—a statistic used to determine some characteristic about a population or sample.
Efficiency—the sampling error associated with a given estimator is less than the sampling error associated with any other method of estimation.
Fit—the degree to which measurement data approximate the assumptions or characteristics of a particular measurement model.
Latent Trait-an ability or characteristic possessed by an individual that cannot be directly observed.
Infit—the weighted total fit statistic.
Logit-a unit of measurement used in the Rasch model.
![Page 22: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/22.jpg)
13
Logit Residual Index-the sum of the chi-squares for an individual item. Provides an indication of variations (steepness or flatness) in the item characteristic curve.
Measurement-the assignment of numerals to objects or events according to rules.
Measurement Disturbances-conditions that interfere with the measurement of some underlying psychological construct (aptitude, ability, or attitude).
Outfit-the unweighted total fit statistic.
Overfit—items with negative total fit statistics and steeper observed item characteristic curves than predicted.
Parameters—characteristics about a population.
Person ability-the amount of a specific trait possessed by an individual that enables that person to answer a test question correctly.
Plodding—to work slowly on a test and run out of time before attempting all items.
Point biserial correlation-the correlation between a continuous variable and a dichotomous variable (correlation between an item score and the total test score).
Population-all members in an entire group having some common characteristic.
PROX-a normal approximation estimation procedure that expresses item difficulty calibrations and person ability measures on a common linear scale.
Start-up-reduced performances at the beginning of a test due to unfamiliarity, anxiety, and so on.
Statistics—characteristics about a sample.
Unbiased-a statistic shows no systematic tendency to be either greater than or less than the population parameter.
Underfit-items with positive total fit statistics and flatter observed item characteristic curves than predicted.
![Page 23: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/23.jpg)
14
Delimitations
For this study, specific parameters related to test lengths, sample sizes
(examinees), and item difficulty distributions were selected based on a review of previous
research. The study is limited to three sample sizes (25, 50, and 100), three test lengths
(25, 50, and 100), three levels of guessing (no guessing [0%], 25%, and 50%), and two
item difficulty distributions (normal and uniform). In addition, the results are based on
experimental conditions that simulate the single administration of four option per item
multiple choice tests as experienced in classroom situations.
![Page 24: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/24.jpg)
CHAPTER 2
REVIEW OF THE LITERATURE
Historical Perspective
"Thorndike (1918) said, whatever exists at all exists in some amount. To know it
thoroughly involves knowing its quantity as well as its quality" (Crocker & Algina 1986,
p. 4). Psychological constructs, however, are hypothetical abstractions that can be
observed only indirectly and the existence of which can never be folly confirmed.
Stevens (1946) defined measurement as the assignment of numerals to objects or events
according to rules—hence, the use of nominal, ordinal, interval, and ratio scales of
measurement. Lord and Novick (1968) and Torgerson (1958) noted that measurement
applied to the properties of objects rather than the objects themselves. Accordingly,
when we measure an individual or object, we are measuring not the object or person, but
rather the properties that define the construct or variable possessed by the person whose
performance is being measured. Thus, the measurement of such abstractions presents
several problems (Crocker & Algina, 1986).
To empirically investigate the existence of a trait or property, it is necessary to
develop a test theory to guide the investigation. Based upon test theory, we develop tests,
the primary tools by which we collect information about individual differences.
However, before any measurement can be made, an operational definition of the variable
![Page 25: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/25.jpg)
16
of interest must be established. In other words, we must establish some correspondence
between the test items and the construct being measured. This correspondence is known
as establishing an operational definition. In the literature, this is sometimes referred to as
item-objective congruence (Crocker & Algina, 1986). The operational definition or
common line of inquiry allows the test to define the variable being measured and provide
a means for estimating the location of the person taking the test along an ability
continuum based on his/her test score (Wright & Stone, 1979). Test scores are
meaningful only if (a) they relate to some scale of measurement; (b) they are
generalizable beyond the test; and (c) the response pattern is consistent with expectations.
Binet, Thurstone, Thorndike, Stevens and others were among the first to develop
scales of measurement (Crocker & Algina, 1986). These scales provided rules and
meaningful units of measurement for the comparison of individual items and persons in
the assessment of individual differences. These scales are predominately used in the
assessment of individual differences in the classical true score approach to measurement.
In the classical true score approach, a person's observed score (X) is the sum of two
unobservable scores, a true score (T) and an error score (E). The observed score is
defined as
X = 1 + E.
The true score is defined as the average score resulting from an infinite number of
repeated testing with the same instrument. The error score is the difference between the
observed score and the true score. Measurements obtained using this approach are
![Page 26: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/26.jpg)
17
usually expressed as correlations, percentile ranks, z-scores, t-scores, or scaled scores. It
has also been shown that these statistics are sample specific. The disadvantages of the
classical true score approach are that a single error term (standard error of measurement)
is used for all examinees and that the item difficulties are related to the number of persons
who answered the items correctly. It has also been shown that, as the reference group
changes, so does the measured performance of the person taking the test. To what
degree of certainty then are these measurements valid for generalizations beyond the
reference sample?
What was needed was a test theory or model that allowed for the separate and
independent estimation of both item and person parameters, something the classical true
score approach did not take into account. One such model was introduced by Rasch in
1960 along with a true linear scale of measurement (Rasch, 1980). When using the Rasch
model, generalizations beyond the test are based on several assumptions:
1. The test theory used in the development of the test is appropriate.
2. The items on the test define the variable being measured.
3. The test score gives us some indication of the properties that define the
variable that is possessed by the person taking the test.
4. The scale of measurement is linear.
5. The item and person parameters are independent. Thus, when the Rasch model
is fitted properly, the criteria of independence of sample and items were satisfied, and
generalizations beyond the test can be sufficiently made.
![Page 27: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/27.jpg)
18
Measurement Disturbances
Measurement disturbances are conditions that interfere with the measurement of
some underlying psychological construct, such as aptitude, ability, or attitude (Smith,
1991b). With respect to the Rasch model, only two conditions determine the outcome of
the interaction between the person and any item on the test: (a) the amount of the trait
possessed by the person and (b) the amount of the trait necessary to provide a certain
response to a given stimulus (Smith, 1991b). These conditions are commonly referred to
a s P e r s o n ability and item difficulty. Any other condition that influences measurement is
considered noise in the measurement process.
Measurement disturbances that are characteristics of the person and independent
of the items include, but are not limited to (a) start-up, '(b) plodding, (c) cheating,
(d) illness, (e) external distractions, (f) guessing, (g) boredom, and (h) fatigue (Smith,
1991b). Measurement disturbances associated with the interaction of the person and the
properties of the items are (a) guessing, (b) sloppiness, (c) item content, (d) item type,
and (e) item bias. Examples of the third type of measurement disturbance may include
such things as typographical errors, unrelated answer choices, and items unrelated to
content. The most common types of measurement disturbances are cheating and guessing
(Smith, 1991b).
Thorndike (1949) developed a list of possible disturbances to the measurement
process. Smith (1985) later classified measurement disturbances into three general
categories: (a) disturbances that are the results of characteristics of the person that are
![Page 28: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/28.jpg)
19
independent of the items, (b) disturbances that are the interaction between the
characteristics of the person and the properties of the items, and (c) disturbances that are
the results of the properties of the items that are independent of the characteristics of the
person. The classification of measurement disturbances is important in that the source of
measurement disturbances dictates the techniques necessary to detect its presence
(Smith, 1991b).
Glaser (1949,1952) and Mosier (1941) felt that a person would exhibit
consistently correct answers to relatively easy items, consistently incorrect responses to
difficult items, and inconsistent responses to items centered on their ability level. Since
inconsistent responses could be associated with measurement disturbances, Thurstone and
Chave (1929) believed that some criterion should be established such that inconsistent
records should be eliminated from the tabulation. Thus, persons with perfect scores (all
correct) and persons with no items correct (score of zero) are eliminated from Rasch item
and person analysis.
The detection of measurement disturbances can be divided into two general
categories: an investigation of the structure of the entire response matrix (an investigation
of the fit of the responses to individual items [item fit]), and an investigation of the fit of
the responses for an individual person (person fit) (Smith, 1991b). Once a measurement
disturbance has been detected, there are four possible responses: (a) ignore the problem,
(b) assume everyone has the problem and make a correction, (c) use some method of
robust estimation, or (d) use the available information about the items and the people to
![Page 29: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/29.jpg)
20
make a systematic analysis of each individual's response patterns. If a measurement
disturbance is noticed with person analysis, there are four possible alternative actions:
(a) accept the original estimate, (b) modify the response pattern and re-estimate ability,
(c) report only subset ability estimates and no total ability, or (d) decide that there is not
enough information in the responses to report any ability estimate (Smith, 1991b).
An analysis of fit for the entire response matrix does not require additional
information about the items or persons. It can be based solely on the observed responses,
but it is more useful when based on some characteristic of the persons (age, gender,
native language, or ethnic origin). These characteristics can be used to create subgroups
of persons that can be used to test the invariance of the item difficulty parameters. Person
fit analysis is more useful when based on groups of items that have the potential to evoke
measurement noise in certain groups of persons. However, there are some measurement
disturbances that cannot be easily identified in either items or persons (Smith, 1991b).
In 1982 Smith compared the weighted and unweighted between fit statistics as
applied to persons and found that (a) the mean and the standard deviation of the two
statistics were almost identical; (b) the correlation between the two fit statistics was very
high (r = .99); (c) the Type I error rates were almost identical; and (d) there was high
correspondence between items and persons identified as misfitting by the two fit statistics
(Smith, 1991b).
Anderson (1973), Gustafsson (1980), and Wollenberg (1982) suggested the use of
the likelihood ratio chi-square test as an alternative to the between fit statistic because the
![Page 30: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/30.jpg)
21
distributional properties of the Pearson chi-square are not known. Smith and Hedges
(1982) demonstrated that the distributions of the Pearson chi-square and the likelihood
ratio chi-square were almost identical in data simulated to fit the Rasch model and data
designed to simulate various forms of measurement disturbances.
Smith (1991a) examined the effects of test length, number of persons, item
difficulty distributions, person abilities, and the number of steps in each item on the mean
squares. The results suggested that (a) item responses are discrete rather than continuous
variables and have little influence on the distribution of the fit statistics; and (b) estimated
item and person parameters appear to have little effect on the mean of the fit statistics, but
seem to reduce the standard deviation. Further simulations by Smith (1991a) studied the
effects of test length, number of persons, range of item difficulties, and offset between
person ability and item difficulty distributions. The results suggested correction factors
for the restrictions imposed by the use of both estimated item difficulties and person
abilities in the fit analysis. Because there was a magnitude of difference between the
weighted and unweighted versions of the fit statistics, two correction factors were
proposed.
Smith (1988b) performed several simulations to assess the distributional
properties of the weighted and unweighted item between fit statistics. These simulations
involved 10 replications of 1,000 persons taking a 20-item test, with the item difficulties
uniformly distributed from -1 to +1 logits. The results showed that, as the number of
ability groups increased, the mean and standard deviation of the transformed fit values
![Page 31: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/31.jpg)
22
approached the hypothesized values of 0,1. Additional simulations studied the effect of
increasing the number of persons and number of items, varying the dispersion of item
difficulties, and varying the offset between the mean of the item and the person
distributions. The results indicated that, within the ranges studied, varying these factors
had little effect on the distribution of the transformed fit values. Thus, there appears to be
no reason to develop correction factors such as those developed for the weighted and
unweighted total fit statistic to correct for the influence of these factors on the distribution
and Type I error rate of the item between fit statistics (Smith 1991a).
Smith (1988a, 1991a) also studied the power of the total and between item fit
statistics to detect two types of measurement disturbances, item bias and guessing when
unknown. These studies found that the total weighted, unweighted, and between fit
statistics were capable of detecting different types of measurement disturbances. The
between fit statistic was more efficient at detecting item bias than either the unweighted
or weighted total fit statistic. The unweighted and weighted total fit statistics were more
sensitive to disturbances such as guessing and start-up. The primary difference between
the two statistics is that the unweighted version is based on the sum of the standardized
residuals, whereas the weighted version is based on the sum of the standardized residuals
that have been weighted by the information function. For items far away from the
person's estimated ability, the weighting process makes the weighted total fit statistic less
sensitive to residuals from those items. Systematic identification and evaluation of
measurement disturbances were also demonstrated by Wright (1977), Wright and Stone
![Page 32: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/32.jpg)
23
(1979), and Wright and Masters (1982). Unless one is looking for a specific type of
measurement disturbance, it seems necessary to use both the total and between fit
statistics in the analysis of item fit information.
Smith (1986,1988a, 1991b) and Smith and Hedges (1982) studied the power of
the total and between item fit statistics to detect two types of measurement disturbances,
item bias and guessing. These studies found that the total and between item fit statistics
were capable of detecting different types of measurement disturbances. The between fit
statistic was more efficient at detecting item bias, and the total fit statistics were more
sensitive in detecting disturbances such as guessing and start-up.
Rasch Fit Statistics
Prior to the development of computers, the calculation of Rasch fit indexes were
not practical. In fact, the first fit statistic developed for use with Rasch item calibration
computer programs was the overall chi-square statistic (Smith, 1991b).
This statistic was based on the Pearson chi-square typically used in the fit statistics
developed by Wright (1977). The overall chi-square fit statistic was developed to be used
with dichotomously scored test items to assess the fit of the entire data matrix to the
Rasch measurement model rather than assessing the fit of individual items or persons
(Smith, 1991b). The overall chi-square
L m
i=1 j=l
![Page 33: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/33.jpg)
24
is formed by summing a version of the squared standardized residual for the entire matrix
where in is the number of raw score persons or groups L -1) and L is the number of
items on the test with (L - l)(m -1) degrees of freedom (Smith, 1991b; Wright &
Panchapakesan, 1969). The standardized residual is defined as
y aij - (n)(Pij)
where a is the observed number of correct responses for persons with a raw score j, rj is
the number of persons with raw score j, and Py is the probability of a correct response on
that item for group j (Smith, 1991b, p. 165).
Anderson (1973) developed an additional index of overall fit based on the
likelihood-ratio chi-square. Wright and Panchapakesan (1969) also developed a fit
statistic known as the item chi-square, which can be used to test the fit of responses to
individual items. These statistics are referred to as between ability group fit statistics.
Traditionally, the point biserial correlation offered a rough estimate of item fit; however,
this statistic is sample specific. That is, it is dependent upon the score distribution of the
sample. Anderson (1973) and Bamdorff-Nielson (1978) have shown that only item
difficulty is necessary for consistent and sufficient estimates.
Rasch suggested several fit statistics to assess the fit of data to his measurement
models, the weighted total fit and the unweighted total fit statistics. The weighted total
fit statistic is referred to as infit and the unweighted total fit statistic is referred to as
outfit. In the weighted version, a greater weight is given to unexpected responses to
![Page 34: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/34.jpg)
25
items near the person's logit measure (ability), and in the unweighted version, a greater
weight is given to unexpected responses that are farther away from the person's logit
measure (Wright & Stone, 1979). The total fit statistics evaluate the general agreement
between the variable defined by the item and the variable defined by all other items over
the whole sample. The weighted item total fit statistic was developed to diminish the
effect of anomalous outliers, which results in unusually large mean squares (Smith,
1991b). This is evident when an unexpected number of low-ability persons answer
difficult items correctly and an unexpected number of high-ability persons answer easy
items incorrectly at the beginning of the test. These fit statistics are sensitive to
measurement disturbances, such as guessing, start-up, highly discriminating items, and
very easy items, but are relatively insensitive to systematic disturbances such as bias
(Smith, 1991b).
BICAL, a computer program used to test item fit, was developed by Wright and
Mead (1978). This program uses the unweighted versions of two item fit statistics, the
total and between fit statistics. The between fit statistic is based on the number of ability
groups, and the total fit statistic is based on the item/person residual rather than the
item/ability level residual (Smith, 1991b). Later versions of BICAL introduced a log
transformation in an attempt to standardize the fit statistics to an approximate unit normal
distribution. These transformations were introduced because the mean squares that
indicated possible misfit varied from item to item and analysis to analysis, depending i on
![Page 35: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/35.jpg)
26
the number of persons, item difficulty distributions, and the distribution of person
abilities (Smith, 1991b).
The latest version of BICAL uses a cube root transformation to convert the mean
squares of the unweighted total and between fit statistics into approximate unit normals
(Smith, 1991b). However, these statistics are sensitive to start-up, guessing, large ranges
of item difficulties, person abilities, and easy items, thereby producing large mean
squares (misfit). This latest version also introduces the weighted version of the total item
fit statistic, which replaced the unweighted version. Wright and Masters (1982) expanded
the notion of item fit to two polychotomous Rasch models, the rating scale and partial
credit models. With this addition, Rasch fit statistics are now available for models with
other than dichotomously scored items.
MSCALE, CREDIT, BIGSCALE, and BIGSTEPS are among the most recent
Rasch calibration programs. The primary purpose of these programs is to estimate item
and person parameters from a collection of responses to the items (Smith, 1991b). These
programs contain both the unweighted and weighted item total fit statistics. These two
statistics accentuate different parts of the item-person relationship. Although there is a
high correlation between the two fit statistics, the difference between the two can help
diagnose different types of measurement disturbances. The total fit statistics are more
sensitive to measurement disturbances such as guessing, where unusual numbers of
low-ability examinees give correct answers to difficult items, and start-up, where an
unusually high number of high-ability examinees give incorrect responses to easy items
![Page 36: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/36.jpg)
27
at the beginning of the test. The total fit statistics are also sensitive to variation in the
item characteristic curve.
IP ARM (item and person analysis with the Rasch model) is an item analysis and
person analysis computer program for dichotomous and rating scale data. The major
advantage of IP ARM is that it constructs between fit statistics based on characteristics of
the person for item analysis or properties of the items for person analysis. It is the only
software program that provides between fit statistics (unweighted version) for
biographical subpopulations (Smith, 1991b). When biographical data are used in the
analysis, it is a direct test of the invariance of the estimation of the item difficulty
parameter over ability groups (Smith, 1991b). When demographic data are used in the
analysis to create subgroups (sex, race, and age), the resulting statistic will give an
indication of the presence of bias, or differential item familiarity, in response patterns for
the items (Smith, 1991b).
In item analysis with IPARM, the software first calculates the item mean squares
associated with the Rasch fit statistics then converts them to their associated fit statistic
with a cube root transformation. The unweighted mean square item (UMSj) is defined as
UMSJ= f Z2
N^lPn(\-Pn)\ Ntt
where N is the number of people, Xn is the observed response, and Pn is the response
predicted from the logit difficulty of the item and the logit ability of the person (Smith,
![Page 37: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/37.jpg)
28
1991b). In the Rasch model, the probability of a correct response Xvi by person v with
ability pv to item i with difficulty (8;) can be found as
P K i = 1 I M i } = e x P (Pv" 8i)/[l + exp((3v - 8j)]
(Wright & Stone, 1979), or
_ exp (bj-di) lij — ,
1 + Qxp(bj - di)
where bj is the ability measure for persons in score group j (Smith, 1991b, p. 153).
Although these formulas appear to be considerably different, they yield the same results
provided the person's ability measure is the same. The standard deviation of the
unweighted total mean square items can be found as
S[MS(UT)i\ =
N i
iZi w™
1/2
TV
The weighted mean square item (WMSj) can be calculated as
j^iXn-Pnf WMSj = —
YWn n=1
where W, the weighting function, can be calculated as
W = [ P ( l - P ) ]
with a standard deviation of
![Page 38: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/38.jpg)
29
$[MS(WT)i\ =
" N N 1/2
Wni - 4 2 > * _«=1 n=1
2 X n=1
(Smith, 1991b). The unweighted between mean square item (UBMSj) is defined as
UBMSj = l 2L w; ( J - l) ^
£ P„(l- P„)
where / is the number of score groups, N, is the number of persons in each score group,
Xn is the observed response for person n, and P„ is the predicted response for person v
(Smith 1991b, p. 32). The unweighted between standard deviation can be approximated
by
L C - i ) J
1/2
Once calculated, the mean squares are converted to unit normal fit statistics by the
following cube root transformation formula:
where V, the mean square, and S, the standard deviation, are the values associated with
the mean square under consideration (Smith, 1991b; Wright & Masters, 1982). The
resulting fit statistics have expected values of 0,1 (mean of 0 and a standard deviation of
one 1).
![Page 39: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/39.jpg)
30
Logit Residual Index
The Logit Residual Index (LRI) provides a reference as to the flatness or steepness
of the item characteristic curve (ICC). It indicates the linear trend of the residuals summed
over persons for each items. The LRI can be calculated as
N
£ ( Y n i - Y . i ) ( b n ~ d i )
LRI i = — ji "Zibn-dif
n-1
where dj is the difficulty of the item, b„ is the ability of the person, N is the number of
persons, Xni is the observed response, and Pni is the predicted response
and a standardized residual Yni
Yni = { X n i - P n i ) v ^ Yni
— ~ ~ /
P n i ( \ - P n i ) t t N '
(Smith, 1991b, p. 30).
As one of the output variables from IP ARM, the LRI is a measure of how far an
item deviates from the common slope that is fitted for all items. The index has an
expected value of zero (Smith, 1991b). That is, an item with an ICC that fits the modeled
common curve will have an LRI value of zero. Therefore, an item with an LRI value
greater than zero will have an ICC that is steeper than the modeled curve, and items with
an LRI value less than zero will have an ICC that is flatter than the modeled curve.
(Note: In the traditional classical true score approach, the point biserial correlation is used
to provide an estimate of the slope of the observed item characteristic curve. However,
![Page 40: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/40.jpg)
31
the point biserial correlation has been found to be sample specific, and no discrete values
have been established to provide an accurate interpretation of the correlation coefficient
as related to the slope of the item characteristic curve).
IP ARM automatically assigns group membership (ability groups) based on the
performance of each person on each item. The program attempts to place an equal
number of persons in each ability group. Negative LRI values indicate that low-ability
groups should have positive residuals and high-ability groups should have negative
residuals. This indicates that low-ability persons performed better than expected and
high-ability persons performed less well than expected. Positive LRI values indicate that
low-ability groups should have negative residuals and high-ability groups should have
positive residuals. This indicates that high-ability persons performed better than expected
and low-ability persons performed less well than expected. Items with negative total fit
statistics tend to have steeper observed ICCs than predicted, indicating an overfit to the
model, and items with positive total fit statistics tend to have flatter observed ICCs than
predicted, indicating an underfit to the model.
Purpose of Study
Smith (1991b) proposed that item fit statistics in Rasch calibration programs
provide a frame of reference forjudging item performance and that one way of
establishing this frame of reference is to simulate data that fit the Rasch model over a
variety of test conditions. The present Monte Carlo study investigated the effects of
varying item difficulty distributions, number of persons, number of items, and levels of
![Page 41: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/41.jpg)
32
guessing on Rasch fit statistics and the LRI. SIMTEST, a data simulation program, and
IP ARM, a Rasch item and person analysis program, were used to simulate and test the
effects of varying test conditions on Rasch fit statistics and the LRI.
![Page 42: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/42.jpg)
CHAPTER 3
METHODS AND PROCEDURES
Data Set Construction
For the purpose of this investigation, synthetic data sets were generated using
SIMTEST version 2.1, a software program developed by Stuart Luppescu (1992) for
simulating dichotomous test data. The program allows the user to adjust person abilities,
number of persons, number of items, item difficulty, bias, start-up, guessing, and slope
(discrimination) parameters. Once the parameters have been set, the program randomly
generates dichotomous data sets based on the input parameters. If bias or start-up is used
in the analyses, bias is added, or start-up (i.e., reduced performance at the beginning of
the test due to unfamiliarity, anxiety, etc.) is subtracted from some of the interactions.
When guessing is introduced into the analyses, expected values are calculated according
to the Rasch model (with weighting if there are slopes not equal to 1.0), and dichotomous
test item responses are produced.
Simulated Data Sets
Dichotomous test data were simulated in varying test length, number of persons,
and item difficulty distributions. A 3 by 3 by 2 design was used (three test lengths (25,
50, and 100 items), three person parameters (25, 50, and 100 persons) and two item
difficulty distributions (normal and uniform). To test the effect of guessing on item fit
![Page 43: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/43.jpg)
34
statistics, the design was subjected to three experimental conditions: (a) no guessing, (b) a
25% chance of guessing correctly, and (c) a 50% chance of guessing correctly.
A SIMTEST batch file was written for each of 54 experiments. The first
experimental condition involved 25 items, 25 persons, normally distributed item
difficulties, and no guessing. The control file for this experiment was as follows:
SIMTEST -H -OEXPl.DAT -S2.0 -NO -DN -125- -P25 -F0 -M0.0 -A0.0 -U0.0 -T-1E200 -CO -L1.0.
The following is a description of the parameters used in the SIMTEST control file
to simulate the data: (a) the batch mode (-H or hands-off switch); (b) output file name (-
0EXP1.DAT); (c) a standard deviation of 2 (-S2.0); (d) no biased items (-NO); (e) normal
distribution of item difficulties (-DN); (f) 25 items (-125); (g) 25 persons (-P25); (h) no
persons with biased scores (-F0); (i) a mean person ability measure of zero (-M1.0); (j) no
startup reduction (-AO); (k) no persons with startup (-U0); (1) threshold for guessing (-T-
1E200); (m) the chance of guessing correctly (-CO); and (n) a slope of 1 [the Rasch
model] (-L1.0).
The item difficulty distributions for the simulation were set at normal and
uniform. The parameters used to simulate the normally distributed data sets were (a) a
mean person ability of 1; (b) a standard deviation of 2; (c) a slope of 1; (d) three test
lengths (25, 50, and 100); and (e) three levels of guessing (no guessing [0%], 25%, and
50%). For the uniformly distributed data sets, the parameters were (a) a mean person
ability of 1; (b) a standard deviation of 1; (c) a slope of 1; (d) three test lengths (25, 50,
![Page 44: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/44.jpg)
35
and 100); and (e) three levels of guessing (no guessing [0%], 25%, and 50%). Given
these parameters, 95% of the item difficulties for both the normal and uniform
distributions will fall between ±2 standard deviations, with the greatest concentration
centered on the mean of zero (0). The parameters used to simulate the chances of
guessing the correct response to an item were 25% and 50%. This is equivalent to having
a l-in-4 and a 2-in-4 chance of guessing the correct answer to an item on a four-option
per item multiple choice test. The discrimination or slope parameter was set at 1,
invoking the program to simulate data appropriate for the Rasch model. Bias as a
measurement disturbance was not used as an experimental condition in the data
simulation; therefore, the control variable used to invoke bias was not entered as a control
variable.
The experiments are referred to by number, as shown in Table 1. For example,
Experiment 1 consists of a 25-item test with normally distributed item difficulties,
administered to 25 persons with a slope of 1 and no guessing.
Table 1
Definition of Experiments
Experiment Measurement disturbance
Difficulty distribution Items Persons
1 No guessing Normal 25 25 2 No guessing Normal 50 25 3 No guessing Normal 100 25 4 No guessing Normal 25 50 5 No guessing Normal 50 50 6 No guessing Normal 100 50 7 No guessing Normal 25 100 8 No guessing Normal 50 100
Ctable continues^
![Page 45: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/45.jpg)
36
Measurement Difficulty Experiment disturbance distribution Items Persons
9 No guessing Normal 100 100 10 Guessing (.25) Normal 25 25 11 Guessing (.25) Normal 50 25 12 Guessing (.25) Normal 100 25 13 Guessing (.25) Normal 25 50 14 Guessing (.25) Normal 50 50 15 Guessing (.25) Normal 100 50 16 Guessing (.25) Normal 25 100 17 Guessing (.25) Normal 50 100 18 Guessing (.25) Normal 100 100 19 Guessing (.50) Normal 25 25 20 Guessing (.50) Normal 50 25 21 Guessing (.50) Normal 100 25 22 Guessing (.50) Normal 25 50 23 Guessing (.50) Normal 50 50 24 Guessing (.50) Normal 100 50 25 Guessing (.50) Normal 25 100 26 Guessing (.50) Normal 50 100 27 Guessing (.50) Normal 100 100 28 No guessing Uniform 25 25 29 No guessing Uniform 50 25 30 No guessing Uniform 100 25 31 No guessing Uniform 25 50 32 No guessing Uniform 50 50 33 No guessing Uniform 100 50 34 No guessing Uniform 25 100 35 No guessing Uniform 50 100 36 No guessing Uniform 100 100 37 Guessing (.25) Uniform 25 25 38 Guessing (.25) Uniform 50 25 39 Guessing (.25) Uniform 100 25 40 Guessing (.25) Uniform 25 50 41 Guessing (.25) Uniform 50 50 42 Guessing (.25) Uniform 100 50 43 Guessing (.25) Uniform 25 100 44 Guessing (.25) Uniform 50 100 45 Guessing (.25) Uniform 100 100 46 Guessing (.50) Uniform 25 25 47 Guessing (.50) Uniform 50 25 48 Guessing (.50) Uniform 100 25 49 Guessing (.50) Uniform 25 50 50 Guessing (.50) Uniform 50 50 51 Guessing (.50) Uniform 100 50 52 Guessing (.50) Uniform 25 100 53 Guessing (.50) Uniform 50 100 54 Guessing (.50) Uniform 100 100
![Page 46: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/46.jpg)
37
Rasch Analysis
SIMTEST version 2.1, a software program developed by Stuart Luppescu (1992)
for simulating dichotomous test data, was used to generate the 54 simulated experimental
data sets. Once the data were simulated, a BIGSTEPS program was written for each
experimental condition to output a data file containing item difficulty parameters to be
read by IP ARM. The BIGSTEPS control program for Experiment 1 (25 items, 25
persons, no guessing, and normally distributed item difficulties) was as follows:
TITLE='EXP1 - ND, NG, 25 ITEMS, 25 PERSONS' NI=10 NAME1=1 ITEM1=13 DATA=C:\EXP1\EXP1 .DAT CODES-Ol TABLES=1010001000100100000000 STBIAS=N INUM=Y IFILE=C :\EXP 1 \EXP 1 BIG.D AT &END
The control parameters consisted of (a) the title; (b) the number of items (NI=25);
(c) the beginning column for identification information (NAME=1); (d) the beginning
column for the data (ITEM1=13); (e) the name of the data file (C:\EXP1\EXP1.DAT); (f)
the possible item response values (CODES=01); (g) the tables included in the output
(TABLES=1010001000100100000000); (h) no statistical bias correction factor
(STBIAS=N), (i) automatic generation of item names (INUM=Y); (j) name of the item
![Page 47: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/47.jpg)
38
difficulty output file (IFILE=C r\EXPl\EXPlBIG.DAT), and (k) the end statement
(&END).
IP ARM was used to produce Rasch fit statistics for the experimental conditions
and the logit residual indices (See Appendix A for a description of the input parameters
required to construct the data analysis control file and initiate the IP ARM program.)
Statistical Analysis
In order to determine whether there were significant differences between mean fit
statistics across and within experimental conditions involving varying levels of guessing
and distribution types, two-tailed independent t-tests were conducted using the formula
for independent samples at the .05 level of significance (Ferguson, 1981). The
experimental conditions were no guessing (0%), 25%, and a 50% chance of guessing the
correct answer in normal and uniformly distributed item difficulty distributions. The
sample n used in the t-test analyses was the number of experimental conditions (9 total)
involved at each level of guessing.
Chi-square was used to determine whether an increased probability of guessing
the correct answer increased the frequency of misfitting items detected by Rasch fit
statistics (Ferguson, 19981). The sample n used in the (y2) analyses was the levels of
guessing (3 total).
![Page 48: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/48.jpg)
CHAPTER 4
RESULTS
Effects of Guessing on Rasch Fit Statistics
To test the effects of guessing on Rasch fit statistics and the Logit Residual Index
(LRI), experimental conditions were simulated, with varying test lengths (25, 50, and
100), persons (25,50, and 100), levels of guessing (no guessing [0%], 25%, and 50%),
and distribution types (normal aid uniform), resulting in 54 experiments. Item and
person parameters were not considered as effect factors but were used to simulate
test-taking conditions. Experiments 1-27 involved normally distributed item difficulty
distributions with varying levels of guessing (1-9, no guessing [0%]; 10-18, 25%; and 19-
27, 50%), and Experiments 28-54 involved uniformly distributed item difficulty
distributions with varying levels of guessing (28-36, no guessing [0%]; 37-45,25%; and
46-54, 50%). The LRI has an expected value = 0; therefore, the mean and S. D. for this
index are not reported in the summary tables. A summary of item fit information by
experiment is presented in Appendix C. Summaries of mean item fit information by
experiment and level of guessing are presented in Tables 2 through 8. A brief description
of the table format used to display the data is presented in Figure 2.
![Page 49: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/49.jpg)
40
Column 1 - Experiment number Column 2 - .Number of items Column 3 - Logit item difficulty. Column 4 - Point biserial correlation Column 5 - Unweighted total fit statistic (outfit statistic). Column 6 - Weighted total fit statistic (infit statistic). Column 7 - Ability between fit statistic (ability groups). Column 8 - Mean item score (proportion correct). Column 9 - Logit Residual Index (indicate variations in the ICC). Mean - Average value for each column is located at the base of the table S.D. - Standard deviation
Figure 2. Table format for the display of experimental data.
Table 2
Mean Summary of Item Fit Information for Experiments 1-9 (Normally Distributed Item Difficulty Distributions and No Guessing)
Logit Point. Unwt. Wt. Ability Mean Logit #of item bis. total total between item residual
Exp. # items diff. corr. fit fit fit score index
25 Persons 1 25 -0.16 0.41 0.20 -0.05 0.20 0.61 -0.07 2 50 -0.10 0.41 0.10 0.01 -0.03 0.66 0.00 3 100 0.29 0.39 0.15 0.00 0.03 0.67 -0.01
50 Persons 4 25 -0.37 0.41 0.12 -0.05 0.01 0.64 0.00 5 50 -0.09 0.39 0.05 0.08 0.00 0.65 0.03 6 100 -0.13 0.38 0.00 0.06 0.04 0.62 0.01
100 Persons 7 25 -0.21 0.42 0.11 -0.03 0.04 0.65 0.00 8 50 -0.09 0.31 0.13 0.06 0.08 0.67 -0.03 9 100 0.00 0.40 0.07 0.05 0.01 0.63 0.00
Mean -0.10 0.39 0.10 0.01 0.04 0.64 S.D. 0.18 0.03 0.06 0.05 0.07 0.02
![Page 50: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/50.jpg)
41
A 25% chance of guessing the correct answer was randomly introduced into
Experiments 10-19 as a measurement disturbance, while the item and person parameters
(test lengths and group membership) remained constant. A summary of mean item fit
information is shown in Table 3.
Table 3
Mean Summary of Item Fit Information for Experiments 10-18 (Normally Distributed Item Difficulty Distributions and a 25% Chance of Guessing Correctly)
Logit Point. Unwt. Wt. Ability Mean Logit #of item bis. total total between item residual
Exp. # items diff. corr. fit fit fit score index
25 Persons 10 25 -0.30 0.40 0.19 -0.08 0.15 0.65 -0.06 11 50 -0.21 0.33 0.03 -0.00 0.22 0.61 0.03 12 100 -0.32 0.29 0.01 -0.05 0.18 0.62 0.02
50 Persons 13 25 -0.16 0.37 0.16 0.06 0.34 0.68 -0.04 14 50 -0.17 0.36 0.05 0.01 0.13 0.64 0.01 15 100 -0.22 0.35 0.04 0.03 -0.06 0.65 0.02
100 Persons 16 25 -0.19 0.34 -0.04 -0.05 0.21 0.64 0.04 17 50 -0.10 0.33 0.07 0.01 0.19 0.64 0.00 18 100 -0.15 0.34 0.08 -0.02 0.22 0.64 -0.01
Mean -0.20 0.35 0.07 -0.01 0.18 0.64 S.D. 0.07 0.03 0.07 0.04 0.11 0.02
A 50% chance of guessing the correct answer was randomly introduced into
Experiments 19-27 while holding all other conditions constant. A summary of mean fit
information is presented in Table 4.
![Page 51: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/51.jpg)
42
Table 4
Summary of Mean Item Fit Information for Experiments 19-27 (Normally Distributed Item Difficulty Distributions and a 50% Chance of Guessing Correctly')
Logit Point. Unwt. Wt. Ability Mean Logit #of item bis. total total between item residual
Exp. # Items diff. corr. fit fit fit score index
25 Persons 19 25 -0.31 0.40 0.08 -0.04 0.29 0.63 0.01 20 50 -0.28 0.37 0.07 0.06 0.12 0.67 0.00 21 100 -0.14 0.27 0.06 0.01 0.24 0.62 0.02
50 Persons 22 25 -0.36 0.38 0.12 -0.04 0.13 0.64 0.02 23 50 -0.18 0.38 0.18 -0.01 0.06 0.61 -0.05 24 100 -0.09 0.38 0.10 0.01 0.14 0.64 0.00
100 Persons 25 25 0.00 0.36 0.12 -0.02 0.19 0.64 0.01 26 50 -0.09 0.33 0.06 -0.01 0.13 0.66 0.00 27 100 -0.10 0.32 0.08 0.01 0.00 0.63 0.00
Mean -0.17 0.35 0.10 0.00 0.14 0.64 S.D. 0.12 0.04 0.04 0.03 0.09 0.02
To determine whether levels of guessing had an effect on mean Rasch fit statistics
within normally distributed experimental conditions, comparisons were made between
the mean fit values obtained at each level of guessing using an independent t-test at the
.05 level of significance. Item and person parameters remained constant across
experimental conditions. Mean fit values for Rasch fit statistics by experimental
conditions (experiments and levels of guessing) are presented in Table 5.
![Page 52: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/52.jpg)
43
Table 5
Mean Differences for Experiments With Normally Distributed Item Difficulties With No Guessing. 25%. and a 50% Chance of Guessing Correctly
Experiments % Guessing
Unwt. total Wt. total Between
Experiments % Guessing Mean S.D. Mean S.D. Mean S.D.
1-9 0 .10 .06 .01 .05 .04* .07 10-18 25 .07 .07 -.01 .04 .18* .11 19-27 50 .10 .04 .00 .03 .14* .09
Note. Independent t-tests (two tailed) were calculated based upon the number of experimental conditions (n = 9, df=16), with significance set at the .05 level (t > 2.120). An asterisk (*) indicates significance at the .05 level.
No significant differences were observed across levels of guessing between mean
fit values associated with the unweighted and weighted total fit statistics at the .05 level.
For the between fit statistics, the mean values showed significant differences at the .05
level between experimental conditions involving no guessing and those involving a 25%
and 50% chance of guessing correctly.
Experiments 28-54 involved uniformly distributed item difficulty distributions
across the same conditions that were used in the normally distributed conditions. A
summary of mean fit information for experimental conditions involving no guessing in
uniformly distributed conditions (Experiments 28-36) is presented in Table 6.
![Page 53: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/53.jpg)
44
Table 6
Summary of Mean Item Fit Information for Experiments 28-36 (Uniformly Distributed Item Difficulty Distributions and No Guessing)
Logit Point. Unwt. Wt. Ability Mean Logit # of item bis. total total between item residual
Exp. # items diff. corr. fit fit fit score index
25 Persons 28 25 0.00 0.44 0.09 0.02 -0.20 0.68 0.02 29 50 -0.00 0.34 0.11 0.03 -0.16 0.71 -0.01 30 100 0.00 0.36 0.05 0.03 0.10 0.62 -0.02
50 Persons 31 25 0.00 0.33 -0.05 0.05 0.01 0.71 0.05 32 50 -0.00 0.41 0.03 -0.02 0.17 0.68 0.03 33 100 -0.00 0.40 0.02 0.01 -0.07 0.67 0.00
100 Persons 34 25 0.00 0.39 0.01 0.03 0.02 0.70 0.00 35 50 -0.00 0.42 0.07 -0.01 -0.07 0.66 -0.01 36 100 0.00 0.43 0.07 -0.02 -0.04 0.68 -0.01
Mean 0.00 0.39 0.04 0.01 -0.03 0.68 S.D. 0.00 0.04 0.05 0.03 0.12 0.03
A 25% chance of guessing the correct answer was randomly introduced into
Experiments 37-45 as a measurement disturbance, while the item (test lengths) and
person (group membership) parameters remained constant. A summary of mean fit
information is shown in Table 7.
![Page 54: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/54.jpg)
45
Table 7
Summary of Mean Ttem Fit Information for Experiments 37-45 (Uniformly Distributed Ttem Difficulty Distributions and a 25% Chance of Guessing Correctly)
Exp. #
Logit Point. # of item bis.
items diff. corr.
Unwt. Wt. total total fit fit
Ability between
fit
Mean item score
Logit residual index
37 38 39
40 41 42
43 44 45
25 Persons 25 -0.00 0.48 0.07 0.02 0.26 0.70 0.06 50 -0.00 0.37 0.08 0.06 0.21 0.71 -0.02
100 0.00 0.36 0.05 0.06 0.28 0.72 0.00
50 Persons 25 -0.00 0.41 0.07 -0.00 0.03 0.71 -0.01 50 0.00 0.34 0.04 0.02 -0.04 0.66 0.04
100 0.00 0.42 0.06 0.04 0.14 0.72 -0.01
100 Persons 25 -0.00 0.40 0.11 -0.03 0.13 0.71 -0.02 50 0.00 0.37 -0.00 0.03 -0.16 0.68 0.01
100 -0.00 0.40 0.00 0-02 0.08 0.67 0.01
Mean S.D.
0.00 0.00
0.39 0.04
0.05 0.04
0.02 0.03
0.10 0.14
0.70 0.02
A 50% chance of guessing the correct answer was randomly introduced into
Experiments 46-54 while holding all other experimental conditions constant. A summary
of mean fit information is presented in Table 8.
![Page 55: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/55.jpg)
46
Table 8
Summary of Mean Item Fit Information for Experiments 46-54 (Uniformly Distributed Item Difficulty Distributions and a 50% Chance of Guessing Correctly*
Exp.#
Logit Point. Unwt. Wt. Ability # of item bis. total total between
Items diff. corr. fit fit fit
Mean item score
Logit residual index
46 47 48
49 50 51
52 53 54
25 Persons 25 50
100
25 50
100
25 50
100
0.00 -0.00 -0.06
0.00 0.00
-0.00
0.00 0.00 0.00
0.52 0.44 0.34
0.45 0.40 0.41
0.39 0.37 0.37
0.18 0.04 0.05
-0.00 0.02 0.07
50 Persons 0.00 0.05 0.04 -0.03 0.02 0.03
100 Persons 0.04 -0.04
-0.01 0.05 0.02 0.03
0.08 0.30 0.15
0.23 0.23 0.26
0.60 0.17 0.04
0.70 0.72 0.71
0.68 0.69 0.72
0.72 0.68 0.73
-0.05 -0.01 -0.03
0.00 0.03 0.00
0.00 -0.01 0.00
Mean S.D.
-0.01 0.02
0.41 0.05
0.04 0.06
0.02 0.04
0.23 0.16
0.71 0.02
To determine whether levels of guessing had an effect on Rasch fit statistics in
uniformly distributed experimental conditions, comparisons were made between the mean
fit values obtained at each level of guessing with an independent t-test at the .05 level of
significance. The item and person parameters remained constant across experimental
conditions. Mean fit values for each experimental condition are presented in Table 9.
![Page 56: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/56.jpg)
47
Table 9
Mean Differences for Experiments With Uniformly Distributed Item Difficulties With No Guessing, 25%. and a 50% Chance of Guessing Correctly
Experiments % Guessing
Unwt. total Wt. total Between Experiments % Guessing Mean S.D. Mean S.D. Mean S.D. 28-38 0 .04 .05 .01 .03 -.03* .12 37-45 25 .05 .04 .02 .03 .10* .14 46-54 50 .04 .06 .02 .04 .23* .16 Note. Independent t-tests (two tailed) were calculated based upon the number of experimental conditions (n = 9, df=16), with significance set at the .05 level (t > 2.120). An asterisk (*) indicates significance at the .05 level.
No significant differences were found between the mean fit values associated with
the unweighted and weighted total fit statistics across guessing levels at the .05 level. For
the between fit statistic, a significant difference was found between the mean values for
experimental conditions involving no guessing and those involving a 25% and 50%
chance of guessing correctly.
To determine whether distribution types had a significant effect on Rasch fit
statistics, comparisons were made between the mean fit values obtained at the same levels
of guessing in each distribution type using an independent t-test (two-tailed) at the .05
level of significance. A summary of mean fit values by experimental conditions
(experiments and levels of guessing) and distribution type is presented in Table 10.
![Page 57: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/57.jpg)
48
Table 10
Mean Differences for Experiments With Normal and Uniformly Distributed Item Difficulties at the Same Levels of Guessing
Unwt. total Wt. total Between
Normal item difficulty distributions
1-9 0 .10 .06 .10* .05 .04 .07 10-18 25 .07 .07 -.01 .04 .18 .11 19-27 50 .10* .04 .00 .03 .14 .09 Uniform item difficulty distributions
28-38 0 .04 .05 .01* .03 -.03 .12 37-45 25 .05 .04 .02 .03 .10 .14 46-54 50 .04* .06 .02 .04 .23 .16
Note. Independent t-tests (two tailed) were calculated based upon the number of experimental conditions (n = 9,df=16), with significance set at the .05 level (t > 2.120). An asterisk (*) indicates significance at the .05 level.
The mean fit values for the unweighted total fit statistic showed a significant
difference at the .05 level between experimental conditions involving a 50% chance of
guessing the correct answer in normal and uniformly distributed item difficulty
distributions. No significant differences were found between mean fit values for the
unweighted total fit statistic at the 0% and 25% levels of guessing. The mean weighted
total fit values showed a significant difference at the .05 level between experimental
conditions involving no guessing (0%) in normal and uniform item difficulty
distributions. No significant differences were observed between the mean fit values for
the weighted total fit statistic at the 25% and 50% levels of guessing correctly. No
![Page 58: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/58.jpg)
49
significant differences were found between the mean fit values for the various levels of
guessing (no guessing, 25%, and 50%) for the between fit statistic at the .05 level.
Detection of Guessing by Rasch Fit Statistics
In order to determine whether the frequency of misfitting items detected by Rasch
fit statistics differed by distribution types and levels of guessing, an analysis of observed
and expected frequencies were conducted using chi-square (x2). The sum of misfitting
items detected in experimental conditions involving no guessing on tests of 25, 50 and
100 items in each distribution type was used as the baseline or theoretical frequencies
(expected) against which those observed on equivalent tests at the 25% and 50% levels of
guessing were compared. Because these values occurred in the absence of measurement
disturbance, they were purely a matter of chance and provided a realistic frame of
reference against which conditions involving different levels of guessing could be
compared. (See Appendix B for a summary of misfitting items.) Comparisons were
made within and across distribution types. Comparisons of observed and expected
frequencies within normally distributed conditions are presented in Table 11.
![Page 59: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/59.jpg)
50
Table 11
A Comparison of the Frequency of Misfitting Items Detected by Rasch Fit Statistics in a Normal Distribution of Item Difficulties at Varying Test Lengths (Items') and Levels of Guessing Using y-.
Statistic % Items 0 E {O-E) 0O-E)2 (o - ey E
Sig.
Unwt. 25 25 3 3 0.00 0.00 0.00 Unwt. 50 25 2 3 -LOO 1.00 0.33 Unwt. 25 50 2 3 -1.00 1.00 0.33 Unwt. 50 50 3 3 0.00 0.00 0.00 Unwt. 25 100 6 5 1.00 1.00 0.20 Unwt. 50 100 5 5 0.00 0.00 0.00 Wt. 25 25 1 0 1.00 1.00 Wt. 50 25 3 0 3.00 9.00 __ Wt. 25 50 3 3 0.00 0.00 0.00 Wt. 50 50 4 3 1.00 1.00 0.33 Wt. 25 100 3 0 3.00 9.00 Wt. 50 100 2 0 2.00 4.00 . .
Bet. 25 25 3 1 2.00 4.00 4.00 Bet. 50 25 3 1 2.00 4.00 4.00 Bet. 25 50 7 1 6.00 36.00 36.00 *
Bet. 50 50 6 1 5.00 25.00 25.00 *
Bet. 25 100 7 7 0.00 0.00 0.00 Bet. 50 100 6 7 -LOO 1.00 0.17
Note. Percent (%) indicates the chance of guessing correctly. Statistic indicates Rasch fit statistics (Unwt. = unweighted total, Wt. = weighted total, and Bet. = unweighted ability between). "Items" indicate the number of items on each test. The symbol O indicates observed frequencies and E expected frequencies Significance (Sig.) level for x at .05 is > 5.991 with df= 2. An asterisk (*) indicates significance at the .05 level. Chi-square values indicated by are undefined values (division by zero).
![Page 60: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/60.jpg)
51
In experimental conditions involving normally distributed conditions, significant
differences were found at the .05 level between the frequency of observed and expected
misfitting items detected by the between fit statistic on tests simulated with 50 items and
a 25% chance of guessing the correct answer and on tests with 50 items and a 50%
chance of guessing the correct answer. No significant differences were found between
the frequency of observed and expected misfitting items detected by the weighted and
unweighted total fit statistics in normally distributed experimental conditions.
Comparisons of expected and observed frequencies in uniformly distributed conditions
are presented in Table 12.
On tests with uniformly distributed item difficulties, significant differences were
observed at the .05 level between the frequency of expected and observed misfitting items
detected by the weighted total and between fit statistics on tests simulated with 100 items
and a 25% chance of guessing the correct answer and on tests with 25 items and a 50%
chance of guessing the correct answer, respectively. No significant differences were
observed between the expected and observed frequency of misfitting items detected by
the weighted and unweighted total fit statistics in the uniformly distributed experimental
conditions.
![Page 61: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/61.jpg)
52
Table 12
A Comparison of the Frequency of Misfitting Items Detected by Rasch Fit Statistics in a Uniform Distribution of Item Difficulties at Varying Test Lengths and Levels of Guessing Using y-
Statistic % Items O E (O-E) {O-Ef (O - Ef E
Sig.
Unwt. 25 25 2 2 0.00 0.00 0.00 Unwt. 50 25 2 2 0.00 0.00 0.00 Unwt. 25 50 3 3 0.00 0.00 0.00 Unwt. 50 50 5 3 2.00 4.00 1.33 Unwt. 25 100 4 9 -5.00 25.00 2.78 Unwt. 50 100 7 9 -2.00 4.00 0.44 Wt. 25 25 1 3 -2.00 4.00 1.33 Wt. 50 25 0 3 -3.00 9.00 3.00 Wt. 25 50 3 3 0.00 0.00 0.00 Wt. 50 50 6 3 3.00 9.00 3.00 Wt. 25 100 7 2 5.00 25.00 12.50 *
Wt. 50 100 2 2 0.00 0.00 0.00 Bet. 25 25 2 1 1.00 1.00 1.00 Bet. 50 25 4 1 3.00 9.00 9.00 *
Bet. 25 50 3 3 0.00 0.00 0.00 Bet. 50 50 6 3 3.00 9.00 3.00 Bet. 25 100 9 6 3.00 9.00 1.50 Bet. 50 100 5 6 -1.00 1.00 0.17
Note. Percent (%) indicates the chance of guessing the correct answer. Statistic indicates Rasch fit statistics (Unwt. = unweighted total, Wt. = weighted total, and Bet. = unweighted ability between). "Items" indicate the number of items on each test. The symbol O indicates observed frequencies and E expected frequencies. Significance (Sig.) level for %2 at .05 is > 5.991 with df= 2. An asterisk (*) indicates significance at the .05 level. Chi-square values indicated by
are undefined values (division by zero).
![Page 62: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/62.jpg)
53
In order to determine whether normal or uniform distribution types had an effect
on the frequency of misfitting items detected by Rasch fit statistics, a chi-square analysis
of the frequency of misfitting items across distribution types was conducted at the .05
level of significance. Assuming that no significance differences would occur (null
hypothesis), the frequency of misfitting items detected in experiments involving normally
distributed conditions were used as the expected frequencies, and those detected in
experiments involving uniformly distributed conditions were used as observed
frequencies. The results of the % analyses are presented in Table 13. No significant
differences were found between the frequency of misfitting items detected by Rasch fit
statistics in normal and uniformly distributed conditions when varying the levels of
guessing.
![Page 63: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/63.jpg)
54
Table 13
A Comparison of the Frequency of Misfitting Items Detected bv Rasch Fit Statistics in Normal and Uniformly Distributed Item Difficulties Using v~.
Statistic % Items O E (O-E) (O-E)2 o o - E y sig. E
Unwt. 0 25 2 3 -1.00 1.00 0.33 Unwt. 0 50 3 3 0.00 0.00 0.00 Unwt. 0 100 9 5 4.00 16.00 3.20 Unwt. 25 25 2 3 -1.00 1.00 0.33 Unwt. 25 50 3 2 1.00 1.00 0.50 Unwt. 25 100 4 6 -2.00 4.00 0.67 Unwt. 50 25 2 2 0.00 0.00 0.00 Unwt. 50 50 5 3 2.00 4.00 1.33 Unwt. 50 100 7 5 2.00 4.00 0.80 Wt. 0 25 3 0 3.00 9.00 . . .
Wt. 0 50 3 3 0.00 0.00 0.00 Wt. 0 100 2 0 2.00 4.00 Wt. 25 25 1 1 0.00 0.00 0.00 Wt. 25 50 3 3 0.00 0.00 0.00 Wt. 25 100 7 3 4.00 16.00 5.33 Wt. 50 25 0 3 -3.00 9.00 3.00 Wt. 50 50 6 4 2.00 4.00 2.00 Wt. 50 100 2 2 0.00 0.00 0.00 Bet. 0 25 1 1 0.00 0.00 0.00 Bet. 0 50 3 1 2.00 4.00 4.00 Bet. 0 100 6 7 -1.00 1.00 0.14 Bet. 25 25 2 3 -1.00 1.00 0.33 Bet. 25 50 3 7 -4.00 16.00 2.29 Bet. 25 100 9 7 2.00 4.00 0.57 Bet. 50 25 4 3 1.00 1.00 0.33 Bet. 50 50 6 6 0.00 0.00 0.00 Bet. 50 100 5 6 -1.00 1.00 0.17
Note- Percent (%) indicates the levels of guessing. Statistic indicates Rasch fit statistics (Unwt. = unweighted total, Wt. = weighted total, and Bet. = unweighted ability between). "Items" indicate the number of items on each test. The symbols O and E represent observed and expected frequencies, respectively. Significance (Sig.) level for x at .05 is > 5.991 with df= 2. An asterisk (*) indicates significance at the .05 level. Chi-square values indicated by are undefined values (division by zero).
![Page 64: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/64.jpg)
55
Guessing and the Logit Residual Index
Misfitting item data were investigated to determine the effects of guessing on the
Logit Residual Index (LRI). (See Appendix B for a summary of misfitting item
information by experimental conditions.) Presented in Table 14 are the total number of
misfitting items detected by experimental condition and the number and percentage of
those items detected by each fit statistic.
Table 14
Number and Percentage of Misfitting Items Detected bv Rasch Fit Statistics Across Experiments Involving Normal and Uniformly Distributed Item Difficulty Distributions
% Number (%) detected Experiments Guessing N Unweighted. Weighted. Between
Normal item difficulty distributions 1-9 0 17 10 (.59) 2 (.12) 7 (.41) 10-18 25 23 11 (.48) 7 (.30) 17 (.74) 19-27 50 26 10 (.38) 8 (.31) 15 (.58) Uniform item difficulty distributions 28-36 0 24 14 (.58) 7 (.29) 9 (.38) 37-45 25 29 9 (.31) 9 (.31) 16 .(55) 46-54 50 27 13 (.48) 8 (.30) 15 (.56)
Note. Percentages may not sum to 100% because one or more fit statistics may have detected the same items as misfitting. The sum of the items detected by Rasch fit statistics may not sum to the total number of misfitting items at each level of guessing "N" because one or more statistics may have detected the same items. Percent (%) indicates the levels of guessing.
The sum of the items detected by each fit statistic may not sum to the total
number of items detected at each level of guessing because one or more statistics
detecting the same items as misfitting. Presented in Table 15 are the total number of
![Page 65: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/65.jpg)
56
misfitting items by experimental condition (N) and the number and percentage of those
items producing positive and negative LRI values.
Table 15
Number of Misfitting Items and Number and Percentage of LRI Values bv Experimental Conditions in Normal and Uniform Item Difficulty Distributions
—
Positive Negative
Experiments % Guessing Misfitting items N % N %
Normal item difficulty distributions 1-9 0 17 1 6 16 94 10-18 25 23 4 22 18 78 19-27 50 26 5 19 21 81 Uniform item difficulty distributions 28-36 0 24 4 17 20 83 37-45 25 29 7 24 22 76 46-54 50 27 2 7 25 93
In experiments involving no guessing with normally distributed conditions,
approximately 60% of misfitting items were detected by the unweighted total fit statistic.
Of the 17 items detected in the normally distributed conditions, 94% (16 of 17) produced
negative LRI values. The logit item difficulties for these items ranged from -4.37 to 3.14.
The greatest magnitude of change (negative) in LRI values was observed when the misfit
values of the unweighted total statistic was > 2.69 and the associated item difficulties
were either very easy or very difficult. The negative LRI values for these items indicated
a negative trend in the residuals.
In the uniformly distributed conditions with no guessing, 58% (14 of 24) of the
misfitting items were detected by the unweighted total fit statistic. Of the total detected,
![Page 66: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/66.jpg)
57
83% (20 of 24) had negative LRI values. The logit item difficulties for these items
ranged from -1.32 to .99. The greatest magnitude of change in LRI values (negative) was
observed when the item difficulties of the misfitting items detected by the unweighted
total fit statistic were either very easy or very difficult. The negative LRI values
indicated a negative trend in the residuals. That is, persons in the low-ability groups
performed better than expected on these items.
As the levels of guessing increased (25% and 50%), the between and unweighted
total fit statistics detected the greatest number of misfitting items, respectively. At the
25% levels of guessing in the normally distributed conditions, 74% (17 of 23) of
misfitting items were detected by the between fit statistic and 48% (11 of 23) by the
unweighted total fit statistic. Of the items detected, 78% (18 of 23) produced negative
LRI values. The logit item difficulties associated with these items ranged from -4.16 to
5.11. Again, the greatest change (negative) in LRI values was observed when the
misfitting items detected by the unweighted total fit statistic was > 2.69 and associated
with either easy or difficult items.
In the uniformly distributed conditions at the 25% level, 55% (16 of 29) of the
misfitting items were detected by the between fit statistic, and 31% (9 of 29) were
detected by the unweighted total fit statistic. The logit item difficulties associated with
these items ranged from -1.55 to 1.80. Of the misfitting items detected, 76% (22 of 29)
produced negative LRI values. Again, the greatest change (negative) in LRI values was
![Page 67: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/67.jpg)
58
when the unweighted total fit statistic was > 2.69 and the misfitting items were either
very easy or very difficult.
At the 50% level in the normally distributed conditions, more than half (58%) of
the misfitting items were detected by the between fit statistic, 38% by the unweighted
total fit statistic, and 31% by the weighted total fit statistic. Of the total number of
misfitting items detected in the normally distributed conditions (26), 81% (21 of 26) had
negative LRI values. The logit item difficulties for these items ranged from -3.29 to 4.74.
As in previous experiments, the greatest change in LRI values was in the negative
direction when associated with large positive unweighted total misfit values and very
easy or difficult items.
In uniformly distributed conditions at the 50% level of guessing, 56% of the
misfitting items were detected by the between fit statistic, 48% by the unweighted total,
and 30% by the weighted total fit statistic. Of the total misfitting items detected in the
uniformly distributed conditions, 93% (25 of 29) had negative LRI values. The logit item
difficulties associated with these items ranged from -1.77 to 1.21. Large negative
changes in the LRI resulted from large unweighted total misfit values associated with
very easy or very difficult items.
![Page 68: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/68.jpg)
CHAPTER 5
FINDINGS AND CONCLUSIONS
Effects of Guessing on Rasch Fit Statistics
Independent t-tests were used to determine whether item distribution types and
levels of guessing affected the Rasch fit statistics. Item and person parameters were not
interpreted as cause and effect factors, but were used to establish the experimental
conditions. This was justified based upon previous research that has shown that these
parameters have minimal to no effect on Rasch fit statistics.
In experimental conditions involving normally distributed item difficulty
distributions, the mean unweighted and weighted total fit statistics were robust (no
significant differences found) to varying levels of guessing at the .05 level of significance
within normally distributed conditions. As the level of guessing increased (25% and 50%
levels), the mean between fit statistics showed significant differences between mean fit
values observed in conditions involving no guessing and conditions involving a 25% and
a 50% chance of guessing correctly. These findings indicate that the mean unweighted
and weighted total fit statistics were robust to the effects of guessing in normally
distributed item distributions. The significant differences observed with the between fit
statistic indicate that, as the level of guessing increased, the statistic detected a bias in
item function (item familiarity) between ability groups. That is, low-ability groups
![Page 69: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/69.jpg)
60
tended consistently to guess the correct answer as the probability of guessing increased.
An inspection of the data indicated that low-ability persons performed better than
expected on the majority of misfitting items detected in these experimental conditions.
In experiments involving uniformly distributed conditions, no significant
differences were noted between the mean unweighted and weighted total fit statistic at the
.05 level when varying the levels of guessing. This finding indicates that the unweighted
and weighted total fit statistics are robust to the effects of guessing in uniformly
distribution item distributions. The mean between fit statistic showed a significant
difference at the .05 level between experimental conditions involving no guessing and
those involving a 50% chance of guessing correctly. The difference observed with the
between fit statistic is believed to be related to a detection of item bias that favored
low-ability groups, a condition that was also observed in the normally distributed
experimental conditions.
To determine whether distribution types had an effect on Rasch fit statistics,
comparisons were made between mean fit statistics obtained at the same level of guessing
in each distribution type (normal and uniform). A significant difference was observed at
the .05 level between the mean unweighted total fit statistic in normal and uniform
experimental conditions involving no guessing. Significant differences were also found
at the .05 level between mean weighted fit values for experimental conditions involving
no guessing by distribution type and conditions involving a 25% chance of guessing
correctly by distribution type. No significant differences were found between mean fit
![Page 70: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/70.jpg)
61
values associated with the between fit statistic in normal and uniform experimental
conditions.
These results indicate that, within distribution types, the unweighted and weighted
total fit statistics were robust (no significant differences) to varying levels of guessing,
and the between fit statistic tended to detect item bias as the level of guessing increased.
Thus, the between fit statistic was sensitive to item bias as the probability of guessing
correctly increased within distribution types. When comparisons were made between
distribution types, significant differences were found between the mean fit values
associated with the weighted and unweighted total fit values, but no differences were
observed in the between fit mean values.
Simulation Design Effects
The differences observed with the weighted and unweighted total fit values may
have been influenced by the parameters used when simulating the data. The normally
distributed conditions had an S.D. of 2, whereas the data simulated for the uniformly
distributed conditions had an S.D. of 1. This restriction significantly reduced the range of
item difficulties in the uniformly distributed data to > -2.0 to < 2.0, while the item
difficulties in the normally distributed conditions ranges from < -4.0 to > 5.0. Another
factor to be considered is the mean person ability used in the simulation. A mean person
ability of+1 was used in the simulation of all experimental conditions (normal and
uniform). Therefore, the item difficulties in the uniformly distributed conditions
centered on the average person ability. Thus, the difference observed with the weighted
![Page 71: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/71.jpg)
62
fit statistic may have been influenced by its sensitivity to fit problems centered on the
item difficulty, which was, in the uniformly distributed conditions, centered on the
average persons' ability.
The differences observed with the between fit statistic within distribution types as
the level of guessing increased may have been influenced by differences in item difficulty
ranges between the two distribution types, and the number of ability groups used in the
analysis. The between fit statistic is sensitive to group membership (number of ability
groups). The greater the number of ability groups (maximum allowed by IP ARM = 5),
the less likely low- and medium-ability persons will be forced into the same group,
therefore reducing the influence of group membership on the between fit statistic. This is
especially important for persons using IP ARM for data analysis; the software attempts to
place an equal number of persons in each ability group.
Multiple choice tests constructed with five options per item and normally
distributed item difficulties will substantially reduce the probability of low-ability
persons consistently guessing the correct answer. In addition, the total test variance will
increase, thereby increasing the test's ability to differentiate between ability groups.
Detection of Guessing by Rasch Fit Statistics
In order to determine whether the frequency of misfitting items detected by Rasch
fit statistics differed by distribution type as a result of varying the levels of guessing, an
analysis of observed and expected misfitting frequencies was conducted using chi-square
at the .05 level of significance. Assuming no significant difference, the frequency of
![Page 72: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/72.jpg)
63
misfitting items obtained in experimental conditions involving no guessing in normal and
uniform conditions was used as the baseline or theoretical frequency (expected) against
which those observed at the 25% and 50% levels were compared. By inducing no
measurement noise, the results were purely a matter of chance and were assumed to
closely approximate what could be expected on the single administration of four-option
per item multiple choice tests of various lengths administered to relatively high ability
groups of various sizes. Comparisons were made within and across distribution types.
In normally distributed experiments, significant differences were observed at the
.05 level between the theoretical and observed frequencies of misfitting items detected by
the between fit statistic. Significant differences were observed on a test simulated with
50 items and a 25% chance of guessing correctly, and a test involving a 50% chance of
guessing the correct answer. No significant differences were found between the observed
and expected frequencies detected by the unweighted and weighted total fit statistics.
In tests with uniformly distributed item difficulties, significant differences were
observed at the .05 level between expected and observed frequencies detected by the
between fit statistic and the weighted total fit statistics. On a test simulated with 25 items
and a 50% chance of guessing correctly, a significant difference was found at the .05
level. A significant difference was also found at the .05 level between observed and
expected frequencies detected by the weighted total fit statistic on a test with 100 items
and a 25% chance of guessing correctly. No significant differences were observed
between the expected and observed frequency of misfitting items detected by the
![Page 73: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/73.jpg)
64
unweighted total fit statistic. Indications are that the unweighted between fit statistic was
more sensitive at detecting random guessing, because no significant differences were
observed between the expected and observed frequencies. However, as the probability of
guessing correctly increased, the between fit statistic was able to detect bias in the
function of the items. In order to determine whether distribution types had an effect on
the frequency of misfitting items detected by Rasch fit statistics, a chi-square analysis of
the frequency of misfitting items was conducted at the .05 level of significance.
Assuming no significance (null hypothesis), comparisons were made across distribution
types. The frequency of misfitting items detected in experiments involving normally
distributed conditions was used as the expected frequencies, and those detected in
experiments involving uniformly distributed conditions were used as the observed
frequencies. No significant differences were found between the frequency of misfitting
items detected by Rasch fit statistics in normal and uniformly distributed experimental
conditions. Thus, the unweighted total, weighted total, and between fit statistics were
robust to changes in item and person parameters, levels of guessing, and distribution
types (normal and uniform).
Guessing and the Logit Residual Index
Misfitting item data were investigated to determine the effects of guessing on the
Logit Residual Index (LRI). In experiments involving no guessing with normally
distributed conditions, approximately 60% of misfitting items were detected by the
unweighted total fit statistic compared to 12% by the weighted total and 41% by the
![Page 74: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/74.jpg)
65
between fit statistic. In experiments involving uniformly distributed conditions, again,
about 60% of the misfitting items were detected by the unweighted total fit statistic
compared to 29% by the weighted total and 38% by the between fit statistic. Thus, the
unweighted total fit statistic was more sensitive to measurement disturbances of a random
nature than were the weighted total or between fit statistics. The sum of the percentages
may not total 100% due to one or more fit statistics identifying the same items as
misfitting. Likewise, the sum of the items detected by each fit statistic will not add up to
the total number of misfitting items at each level of guessing due to one or more fit
statistics detecting the same items.
Of the 17 items detected in the normally distributed conditions, 94% (16 of 17)
produced negative LRI values. The logit item difficulties for these items ranged from -
4.37 to 3.14. The greatest magnitude of change (negative) in LRI values was observed
when item misfit values for the unweighted total statistic were > 2.69 and the associated
item difficulty values indicated very easy or very difficult items. In the uniformly
distributed conditions with no guessing, 58% (14 of 24) of the misfitting items were
detected by the unweighted total fit statistic. Of the total detected (24), 83% (20 of 24)
had negative LRI values. The logit item difficulties for these items ranged from -1.32 to
.99. The greatest magnitude of change observed in the LRI values was when the
misfitting item difficulties were either very difficult or very easy, producing negative LRI
values. The negative LRI values indicated a negative trend in the residuals in both
![Page 75: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/75.jpg)
66
distribution types. That is, persons in the low-ability groups performed better than
expected, but not well enough to cause a bias in the item function.
As the levels of guessing increased (25% and 50%), the between and unweighted
total fit statistics detected the greatest number of misfitting items, respectively. At the
25% level of guessing in normally distributed conditions, 74% (17 of 23) of misfitting
items were detected by the between fit statistic and 48% (11 of 23) by the unweighted
total fit statistic. Of these items, 78% (18 of 23) produced negative LRI values. The logit
item difficulties associated with these items ranged from -4.16 to 5.11. Again, the
greatest change (negative) in LRI values occurred when the unweighted total fit statistic
was > 2.69. When the between fit statistic was > 2.69, a negative LRI value was
observed, but the magnitude of change was not as great as the change observed with
unweighted total fit statistic. In the uniformly distributed conditions at the 25% level of
guessing, 55% (16 of 29) of the misfitting items were detected by the between fit statistic
and 31% (9 of 29) were detected by the weighted and unweighted total fit statistics,
respectively . The logit item difficulties associated with these items ranged from -1.55 to
1.80. Of the misfitting items detected, 76% (22 of 29) produced negative LRI values.
Again, the greatest change (negative) in LRI values occurred when the unweighted total
fit statistic was >2.69.
At the 50% level of guessing, more than half of the misfitting items were detected
by the between fit statistic in the normal and uniform experimental conditions. In the
normally distributed conditions, the unweighted total fit statistic detected 38% of the
![Page 76: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/76.jpg)
67
misfitting items, and the weighted total detected 31%. Of the total number of misfitting
items detected in the normally distributed conditions, 81% (21 of 26) had negative LRI
values. The logit item difficulties for these items ranged from -3.29 to 4.74. As in
previous experiments, the greatest change in LRI values was observed when the
misfitting items detected by the unweighted total fit statistic were either very easy or very
difficult.
In the uniform conditions, 48% of the misfitting items were detected by the
unweighted total fit statistic, 30% by the weighted total, and 56% by the between fit
statistic. Of the total number of misfitting items detected, 93% (25 of 29) had negative
LRI values. The logit item difficulties associated with these items ranged from -1.77 to
1.21. The LRI showed large negative changes when misfitting items detected by the
unweighted total fit statistic were either very easy or very difficult. The magnitude of
change was greatest when the misfit values of the unweighted total fit statistic increased
above 2.69 than with similar values associated with the weighted and between fit
statistics. As the level of guessing increased, the percentage of misfitting items
producing negative LRI values also increased. However, in all situations the LRI was
able to identify group membership for the misfit problem. This is of particular interest
for persons using IP ARM for data analysis, because the software allows the user to
identify group membership based on demographic characteristics.
![Page 77: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/77.jpg)
68
Summary
No significant differences were noted between mean fit values for the unweighted
and weighted total fit statistics within distribution types. Thus, these statistics were robust
to varying levels of guessing within distribution types (normal and uniform). The
differences observed across distribution types reflect the sensitivity of these statistics to the
magnitude of the differences between the ranges of item logit difficulties in each
distribution type as related to fit problems far away (unweighted total) and centered on
(weighted total) the logit difficulty of the items relative to the ability of the groups in which
the fit problems occurred. In the normally distributed conditions, the item difficulties
ranged from about -5 to +5 logits and from about -2 to +2 logits in the uniformly distributed
conditions.
The between fit statistics showed significant differences between mean fit values
within distributions types, but no significant differences were found across distribution
types. The differences observed within distribution types were due to low-ability persons
consistently guessing the correct answer as the probability of guessing the correct answer
increased. Consequently, the between fit statistic detected bias (item familiarity) in the item
functions. Because the experimental conditions (items, persons, and levels of guessing)
were the same in both distribution types, the detection of biased items by the between fit
statistic was relatively the same in both distribution types, resulting in no significant
differences across distribution types. Therefore, Rasch fit statistics appear to be robust to
varying levels of guessing within and across distribution types.
![Page 78: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/78.jpg)
69
Of the misfitting items detected, the majority produced negative LRI values,
indicating a negative trend in the residuals. That is, low-ability persons performed better
than expected. The LRI was more sensitive to large positive unweighted total misfit
values than to the same or similar values observed for the weighted total or between fit
statistics. This was especially evident when the misfitting items were very easy or very
difficult. In all situations involving fit problems, the LRI was able to identify group
membership of persons in which the fit problems occurred. Since IP ARM allows the user
to identify group membership based on demographic characteristics, the implications for
the use of this statistic are phenomenal. Therefore, in the assessment of individual
differences, it is necessary to use Rasch fit statistics (unweighted total, weighted total,
and unweighted between fit statistics) in conjunction with the LRI in order to make more
accurate decisions about the classification, selection, and placement of the most able
persons.
Conclusions
The results of this study indicate that Rasch fit statistics yield valid and reliable
results on the single administration of tests of varying lengths and levels of guessing
within and across distributions types.
1. The Rasch fit statistics were robust to varying levels of guessing within and
across distribution types (uniform and normal), making the Rasch model an ideal
measurement tool for the assessment of nonnormal populations.
![Page 79: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/79.jpg)
70
2. The unweighted total fit statistic was more sensitive to measurement
disturbances (random guessing)in items far away from the average ability of the groups in
which the misfit problem occurred.
3. The weighted total fit statistic was more sensitive to fit problems centered on
the logit difficulty of the misfitting items. Stated differently, the statistic was sensitive to
fit problems in persons with a logit ability centered on the logit difficulty of the misfitting
items.
4. As the probability of guessing correctly increased, low-ability persons tended
consistently to guess the correct answer, thereby inducing systematic item familiarity bias
in the items that was detected by the between fit statistic.
5. The Logit Residual Index was more sensitive to large positive misfit values
associated with the unweighted total statistic than to similar values associated with the
weighted total or between fit statistics.
6. The Logit Residual Index simplifies the interpretation and identification of
group membership in which misfit problems occurred.
Further Study
The Rasch measurement model should be applied to known published results of
tests constructed under other measurement models to determine whether the results
obtained by Rasch analysis differ significantly from the published results. In addition,
the information included in Appendix C of this study will provide an excellent data
source for establishing acceptable limits for the point biserial correlation as a fit index.
![Page 80: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/80.jpg)
71
Recommendations
The Rasch measurement model has been found to be sufficient, efficient,
unbiased, and consistent when applied to the measurement of individual differences.
Given these characteristics, it is strongly recommended that these statistics be applied to
the analysis of test data obtained under other measurement models. This will reduce the
one-size-fits-all approach to measurement and insure that decisions made about the
selection, placement, and classification of individuals are based on "true scores" (test
results) that are unbiased, efficient, consistent, and efficient.
![Page 81: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/81.jpg)
APPENDIX A
IP ARM DATA CONTROL FILE PARAMETERS
![Page 82: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/82.jpg)
73
IP ARM was used to analyze each of the 54 data sets used in the 54 experimental
conditions. To perform an analysis, you must construct a control file. All information
entered into the control program will be used as a scratch file. The scratch file is then
used by all subprograms to perform the selected analyses. In order to use the control
information in subsequent analyses, a permanent control file must be created. To create a
permanent control file, the following information must be entered and saved on a
computer diskette or the hard disk of your computer:
Test information
1. Test name (60 characters maximum).
2. Model selection (dichotomous or rating scale).
3. Analysis Type (item fit, person fit, both item and person analysis).
4. Test information (number of items, number of people, starting location of items in record).
Item information 1. Item difficulty information (output files from MSCALE, BIGSCALE,
BIGSTEPS, or hand entered).
2. Item name information (eight characters maximum).
Item fit information
1. Item fit information (omit misfitting persons; Y or N).
2. Enter maximum person t value for inclusion (99 for all positive misfit persons and -99 for all negative misfit persons). By default, item between fit information is provided.
Item between fit information
1. How many additional between fit statistic do you want? (item analysis always provide between fit statistics based on ability groups as a default. A maximum of four additional between fit statistics can be requested).
(appendix continues'!
![Page 83: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/83.jpg)
74
Appendix A (continued)
If additional between fit statistics were requested, you will be prompted to enter
the title of the report, number of subgroups (maximum of 5), and the location of the
variable in the data vector. You will then be asked to enter a value for each group, the
title for each group, and if you want a listing of the group information and whether you
want to make changes to any item fit values.
Person fit information
1. Enter the number of person between fit statistics waited (5 maximum). 2. Enter the title for each group.
3. Enter the number of items in each group.
4. Identify group membership for each item.
5. Do you want a listing of the information (Y or N)?
6. Do you want to add an analysis (Y or N)?
7. Do you want to modify any item assignment.
Scoring information
1. Enter the name of the response file (list disk drive and file name)
2. Enter drive name for scratch files (temporary files used by program during analysis).
3. Do you want a listing of the logit ability scale (Y or N)?
4. Enter choice for person plot (Score resid. =1, Std. resid. =2).
5. Enter valid responses.
6. Enter the omit character.
7. Should omits be scored as incorrect (Y or N)?
8. Do the responses need to be scored (Y or N)?
9. Do you want to make changes to the score information values (Y or N).
(appendix continues)
![Page 84: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/84.jpg)
75
Appendix A (continued)
Saving control information
All information entered into the program up to this point is used to create the
control file. It is used as a scratch file by the subprogram to perform all analysis. In
order to use this information in future analyses, you must create a permanent file. You
will be prompted to answer the following questions:
1. Do you want to save all the control information (Y or No)?
2. Enter the drive specification and the name of the file to be created.
Depending on the number of item between fit analyses you request, IP ARM will
produce one to two pages of item information. However, the greater the number of items
and groups used in the analysis, the greater the amount of output.
![Page 85: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/85.jpg)
APPENDIX B
A SUMMARY OF MISFITTING ITEM STATISTICS BY EXPERIMENT
![Page 86: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/86.jpg)
77
A Summary of Misfitting Item Statistics by Experiment
Item #
Logit item diff.
Point, bis. corr.
Unwt. Wt. total total fit fit
Ability between
fit
Mean item score
Logit Residual
Index
17 0.88 Experiment 1
17 0.88 0.30 2.02 0.98 -1.00 0.48 -1.49 22 3.14 0.05 1.07 0.46 2.44 0.12 -0.16 24 3.14 0.05 2.21 0.23 0.43 0.12 -1.17
Experiment 2 (none)
69 1.49 Experiment 3
69 1.49 0.23 0.75 1.76 2.05 0.40 -0.21 87 2.17 0.13 2.83 1.02 1.28 0.28 -1.96 92 3.05 0.07 3.13 0.59 -0.40 0.16 -2.06 94 2.17 0.18 2.69 0.91 -0.28 0.28 -1.89
18 0.88 Experiment 4
18 0.88 0.28 2.41 1.81 -1.12 0.48 -1.13
15 -1.47 Experiment 5
15 -1.47 0.26 0.71 0.58 2.07 0.88 -0.03 Experiment 6
11 -3.68 -0.23 2.75 0.46 2.34 0.98 -1.00 62 0.31 0.24 1.48 1.77 2.11 0.62 -0.40 74 1.02 0.30 0.68 1.62 2.39 0.48 -0.14
Experiment 7(none)
Experiment 8 36 1.04 0.32 3.13 0.88 -0.31 0.53 -1.53 41 1.72 0.60 -2.23 -2.41 1.48 0.39 0.48 44 2.04 0.20 2.27 2.10 0.68 0.33 -0.59
Experiment 9 3 -4.37 -0.09 1.98 0.38 2.70 0.99 -0.26
36 -0.80 0.36 0.94 0.32 2.14 0.80 -0.10 Experiment 10
10 -0.42 0.15 3.11 1.16 0.57 0.72 -2.27 23 2.59 0.04 1.55 1.40 2.01 0.20 -0.46
22 -1.15 Experiment 11
22 -1.15 -0.04 1.40 1.07 2.26 0.80 -0.36 24 -0.65 0.19 0.55 0.87 2.12 0.72 -0.07
18 -1.72 Experiment 12
18 -1.72 -0.28 2.29 0.61 1.43 0.88 -0.68 51 0.03 0.15 0.88 1.19 2.11 0.60 -0.37
(appendix continues!
![Page 87: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/87.jpg)
78
Appendix C (continued)
Item #
Logit item diff.
Point, bis. corr.
Unwt. Wt. total total fit fit
Ability between
fit
Mean item score
Logit Residual
Index
-3.33 Experiment 13
2 -3.33 -0.03 1.21 0.41 2.80 0.98 -0.03 19 1.12 0.24 2.01 2.43 2.56 0.52 -0.70 25 5.11 0.00 2.41 0.33 1.18 0.04 -0.77
37 0.94 Experiment 14
37 0.94 0.65 -2.02 -2.58 2.29 0.48 0.67
10 -2.83 Experiment 15
10 -2.83 -0.01 0.92 0.50 2.89 0.96 -0.01 32 -1.31 0.18 0.88 0.54 2.24 0.86 -0.09 98 4.13 0.09 1.39 0.20 2.16 0.06 -0.16
Experiment 16 (none)
-4.16 Experiment 17
3 -4.16 0.01 0.86 0.35 2.02 0.99 0.00 17 -1.08 0.22 0.52 0.44 2.28 0.84 -0.02 35 0.73 0.57 -1.62 -2.44 2.33 0.53 0.39 36 1.33 0.26 2.51 2.22 2.72 0.41 -0.79
14 -2.32 Experiment 18
14 -2.32 0.22 2.05 -0.11 -1.52 0.94 -0.45 51 0.15 0.27 2.47 1.33 1.36 0.64 -0.67 66 0.41 0.23 2.64 2.20 1.28 0.59 -0.76 72 1.36 0.38 0.40 0.84 2.09 0.40 -0.06 75 1.36 0.60 -2.03 -2.01 2.01 0.40 0.43 84 1.46 0.66 -2.72 -2.96 2.60 0.38 0.50
23 1.54 Experiment 19
23 1.54 -0.07 2.87 2.40 1.42 0.32 -1.79
28 0.27 Experiment 20
28 0.27 0.06 1.67 2.27 2.86 0.64 -0.64
40 -0.66 Experiment 21
40 -0.66 -0.22 1.77 1.53 2.01 0.76 -0.54 48 -1.22 -0.08 1.22 0.75 2.07 0.84 -0.22 49 -0.21 0.43 -0.57 -0.17 2.15 0.68 0.22 54 0.74 -0.03 2.00 2.08 0.28 0.48 -1.30 72 -0.01 0.72 -1.83 -1.87 2.47 0.64 0.72 92 2.49 -0.12 1.26 0.70 2.23 0.16 -0.25
21 1.63 Experiment 22
21 1.63 0.22 1.80 2.06 2.14 0.34 -0.52
(appendix continues)
![Page 88: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/88.jpg)
79
Appendix B (continued)
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
9 -2.61 Experiment 23
9 -2.61 0.13 3.96 -0.12 0.66 0.94 -2.81 31 0.19 0.36 0.35 1.10 2.03 0.60 -0.01 32 0.60 0.22 1.74 2.35 1.50 0.52 -0.64 33 1.01 0.61 -1.20 -1.69 2.37 0.44 0.40 46 2.16 0.28 0.89 0.80 2.21 0.24 -0.10
25 -1.35 Experiment 24
25 -1.35 0.42 2.09 -0.75 -1.21 0.86 -0.72 34 -1.35 0.31 2.24 -0.11 -1.21 0.86 -0.74 39 -1.00 0.12 2.08 0.96 2.15 0.82 -0.42 45 -0.06 0.30 2.01 1.05 -0.51 0.68 -0.73
-2.70 Experiment 25
3 -2.70 0.10 1.30 0.23 2.02 0.96 -0.12 5 -1.50 0.25 2.69 0.07 -0.64 0.89 -0.74 7 -1.62 0.19 1.18 0.35 2.13 0.90 -0.10 19 1.60 0.31 1.25 2.07 1.06 0.40 -0.26
-3.29 Experiment 26
3 -3.29 -0.01 0.94 0.32 3.13 0.98 -0.01 34 0.89 0.58 -2.13 -2.28 1.10 0.54 0.52 40 1.53 0.59 -2.12 -2.15 0.77 0.41 0.48 49 4.74 -0.07 1.57 0.47 2.81 0.04 -0.09
Experiment 27(none)
Experiment 28 (none)
-0.74 Experiment 29
2 -0.74 -0.01 2.63 0.60 -1.38 0.84 -1.33
40 -0.45 Experiment 30
40 -0.45 -0.17 1.82 1.99 2.24 0.72 -0.71 58 -0.23 0.05 2.73 0.86 -1.13 0.68 -1.65 66 0.74 -0.00 2.51 2.16 0.93 0.48 -1.77 92 -0.03 -0.15 2.09 2.58 2.14 0.64 -1.06 98 0.93 0.03 2.20 1.95 1.40 0.44 -1.38
19 0.25 Experiment 31
19 0.25 0.07 1.38 2.15 1.64 0.68 -0.38
30 -0.19 Experiment 32
30 -0.19 0.12 1.09 2.23 2.59 0.72 -0.23 37 0.16 0.63 -1.34 -2.05 1.98 0.66 0.41 45 0.99 0.25 1.21 2.02 1.77 0.50 -0.42
(appendix continues)
![Page 89: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/89.jpg)
80
Appendix B (continued*)
Item #
Logit item diff.
Point, bis. corr.
Unwt. Wt. total total fit fit
Ability between
fit
Mean item score
Logit Residual
Index
23 -0.73 Experiment 33
23 -0.73 0.17 0.75 1.07 2.93 0.80 -0.05 57 0.36 0.24 2.09 1.34 1.30 0.62 -0.77 61 0.46 0.60 -1.51 -1.39 2.30 0.60 0.45 73 0.56 0.33 0.68 1.04 2.06 0.58 -0.14 84 0.86 0.22 2.00 1.91 0.95 0.52 -0.79
8 -0.06 Experiment 34
8 -0.06 0.23 2.44 1.29 1.16 0.73 -0.58 17 0.00 0.20 1.26 2.06 2.66 0.72 -0.17 21 0.97 0.61 -2.07 -2.57 1.79 0.54 0.49
8 -0.09 Experiment 35
8 -0.09 0.34 2.57 0.76 0.37 0.69 -0.88 11 -1.32 0.12 2.74 0.89 3.22 0.86 -0.53 36 -0.20 0.54 -1.42 -1.18 2.10 0.71 0.23
11 -1.15 Experiment 36
11 -1.15 0.26 2.22 0.53 0.66 0.85 -0.41 67 0.34 0.29 3.21 1.50 0.99 0.63 -1.06 76 0.23 0.39 2.20 0.43 -0.04 0.65 -0.77
21 1.41 Experiment 37
21 1.41 0.30 0.98 2.19 2.36 0.48 -0.28
-0.77 Experiment 38
5 -0.77 -0.01 2.42 0.68 -1.05 0.84 -1.12 9 -0.19 0.04 1.28 1.24 2.50 0.76 -0.34
31 0.49 -0.11 2.54 2.60 2.03 0.64 -1.33
13 -0.67 Experiment 39
13 -0.67 -0.07 1.72 0.80 2.16 0.84 -0.38 36 -1.55 -0.17 1.44 0.65 2.44 0.92 -0.19 84 0.80 0.70 -1.79 -1.72 2.07 0.60 0.75 89 1.75 0.71 -1.86 -2.47 1.00 0.40 0.80 90 0.80 0.73 -1.97 -1.95 2.07 0.60 0.82
10 -0.62 Experiment 40
10 -0.62 0.31 2.12 -0.12 -1.04 0.82 -0.62
28 0.46 Experiment 41
28 0.46 0.18 0.89 1.88 2.46 0.58 -0.33 42 0.65 0.54 -1.10 -2.00 0.45 0.54 0.53
(appendix continues^
![Page 90: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/90.jpg)
Appendix B (continued^)
81
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
38 Experiment 42
38 0.20 0.34 3.99 0.26 -0.47 0.70 -2.71 42 -0.47 0.15 1.48 1.31 2.53 0.80 -0.31 47 0.08 0.21 2.16 1.22 1.05 0.72 -0.75 56 0.55 0.21 1.62 1.96 2.09 0.64 -0.53 64 -0.05 0.26 0.75 1.20 2.28 0.74 -0.08 88 1.28 0.63 -1.64 -2.20 1.40 0.50 0.57
23 1.22 Experiment 43
23 1.22 0.34 1.38 1.92 2.08 0.51 -0.33 24 1.80 0.33 3.16 1.32 -0.11 0.40 -1.10
42 0.93 Experimemt 44
42 0.93 0.22 2.36 2.23 1.11 0.50 -0.73
32 -0.24 Experiment 45
32 -0.24 0.25 0.71 1.61 2.08 0.72 -0.07 33 -0.24 0.29 0.54 1.21 2.08 0.72 -0.04 49 0.25 0.29 1.54 1.66 2.21 0.63 -0.34 76 0.76 0.26 1.62 2.49 1.92 0.53 -0.36 86 0.91 0.30 2.09 1.49 -0.08 0.50 -0.64 87 0.81 0.57 -1.58 -2.02 0.04 0.52 0.38 90 0.86 0.29 2.13 1.70 0.26 0.51 -0.67 99 0.96 0.26 1.73 2.40 2.40 0.49 -0.42
Experiment 46 (none)
-0.50 Experiment 47
7 -0.50 0.14 2.35 0.79 1.00 0.80 -1.04 10 -1.77 -0.20 1.80 0.90 3.41 0.92 -0.31 23 -0.50 0.20 0.87 0.93 2.44 0.80 -0.13 28 0.96 -0.01 2.75 2.71 1.49 0.56 -1.74 39 0.53 0.77 -1.83 -2.13 1.86 0.64 0.61
33 -0.21 Experiment 48
33 -0.21 -0.20 2.32 1.51 0.83 0.76 -0.95 36 -0.78 -0.07 1.28 0.85 2.11 0.84 -0.26 67 1.21 -0.06 2.35 2.47 2.40 0.48 -1.52 76 0.65 -0.09 2.30 2.36 1.99 0.60 -1.40 96 0.45 -0.03 2.20 1.66 1.50 0.64 -1.14
23 0.57 Experiment 49
23 0.57 0.25 2.39 1.83 0.69 0.58 -0.98
(appendix continues^
![Page 91: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/91.jpg)
82
Appendix B (continued)
Item #
Logit item diff.
Point, bis. corr.
Unwt. Wt. total total fit fit
Ability between
fit
Mean item score
Logit Residual
Index
14 -0.26 Experiment 50
14 -0.26 0.10 1.39 2.04 3.16 0.74 -0.31 17 -0.26 0.15 0.86 1.96 2.39 0.74 -0.13 28 -0.02 0.32 2.11 0.48 -0.95 0.70 -1.15 38 0.54 0.64 -1.37 -2.37 1.76 0.60 0.45 45 0.54 0.26 0.93 2.06 2.39 0.60 -0.27
55 0.30 Experiment 51
55 0.30 0.20 1.41 1.72 2.14 0.68 -0.39 57 -0.06 0.18 1.41 1.38 2.35 0.74 -0.29
-0.61 Experiment 52
3 -0.61 0.17 2.19 1.04 2.02 0.82 -0.36 6 -0.32 0.31 0.95 0.42 2.37 0.78 -0.10
11 0.19 0.23 1.41 1.58 2.78 0.70 -0.26 23 0.67 0.46 -0.09 -0.53 2.24 0.61 -0.04
35 0.42 Experimemt 53
35 0.42 0.14 3.23 2.68 3.82 0.61 -0.99
16 -0.63 Experiment 54
16 -0.63 0.20 2.16 0.56 0.06 0.83 -0.36 45 0.34 0.26 2.46 0.88 -1.42 0.68 -0.73 86 0.76 0.26 1.05 1.74 2.78 0.60 -0.20 89 0.76 0.36 2.10 -0.01 -0.19 0.60 -0.75
![Page 92: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/92.jpg)
APPENDIX C
SUMMARY OF ITEM FIT INFORMATION FOR EXPERIMENT 1-54
![Page 93: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/93.jpg)
84
Summary of Item Fit Information for Experiment 1-54
Experiment 1: Summary of Item Fit Information for a Normally Distributed Item Difficulty Distribution With 25 Items. 25 Persons, and No Guessing
Logit Point Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index 1 -3.31 0.36 0.11 0.27 -0.54 0.96 0.04 2 -3.31 0.07 0.88 0.57 -0.54 0.96 0.03 3 -4.09 -9.99 0.20 -0.92 0.07 1.00 -0.00 4 -2.46 -0.01 1.09 1.02 1.18 0.92 -0.03 5 -2.46 0.20 1.55 0.25 1.18 0.92 -0.30 6 -2.46 0.64 -0.62 -0.70 -0.16 0.92 0.06 7 -1.90 0.72 -0.93 -1.02 0.18 0.88 0.10 8 -0.77 0.68 -0.62 -0.87 -0.50 0.76 0.20 9 -0.22 0.72 -1.21 -1.17 0.69 0.68 0.39
10 -0.77 0.61 -0.64 -0.23 1.18 0.76 0.16 11 -1.46 0.65 -0.38 -0.75 -0.50 0.84 0.11 12 0.02 0.54 -0.24 -0.05 -0.15 0.64 0.06 13 0.02 0.31 1.36 1.08 0.94 0.64 -0.63 14 0.02 0.48 0.25 0.26 -0.55 0.64 -0.01 15 -0.22 0.38 0.83 0.78 0.26 0.68 -0.24 16 0.25 0.72 -1.50 -1.48 0.55 0.60 0.57 17 0.88 0.30 2.02 0.98 -1.00 0.48 -1.49 18 1.09 0.35 0.74 0.78 -0.44 0.44 -0.32 19 1.51 0.51 -0.54 -0.66 0.19 0.36 0.28 20 1.51 0.51 -0.54 -0.66 0.19 0.36 0.28 21 2.20 0.43 -0.30 -0.36 -1.46 0.24 0.15 22 3.14 0.05 1.07 0.46 2.44 0.12 -0.16 23 2.46 0.11 0.57 1.04 1.22 0.20 -0.02 24 3.14 0.05 2.21 0.23 0.43 0.12 -1.17 25 3.14 0.38 -0.25 -0.19 0.09 0.12 0.09 Mean -0.16 0.41 0.20 -0.05 0.20 0.61 S.D. 2.14 0.25 0.99 0.78 0.85 Groups 2 Note: Raw score mean = 15.24 with a S.D. of 4.12. Mean person ability = 0.70 with a S.D. of 1.28. Test reliability (K.R. 20) = 0.80. Reliability of person separation = -0.80.
![Page 94: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/94.jpg)
85
Appendix C (continued)
Experiment 2: Summary of Item Fit Information for a Normally Distributed Item Difficulty Distribution With 50 Items. 25 Persons, anH No Guessing
Ability Mean Logit between item Residual
fit score Index
Logit Point Unwt. Wt. Item item bis. total total
# diff. corr. fit fit 1 -3.54 -9.99 0.15 -0.96 2 -2.78 0.30 0.17 0.28 3 -2.78 0.39 -0.00 0.08 4 -3.54 -9.99 0.15 -0.96 5 -1.97 0.50 -0.40 -0.24 6 -3.54 -9.99 0.15 -0.96 7 -1.97 0.40 -0.10 0.04 8 -1.97 0.47 -0.31 -0.13 9 -1.45 0.49 -0.40 -0.12
10 -2.78 0.30 0.17 0.28 11 -2.78 0.14 0.57 0.45 12 -2.78 0.39 -0.00 0.08 13 -2.78 0.16 0.51 0.43 14 -0.40 0.38 0.49 0.48 15 -2.78 0.07 0.80 0.48 16 -0.70 0.45 -0.09 0.07 17 -1.04 0.53 -0.53 -0.23 18 -0.40 0.41 0.02 0.48 19 -1.04 0.13 1.28 1.03 20 -1.04 0.46 0.14 -0.22 21 0.12 0.69 -1.18 -1.21 22 0.36 0.67 -1.08 -1.03 23 -1.04 0.42 0.07 0.07 24 0.12 0.57 -0.53 -0.31 25 -0.40 0.75 -1.35 -1.71 26 0.12 0.77 -1.65 -1.93 27 0.36 0.39 0.49 0.79 28 0.81 0.58 -0.56 -0.38 29 -0.70 0.56 -0.21 -0.59 30 -0.13 0.10 1.72 1.84 31 0.12 0.47 0.22 0.17 32 -0.40 0.39 0.05 0.61 33 0.59 0.31 0.93 1.31 34 0.36 0.54 0.67 -0.45 35 0.81 0.34 0.96 1.09 36 1.02 0.45 0.31 0.39 37 0.81 0.20 1.86 1.87 38 0.59 0.42 0.49 0.63 39 1.89 0.34 0.25 1.10
0.05 -0.49 -0.49 0.05
-0.11 0.05
-0.11 -0.11 0.21
-0.49 -0.49 -0.49 -0.49 -0.60 -0.49 0.81 0.51 0.17
-0.54 -0.54 0.53 1.03
-0.54 -1.40 1.11 1.77
-0.41 -0.47 -1.52 1.05
-1.40 -0.60 0.49 1.03 1.08
-1.48 1.08 0.49
-0.15
1.00 0.96 0.96 1.00 0.92 1.00 0.92 0.92 0.88 0.96 0.96 0.96 0.96 0.76 0.96 0.80 0.84 0.76 0.84 0.84 0.68 0.64 0.84 0.68 0.76 0.68 0.64 0.56 0.80 0.72 0.68 0.76 0.60 0.64 0.56 0.52 0.56 0.60 0.36
-0.00 0.04 0.03
-0.00 0.06
-0.00 0.06 0.06 0.09 0.04 0.04 0.03 0.04
-0.08 0.03 0.09 0.12 0.05
-0.27 0.04 0.38 0.41 0.05 0.22 0.29 0.46
-0.11 0.27 0.10
-0.63 -0.04 0.05
-0.42 -0.31 -0.36 -0.07 -0.91 -0.10 -0.06
(appendix continues"!
![Page 95: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/95.jpg)
Appendix C (continued)
86
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index 40 2.36 0.53 -0.50 -0.51 -1.31 0.28 0.21 41 2.12 0.38 0.46 0.43 0.15 0.32 -0.07 42 2.12 0.19 0.83 1.70 1.59 0.32 -0.22 43 1.89 0.69 -1.21 -1.74 -0.15 0.36 0.45 44 2.61 0.38 0.72 -0.08 -0.91 0.24 -0.20 45 2.36 0.50 -0.04 -0.56 0.64 0.28 0.10 46 2.61 0.45 0.02 -0.31 -0.91 0.24 0.07 47 3.60 0.13 0.96 0.51 0.83 0.12 -0.13 48 3.21 0.46 -0.40 -0.32 0.14 0.16 0.12 49 3.21 0.47 -0.37 -0.40 0.14 0.16 0.12 50 5.61 -9.99 0.15 -1.03 0.06 0.00 -0.00 Mean S.D. Groups
-0.10 2.12
0.41 0.22
0.10 0.71
0.01 0.86
-0.03 0.79 2
0.66
Note: Raw score mean = 32.96 with a S.D Mean person ability =1.11 with a S.D. of Test reliability (K.R. 20) = 0.90. Reliability of person separation = 0.89
. of 7.92. 1.27.
![Page 96: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/96.jpg)
87
Appendix C (continued^
Experiment 3: Summary of Item Fit Information for a Normally Distributed Item Difficulty Distribution With 100 Items. 25 Persons, and No Guessing
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index 1 -3.68 -9.99 0.20 -0.92 0.06 1.00 -0.00 2 -3.68 -9.99 0.20 -0.92 0.06 1.00 -0.00 3 -3.68 -9.99 0.20 -0.92 0.06 1.00 -0.00 4 -2.92 0.46 -0.05 -0.02 -0.47 0.96 0.03 5 -2.08 0.32 0.47 0.11 1.02 0.92 0.02 6 -3.68 -9.99 0.20 -0.92 0.06 1.00 -0.00 7 -3.68 -9.99 0.20 -0.92 0.06 1.00 -0.00 8 -2.92 0.42 0.01 0.10 -0.47 0.96 0.03 9 -3.68 -9.99 0.20 -0.92 0.06 1.00 -0.00
10 -3.68 -9.99 0.20 -0.92 0.06 1.00 -0.00 11 -2.92 0.42 0.01 0.10 -0.47 0.96 0.03 12 -2.92 0.15 0.57 0.50 -0.47 0.96 0.04 13 -2.08 0.35 0.31 0.09 -0.08 0.92 0.04 14 -2.92 0.14 0.61 0.51 -0.47 0.96 0.04 15 -2.08 0.63 -0.61 -0.78 -0.08 0.92 0.07 16 -1.54 0.60 -0.52 -0.71 0.25 0.88 0.10 17 -2.92 0.42 0.01 0.10 -0.47 0.96 0.03 18 -2.92 0.05 0.88 0.55 -0.47 0.96 0.02 19 -2.92 0.46 -0.05 -0.02 -0.47 0.96 0.03 20 -2.92 0.46 -0.05 -0.02 -0.47 0.96 0.03 21 -1.13 0.36 0.19 0.34 -0.63 0.84 0.05 22 -1.54 0.56 -0.29 -0.59 0.25 0.88 0.09 23 -2.08 0.39 0.09 0.03 -0.08 0.92 0.05 24 -2.92 0.14 0.61 0.51 -0.47 0.96 0.04 25 -3.68 -9.99 0.20 -0.92 0.06 1.00 -0.00 26 -1.54 0.31 0.60 0.29 0.16 0.88 -0.04 27 -2.08 0.45 -0.21 0.01 -0.08 0.92 0.06 28 -0.79 0.29 0.22 0.83 0.86 0.80 0.03 29 -1.13 0.28 0.06 0.86 0.56 0.84 0.05 30 -1.54 0.03 0.99 1.11 0.16 0.88 -0.09 31 -0.23 0.60 -0.73 -0.73 0.06 0.72 0.27 32 -2.92 -0.02 1.11 0.57 2.35 0.96 -0.02 33 -2.08 0.45 -0.21 0.01 -0.08 0.92 0.06 34 -2.92 0.21 0.42 0.46 -0.47 0.96 0.04 35 -1.54 0.36 0.30 0.21 0.16 0.88 0.02 36 -1.13 0.40 0.04 0.22 -0.63 0.84 0.05 37 -1.54 0.31 0.45 0.28 0.16 0.88 0.01 38 -0.49 0.39 -0.01 0.46 -0.51 0.76 0.07 39 -1.13 0.34 0.70 0.25 -0.63 0.84 -0.11
(appendix continues^
![Page 97: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/97.jpg)
Appendix C (continued'!
88
Item #
40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82
Logit Point Unwt. Wt. Ability Mean item bis. total total between item diff. corr. fit fit fit score
-1.54 0.24 1.03 0.34 0.16 0.88 -1.54 0.54 -0.19 -0.55 0.25 0.88 -0.79 0.62 -0.54 -0.92 -1.34 0.80 -1.54 0.61 -0.56 -0.74 0.25 0.88 -0.23 0.63 -0.81 -1.04 0.06 0.72 0.02 0.44 1.46 -0.30 -1.31 0.68
-0.49 0.22 0.60 1.14 1.56 0.76 -0.49 0.15 1.13 1.22 1.56 0.76 -0.23 0.70 -1.22 -1.43 1.46 0.72 0.02 0.42 0.07 0.35 0.38 0.68
-0.23 0.66 -0.98 -1.17 0.06 0.72 -0.49 0.52 -0.22 -0.34 0.08 0.76 -0.49 0.30 0.47 0.73 0.08 0.76 0.02 0.53 -0.44 -0.34 0.55 0.68 0.46 0.23 1.27 1.49 0.50 0.60
-0.49 0.44 0.08 0.08 -0.51 0.76 0.46 0.43 -0.08 0.40 -1.21 0.60
-0.79 0.55 -0.32 -0.48 -1.34 0.80 0.88 0.52 -0.57 -0.22 0.51 0.52 0.67 0.60 -0.90 -1.05 -0.62 0.56 0.46 0.37 0.44 0.74 0.50 0.60 0.24 0.63 -1.03 -1.17 1.01 0.64 0.67 0.38 0.42 0.71 1.11 0.56 0.46 0.17 1.51 1.93 1.66 0.60
-0.23 0.41 0.00 0.42 0.06 0.72 0.67 0.65 -1.27 -1.50 0.81 0.56 0.46 0.24 1.73 1.37 0.50 0.60 0.24 0.53 -0.29 -0.46 -0.41 0.64
-0.79 0.34 -0.04 0.73 0.86 0.80 1.49 0.23 0.75 1.76 2.05 0.40 0.02 0.38 0.38 0.50 0.38 0.68 1.08 0.34 0.54 1.07 1.07 0.48 1.08 0.57 -0.84 -0.80 -0.58 0.48 0.46 0.66 -1.24 -1.49 1.45 0.60 1.29 0.44 -0.07 0.24 0.41 0.44 2.17 0.49 -0.46 -0.16 1.24 0.28 0.88 0.67 -1.44 -1.82 1.36 0.52 1.08 0.44 -0.15 0.35 -0.23 0.48 1.29 0.55 -0.67 -0.58 -1.38 0.44 0.88 0.38 0.42 0.69 -1.13 0.52 1.93 0.49 -0.32 -0.20 0.20 0.32 1.93 0.50 0.08 -0.54 -0.93 0.32 0.46 0.44 0.04 0.27 -1.21 0.60
Logit Residual
Index
-0.16 0.08 0.16 0.10 0.28
-0.79 -0.07 -0.28 0.35 0.10 0.31 0.14
-0.07 0.15
-0.59 0.08 0.16 0.14 0.36 0.46
-0.05 0.41
-0.08 -0.75 0.06 0.57
-0.94 0.20 0.04
-0.21 -0.09 -0.15 0.46 0.51 0.15 0.21 0.64 0.20 0.39
-0.10 0.14 0.04 0.10
(appendix continues')
![Page 98: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/98.jpg)
Appendix C (continued^
89
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index 83 1.93 0.52 -0.39 -0.50 0.20 0.32 0.22 84 2.43 0.22 0.48 1.05 -0.88 0.24 -0.08 85 3.05 0.40 -0.06 0.06 -0.40 0.16 0.08 86 1.71 0.61 -0.74 -1.27 -1.03 0.36 0.35 87 2.17 0.13 2.83 1.02 1.28 0.28 -1.96 88 2.72 0.13 0.47 1.35 -1.19 0.20 -0.03 89 2.17 0.59 -0.17 -1.22 -0.28 0.28 0.10 90 1.71 0.54 -0.44 -0.61 -1.03 0.36 0.26 91 2.72 0.47 -0.35 -0.15 0.70 0.20 0.13 92 3.05 0.07 3.13 0.59 -0.40 0.16 -2.06 93 2.17 0.30 0.58 0.64 -0.22 0.28 -0.12 94 2.17 0.18 2.69 0.91 -0.28 0.28 -1.89 95 2.72 0.51 -0.39 -0.40 0.70 0.20 0.15 96 3.95 0.29 0.32 0.09 -0.15 0.08 0.03 97 2.43 0.29 0.67 0.56 -0.88 0.24 -0.15 98 4.76 0.39 0.02 0.03 -0.51 0.04 0.03 99 4.76 0.27 0.23 0.27 -0.51 0.04 0.04 100 4.76 0.07 0.74 0.45 -0.51 0.04 0.03
Mean S.D. Groups
0.29 2.13
0.39 0.21
0.15 0.78
0.00 0.78
0.03 0.78 2
0.67
Note: Raw score mean = 66.52 with a Mean person ability = 0.97 with a S.D Test reliability (K.R. 20) = 0.94. Reliability of person separation = 0.94
S.D. of 14.91. .of 1.24.
![Page 99: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/99.jpg)
90
Appendix C (continued^
Experiment 4: Summary of Item Fit Information for a Normally Distributed Item Difficulty Distribution With 25 Items. 50 Persons, and No Guessing
Logit Point Item item bis.
# diff. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index 0.45 -1.06 0.04 1.00 -0.00 0.45 -1.06 0.04 1.00 -0.00 0.98 0.31 0.33 0.92 -0.11 0.78 0.54 0.87 0.94 -0.01
-0.05 0.13 -0.34 0.96 0.04 -0.49 -0.39 0.26 0.90 0.08 -0.51 -0.38 0.90 0.82 0.12 0.33 0.53 -0.63 0.88 0.02 1.23 0.69 1.83 0.84 -0.21
-0.12 -0.35 -1.20 0.84 0.07 0.48 -0.28 0.47 0.84 -0.05 0.03 0.24 1.35 0.64 0.09 0.42 0.60 -0.89 0.70 -0.05
-0.03 0.13 0.09 0.68 0.08 0.83 1.05 -1.34 0.52 -0.26
-0.05 -0.46 0.83 0.52 0.02 -0.84 -0.73 -0.63 0.54 0.33 2.41 1.81 -1.12 0.48 -1.13
-1.10 -1.06 0.70 0.44 0.38 -0.33 0.02 -0.85 0.30 0.11 0.54 0.77 -1.10 0.40 -0.11
-0.81 -0.63 -0.63 0.34 0.26 -1.53 -1.90 1.15 0.26 0.29 0.26 0.25 -0.21 0.16 0.00
-0.27 -0.13 0.45 0.10 0.06 0.12 -0.05 0.01 0.64
1 2 3 4 5 6 7 8 9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
-4.57 -4.57 -2.28 -2.62 -3.09 -2.00 -1.18 -1.76 -1.35 -1.35 -1.35 0.02
-0.33 -0.21 0.67 0.67 0.57 0.88 1.10 1.90 1.31 1.65 2.16 2.96 3.64
Mean S.D.
Groups
-0.37 2.16
-9.99 -9.99 0.21 0.11 0.30 0.45 0.48 0.27 0.21 0.44 0.40 0.45 0.38 0.45 0.39 0.52 0.56 0.28 0.59 0.49 0.42 0.56 0.68 0.41 0.45 0.41 0.19 0.82 0.80 0.88
2 Note: Raw score mean = 16.02 with a Mean person ability = 0.81 with a S.D Test reliability (K.R. 20) = 0.79. Reliability of person separation = 0.81
S.D. of 4.05. .of 1.35.
![Page 100: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/100.jpg)
91
Appendix C (continued)
Experiment 5: Summary of Item Fit Information for a Normally Distributed Ttem Difficulty Distribution With 50 Items. 50 Persons, and No Guessing
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index 1 -3.72 0.44 -0.23 -0.15 -0.59 0.98 0.01 2 -4.48 -9.99 0.15 -0.94 0.07 1.00 -0.00 3 -3.72 0.17 0.40 0.47 -0.59 0.98 0.03 4 -3.72 0.24 0.19 0.41 -0.59 0.98 0.03 5 -2.03 0.39 -0.23 0.17 0.27 0.92 0.06 6 -3.72 0.10 0.65 0.50 -0.59 0.98 0.03 7 -2.91 0.29 0.31 0.24 -0.25 0.96 0.02 8 -2.41 0.47 -0.64 0.02 0.03 0.94 0.05 9 -1.47 0.43 -0.22 0.13 -1.56 0.88 0.06
10 -1.47 0.58 -0.91 -0.66 0.70 0.88 0.10 11 -1.25 0.46 -0.50 0.06 0.90 0.86 0.06 12 -2.03 0.31 0.66 0.26 -0.12 0.92 -0.04 13 -0.55 0.40 0.19 0.39 -0.72 0.78 -0.02 14 -2.41 0.42 -0.28 -0.01 0.03 0.94 0.05 15 -1.47 0.26 0.71 0.58 2.07 0.88 -0.03 16 -1.25 0.48 -0.61 0.02 0.90 0.86 0.09 17 -1.47 0.58 -0.98 -0.61 0.70 0.88 0.10 18 -1.25 0.46 0.44 -0.34 0.19 0.86 -0.05 19 -1.05 0.41 0.59 -0.06 1.12 0.84 -0.04 20 -0.55 0.41 -0.11 0.44 0.64 0.78 -0.03 21 -1.25 0.33 0.59 0.56 -0.80 0.86 -0.09 22 0.32 0.45 -0.07 0.08 -0.12 0.64 0.07 23 -0.28 0.34 1.14 0.49 1.17 0.74 -0.20 24 -0.28 0.41 0.18 0.40 -1.57 0.74 0.01 25 -0.55 0.48 -0.07 -0.24 -0.24 0.78 0.06 26 -0.02 0.49 -0.59 -0.06 -0.91 0.70 0.19 27 -0.41 0.39 0.26 0.51 0.56 0.76 0.01 28 0.21 0.48 -0.52 -0.06 -0.62 0.66 0.19 29 -0.02 0.47 0.86 -0.33 -0.29 0.70 -0.27 30 0.21 0.49 -0.38 -0.20 -0.62 0.66 0.15 31 0.84 0.45 0.09 -0.06 0.14 0.54 0.01 32 0.53 0.57 -0.63 -1.23 0.70 0.60 0.20 33 0.74 0.47 -0.26 -0.10 -0.98 0.56 0.17 34 1.04 0.51 -0.65 -0.69 0.17 0.50 0.29 35 1.04 0.37 0.38 0.87 -0.36 0.50 -0.06 36 1.04 0.31 0.82 1.38 0.62 0.50 -0.24 37 0.94 0.35 0.64 1.09 1.06 0.52 -0.18 38 1.24 0.46 -0.03 -0.32 0.28 0.46 0.04 39 1.34 0.55 -1.02 -1.20 -0.15 0.44 0.41
(appendix continues"!
![Page 101: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/101.jpg)
Appendix C (continued)
92
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index 40 1.45 0.32 0.97 0.94 0.42 0.42 -0.34 41 1.55 0.49 -0.70 -0.60 -0.17 0.40 0.31 42 2.20 0.19 0.99 1.33 0.05 0.28 -0.22 43 2.32 0.47 -0.55 -0.70 -0.18 0.26 0.17 44 1.97 0.41 0.11 -0.20 -0.85 0.32 0.01 45 2.45 0.43 -0.18 -0.53 -0.56 0.24 0.09 46 2.87 0.28 0.24 0.29 0.76 0.18 0.02 47 3.21 0.15 0.57 0.68 0.19 0.14 -0.02 48 5.42 0.15 0.39 0.31 -0.48 0.02 0.03 49 3.63 0.24 -0.00 0.25 0.51 0.10 0.06 50 4.68 0.15 0.26 0.28 -0.14 0.04 0.04
Mean S.D. Groups
-0.09 2.20
0.39 0.14
0.05 0.55
0.08 0.58
0.00 0.72 2
0.65
Note: Raw score mean = 32.36 with a S.D. of 7.55. Mean person ability = 1.00 with a S.D. of 1.20. Test reliability (K.R. 20) = 0.89. Reliability of person separation = 0.88
![Page 102: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/102.jpg)
93
Appendix C (continued)
Experiment 6: Summary of Item Fit Information for a Normally Distributed Ttem Difficulty Distribution With 100 Items. 50 Persons, and No Guessing
Logit Point Unwt. Item item bis. total
# diff. corr. fit
1 -4.41 -9.99 -0.02 2 -4.41 -9.99 -0.02 3 -4.41 -9.99 -0.02 4 -3.68 0.06 0.73 5 -3.68 0.35 -0.20 6 -2.08 0.22 0.74 7 -3.68 0.35 -0.20 8 -1.55 0.50 -0.35 9 -2.91 0.39 -0.41
10 -2.91 0.19 0.15 11 -3.68 -0.23 2.75 12 -1.79 0.19 0.36 13 -2.43 0.30 -0.10 14 -1.79 0.56 -0.88 15 -1.79 0.26 0.25 16 -1.55 0.28 0.13 17 -2.43 0.19 0.66 18 -2.91 0.46 -0.65 19 -1.34 0.34 0.16 20 -1.79 0.42 -0.17 21 -1.79 0.40 0.11 22 -1.55 0.48 -0.76 23 -1.55 0.50 -0.71 24 -2.43 0.26 0.22 25 -1.14 0.51 -0.63 26 -1.55 0.41 -0.43 27 -1.55 0.61 -1.15 28 -1.55 0.41 -0.43 29 -0.81 0.40 -0.18 30 -1.79 0.19 0.64 31 -1.34 0.36 0.09 32 -0.81 0.28 0.79 33 -1.14 0.36 -0.20 34 -0.26 0.25 0.98 35 -0.66 0.23 1.22 36 -1.55 0.33 0.58 37 -1.14 0.54 -0.62 38 -1.55 0.30 -0.01 39 -0.14 0.33 0.42
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
-1.04 0.05 1.00 -0.00 -1.04 0.05 1.00 -0.00 -1.04 0.05 1.00 -0.00 0.43 -0.51 0.98 0.02 0.12 -0.51 0.98 0.02 0.31 1.64 0.92 -0.03 0.12 -0.51 0.98 0.02
-0.43 -1.28 0.88 0.05 -0.02 -0.17 0.96 0.04 0.41 -0.17 0.96 0.04 0.46 2.34 0.98 -1.00 0.74 -0.93 0.90 0.02 0.26 0.10 0.94 0.04
-0.81 0.55 0.90 0.08 0.46 -0.93 0.90 0.03 0.56 -1.28 0.88 0.04 0.46 0.34 0.94 -0.03
-0.20 -0.17 0.96 0.03 0.30 -0.67 0.86 0.04
-0.21 -0.93 0.90 0.05 -0.23 -0.93 0.90 0.02 -0.13 0.75 0.88 0.08 -0.41 0.75 0.88 0.09 0.22 0.10 0.94 0.03
-0.51 1.14 0.84 0.11 0.15 0.75 0.88 0.07
-1.00 0.75 0.88 0.11 0.06 0.75 0.88 0.08 0.35 0.41 0.80 0.06 0.59 -0.93 0.90 -0.01 0.20 -0.67 0.86 0.04 0.77 1.29 0.80 -0.10 0.46 -0.25 0.84 0.05 1.35 0.69 0.72 -0.23 1.11 -0.35 0.78 -0.28 0.21 0.56 0.88 -0.06
-0.75 -0.44 0.84 0.10 0.50 0.75 0.88 0.05 1.02 1.23 0.70 -0.08
(appendix continues)
![Page 103: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/103.jpg)
94
Appendix C (continued)
Logit Point Unwt. Wt. Item item bis. total total
# diff. corr. fit fit
Ability Mean between item
fit score
Logit Residual
Index 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82
-0.66 -0.14 -0.26 -0.52 -0.52 0.20
-0.52 -0.39 -0.81 0.42 0.09
-0.02 -0.39 -0.02 -0.66 -0.66 -0.02 -0.26 0.31
-0.02 0.42 0.52 0.31 0.52 0.20 0.72 0.52 1.02 0.72 0.62 0.42 0.20 0.92 1.22 1.02 1.02 1.53 1.32 1.96 1.22 1.32 1.32 1.32
0.37 0.34 0.41 -0.35 0.78 -0.03 0.44 -0.22 0.20 -0.20 0.70 0.03 0.63 -1.34 -1.43 0.53 0.72 0.29 0.46 -0.29 -0.11 -0.90 0.76 0.11 0.45 -0.23 0.00 -0.18 0.76 0.05 0.51 -0.10 -0.65 -0.01 0.64 0.07 0.44 -0.37 0.20 0.96 0.76 0.11 0.34 0.58 0.67 0.03 0.74 -0.11 0.48 -0.39 -0.20 -1.20 0.80 0.08 0.58 -1.33 -1.14 -0.21 0.60 0.43 0.32 0.94 1.11 -0.71 0.66 -0.26 0.47 -0.43 -0.02 -1.14 0.68 0.15 0.43 -0.20 0.25 -1.35 0.74 0.08 0.36 0.85 0.61 -1.14 0.68 -0.22 0.44 -0.22 -0.01 -0.35 0.78 0.06 0.55 -0.73 -0.68 -0.61 -0.78 0.15 0.51 -0.78 -0.33 1.14 0.68 0.20 0.42 0.80 -0.09 -0.45 0.72 -0.20 0.37 0.83 0.61 0.54 -0.62 -0.19 0.47 0.30 -0.33 -1.14 0.68 -0.06 0.42 0.29 0.32 -1.22 0.60 -0.07 0.31 1.38 1.23 -0.43 0.58 -0.49 0.24 1.48 1.77 2.11 0.62 -0.40 0.35 0.78 0.94 0.59 0.58 -0.21 0.53 -0.85 -0.60 -1.51 0.64 0.28 0.36 0.61 0.88 -0.40 0.54 -0.16 0.48 -0.22 -0.33 -0.91 0.58 0.12 0.36 0.94 0.65 -1.22 0.48 -0.39 0.52 -0.88 -0.64 0.21 0.54 0.35 0.50 -0.61 -0.34 -1.10 0.56 0.25 0.54 -1.01 -0.72 0.09 0.60 0.36 0.46 -0.12 -0.05 0.91 0.64 -0.03 0.49 -0.65 -0.32 -0.92 0.50 0.29 0.51 -0.63 -0.86 -0.09 0.44 0.21 0.30 0.68 1.62 2.39 0.48 -0.14 0.41 0.08 0.42 -1.22 0.48 0.03 0.47 -0.59 -0.34 -0.28 0.38 0.23 0.54 -1.09 -1.11 0.49 0.42 0.39 0.44 -0.20 -0.41 -0.75 0.30 0.01 0.47 -0.02 -0.54 0.84 0.44 0.03 0.25 2.49 1.36 0.37 0.42 -1.26 0.40 -0.08 0.45 -0.74 0.42 0.08 0.38 0.05 0.62 1.22 0.42 0.06
(appendix continues'!
![Page 104: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/104.jpg)
Appendix C (continued)
95
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index 83 2.08 0.31 0.05 0.81 0.17 0.28 0.04 84 2.20 0.46 -0.53 -0.55 -0.16 0.26 0.15 85 1.74 0.53 -0.95 -1.08 0.06 0.34 0.29 86 1.96 0.35 -0.06 0.56 -0.41 0.30 0.07 87 2.46 0.34 -0.14 0.29 -0.99 0.22 0.08 88 2.20 0.39 -0.21 0.01 -0.16 0.26 0.09 89 1.63 0.56 -1.12 -1.40 0.40 0.36 0.35 90 2.60 0.36 -0.10 -0.01 0.36 0.20 0.07 91 2.33 0.39 -0.28 -0.00 -0.53 0.24 0.10 92 3.09 0.25 0.24 0.25 0.17 0.14 0.02 93 2.46 0.28 0.79 0.31 1.13 0.22 -0.13 94 3.09 0.36 -0.25 -0.13 -0.81 0.14 0.07 95 3.29 0.48 -0.82 -0.68 0.70 0.12 0.10 96 3.09 0.37 -0.39 -0.15 0.87 0.14 0.08 97 2.46 0.34 -0.15 0.27 -0.99 0.22 0.08 98 4.12 0.26 0.38 -0.04 0.29 0.06 -0.01 99 4.56 0.03 0.71 0.37 1.01 0.04 0.01
100 5.30 0.07 0.59 0.35 -0.48 0.02 0.03 Mean S.D. Groups
-0.13 1.49
0.38 0.15
0.00 0.72
0.06 0.66
0.04 0.85 2
0.62
Note: Raw score mean = 64.40 with a S.D. of 15, Mean person ability = 0.89 with a S.D. of 1.14. Test reliability (K.R. 20) = 0.94. Reliability of person separation = 0.94
30.
![Page 105: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/105.jpg)
96
Appendix C (continued")
Experiment 7: Summary of Item Fit Information for a Normally Distributed Ttem Difficulty Distribution With 25 Items. 100 Persons, and No G u e s s i n g
Logit Point Item item bis.
# diff. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index 1 2 3 4 5 6 7 8 9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
-5.21 -3.31 -2.98 -1.29 -1.85 -2.14 -1.49 -1.01 -1.72 -0.29 -0.49 -0.42 -0.70 -0.04 0.25 0.58 0.90 0.79 1.33 1.22 1.38 1.90 2.64 2.88 3.85
-9.99 0.21 0.12 0.47 0.43 0.26 0.30 0.44 0.40 0.47 0.42 0.34 0.49 0.51 0.51 0.42 0.51 0.45 0.44 0.53 0.47 0.45 0.38 0.49 0.46
0.67 0.07 0.67
-0.59 -0.09 0.13 0.50
-0.16 0.05
-0.56 -0.13 0.90
-0.72 0.33 0.21 0.88 0.59 0.62 0.11
-0.51 -0.04 -0.03 0.68
-0.45 -0.45
-1.11 0.35 0.52
-0.70 -0.81 0.49 0.51
-0.40 -0.58 -0.02 0.32 1.04
-0.49 -0.99 -0.74 0.78
-0.70 0.55 1.00
-0.64 0.34 0.56 0.75
-0.25 -0.45
0.05 -0.14 0.35 0.43
-0.67 -1.60 0.40 1.07
-0.38 0.26
-0.96 0.69 0.70 1.09 0.18
-0.43 0.89
-1.32 1.01
-0.45 -0.75 0.47 1.06
-0.01 -0.89
1.00 0.97 0.96 0.85 0.90 0.92 0.87 0.82 0.89 0.72 0.75 0.74 0.78 0.68 0.63 0.57 0.51 0.53 0.43 0.45 0.42 0.33 0.22 0.19 0.10
-0.00 0.04 0.02 0.10 0.04 0.04
-0.02 0.07 0.03 0.14 0.08
-0.15 0.14
-0.10 -0.06 -0.31 -0.28 -0.14 0.05 0.15 0.07 0.06
-0.06 0.08 0.06
Mean -0.21 0.42 0.11 S.D. 2.08 0.14 0.49 Groups
-0.03 0.67
0.04 0.78 2
0.65
Note: Raw score mean = 16.23 with a Mean person ability = 0.98 with a S.D Test reliability (K.R. 20) = 0.81. Reliability of person separation = 0.83
S.D. of 4.25. .of 1.39.
![Page 106: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/106.jpg)
97
Appendix C (continued)
Experiment 8: Summary of Item Fit Information for a Normally Distributed Item Difficulty Distribution With 50 Items. 100 Persons, and No Guessing
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index 1 -3.86 0.18 -0.01 0.29 -0.31 0.99 0.02 2 -4.56 -9.99 -0.13 -1.21 0.04 1.00 -0.00 3 -3.14 -0.01 0.87 0.31 0.53 0.98 -0.01 4 -3.86 0.06 0.52 0.35 -0.31 0.99 0.02 5 -2.71 0.15 0.26 0.18 -0.24 0.97 0.01 6 -3.14 0.10 0.26 0.26 0.06 0.98 0.02 7 -3.86 0.26 -0.27 0.18 -0.31 0.99 0.01 8 -1.96 0.18 -0.02 0.27 -0.66 0.94 0.03 9 -1.78 0.20 -0.05 0.27 -0.27 0.93 0.04
10 -2.40 0.11 0.30 0.29 -0.95 0.96 0.02 11 -1.49 0.22 -0.18 0.33 0.30 0.91 0.05 12 -0.85 0.33 -0.10 -0.21 0.20 0.85 0.04 13 -1.24 0.38 -0.64 -0.49 -0.69 0.89 0.07 14 -1.49 0.15 0.70 0.37 0.14 0.91 -0.04 15 -2.16 0.15 0.92 0.20 -1.24 0.95 -0.10 16 -0.85 0.28 0.16 0.09 -1.01 0.85 0.02 17 -0.85 0.28 0.06 0.11 -0.35 0.85 0.02 18 -0.94 0.25 0.29 0.16 -0.56 0.86 0.10 19 -0.60 0.42 -0.41 -0.75 1.23 0.82 0.05 20 -0.53 0.43 -0.79 -0.81 -0.18 0.81 0.11 21 -0.53 0.38 -0.51 -0.35 0.69 0.81 0.09 22 -0.68 0.37 -0.62 -0.28 0.22 0.83 0.09 23 -0.13 0.42 -0.42 -0.70 1.24 0.75 0.08 24 -0.53 0.35 -0.28 -0.11 -1.59 0.81 0.06 25 0.11 0.37 -0.29 0.10 -0.01 0.71 0.10 26 0.05 0.40 -0.20 -0.40 -1.17 0.72 0.04 27 0.05 0.41 -0.42 -0.43 -0.14 0.72 0.11 28 0.05 0.46 -0.73 -0.91 1.24 0.72 0.14 29 0.22 0.43 -0.56 -0.56 0.68 0.69 0.14 30 0.38 0.43 0.14 -0.71 0.14 0.66 -0.05 31 0.44 0.29 0.71 1.19 1.48 0.65 -0.14 32 0.84 0.33 0.89 0.86 0.42 0.57 -0.27 33 0.59 0.30 0.78 1.12 -0.81 0.62 -0.20 34 0.89 0.44 -0.43 -0.58 0.51 0.56 0.13 35 0.74 0.37 0.26 0.50 0.41 0.59 -0.02 36 1.04 0.32 3.13 0.88 -0.31 0.53 -1.53 37 0.94 0.35 0.26 0.94 0.41 0.55 -0.01 38 1.47 0.40 0.05 0.38 -0.27 0.44 0.03 39 1.57 0.27 1.34 2.00 1.50 0.42 -0.31
(appendix continues)
![Page 107: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/107.jpg)
Appendix C (continued)
98
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index 40 1.78 0.34 1.18 0.77 0.73 0.38 -0.27 41 1.72 0.60 -2.23 -2.41 1.48 0.39 0.48 42 1.72 0.49 -0.85 -1.00 0.27 0.39 0.23 43 2.33 0.41 0.30 -0.00 -0.31 0.28 -0.03 44 2.04 0.20 2.27 2.10 0.68 0.33 -0.59 45 2.27 0.37 0.17 0.51 0.04 0.29 -0.04 46 2.51 0.33 0.28 0.71 -0.37 0.25 -0.00 47 3.18 0.57 -1.37 -1.25 -0.18 0.16 0.12 48 2.94 0.28 1.70 0.44 0.39 0.19 -0.25 49 3.96 0.34 0.49 -0.13 1.24 0.09 -0.03 50 5.72 0.34 -0.48 -0.08 -0.26 0.02 0.02
Mean S.D. Groups
-0.09 2.15
0.31 0.13
0.13 0.87
0.06 0.79
0.08 0.73 2
0.67
Note: Raw score mean = 33.55 with a Mean person ability =1.21 with a S.D Test reliability (K.R. 20) = 0.83. Reliability of person separation = 0.85
S.D. of 6. .of 1.05.
32.
![Page 108: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/108.jpg)
99
Appendix C (continued')
Experiment 9: Summary of Item Fit Information for a Normally Distributed Ttem Difficulty Distribution With 100 Items. 100 Persons, and No Guessing
Logit Item item
# diff.
1 2 3 4 5 6 7 8 9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39
Point bis. corr.
Unwt. total fit
-3.64 -3.21 -4.37 -4.37 -3.64 -3.64 -2.63 -3.21 -2.63 -2.63 -2.63 -2.63 -1.78 -1.43 -2.07 -1.78 -1.78 -2.23 -2.07 -2.07 -0.97 -1.43 -1.33 -1.43 -2.07 -1.54 -1.54 -1.78 -0.80 -0.51 -1.23 -1.23 -1.33 -1.43 -0.88 -0.80 -0.65 -0.58 -0.65
0.24 0.19
-0.09 0.12 0.13 0.01 0.23 0.15 0.31 0.14 0.29 0.27 0.18 0.39 0.22 0.32 0.34 0.33 0.40 0.42 0.34 0.44 0.37 0.23 0.33 0.36 0.40 0.39 0.38 0.35 0.41 0.42 0.40 0.46 0.47 0.36 0.50 0.49 0.51
-0.24 -0.01 1.98 0.41 0.26 1.04
-0.00 1.07
-0.44 0.49 0.17 1.08 0.34
-0.60 0.58 0.51
-0.45 -0.39 -0.63 -0.58 0.36
-0.49 -0.42 1.36
-0.44 -0.15 -0.30 -0.82 0.82 0.91 0.33
-0.51 0.25
-0.66 -0.62 0.94
-0.99 0.32
-0.17
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index 0.16 -0.33 0.98 0.02 0.28 -0.12 0.97 0.04 0.38 2.70 0.99 -0.26 0.33 -0.62 0.99 0.02 0.30 -0.33 0.98 0.03 0.37 1.49 0.98 -0.00 0.32 0.22 0.95 0.04 0.15 0.80 0.97 -0.07 0.01 0.22 0.95 0.04 0.52 -0.16 0.95 0.02
-0.14 -0.16 0.95 0.02 -0.21 -0.16 0.95 -0.13 1.11 -0.65 0.90 0.03 0.07 1.19 0.87 0.08 0.38 1.91 0.92 0.00 0.14 -0.65 0.90 -0.05 0.20 0.86 0.90 0.06
-0.04 0.49 0.93 0.05 -0.35 0.62 0.92 0.06 -0.65 -1.50 0.92 0.05 0.67 1.09 0.82 0.00
-0.53 0.05 0.87 0.07 0.27 -1.37 0.86 0.07 1.01 -1.15 0.87 -0.20 0.09 0.62 0.92 0.05 0.03 -0.74 0.88 0.05
-0.35 -0.74 0.88 0.05 -0.04 0.86 0.90 0.07 0.27 -0.45 0.80 -0.12 0.87 1.21 0.76 -0.11
-0.20 -0.93 0.85 -0.04 -0.17 0.42 0.85 0.07 -0.31 0.11 0.86 -0.02 -0.67 -1.15 0.87 0.08 -0.49 0.11 0.81 0.10 0.32 2.14 0.80 -0.10
-0.43 1.46 0.78 0.13 -0.93 -0.02 0.77 -0.04 -1.09 -0.27 0.78 0.03
(appendix continues^
![Page 109: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/109.jpg)
100
Appendix C (continued)
Logit Point Unwt. Item item bis. total
# diff. corr. fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82
-0.88 -0.44 -0.31 -0.18 -0.44 -0.73 -0.58 -0.37 -0.58 -0.24 -0.24 0.30
-0.12 0.52
-0.18 0.24 0.30 0.24 0.63 0.63 0.46 0.52 0.35 0.06 0.63 0.79 0.52 1.00 0.84 1.11 0.95 0.95 1.38 1.44 1.27 1.00 1.66 1.66 1.66 1.55 1.38 1.66 1.71
0.36 0.54 0.39 0.47 0.44 0.50 0.45 0.48 0.45 0.46 0.47 0.42 0.45 0.47 0.37 0.50 0.50 0.48 0.57 0.52 0.38 0.40 0.47 0.42 0.35 0.54 0.52 0.38 0.53 0.51 0.57 0.46 0.51 0.55 0.51 0.49 0.49 0.55 0.46 0.56 0.52 0.45 0.45
0.17 -1.22 1.39 0.72 0.07
-0.87 -0.42 0.20
-0.31 -0.62 -0.24 0.46 0.58 0.22 0.85
-0.55 -0.45 0.46
-1.22 -0.52 1.33 0.79 0.96 0.77 1.99 0.00
-0.56 1.82
-0.32 0.16
-1.17 0.32
-0.51 -0.87 -0.32 0.74
-0.26 -1.00 -0.16 -0.87 -0.06 0.07 1.34
0.59 -0.83 0.48
-0.45 0.15
-0.68 -0.01 -0.50 -0.02 0.44
-0.01 1.08 0.22 0.42 1.31
-0.03 0.01
-0.03 -1.03 -0.17 1.48 1.42 0.19 0.71 1.95
-0.66 -0.33 1.53
-0.34 -0.18 -0.98 0.75 0.01
-0.68 -0.03 0.00 0.18
-0.59 0.69
-0.89 -0.34 0.75 0.06
-1.24 0.41
-0.42 -0.07 -0.90 -0.54 -1.31 -1.36 -1.31 -0.76 -0.76 0.03
-0.37 -1.29 -0.07 -0.16 -0.89 -0.49 0.38
-0.37 -1.11 0.96 0.47
-0.40 1.59
-1.04 -0.23 -1.15 0.51 1.14 1.42 1.14 1.05 0.74
-1.26 -1.15 -1.32 0.35 1.16
-0.65 0.35 0.52 0.09
0.81 0.75 0.73 0.71 0.75 0.79 0.77 0.74 0.77 0.72 0.72 0.63 0.70 0.59 0.71 0.64 0.63 0.64 0.57 0.57 0.60 0.59 0.62 0.67 0.57 0.54 0.59 0.50 0.53 0.48 0.51 0.51 0.43 0.42 0.45 0.50 0.38 0.38 0.38 0.40 0.43 0.38 0.37
0.02 0.18
-0.24 -0.15 -0.01 0.13 0.08
-0.04 0.08 0.14 0.06
-0.04 -0.11 -0.07 -0.14 0.13 0.11
-0.09 0.27 0.15
-0.31 -0.15 -0.30 -0.14 -0.48 -0.04 0.17
-0.47 0.09
-0.05 0.29
-0.03 0.16 0.20 0.10
-0.25 0.10 0.22 0.07 0.22 0.01 0.01
-0.33
(appendix continues)
![Page 110: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/110.jpg)
101
Appendix C (continued)
Logit Point Unwt. Wt. Ability Mean Item item bis. total total between item
# (jiff. corr. fit fit _fit score
83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99
100
2.01 2.01 2.01 1.95 1.95 2.25 2.97 2.81 2.81 3.44 3.06 2.74 3.44 3.44 3.66 4.05 4.40 5.60
Mean S.D. Groups
0.00 2.03
0.46 0.53 0.54 0.57 0.50 0.39 0.44 0.41 0.42 0.40 0.33 0.37 0.33 0.44 0.40 0.38 0.27 0.09 0.40 0.13
-0.11 -0.86 -0.33 -1.19 -0.42 0.63 0.52
-0.44 0.50
-0.26 1.89 0.17 0.12
-0.91 -0.38 -0.55 -0.05 1.02 0.07 0.75
0.28 -0.53 -0.90 -1.18 -0.29 0.81
-0.39 0.21
-0.15 -0.41 0.23 0.56 0.07
-0.38 -0.49 -0.40 0.04 0.22
-0.02 1.63 1.01 1.20
-0.25 -0.66 0.85
-0.92 -0.92 -1.10 -0.40 0.38 0.42 1.17
-0.38 -1.38 0.36 1.44
0.05 0.60
0.01 0.94 2
0.32 0.32 0.32 0.33 0.33 0.28 0.18 0.20 0.20 0.13 0.17 0.21 0.13 0.13 0.11 0.08 0.06 0.02 0.63
Logit Residual
Index
0.05 0.17 0.05 0.21 0.11
-0.10 -0.13 0.08
-0.09 0.06
-0.40 0.01 0.03 0.09 0.05 0.05 0.04
-0.04
Note: Raw score mean = 63.49 with a S.D. of 16 Mean person ability = 1.00 with a S.D. of 1.30. Test reliability (K.R. 20) = 0.95. Reliability of person separation = 0.95.
62.
![Page 111: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/111.jpg)
102
Appendix C (continued)
Bvpffrimfint 10: Summary of Item Fit Information for a Normally Distributed Item Difficulty Distribution With 25 Items. 25 Persons, and a 25% Chance of Guessing Correctly
LmS Point Unwt. Wt Ability Mean Ttem item bis. total total between item
# diff. com fit fit fit score
Logit Residual
Index
1 2 3 4 5 6 7 8 9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
-3.75 -3.00 -3.00 -2.21 -3.75 -3.00 -1.70 -0.97 -0.97 -0.42 -0.68 -0.97 -0.42 0.06
-0.68 0.49 0.28 1.13 0.71 1.13 1.57 3.32 2.59 2.92 3.85
Mean -0.30 S.D. 2.16 Groups
-9.99 0.41 0.26 0.11
-9.99 0.26 0.39 0.48 0.09 0.15 0.24 0.60 0.17 0.62 0.72 0.57 0.59 0.41 0.62 0.50 0.57 0.14 0.04 0.66 0.54 0.40 0.25
0.21 0.01 0.27 0.68 0.21 0.27 0.15
-0.28 1.62 3.11 0.44
-0.77 1.44
-0.78 -1.22 -0.60 -0.63 0.34
-0.83 -0.17 -0.48 1.68 1.55
-0.94 -0.45 0.19 1.03
-0.97 0.00 0.30 0.57
-0.97 0.30
-0.12 -0.22 1.11 1.16 1.15
-0.82 1.36
-0.90 -1.83 -0.45 -0.70 0.64
-0.96 -0.02 -0.66 0.28 1.40
-1.19 -0.52
0.05 -0.36 -0.36 0.70 0.05
-0.36 -0.18 -0.81 0.37 0.57
-0.36 1.04 0.57 0.22 1.37
-1.33 0.89
-1.09 1.10
-1.09 0.11 1.18 2.01
-0.01 -0.46
-0.08 0.89
0.15 0.83 2
Note: Raw score mean — 16.16 and a S. Mean person ability = 0.83 with a S.D. Test reliability (K.R. 20) = 0.78. Reliability of person separation = 0.78.
D. of 3.83. of 1.27.
1.00 0.96 0.96 0.92 1.00 0.96 0.88 0.80 0.80 0.72 0.76 0.80 0.72 0.64 0.76 0.56 0.60 0.44 0.52 0.44 0.36 0.12 0.20 0.16 0.08
0.65
-0.00 0.03 0.04 0.01
-0.00 0.04 0.05 0.13
-0.52 -2.27 -0.04 0.18
-0.53 0.33 0.28 0.34 0.31
-0.07 0.42 0.15 0.25
-0.47 -0.46 0.16 0.07
![Page 112: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/112.jpg)
103
Appendic C (continued)
Experiment 11: Summary of Item Fit Information for a Normally Distributed Item Difficulty Distribution With 50 Ttems. 25 Persons, and a 25% Chance of Guessing Correctly
Item #
Logit item diff.
Point bis. corr.
Unwt. Wt. Ability Mean Logit total total between item Residual fit fit fit score Index
1 2 3 4 5 6 7 8 9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39
-3.79 -3.79 -3.06 -3.06 -3.79 -3.06 -3.79 -3.06 -1.82 -1.82 -1.45 -3.06 -1.15 -0.65 -1.82 -1.15 -1.45 -1.15 -0.65 -0.65 -0.65 -1.15 -0.22 -0.65 -0.43 -0.22 -0.22 -0.43 -0.43 0.18 0.38 0.18 0.18 0.18 0.38 0.98 0.77 1.40 1.40
-9.99 -9.99 0.22 0.03
-9.99 0.22
-9.99 0.22 0.33 0.48 0.55
-0.16 0.05 0.28 0.40 0.25 0.43 0.35 0.37 0.31 0.38
-0.04 0.37 0.19 0.51 0.40 0.26 0.48 0.46 0.30 0.45 0.38 0.32 0.55 0.51 0.39 0.65 0.24 0.69
-0.21 -0.21 0.14 0.69
-0.21 0.14
-0.21 0.14
-0.19 -0.59 -0.80 1.45 0.84 0.67
-0.37 0.25
-0.42 -0.34 -0.31 -0.06 -0.19 1.40
-0.08 0.55
-0.74 -0.21 0.42
-0.45 -0.47 0.66
-0.04 0.87 0.24
-0.95 -0.68 0.36
-1.42 0.83
-1.42
-1.05 -1.05 0.29 0.39
-1.05 0.29
-1.05 0.29 0.02
-0.39 -0.72 0.43 0.94 0.26
-0.14 0.36
-0.30 0.14 0.18 0.47 0.01 1.07 0.26 0.87
-0.65 0.12 0.89
-0.57 -0.37 0.73
-0.19 0.00 0.82
-0.88 -0.53 0.26
-1.39 0.92
-1.34
0.05 0.05
-0.06 1.28 0.05
-0.06 0.05
-0.06 0.82 0.82 1.17 1.28 1.03
-0.27 0.82
-0.61 1.17 1.51
-0.27 -0.27 1.01 2.26 0.94 2.12 0.39 0.94 0.94
-1.45 0.35
-0.27 0.54
-0.27 1.89
-0.27 -1.05 -1.31 -0.54 -0.19 1.19
1.00 1.00 0.96 0.96 1.00 0.96 1.00 0.96 0.88 0.88 0.84 0.96 0.80 0.72 0.88 0.80 0.84 0.80 0.72 0.72 0.72 0.80 0.64 0.72 0.68 0.64 0.64 0.68 0.68 0.56 0.52 0.56 0.56 0.56 0.52 0.40 0.44 0.32 0.32
-0.00 -0.00 0.04 0.01
-0.00 0.04
-0.00 0.04 0.08 0.11 0.15
-0.20 -0.12 -0.20 0.10 0.10 0.12 0.13 0.15 0.10 0.14
-0.36 0.13
-0.07 0.30 0.17
-0.09 0.23 0.24
-0.29 0.06
-0.54 -0.02 0.50 0.41
-0.12 0.66
-0.27 0.46
(appendix continues)
![Page 113: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/113.jpg)
104
Appendix C (continued)
Logit Point Unwt. Item item bis. total
# diff. corr. fit
40 41 42 43 44 45 46 47 48 49 50
1.89 1.18 1.89 1.89 2.89 1.64 3.40 4.19 4.19 4.94 4.19
Mean S.D. Groups
-0.21 2.22
0.38 0.39 0.27 0.47 0.46 0.27 0.29 0.43 0.13
-9.99 0.06 0.33 0.21
0.03 0.63
Wt. total fit
0.35 0.68 0.13
-0.14 -0.46 1.11 0.31
-0.26 0.42
-0.23 0.62
0.16 0.11 0.86
-0.16 -0.04 0.66 0.14 0.07 0.40
-0.99 0.43
Ability between
fit
-1.34 0.11 0.71
-1.34 0.00
-0.87 -0.27 -0.60 -0.60 0.06
-0.60 -0.00 0.64
0.22 0.90 2
Mean item score
0.24 0.36 0.24 0.24 0.12 0.28 0.08 0.04 0.04 0.00 0.04 0.61
Logit Residual
Index
-0.08 -0.35 -0.01 0.09 0.09
-0.42 0.02 0.04 0.04
-0.00 0.02
Note: Raw score mean = 30.32 with a S.D. of 6. Mean person ability = 0.52 with a S.D. of 1.03. Test reliability (K.R. 20) = 0.84. Reliability of person separation = 0.85.
46.
![Page 114: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/114.jpg)
105
Appendix C (continued)
Experiment 12: Summary ot item n t in TYififirniltv Distribution With 100 Items,
iormauon iui <x iNuimanv lyiomuuiv/u 25 Persons, and a 25% Chance of Guessing
Correctly
Logit Item item
# diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability Mean Logit between item Residual
fit score Index
1 2 3 4 5 6 7 8 9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39
-2.95 -3.67 -3.67 -3.67 -3.67 -3.67 -3.67 -3.67 -2.19 -2.19 -3.67 -2.95 -1.36 -2.19 -1.72 -2.95 -2.95 -1.72 -3.67 -1.07 -2.95 -3.67 -1.36 -2.19 -2.19 -1.07 -1.36 -1.72 -2.19 -1.72 -1.72 -1.36 -0.36 -0.81 -1.07 -0.81 -1.72 -2.19 -1.07
-0.02 -9.99 -9.99 -9.99 -9.99 -9.99 -9.99 -9.99 -0.18 0.30
-9.99 0.20 0.43
-0.03 0.12 0.17 0.08
-0.28 -9.99 0.46 0.32
-9.99 0.15 0.13 0.10 0.43 0.40 0.15 0.09
-0.18 0.03 0.52 0.29 0.19 0.26 0.46 0.25 0.41 0.39
0.75 -0.50 -0.50 -0.50 -0.50 -0.50 -0.50 -0.50 1.69
-0.23 -0.50 0.10
-0.58 1.55 0.29 0.19 0.42 2.29
-0.50 -0.68 -0.16 -0.50 0.30 0.48 0.77
-0.41 -0.30 0.20 0.42 1.63 0.63
-0.81 0.14 0.35
-0.09 -0.44 -0.10 -0.48 -0.33
0.39 -1.06 -1.06 -1.06 -1.06 -1.06 -1.06 -1.06 0.45 0.14
-1.06 0.31
-0.16 0.29 0.38 0.33 0.36 0.61
-1.06 -0.29 0.22
-1.06 0.41 0.24 0.23
-0.31 -0.17 0.34 0.33 0.62 0.48
-0.43 0.31 0.51 0.31
-0.51 0.22
-0.02 -0.15
1.55 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.30 0.23 0.04
-0.20 0.88 0.30
-0.62 -0.20 -0.20 1.43 0.04 1.17
-0.20 0.04
-1.28 0.30 0.30
-0.46 -1.28 -0.62 0.30 1.43 1.43 0.88
-0.32 -0.70 -0.46 -0.70 0.58 0.23
-0.46
0.96 1.00 1.00 1.00 1.00 1.00 1.00 1.00 0.92 0.92 1.00 0.96 0.84 0.92 0.88 0.96 0.96 0.88 1.00 0.80 0.96 1.00 0.84 0.92 0.92 0.80 0.84 0.88 0.92 0.88 0.88 0.84 0.68 0.76 0.80 0.76 0.88 0.92 0.80
-0.01 -0.00 -0.00 -0.00 -0.00 -0.00 -0.00 -0.00 -0.38 0.07
-0.00 0.04 0.13
-0.36 0.02 0.04 0.02
-0.68 -0.00 0.17 0.04
-0.00 0.01
-0.02 -0.08 0.13 0.09 0.03 0.01
-0.36 -0.04 0.15 0.02
-0.03 0.08 0.16 0.07 0.07 0.12
(appendix continues)
![Page 115: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/115.jpg)
106
Appendix C (continued)
Logit Point Unwt. Wt. Item item bis. total total
# diff. corr. fit fit
Ability Mean between item
fit score
Logit Residual
Index
40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82
-0.81 -1.36 -0.16 -1.07 -0.16 -0.58 -0.81 -0.58 0.03
-2.19 -0.16 0.03
-0.36 0.22 0.41
-0.36 0.78 0.03 0.03 0.03 0.41 0.41
-0.36 0.41
-0.16 0.60
-0.58 1.17
-0.16 0.60 0.60 1.17 0.41
-0.16 0.60 1.59 2.39 1.59 1.37 1.17 1.37 2.39 0.60
0.52 0.07 0.15 0.17 0.41 0.57 0.50 0.37 0.40 0.11 0.04 0.15 0.11
-0.01 0.66 0.44 0.45 0.46 0.25 0.60 0.09 0.57 0.17 0.49 0.50 0.31 0.56 0.09 0.39 0.23 0.42 0.63 0.30 0.29 0.37 0.55 0.15 0.46 0.27 0.49 0.53 0.35 0.47
-0.67 0.75 1.33 0.32
-0.38 -0.98 -0.74 -0.37 -0.33 0.46 0.97 0.88 1.02 1.90
-1.87 -0.37 -0.32 -0.46 0.49
-1.32 1.58
-1.16 0.73
-0.74 -0.59 0.36
-0.94 1.20
-0.41 0.73
-0.11 -1.24 0.28 0.19 0.05
-0.81 0.37
-0.17 0.57
-0.57 -0.46 -0.00 -0.37
-0.70 0.46 0.89 0.48
-0.15 -0.95 -0.53 0.05
-0.06 0.29 1.67 1.19 1.00 1.99
-1.89 -0.41 -0.44 -0.55 0.62
-1.38 1.58
-1.25 0.78
-0.64 -0.82 0.36
-0.91 1.28
-0.04 0.92
-0.30 -1.42 0.50 0.39 0.08
-0.76 0.45
-0.49 0.32
-0.60 -0.87 -0.10 -0.67
0.08 0.70 0.48 0.03 0.17 0.54 0.08 0.54
-0.73 0.30 1.61 2.11 1.04 0.54 0.61
-0.43 -0.69 -0.73 -0.14 0.70 1.11 0.61 1.04 0.61 1.37 0.01 0.54 0.96
-1.16 0.52 0.01 0.99
-0.11 0.48
-0.99 0.16
-1.04 0.16 0.21 0.99 0.59 1.06 0.01
0.76 0.84 0.64 0.80 0.64 0.72 0.76 0.72 0.60 0.92 0.64 0.60 0.68 0.56 0.52 0.68 0.44 0.60 0.60 0.60 0.52 0.52 0.68 0.52 0.64 0.48 0.72 0.36 0.64 0.48 0.48 0.36 0.52 0.64 0.48 0.28 0.16 0.28 0.32 0.36 0.32 0.16 0.48
0.20 -0.10 -0.70 0.00 0.21 0.31 0.22 0.16 0.21
-0.01 -0.37 -0.37 -0.38 -1.11 0.99 0.18 0.15 0.27
-0.17 0.63
-0.95 0.66
-0.22 0.45 0.27
-0.19 0.31
-0.50 0.23
-0.35 0.08 0.51
-0.08 -0.01 0.02 0.27
-0.02 0.08
-0.19 0.27 0.17 0.04 0.16
(appendix continied)
![Page 116: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/116.jpg)
107
Appendix C (continued)
Logit Point Unwt. Item item bis. total
# diff. corr. fit
Wt. Ability Mean Logit total between item Residual fit fit score Index
83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99
100
1.17 1.59 1.83 2.39 1.37 1.83 3.24 2.76 1.83 1.59 4.01 2.76 3.24 3.24 2.39 2.39 4.01 4.74
0.30 0.28 0.20 0.31 0.45 0.45 0.17
-0.04 0.54 0.48 0.01 0.21 0.37 0.17 0.52 0.22 0.24
-9.99 Mean -0.32 S.D. 2.03 Groups
0.29 0.23
0.36 -0.00 0.25
-0.15 -0.52 -0.46 0.12 1.16
-0.86 -0.23 0.65 0.12
-0.37 0.12
-0.77 0.28
-0.00 -0.52 0.01 0.74
0.32 0.45 0.66 0.15
-0.32 -0.30 0.30 0.54
-0.62 -0.60 0.38 0.29 0.07 0.30
-0.39 0.27 0.28
-1.05
-0.36 -0.89 -0.32 -1.04 -1.09 -0.32 0.06 1.76 1.21
-0.89 -0.34 -0.22 0.06 0.06 0.66 1.06
-0.34 0.06
-0.05 0.72
0.18 0.72 2
Note: Raw score mean = 62.00 with a Mean person ability = 0.53 with a S.D Test reliability (K.R. 20) = 0.89. Reliability of person separation = 0.89
S.D.of 11 of 0.84.
.22.
0.36 0.28 0.24 0.16 0.32 0.24 0.08 0.12 0.24 0.28 0.04 0.12 0.08 0.08 0.16 0.16 0.04 0.00 0.62
-0.11 0.03
-0.03 0.07 0.20 0.15 0.04
-0.19 0.23 0.10 0.00 0.03 0.07 0.04 0.14
-0.01 0.04
-0.00
![Page 117: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/117.jpg)
108
Appendix C (continued)
Pvpprimpnt 1V Summary of Item Fit Information for a Normally Distributed Item Difficulty Distribution With 25 Items. 50 Persons, and a 25% Chance of Guessing Correctly
Logit Item item
# diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
1 2 3 4 5 6 7 8 9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
-4.04 -3.33 -2.58 -3.33 -1.79 -1.79 -1.79 -1.79 -2.12 -0.71 -0.56 -0.88 -0.71 -0.13 0.80 0.69 1.02 0.91 1.12 1.44 1.88 2.74 2.74 3.04 5.11
-9.99 -0.03 0.31 0.22 0.35 0.24 0.29 0.38 0.33 0.47 0.50 0.43 0.21 0.56 0.45 0.32 0.56 0.58 0.24 0.49 0.53 0.46 0.37 0.51 0.00
0.18 1.21
-0.22 0.20
-0.30 0.08
-0.08 -0.32 -0.22 -0.62 -0.31 -0.49 1.71
-0.97 1.14 0.88
-0.95 -0.87 2.01 0.01 0.25
-0.46 0.22
-0.58 2.41
-1.10 0.41 0.05 0.27
-0.00 0.32 0.16
-0.19 -0.02 -0.33 -0.75 -0.17 0.79
-0.86 0.17 1.64
-0.58 -0.99 2.43 0.09
-0.49 0.14 0.64
-0.47 0.33
0.05 2.80
-0.34 -0.63 0.08 0.08 0.08 0.08
-0.12 0.87
-0.26 0.72 0.13 1.48
-0.72 0.79
-0.44 0.90 2.56
-0.19 0.18
-0.43 -0.55 0.24 1.18
1.00 0.98 0.96 0.98 0.92 0.92 0.92 0.92 0.94 0.82 0.80 0.84 0.82 0.74 0.58 0.60 0.54 0.56 0.52 0.46 0.38 0.24 0.24 0.20 0.04
-0.00 -0.03 0.04 0.03 0.06 0.05 0.06 0.06 0.05 0.12 0.10 0.10
-0.41 0.20
-0.50 -0.24 0.35 0.27
-0.70 0.06
-0.11 0.13 0.00 0.12
-0.77
Mean -0.16 0.37 0.16 0.06 S.D. 2.27 0.18 0.91 0.78 Groups
0.34 0.91 2
Note: Raw score mean = 16.92 with a S.D. of 3 Mean person ability = 1.25 with a S.D. of 1.28. Test reliability (K.R. 20) = 0.76. Reliability of person separation = 0.76.
57.
0.68
![Page 118: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/118.jpg)
109
Appendix C (continued)
F.xneriment 14- Summary of Ttem Fit Information for a Normally Distributed Item Difficulty Distribution With 50 Items. 50 Persons, and a 25 % Chance of Guessing Correctly
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
1 -4.34 -9.99 -0.05 -1.06 0.05 2 -4.34 -9.99 -0.05 -1.06 0.05 3 -2.39 0.32 -0.21 0.05 0.13 4 -3.61 0.06 0.64 0.40 -0.48 5 -3.61 0.36 -0.24 0.08 -0.48 6 -2.86 0.14 0.27 0.41 -0.13 7 -1.77 0.30 -0.16 0.20 0.56 8 -1.77 0.04 0.73 0.93 1.02 9 -2.86 0.17 1.06 0.08 1.02
10 -2.39 0.18 0.63 0.20 0.27 11 -2.39 0.40 -0.07 -0.38 0.27 12 -1.33 0.35 0.01 -0.04 0.07 13 -0.99 0.37 0.09 -0.07 0.62 14 -0.99 0.32 -0.01 0.39 0.06 15 -0.99 0.44 -0.12 -0.52 -0.94 16 -0.83 0.40 -0.52 0.13 1.46 17 -0.83 0.40 -0.21 -0.11 -1.38 18 -0.83 0.18 0.62 1.11 0.21 19 -1.33 0.27 0.16 0.45 -0.68 20 -0.43 0.42 0.02 -0.20 1.09 21 -0.56 0.28 1.17 0.40 1.59 22 -0.69 0.45 -0.33 -0.46 -0.76 23 -0.69 0.39 0.43 -0.32 -0.22 24 -1.33 0.28 -0.05 0.42 -0.68 25 -0.43 0.42 -0.33 -0.07 -1.32 26 0.14 0.50 -0.95 -0.65 -0.47 27 0.03 0.35 0.91 0.29 -0.30 28 0.24 0.44 -0.50 -0.04 0.90 29 0.94 0.49 0.16 -0.86 0.85 30 0.03 0.30 0.37 1.03 -0.30 31 0.14 0.51 -0.81 -0.93 1.44 32 -0.08 0.57 -1.07 -1.47 0.92 33 0.84 0.38 0.99 0.28 -0.59 34 0.34 0.43 -0.43 0.10 -0.51 35 1.03 0.41 0.23 0.18 -0.70 36 1.03 0.40 0.67 0.12 -0.63 37 0.94 0.65 -2.02 -2.58 2.29 38 1.54 0.40 0.05 0.25 -0.26 39 1.13 0.24 1.55 1.69 1.58
item score
Logit Residual
Index
1.00 1.00 0.94 0.98 0.98 0.96 0.90 0.90 0.96 0.94 0.94 0.86 0.82 0.82 0.82 0.80 0.80 0.80 0.86 0.74 0.76 0.78 0.78 0.86 0.74 0.64 0.66 0.62 0.48 0.66 0.64 0.68 0.50 0.60 0.46 0.46 0.48 0.36 0.44
-0.00 -0.00 0.05 0.02 0.02 0.04 0.06
-0.01 -0.10 -0.02 0.03 0.05 0.04 0.04 0.06 0.11 0.08
-0.08 0.01
-0.01 -0.27 0.10
-0.05 0.06 0.10 0.31
-0.32 0.23
-0.15 -0.02 0.28 0.30
-0.51 0.21
-0.06 -0.35 0.67 0.01
-0.65
(appendix continues)
![Page 119: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/119.jpg)
110
Appendix C (continued')
Item Logit Point Unwt. Wt. Ability Mean Logit
Item item bis. total total between item Residual # diff. corr. fit fit fit score Index
40 1.76 0.36 0.05 0.63 -0.44 0.32 0.01 41 1.34 0.26 1.12 1.46 -0.16 0.40 -0.38 42 1.76 0.51 -0.46 -0.85 -0.79 0.32 0.17 43 1.88 0.52 -0.85 -0.69 0.99 0.30 0.24 44 2.12 0.38 -0.16 0.31 -0.96 0.26 0.07 45 2.38 0.38 0.10 0.04 0.67 0.22 0.04 46 2.00 0.29 0.43 0.92 1.40 0.28 -0.04 47 3.04 0.27 0.50 0.29 -0.16 0.14 -0.02 48 3.48 0.31 -0.28 0.34 0.64 0.10 0.06 49 3.76 0.26 0.36 0.18 -0.52 0.08 -0.00 50 4.10 0.54 -0.98 -0.65 0.18 0.06 0.05
Mean -0.17 0.36 0.05 0.01 0.13 0.64 S.D. 2.02 0.15 0.65 0.74 0.84 Groups 2 Note: Raw score mean = 31.90 with a S.D. of 7.10. Mean person ability = 0.85 with a S.D. of 1.10. Test reliability (K.R. 20) = 0.86. Reliability of person separation = 0.87
![Page 120: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/120.jpg)
I l l
Appendix C (continued)
F.xneriment 15- Summary of Ttem Fit Information for a Normally Distributed Item Difficulty Distribution With 100 Items. 50 Persons, and a 25% Chance of Guessing Correctly
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
1 -4.32 -9.99 -0.05 -1.04 0.05 2 -4.32 -9.99 -0.05 -1.04 0.05 3 -4.32 -9.99 -0.05 -1.04 0.05 4 -3.59 0.13 0.37 0.38 -0.40 5 -3.59 0.24 0.05 0.30 -0.40 6 -4.32 -9.99 -0.05 -1.04 0.05 7 -2.37 0.23 0.19 0.29 0.01 8 -4.32 -9.99 -0.05 -1.04 0.05 9 -3.59 0.11 0.47 0.40 -0.40
10 -2.83 -0.01 0.92 0.50 2.89 11 -2.83 0.46 -0.65 -0.31 -0.03 12 -2.83 0.07 0.54 0.46 -0.03 13 -2.02 0.23 0.32 0.15 -0.69 14 -2.02 0.25 -0.05 0.34 0.49 15 -3.59 0.02 0.86 0.42 -0.40 16 -3.59 0.25 -0.00 0.28 -0.40 17 -1.75 0.33 0.02 -0.05 0.72 18 -2.37 0.34 -0.19 -0.08 0.26 19 -1.51 0.30 0.61 -0.07 0.19 20 -1.51 0.39 -0.50 -0.08 0.92 21 -2.02 0.25 0.26 0.09 -0.69 22 -1.13 0.44 -0.50 -0.33 0.05 23 -1.51 0.20 0.24 0.61 1.58 24 -2.37 0.22 0.19 0.31 0.01 25 -1.51 0.21 0.02 0.68 -0.78 26 -1.31 0.26 0.09 0.44 -0.32 27 -1.75 0.31 0.09 -0.03 0.72 28 -1.13 0.41 -0.36 -0.26 0.05 29 -1.51 0.24 0.05 0.52 -0.78 30 -0.81 0.25 0.19 0.79 0.97 31 -1.13 0.38 -0.29 -0.05 -0.90 32 -1.31 0.18 0.88 0.54 2.24 33 -0.96 0.29 0.41 0.30 -1.36 34 -0.96 0.46 -0.60 -0.51 -1.36 35 -0.67 0.39 -0.35 0.07 -0.26 36 -1.13 0.25 0.17 0.55 0.67 37 -0.67 0.50 -0.90 -0.64 1.86 38 -1.31 0.57 -1.07 -1.07 1.12 39 -1.13 0.46 -0.53 -0.54 -0.90
Mean Logit item Residual
Index score
1.00 1.00 1.00 0.98 0.98 1.00 0.94 1.00 0.98 0.96 0.96 0.96 0.92 0.92 0.98 0.98 0.90 0.94 0.88 0.88 0.92 0.84 0.88 0.94 0.88 0.86 0.90 0.84 0.88 0.80 0.84 0.86 0.82 0.82 0.78 0.84 0.78 0.86 0.84
-0.00 -0.00 -0.00 0.03 0.03
-0.00 0.02
-0.00 0.03
-0.01 0.04 0.02 0.03 0.05 0.01 0.03 0.04 0.05
-0.06 0.09 0.03 0.10 0.02 0.03 0.05 0.03 0.04 0.09 0.04 0.04 0.09
-0.09 -0.03 0.12 0.11 0.04 0.14 0.13 0.10
(appendix continues)
![Page 121: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/121.jpg)
112
Appendix C (continued)
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index 40 -0.41 0.17 1.52 1.19 1.71 0.74 -0.34 41 -0.81 0.47 -0.16 -0.70 0.65 0.80 0.04 42 -0.41 0.38 -0.01 0.10 -0.84 0.74 0.00 43 -0.54 0.37 0.12 0.08 -1.48 0.76 0.02 44 -0.41 0.57 -1.19 -1.29 1.40 0.74 0.26 45 -1.13 0.43 -0.20 -0.44 0.05 0.84 0.06 46 0.36 0.36 0.56 0.45 -1.54 0.60 -0.15 47 -0.29 0.46 -0.65 -0.33 -0.36 0.72 0.19 48 -0.67 0.25 0.46 0.80 -0.76 0.78 -0.08 49 -0.29 0.37 0.55 -0.03 -0.82 0.72 -0.11 50 -0.54 0.31 0.63 0.39 0.11 0.76 -0.10 51 -0.06 0.39 -0.11 0.22 -0.77 0.68 -0.00 52 -0.17 0.34 1.94 0.10 0.01 0.70 -0.85 53 0.36 0.42 0.32 -0.17 0.85 0.60 -0.21 54 -0.54 0.31 0.45 0.42 0.15 0.76 -0.04 55 0.56 0.33 0.73 0.97 0.05 0.56 -0.22 56 -0.17 0.33 0.29 0.55 -1.54 0.70 -0.02 57 -0.06 0.46 -0.59 -0.34 0.40 0.68 0.20 58 0.05 0.41 -0.27 0.07 -1.03 0.66 0.13 59 0.15 0.41 -0.15 0.13 -0.22 0.64 0.12 60 0.56 0.36 0.86 0.46 -1.23 0.56 -0.34 61 0.36 0.49 -0.56 -0.77 -1.54 0.60 0.24 62 0.26 0.43 -0.49 0.05 -0.56 0.62 0.23 63 0.36 0.44 -0.47 0.02 0.87 0.60 0.23 64 0.26 0.41 -0.13 0.16 -0.74 0.62 0.08 65 0.26 0.30 0.40 1.29 -0.74 0.62 -0.08 66 0.85 0.50 -0.80 -0.69 -0.39 0.50 0.34 67 0.76 0.30 1.14 1.31 0.99 0.52 -0.42 68 0.95 0.48 -0.58 -0.38 -0.35 0.48 0.27 69 1.05 0.39 0.33 0.42 0.18 0.46 -0.10 70 0.76 0.36 0.67 0.73 0.12 0.52 -0.23 71 0.46 0.54 -1.10 -1.22 -0.77 0.58 0.37 72 1.15 0.33 0.58 1.08 0.12 0.44 -0.15 73 0.66 0.54 -1.18 -1.21 0.22 0.54 0.45 74 1.05 0.49 -0.37 -0.60 -0.99 0.46 0.17 75 0.95 0.41 0.88 -0.05 -0.35 0.48 -0.49 76 0.66 0.41 0.28 0.11 -0.44 0.54 -0.08 77 1.78 0.48 -0.53 -0.32 -1.03 0.32 0.19 78 1.67 0.48 -0.56 -0.40 -0.79 0.34 0.19 79 1.78 0.39 0.03 0.42 0.25 0.32 -0.03 80 1.90 0.39 0.37 0.12 0.20 0.30 -0.07 81 0.95 0.34 0.81 0.86 -1.05 0.48 -0.28 82 1.67 0.29 0.54 1.17 0.38 0.34 -0.11
(appendix continues)
![Page 122: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/122.jpg)
Appendix C (continued')
113
Logit Point Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
83 1.90 0.43 -0.27 0.05 -0.14 0.30 0.11 84 1.56 0.60 -1.34 -1.44 1.73 0.36 0.40 85 1.90 0.34 0.52 0.54 0.20 0.30 -0.15 86 1.90 0.47 -0.56 -0.21 -0.14 0.30 0.16 87 1.90 0.43 -0.17 0.04 -1.16 0.30 0.09 88 1.78 0.33 0.44 0.82 -1.03 0.32 -0.05 89 1.35 0.43 0.15 -0.09 -0.19 0.40 -0.10 90 1.90 0.48 -0.52 -0.31 -0.14 0.30 0.17 91 2.56 0.28 0.66 0.55 0.02 0.20 -0.08 92 2.56 0.27 0.36 0.80 0.02 0.20 -0.01 93 2.56 0.54 -0.53 -0.96 0.48 0.20 0.11 94 2.88 0.45 -0.16 -0.42 -0.16 0.16 0.05 95 2.56 0.63 -1.13 -1.49 0.48 0.20 0.17 96 3.51 0.29 0.26 0.04 -1.12 0.10 0.02 97 3.79 0.27 0.01 0.30 0.39 0.08 0.04 98 4.13 0.09 1.39 0.20 2.16 0.06 -0.16 99 4.13 0.43 -0.70 -0.14 0.16 0.06 0.05
100 5.34 0.16 0.26 0.34 -0.46 0.02 0.03 Mean -0.22 0.35 0.04 0.03 -0.06 0.65 S.D. 2.05 0.15 0.60 0.62 0.86 Groups 2 Note: Raw score mean = 64.62 with a S.D. of 14. Mean person ability = 0.87 with a S.D. of 1.11. Test reliability (K.R. 20) = 0.93. Reliability of person separation = 0.93.
13.
![Page 123: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/123.jpg)
114
Appendic C (continued)
Experiment 16: Summary of Item Fit Information for a Normally Distributed Item Difficulty Distribution With 25 Items. 100 Persons, and a 25% Chance of Guessing Correctly
Logit Point Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
1 -4.84 -9.99 -0.05 -1.21 0.04 1.00 -0.00 2 -2.99 0.27 -0.39 -0.05 0.29 0.97 0.03 3 -3.42 0.10 0.54 0.24 0.61 0.98 0.00 4 -3.42 0.16 0.04 0.21 0.03 0.98 0.03 5 -2.23 0.07 0.74 0.48 0.14 0.94 -0.01 6 -2.23 0.12 0.43 0.38 0.14 0.94 0.02 7 -1.89 0.19 0.71 0.16 -0.82 0.92 -0.05 8 -1.30 0.12 1.32 0.88 1.82 0.87 -0.13 9 -0.50 0.22 0.90 1.16 0.47 0.77 -0.10
10 -0.71 0.39 -0.53 -0.24 0.79 0.80 0.09 11 -0.50 0.37 -0.22 -0.10 -0.10 0.77 0.07 12 -0.02 0.39 0.32 0.04 0.32 0.69 -0.06 13 -0.08 0.52 -1.34 -1.50 -0.56 0.70 0.25 14 0.30 0.39 0.10 0.43 1.16 0.63 0.04 15 0.51 0.45 -0.75 -0.05 -0.03 0.59 0.23 16 0.51 0.55 -1.75 -1.80 1.24 0.59 0.41 17 0.66 0.52 -0.99 -1.46 1.47 0.56 0.22 18 1.06 0.35 1.17 1.26 0.23 0.48 -0.30 19 0.96 0.45 0.15 -0.08 -0.35 0.50 -0.06 20 1.74 0.51 -0.75 -0.51 -0.39 0.35 0.19 21 2.03 0.47 -0.47 -0.11 -1.11 0.30 0.10 22 1.85 0.47 -0.35 -0.18 -1.14 0.33 0.05 23 2.63 0.25 1.60 1.35 0.24 0.21 -0.24 24 3.57 0.49 -0.92 -0.67 -0.35 0.11 0.07 25 3.45 0.39 -0.41 0.11 1.10 0.12 0.06
Mean -0.19 0.34 -0.04 -0.05 0.21 0.64 S.D. 2.20 0.17 0.84 0.82 0.76 Groups 2 Note: Raw score mean =16.10 with a S.D. of 3 Mean person ability = 1.00 with a S.D. of 1.14. Test reliability (K.R. 20) = 0.72. Reliability of person separation = 0.74.
.49.
![Page 124: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/124.jpg)
115
Appendix C (continued)
Experiment 17: Summary of Item Fit Information for a Normally Distributed Item Difficulty Distribution With 50 Items. 100 Persons, and a 25% Chance of Guessing Correctly
Logit Point Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
1 -4.16 0.13 0.20 0.31 -0.42 0.99 0.02 2 -4.87 -9.99 -0.01 -1.22 0.04 1.00 -0.00 3 -4.16 0.01 0.86 0.35 2.02 0.99 0.00 4 -2.72 0.22 -0.27 0.10 0.36 0.96 0.04 5 -2.72 0.20 -0.20 0.13 0.36 0.96 0.04 6 -4.16 0.06 0.55 0.34 -0.42 0.99 0.02 7 -2.27 0.18 0.71 -0.03 0.53 0.94 -0.05 8 -2.27 0.28 -0.59 -0.03 0.69 0.94 0.05 9 -1.80 0.27 -0.29 -0.02 -0.17 0.91 0.05
10 -2.09 0.21 -0.13 0.10 0.83 0.93 0.04 11 -1.45 0.26 1.91 -0.18 -0.24 0.88 -0.49 12 -1.45 0.36 -0.76 -0.34 1.46 0.88 0.08 13 -1.56 0.34 -0.55 -0.25 0.29 0.89 0.06 14 -1.08 0.16 0.89 0.85 1.55 0.84 -0.07 15 -1.35 0.23 -0.01 0.43 -0.52 0.87 0.04 16 -1.26 0.15 0.64 0.71 1.26 0.86 -0.01 17 -1.08 0.22 0.52 0.44 2.28 0.84 -0.02 18 -1.26 0.23 0.21 0.33 -0.99 0.86 0.02 19 -0.64 0.42 -0.81 -0.61 -1.17 0.78 0.14 20 -0.77 0.36 -0.50 -0.19 -1.02 0.80 0.10 21 -0.70 0.38 0.23 -0.65 -0.66 0.79 -0.04 22 -0.57 0.42 -0.81 -0.60 -1.34 0.77 0.14 23 -0.77 0.20 0.52 1.01 -0.42 0.80 -0.02 24 0.02 0.37 -0.13 0.30 -1.12 0.67 0.06 25 -0.44 0.35 -0.36 0.22 0.05 0.75 0.10 26 -0.04 0.39 -0.35 0.10 -0.40 0.68 0.11 27 0.23 0.30 0.95 1.34 -1.06 0.63 -0.21 28 0.02 0.48 -1.07 -1.19 0.61 0.67 0.24 29 0.28 0.45 -0.38 -0.81 0.42 0.62 0.12 30 0.48 0.45 0.21 -0.80 0.09 0.58 -0.10 31 0.48 0.36 0.42 0.79 -0.10 0.58 -0.06 32 0.48 0.38 0.26 0.61 0.58 0.58 -0.03 33 0.48 0.29 1.99 1.57 -0.10 0.58 -0.71 34 0.73 0.52 -1.22 -1.40 -0.64 0.53 0.34 35 0.73 0.57 -1.62 -2.44 2.33 0.53 0.39 36 1.33 0.26 2.51 2.22 2.72 0.41 -0.79 37 1.23 0.43 0.44 0.24 -0.89 0.43 -0.15 38 1.12 0.47 -0.43 -0.40 -0.89 0.45 0.15 39 1.33 0.37 1.00 0.99 0.66 0.41 -0.22
(appendix continues)
![Page 125: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/125.jpg)
Appendix C (continued)
116
Groups
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
40 1.28 0.50 -0.92 -0.68 1.50 0.42 0.24 41 2.47 0.57 -1.01 -1.13 1.36 0.22 0.14 42 1.76 0.55 -1.42 -0.95 -0.37 0.33 0.28 43 2.00 0.45 -0.27 0.11 -0.66 0.29 0.08 44 2.32 0.36 0.92 0.66 0.16 0.24 -0.13 45 2.12 0.50 -0.46 -0.47 0.27 0.27 0.07 46 3.05 0.45 -0.10 -0.34 0.18 0.15 0.03 47 2.87 0.37 0.30 0.39 0.54 0.17 0.01 48 3.37 0.39 0.27 0.01 -1.18 0.12 -0.02 49 5.09 0.05 0.77 0.55 -0.01 0.03 0.01 50 5.54 0.16 0.90 0.19 1.25 0.02 -0.03
Mean S.D.
-0.10 2.23
0.33 0.15
0.07 0.84
0.01 0.81
0.19 1.00
0.64
Note: Raw score mean = 31.83 with a S.D. of 6. Mean person ability = 0.93 with a S.D. of 1.13. Test reliability (K.R. 20) = 0.85. Reliability of person separation = 0.87.
75.
![Page 126: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/126.jpg)
117
Appendix (continued)
Exper iment 18- Summary ofTtem Fit Information for a Normally Distributed Item Difficulty Distribution With 100 Items. 100 Persons, and a 25% Chance of Guessing Correctly
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
1 -4.93 -9.99 -0.18 -1.21 0.04 1.00 -0.00 2 -4.93 -9.99 -0.18 -1.21 0.04 1.00 -0.00 3 -4.93 -9.99 -0.18 -1.21 0.04 1.00 -0.00 4 -3.51 0.10 0.20 0.27 -0.12 0.98 0.03 5 -3.08 0.03 1.22 0.30 0.22 0.97 -0.09 6 -4.22 0.06 0.52 0.34 -0.45 0.99 0.02 7 -4.22 0.06 0.52 0.34 -0.45 0.99 0.02 8 -3.51 0.11 0.47 0.22 0.93 0.98 0.00 9 -2.77 0.01 1.11 0.37 1.55 0.96 -0.05
10 -2.77 0.22 -0.28 0.10 0.32 0.96 0.04 11 -3.08 0.26 -0.37 0.01 0.12 0.97 0.03 12 -3.08 0.17 0.02 0.18 0.12 0.97 0.03 13 -2.32 0.10 0.80 0.41 -1.52 0.94 -0.04 14 -2.32 0.22 2.05 -0.11 -1.52 0.94 -0.45 15 -2.77 0.15 0.44 0.14 1.55 0.96 0.00 16 -1.50 0.23 0.63 0.40 -0.98 0.88 -0.07 17 -2.15 0.28 -0.50 0.02 0.80 0.93 0.05 18 -1.30 0.31 -0.27 0.14 -0.33 0.86 0.06 19 -2.53 0.13 0.58 0.22 -0.90 0.95 -0.01 20 -1.85 0.22 1.81 -0.00 -0.56 0.91 -0.36 21 -1.85 0.17 0.86 0.34 -0.56 0.91 -0.07 22 -1.04 0.36 0.38 -0.36 0.36 0.83 -0.06 23 -1.30 0.25 0.59 0.26 1.34 0.86 -0.04 24 -1.40 0.40 -0.76 -0.49 -0.62 0.87 0.08 25 -1.61 0.40 -1.12 -0.31 1.30 0.89 0.09 26 -1.12 0.26 0.19 0.47 -0.20 0.84 0.02 27 -1.30 0.38 -0.77 -0.32 0.76 0.86 0.09 28 -1.85 0.13 0.63 0.54 -0.56 0.91 -0.01 29 -1.21 0.28 0.54 0.12 0.11 0.85 -0.04 30 -1.04 0.23 1.90 0.41 -0.53 0.83 -0.38 31 -0.96 0.35 0.55 -0.12 -0.45 0.82 -0.11 32 -1.72 0.31 -0.27 -0.18 -0.99 0.90 0.04 33 -1.30 0.35 -0.40 -0.22 -0.86 0.86 0.06 34 -0.67 0.54 -1.57 -1.56 1.90 0.78 0.19 35 -0.96 0.27 0.13 0.57 1.11 0.82 0.02 36 -0.41 0.32 0.83 0.45 -0.20 0.74 -0.13 37 -0.54 0.46 -0.48 -0.99 -0.96 0.76 0.09 38 -0.89 0.33 -0.08 0.16 -0.18 0.81 0.04 39 -0.81 0.25 0.89 0.58 -0.34 0.80 -0.10
(appendix continues)
![Page 127: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/127.jpg)
118
Appendix C (continued)
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
40 -0.61 0.50 -1.11 -1.21 1.40 0.77 0.15 41 -0.61 0.32 0.00 0.52 -1.45 0.77 0.03 42 -0.29 0.43 -0.75 -0.19 0.93 0.72 0.14 43 -0.61 0.38 0.14 -0.16 -0.16 0.77 -0.02 44 -0.67 0.42 -0.51 -0.46 -1.07 0.78 0.09 45 -0.29 0.29 1.13 0.77 1.45 0.72 -0.19 46 -0.54 0.25 0.91 0.97 1.10 0.76 -0.09 47 -0.67 0.52 -1.51 -1.32 1.24 0.78 0.18 48 -0.07 0.53 -1.54 -1.40 1.65 0.68 0.28 49 -0.07 0.44 -0.76 -0.16 -0.37 0.68 0.15 50 -0.01 0.48 -0.89 -0.82 0.64 0.67 0.19 51 0.15 0.27 2.47 1.33 1.36 0.64 -0.67 52 -0.07 0.27 0.91 1.60 1.71 0.68 -0.15 53 -0.18 0.46 -0.82 -0.60 1.30 0.70 0.16 54 -0.35 0.22 0.52 1.85 1.06 0.73 -0.05 55 0.15 0.34 1.17 0.90 1.36 0.64 -0.29 56 -0.01 0.45 -0.55 -0.44 -0.62 0.67 0.13 57 0.20 0.47 -0.94 -0.51 -0.22 0.63 0.24 58 0.20 0.39 1.13 -0.02 -0.53 0.63 -0.30 59 0.41 0.36 1.39 0.61 -0.80 0.59 -0.42 60 0.31 0.37 0.44 0.67 -0.73 0.61 -0.06 61 0.25 0.54 -1.37 -1.65 1.12 0.62 0.30 62 0.46 0.35 0.69 1.11 0.42 0.58 -0.12 63 0.71 0.52 -1.08 -1.06 0.40 0.53 0.27 64 1.06 0.43 -0.07 0.22 -0.10 0.46 0.05 65 0.31 0.29 1.26 1.66 1.74 0.61 -0.28 66 0.41 0.23 2.64 2.20 1.28 0.59 -0.76 67 0.61 0.40 0.93 0.18 -0.24 0.55 -0.28 68 0.66 0.45 -0.34 -0.10 -1.04 0.54 0.12 69 0.81 0.34 1.33 1.33 1.43 0.51 -0.37 70 0.56 0.57 -1.91 -1.80 0.80 0.56 0.46 71 1.16 0.57 -1.68 -1.79 0.76 0.44 0.40 72 1.36 0.38 0.40 0.84 2.09 0.40 -0.06 73 1.06 0.53 -1.37 -1.11 0.06 0.46 0.36 74 1.26 0.43 0.18 0.08 0.49 0.42 -0.01 75 1.36 0.60 -2.03 -2.01 2.01 0.40 0.43 76 1.06 0.41 0.50 0.31 0.57 0.46 -0.12 77 1.26 0.46 -0.43 -0.18 -0.59 0.42 0.11 78 1.52 0.46 0.19 0.27 -0.66 0.37 0.02 79 1.62 0.34 1.17 0.97 -1.30 0.35 -0.26 80 1.73 0.40 0.14 0.47 0.35 0.33 0.01 81 1.46 0.37 0.97 0.73 0.32 0.38 -0.22 82 1.68 0.52 -1.02 -0.98 0.07 0.34 0.21
(appendix continues)
![Page 128: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/128.jpg)
Appendix C (continued)
119
Logit Point Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
83 1.52 0.30 1.06 1.66 -0.10 0.37 -0.19 84 1.46 0.66 -2.72 -2.96 2.60 0.38 0.50 85 2.09 0.48 -0.50 -0.56 -1.44 0.27 0.10 86 2.28 0.42 -0.39 -0.06 -0.09 0.24 0.07 87 1.91 0.38 1.10 0.27 1.13 0.30 -0.20 88 2.28 0.39 0.25 0.12 -1.29 0.24 -0.04 89 2.35 0.55 -1.38 -1.16 1.93 0.23 0.18 90 2.15 0.36 0.11 0.68 -0.50 0.26 -0.01 91 2.28 0.46 -0.48 -0.35 0.73 0.24 0.08 92 2.49 0.44 -0.35 -0.41 -0.41 0.21 0.07 93 2.42 0.40 0.82 -0.26 -0.76 0.22 -0.13 94 2.80 0.36 -0.28 0.24 0.18 0.17 0.06 95 3.51 0.38 -0.28 -0.26 -0.65 0.10 0.03 96 3.94 0.17 0.37 0.50 0.46 0.07 0.01 97 3.38 0.31 -0.44 0.30 1.19 0.11 0.06 98 3.94 0.31 -0.34 -0.01 -1.25 0.07 0.04 99 4.58 -0.02 1.73 0.44 1.78 0.04 -0.15
100 4.90 0.34 -0.66 -0.14 0.03 0.03 0.03 Mean -0.15 0.34 0.08 -0.02 0.22 0.64 S.D. 2.08 0.15 0.99 0.86 0.97 Groups 2 Note: Raw score mean = 63.57 with a S.D. of 14. Mean person ability = 0.88 with a S.D. of 1.08. Test reliability (K.R. 20) = 0.93. Reliability of person separation = 0.93.
08.
![Page 129: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/129.jpg)
120
Appendix C (continued)
Experiment 19: Summary of Ttem Fit Information for a Normally Distributed Item Difficulty Distribution With 25 Items. 25 Persons, and a 50% Chance of Guessing Correctly
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
1 -3.92 -9.99 0.20 -0.96 0.05 1.00 -0.00 2 -3.92 -9.99 0.20 -0.96 0.05 1.00 -0.00 3 -1.46 0.47 -0.33 -0.18 0.97 0.84 0.11 4 -3.17 0.06 0.80 0.47 -0.24 0.96 0.03 5 -1.13 0.32 0.40 0.35 -0.14 0.80 -0.03 6 -0.85 0.41 0.14 0.04 0.36 0.76 0.05 7 -1.46 0.37 0.09 0.07 -1.12 0.84 0.06 8 -1.13 0.46 -0.11 -0.25 -0.14 0.80 0.11 9 -0.59 0.62 -0.79 -1.21 0.89 0.72 0.29
10 -0.85 0.48 -0.33 -0.17 0.36 0.76 0.17 11 -1.13 0.49 -0.11 -0.52 -0.26 0.80 0.10 12 -0.85 0.53 -0.50 -0.50 0.36 0.76 0.18 13 -1.13 0.26 0.52 0.60 -0.14 0.80 -0.04 14 -0.59 0.37 0.24 0.28 1.34 0.72 0.05 15 -1.13 0.06 0.95 1.34 -0.14 0.80 -0.17 16 0.49 0.57 -0.85 -0.91 0.34 0.52 0.47 17 1.31 0.54 -0.46 -0.69 0.01 0.36 0.26 18 0.49 0.43 -0.18 0.37 1.52 0.52 0.23 19 0.69 0.26 1.29 1.36 0.83 0.48 -0.73 20 1.54 0.59 -0.88 -0.78 1.15 0.32 0.34 21 1.10 0.64 -1.22 -1.52 0.41 0.40 0.53 22 1.54 0.50 -0.02 -0.37 -0.44 0.32 0.07 23 1.54 -0.07 2.87 2.40 1.42 0.32 -1.79 24 2.33 0.30 0.21 0.71 0.44 0.20 0.02 25 4.47 0.46 -0.10 -0.06 -0.67 0.04 0.03
Mean S.D. Groups
-0.31 1.92
0.40 0.23
0.08 0.82
-0.04 0.89
0.29 0.67 2
0.63
Note: Raw score mean = 15.84 and a S.D. of 4.03. Mean person ability = -0.31 with a S.D. of 1.92. Test reliability (K.R. 20) = 0.78. Reliability of person separation = 0.78.
![Page 130: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/130.jpg)
121
Appendix C (continued)
Experiment 20: Summary of Item Fit Information for a Normally Distributed Item Difficulty Distribution With 50 Items. 25 Persons, and a 50% Chance of Guessing Correctly
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index 1 -3.47 -9.99 0.05 -LOO 0.05 1.00 -0.00 2 -2.73 0.09 0.65 0.43 -0.34 0.96 0.03 3 -2.73 0.32 0.07 0.21 -0.34 0.96 0.04 4 -3.47 -9.99 0.05 -1.00 0.05 1.00 -0.00 5 -3.47 -9.99 0.05 -1.00 0.05 1.00 -0.00 6 -1.94 0.05 1.09 0.49 0.67 0.92 -0.12 7 -1.94 0.31 0.02 0.17 0.06 0.92 0.06 8 -2.73 0.37 -0.02 0.11 -0.34 0.96 0.04 9 -1.94 0.42 -0.27 -0.09 0.06 0.92 0.07
10 -2.73 0.27 0.18 0.28 -0.34 0.96 0.04 11 -3.47 -9.99 0.05 -1.00 0.05 1.00 -0.00 12 -1.94 0.24 0.22 0.32 0.06 0.92 0.05 13 -2.73 0.11 0.57 0.42 -0.34 0.96 0.04 14 -1.94 0.03 0.83 0.60 0.67 0.92 -0.01 15 -1.44 0.51 -0.56 -0.40 0.40 0.88 0.10 16 -1.44 0.25 0.22 0.42 -0.19 0.88 0.05 17 -1.44 0.53 -0.62 -0.44 0.40 0.88 0.10 18 -1.94 0.49 -0.45 -0.31 0.06 0.92 0.07 19 -1.94 -0.03 1.11 0.63 0.67 0.92 -0.08 20 -1.05 0.37 -0.16 0.22 0.71 0.84 0.10 21 -0.45 0.50 -0.49 -0.24 -0.19 0.76 0.17 22 -0.45 0.54 -0.59 -0.57 -0.19 0.76 0.20 23 -1.05 0.31 0.14 0.38 -1.07 0.84 0.05 24 -0.19 0.60 -0.83 -0.91 0.34 0.72 0.27 25 -1.05 0.45 -0.35 -0.13 0.71 0.84 0.11 26 -0.73 0.67 -1.02 -1.36 1.01 0.80 0.21 27 -0.45 0.51 -0.50 -0.32 -0.19 0.76 0.18 28 0.27 0.06 1.67 2.27 2.86 0.64 -0.64 29 -0.45 0.54 -0.37 -0.70 -0.28 0.76 0.17 30 0.27 0.42 0.00 0.44 -0.78 0.64 0.05 31 1.12 0.35 0.62 1.00 0.61 0.48 -0.20 32 1.76 0.63 -0.86 -1.05 -0.74 0.36 0.36 33 0.70 0.32 0.72 1.11 0.74 0.56 -0.29 34 0.27 0.26 0.54 1.43 -0.78 0.64 -0.20 35 0.91 0.43 0.83 0.16 -0.88 0.52 -0.44 36 1.12 0.32 1.72 0.74 0.10 0.48 -1.08 37 1.98 0.48 0.27 -0.21 -0.39 0.32 -0.06 38 0.70 0.29 0.69 1.44 1.84 0.56 -0.21 39 1.76 0.51 -0.33 -0.18 -0.74 0.36 0.20
(appendix continues)
![Page 131: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/131.jpg)
122
Appendix C ('continued')
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
40 1.32 0.65 -1.23 -1.19 -0.51 41 1.76 0.44 -0.10 0.36 0.08 42 1.76 0.63 -0.97 -1.02 0.08 43 1.98 0.61 -0.89 -0.81 1.17 44 1.32 0.37 0.65 0.74 1.13 45 1.98 0.54 -0.48 -0.35 -0.39 46 1.32 0.23 1.42 1.59 1.13 47 2.78 0.26 0.27 0.80 0.44 48 4.80 0.25 0.19 0.29 -0.63 49 3.50 0.32 0.00 0.30 -0.04 50 4.01 0.37 -0.17 0.09 -0.30
Mean S.D. Groups
-0.28 2.08
0.37 0.21
0.07 0.68
0.06 0.79
0.12 0.73 2
Mean item score
Logit Residual
Index
0.44 0.36 0.36 0.32 0.44 0.32 0.44 0.20 0.04 0.12 0.08
0.55 0.10 0.39 0.30
-0.24 0.22
-0.73 0.03 0.04 0.07 0.06
0.67
Mean person ability = 1.04 with a S.D. of 1.20. Test reliability (K.R. 20) = 0.88. Reliability of person separation = 0.87.
08.
![Page 132: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/132.jpg)
123
Appendix C (continued)
Kvpftriment 21: Summary of Item Fit Information for a Normally Distributed Item Difficulty Distribution With 100 Items. 25 Persons, and a 50% Chance of Guessing Correctly
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
1 -2.83 -0.19 1.36 0.43 1.52 0.96 -0.16
2 -2.83 0.35 -0.23 0.21 -0.19 0.96 0.04
3 -3.56 -9.99 -0.55 -1.05 0.05 1.00 -0.00
4 -3.56 -9.99 -0.55 -1.05 0.05 1.00 -0.00
5 -3.56 -9.99 -0.55 -1.05 0.05 1.00 -0.00
6 -2.83 0.37 -0.27 0.19 -0.19 0.96 0.04
7 -2.83 0.37 -0.27 0.19 -0.19 0.96 0.04 8 -2.83 0.02 0.60 0.39 -0.19 0.96 0.01 9 -2.83 -0.13 1.12 0.42 1.52 0.96 -0.09
10 -1.22 0.53 -0.71 -0.44 0.93 0.84 0.13 11 -2.83 0.37 -0.27 0.19 -0.19 0.96 0.04 12 -2.83 0.13 0.29 0.35 -0.19 0.96 0.03 13 -0.92 0.20 0.98 0.26 -0.35 0.80 -0.30 14 -2.83 0.35 -0.23 0.21 -0.19 0.96 0.04 15 -2.83 0.07 0.44 0.37 -0.19 0.96 0.02 16 -1.22 0.43 -0.45 -0.16 0.93 0.84 0.11 17 -1.22 0.18 0.48 0.30 -1.14 0.84 -0.05 18 -2.06 0.13 0.20 0.35 0.25 0.92 0.04 19 -3.56 -9.99 -0.55 -1.05 0.05 1.00 -0.00 20 -1.58 0.29 -0.10 0.15 -0.70 0.88 0.06 21 -2.83 0.06 0.49 0.38 -0.19 0.96 0.02 22 -2.06 0.31 -0.14 0.13 0.25 0.92 0.05 23 -2.06 -0.12 1.12 0.48 0.25 0.92 -0.12 24 -1.22 0.36 0.13 -0.12 -1.14 0.84 0.00 25 -2.06 0.04 0.46 0.41 0.25 0.92 0.01 26 -1.22 0.24 0.46 0.15 -1.14 0.84 -0.07 27 -1.22 0.46 -0.21 -0.36 -1.14 0.84 0.06 28 -0.92 0.37 0.45 -0.25 -0.35 0.80 -0.13 29 -1.58 0.11 0.71 0.31 1.36 0.88 -0.08 30 -0.92 0.46 -0.21 -0.42 -0.35 0.80 0.07 31 -0.43 0.13 1.13 0.63 0.17 0.72 -0.40 32 -0.43 0.61 -1.12 -0.94 1.84 0.72 0.34 33 -0.92 0.22 0.41 0.32 -0.35 0.80 -0.07 34 -1.22 0.42 -0.38 -0.13 -1.14 0.84 0.09 35 -2.83 -0.15 1.19 0.42 1.52 0.96 -0.11 36 -1.22 0.39 -0.32 -0.05 -1.14 0.84 0.09 37 -2.06 0.30 -0.03 0.09 0.25 0.92 0.04 38 -1.22 0.32 -0.25 0.15 0.93 0.84 0.09 39 -1.58 0.18 0.07 0.36 0.61 0.88 0.04
(appendix continues)
![Page 133: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/133.jpg)
124
Appendix C (continued)
Logit Point Unwt. Wt. Ability Item item bis. total total between
# diff. corr. fit fit fit
Mean Logit item Residual score Index
40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82
-0.66 -0.43 -0.92 0.37
-0.21 0.37
-0.66 -0.43 -1.22 -0.21 -0.21 0.18
-0.01 -0.66 0.74 0.37 0.37
-0.01 0.37 0.55
-0.01 -0.21 0.37 0.37 0.74 0.37 0.37 0.18 1.50 0.55 1.30 0.18
-0.01 1.50 1.11 1.30 1.30 1.71 0.18 1.30 1.11 0.74 1.94
-0.22 0.24 0.24 0.47 0.35 0.65 0.14 0.50
-0.08 0.43 0.35 0.43 0.71 0.45
-0.03 0.46 0.52 0.20 0.51 0.42 0.26 0.37 0.64 0.27 0.42 0.42 0.63 0.11 0.55 0.66 0.23 0.07 0.72 0.26 0.35 0.16 0.30 0.15 0.38 0.15 0.48 0.30 0.10
1.77 0.27 0.17
-0.70 -0.04 -1.69 0.62
-0.61 1.22
-0.57 0.42
-0.37 -1.80 -0.54 2.00
-0.62 -0.91 0.60
-0.58 -0.31 0.83
-0.31 -1.63 0.33
-0.29 -0.46 -1.55 1.09
-0.97 -1.89 0.80 1.52
-1.83 0.69 0.16 0.89 0.12 0.51
-0.19 0.85
-0.78 0.58 1.27
1.53 0.48 0.33
-0.53 0.04
-1.75 0.63
-0.54 0.75
-0.17 -0.14 -0.34 -1.83 -0.23 2.08
-0.49 -0.91 0.75
-0.99 -0.36 0.28 0.08
-1.70 0.58
-0.43 -0.27 -1.60 1.24
-1.06 -1.92 0.39 1.36
-1.87 0.13
-0.10 0.81 0.22 0.67
-0.03 0.90
-0.73 0.26 0.42
2.01 0.17
-0.35 -1.40 -0.53 1.36
-0.89 0.67 2.07 2.15
-0.53 -0.39 1.50 1.53 0.28
-1.40 -1.40 -1.49 0.22
-0.50 0.30
-0.53 0.22 1.44 0.25 0.22 0.22 1.95
-0.68 0.81
-0.78 0.92 2.47
-0.09 -1.19 1.79
-0.78 0.49
-0.47 -0.78 0.45 0.25
-0.52
0.76 0.72 0.80 0.56 0.68 0.56 0.76 0.72 0.84 0.68 0.68 0.60 0.64 0.76 0.48 0.56 0.56 0.64 0.56 0.52 0.64 0.68 0.56 0.56 0.48 0.56 0.56 0.60 0.32 0.52 0.36 0.60 0.64 0.32 0.40 0.36 0.36 0.28 0.60 0.36 0.40 0.48 0.24
-0.54 -0.04 0.01 0.44 0.07 0.88
-0.13 0.21
-0.22 0.22
-0.14 0.25 0.71 0.17
-1.30 0.39 0.50
-0.22 0.31 0.21
-0.34 0.16 0.86
-0.14 0.19 0.32 0.83
-0.47 0.38 1.03
-0.39 -0.84 0.72
-0.27 -0.05 -0.39 -0.01 -0.12 0.09
-0.38 0.40
-0.39 -0.43
(appendix continues)
![Page 134: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/134.jpg)
Appendix C (continued)
125
Logit Point Unwt. Item item bis. total
# diff. corr. fit
Wt. Ability Mean Logit total between item Residual fit fit score Index
83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99
100 Mean S.D. Groups
2.19 1.30 2.19 2.49 2.19 2.84 3.31 1.50 3.31 2.49 2.84 2.84 3.31 2.19 3.31 4.07 3.31 3.31
-0.14 1.91
0.22 0.28 0.19 0.13 0.48 0.44 0.38 0.14 0.09
-0.12 0.24
-0.11 0.05 0.20 0.25
-0.02 0.25 0.37 0.27 0.22
-0.01 0.03 0.28 0.42
-0.67 -0.59 -0.38 0.57 0.32 1.26
-0.06 1.49 0.60 0.18
-0.14 0.69
-0.14 -0.42
0.31 0.43 0.31 0.31
-0.49 -0.22 -0.03 0.86 0.30 0.70 0.15 0.47 0.30 0.28 0.17 0.36 0.17 0.00
-0.61 1.79 0.14 0.79 1.09 0.52 0.19 1.21 0.35 2.23
-0.52 1.49 0.35 1.60 0.19 1.58 0.19 0.19
0.06 0.80
0.01 0.71
0.24 0.96 2
0.20 0.36 0.20 0.16 0.20 0.12 0.08 0.32 0.08 0.16 0.12 0.12 0.08 0.20 0.08 0.04 0.08 0.08 0.62
0.05 0.07
-0.02 -0.02 0.17 0.10 0.07
-0.17 0.02
-0.25 0.05
-0.35 -0.04 0.02 0.06
-0.01 0.06 0.07
Note: Raw score mean = 61.60 with a S.D. of 11.30. Mean person ability = 0.64 with a S.D. of 0.80. Test reliability (K.R. 20) = 0.89. Reliability of person separation = 0.89.
![Page 135: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/135.jpg)
126
Appendix (continued)
Experiment 22: Summary of Ttem Fit Information for a Normally Distributed Item Difficulty Distribution With 25 Items. 50 Persons, and a 50% Chance of Guessing Correctly
Logit Point Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
1 -4.50 -9.99 0.41 -1.08 0.04 1.00 -0.00 2 -4.50 -9.99 0.41 -1.08 0.04 1.00 -0.00 3 -3.78 0.05 0.88 0.40 -0.57 0.98 0.02 4 -3.03 0.16 0.40 0.33 -0.26 0.96 0.04 5 -1.96 0.42 -0.28 -0.41 0.36 0.90 0.07 6 -1.96 0.22 0.96 0.12 1.47 0.90 -0.12 7 -1.52 0.36 -0.29 0.22 0.70 0.86 0.08 8 -1.33 0.41 -0.35 -0.14 0.86 0.84 0.10 9 -1.01 0.54 -0.76 -1.08 -0.11 0.80 0.16
10 -1.96 0.37 -0.18 -0.21 0.36 0.90 0.07 11 -0.86 0.31 0.95 0.24 1.53 0.78 -0.17 12 -0.10 0.45 0.01 0.13 -0.85 0.66 0.06 13 -0.34 0.44 0.06 -0.01 -1.48 0.70 0.05 14 0.01 0.43 0.08 0.37 0.05 0.64 -0.04 15 0.23 0.31 0.88 1.57 0.20 0.60 -0.24 16 -0.10 0.44 0.10 0.10 0.39 0.66 0.05 17 0.34 0.48 0.15 -0.16 -1.09 0.58 -0.09 18 0.12 0.51 -0.37 -0.40 -0.75 0.62 0.17 19 1.63 0.64 -1.52 -1.68 0.71 0.34 0.38 20 1.63 0.54 -0.33 -0.75 -0.38 0.34 0.12 21 1.63 0.22 1.80 2.06 2.14 0.34 -0.52 22 2.41 0.61 -1.26 -1.16 1.31 0.22 0.21 23 1.63 0.42 0.04 0.74 -0.80 0.34 0.04 24 3.56 0.27 0.80 0.32 -0.27 0.10 -0.05 25 4.75 0.19 0.52 0.56 -0.36 0.04 0.02
Mean -0.36 0.38 0.12 -0.04 0.13 0.64 S.D. 2.33 0.19 0.73 0.84 0.88 Groups 2 Note: Raw score mean = 16.10 and a S.D. of 3.77. Mean person ability = 0.79 with a S.D. of 1.30. Test reliability (K.R. 20) = 0.77. Reliability of person separation = 0..78.
![Page 136: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/136.jpg)
127
Appendic C (continued"!
Experiment 23: Summary of Item Fit Information for a Normally Distributed Ttem Difficulty Distribution With 50 Items. 50 Persons, and a 50%'Chanr.e nf f i n i n g Correctly *
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit 1 -3.84 0.23 0.17 0.31 -0.59 2 -3.84 0.07 0.69 0.41 -0.59 3 -4.57 -9.99 0.14 -1.04 0.05 4 -4.57 -9.99 0.14 -1.04 0.05 5 -3.08 0.30 -0.14 0.15 -0.27 6 -2.61 0.26 -0.08 0.33 -0.04 7 -3.84 0.16 0.38 0.37 -0.59 8 -1.54 0.21 0.71 0.67 0.60 9 -2.61 0.13 3.96 -0.12 0.66
10 -1.98 0.26 0.39 0.31 -0.42 11 -1.54 0.57 -1.08 -1.03 0.70 12 -1.54 0.38 -0.16 -0.06 0.70 13 -1.18 0.34 0.31 0.22 -0.22 14 -1.75 0.16 0.51 0.86 -0.95 15 -0.88 0.45 -0.46 -0.23 -1.21 16 -1.18 0.44 -0.42 -0.18 -0.44 17 -1.18 0.54 -0.79 -0.95 1.02 18 -0.25 0.43 -0.15 0.17 -0.24 19 -0.74 0.37 1.06 0.18 -1.14 20 -1.18 0.38 0.54 -0.12 -0.22 21 -1.35 0.33 0.33 0.19 0.19 22 -0.49 0.40 0.30 0.21 -0.79 23 -1.03 0.56 -1.03 -0.95 1.18 24 -0.36 0.51 -0.69 -0.55 0.06 25 -0.02 0.54 -0.55 -1.00 -0.09 26 -0.25 0.60 -1.14 -1.52 0.36 27 -0.61 0.32 0.55 0.63 -0.34 28 -0.61 0.46 -0.39 -0.20 -0.34 29 -0.02 0.49 -0.43 -0.36 -0.09 30 0.60 0.48 -0.19 -0.14 0.14 31 0.19 0.36 0.35 1.10 2.03 32 0.60 0.22 1.74 2.35 1.50 33 1.01 0.61 -1.20 -1.69 2.37 34 1.01 0.48 -0.23 -0.21 -1.40 35 0.81 0.51 -0.46 -0.43 0.03 36 1.44 0.39 0.33 0.68 -1.15 37 0.70 0.47 -0.07 -0.12 -0.85 38 1.33 0.42 1.00 0.15 -0.48 39 1.90 0.25 0.83 1.46 -0.65
Mean item score
Logit Residual
Index
0.98 0.98 1.00 1.00 0.96 0.94 0.98 0.86 0.94 0.90 0.86 0.86 0.82 0.88 0.78 0.82 0.82 0.68 0.76 0.82 0.84 0.72 0.80 0.70 0.64 0.68 0.74 0.74 0.64 0.52 0.60 0.52 0.44 0.44 0.48 0.36 0.50 0.38 0.28
0.03 0.03
-0.00 -0.00 0.04 0.05 0.03
-0.07 -2.81 -0.01 0.13 0.07
-0.01 0.00 0.13 0.10 0.14 0.11
-0.33 -0.08 -0.00 -0.03 0.17 0.20 0.07 0.31
-0.05 0.13 0.19 0.12
-0.01 -0.64 0.40 0.15 0.21
-0.04 0.06
-0.37 -0.23
(appendix continues"!
![Page 137: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/137.jpg)
Appendix C (continued^
128
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit 40 1.66 0.58 -1.07 -1.05 0.36 41 1.90 0.43 0.15 0.04 -0.48 42 2.16 0.29 1.12 0.73 0.32 43 2.03 0.51 -0.27 -0.67 0.33 44 2.97 0.29 0.38 0.44 -0.07 45 2.45 0.39 0.95 -0.09 -0.02 46 2.16 0.28 0.89 0.80 2.21 47 2.97 0.19 1.71 0.60 1.33 48 2.97 0.36 0.55 -0.03 -0.54 49 4.06 0.52 -0.81 -0.63 0.15 50 4.53 0.16 0.53 0.35 0.99
Mean S.D. Groups
-0.18 2.18
0.38 0.16
0.18 0.89
-0.01 0.75
0.06 0.85 2
Mean item score
0.32 0.28 0.24 0.26 0.14 0.20 0.24 0.14 0.14 0.06 0.04
Logit Residual
Index
0.29 0.00
-0.24 0.10
-0.00 -0.23 -0.10 -0.39 -0.07 0.05 0.01
Note: Raw score mean = 30.72 with a S.D. of 7.53. Mean person ability = 0.72 with a S.D. of 1.21. Test reliability (K.R. 20) = 0.88. Reliability of person separation = 0.89.
0.61
![Page 138: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/138.jpg)
129
Appendix C (continued)
Experiment 24: Summary of Item Fit Information for a Normally Distributed Item Difficulty Distribution With 100 Items. 50 Persons, and a 50% Chanpp nf Guessing Correctly
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index 1 -4.33 -9.99 -0.02 -1.09 0.05 1.00 -0.00 2 -4.33 -9.99 -0.02 -1.09 0.05 1.00 -0.00 3 -2.86 0.34 -0.30 -0.06 -0.24 0.96 0.04 4 -2.40 0.26 0.12 0.07 0.00 0.94 0.04 5 -2.86 0.24 -0.12 0.24 -0.24 0.96 0.04 6 -2.86 0.10 1.30 0.23 1.28 0.96 -0.15 7 -2.40 0.27 -0.08 0.18 0.00 0.94 0.05 8 -3.61 0.27 -0.03 0.22 -0.56 0.98 0.02 9 -2.86 0.19 0.20 0.27 -0.24 0.96 0.03
10 -3.61 0.27 -0.03 0.22 -0.56 0.98 0.02 11 -2.40 0.11 1.14 0.31 0.55 0.94 -0.11 12 -2.06 0.23 0.04 0.40 -0.02 0.92 0.04 13 -1.00 0.28 0.62 0.64 -0.36 0.82 -0.06 14 -2.86 0.07 1.23 0.32 1.28 0.96 -0.12 15 -2.06 0.25 -0.12 0.40 0.21 0.92 0.06 16 -1.55 0.24 0.43 0.41 0.91 0.88 -0.00 17 -2.06 0.34 0.07 -0.19 -0.02 0.92 0.03 18 -2.40 0.34 -0.42 0.06 0.00 0.94 0.05 19 -1.35 0.46 -0.51 -0.46 -1.21 0.86 0.09 20 -1.55 0.41 -0.56 -0.15 0.57 0.88 0.09 21 -1.79 0.41 -0.42 -0.30 0.40 0.90 0.07 22 -2.40 0.26 -0.14 0.26 0.00 0.94 0.05 23 -2.06 0.30 -0.05 0.06 0.21 0.92 0.05 24 -1.00 0.29 1.09 0.40 -0.36 0.82 -0.21 25 -1.35 0.42 2.09 -0.75 -1.21 0.86 -0.72 26 -1.35 0.29 1.26 0.09 0.48 0.86 -0.25 27 -2.06 0.04 1.38 0.67 1.91 0.92 -0.15 28 -2.06 0.24 0.28 0.25 -0.02 0.92 0.02 29 -1.35 0.37 -0.49 0.30 0.74 0.86 0.09 30 -1.35 0.29 0.33 0.35 0.48 0.86 0.01 31 -2.06 0.26 0.20 0.12 -0.02 0.92 0.03 32 -1.16 0.16 1.14 1.04 0.06 0.84 -0.16 33 -0.30 0.37 0.45 0.58 0.34 0.72 -0.08 34 -1.35 0.31 2.24 -0.11 -1.21 0.86 -0.74 35 -1.16 0.29 -0.11 0.72 -0.70 0.84 0.06 36 -1.00 0.60 -1.32 -1.20 1.07 0.82 0.16 37 -0.56 0.36 0.44 0.46 -0.10 0.76 -0.05 38 -0.42 0.58 -1.23 -0.99 0.77 0.74 0.24 39 -1.00 0.12 2.08 0.96 2.15 0.82 -0.42
(appendix continues)
![Page 139: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/139.jpg)
Appendix C (continuedJ
130
Item #
Logit item diff.
Point bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index 40 -1.16 0.18 1.21 0.82 1.43 0.84 -0.18 41 -0.84 0.46 -0.43 -0.30 0.67 0.80 0.11 42 -0.69 0.23 1.76 0.83 0.29 0.78 -0.41 43 -0.30 0.38 0.68 0.35 0.34 0.72 -0.11 44 -1.16 0.53 -0.68 -0.92 -0.70 0.84 0.10 45 -0.06 0.30 2.01 1.05 -0.51 0.68 -0.73 46 0.06 0.57 -0.43 -1.20 0.84 0.66 0.13 47 -0.69 0.36 0.96 0.16 -1.52 0.78 -0.20 48 -0.18 0.55 -0.56 -0.86 1.24 0.70 0.14 49 -0.06 0.44 0.07 0.15 -0.64 0.68 -0.03 50 0.06 0.54 -0.98 -0.51 0.84 0.66 0.26 51 0.39 0.41 0.54 0.52 0.05 0.60 -0.11 52 -0.56 0.48 -0.39 -0.40 -0.90 0.76 0.11 53 -0.42 0.42 0.19 0.03 -0.55 0.74 -0.04 54 0.49 0.45 0.13 0.26 -0.43 0.58 -0.02 55 -0.42 0.53 -1.02 -0.51 0.77 0.74 0.19 56 -0.18 0.38 0.01 0.83 -1.20 0.70 0.04 57 0.17 0.50 -0.28 -0.29 -1.17 0.64 0.12 58 0.39 0.45 0.43 0.21 0.05 0.60 -0.14 59 0.60 0.68 -1.85 -2.12 1.50 0.56 0.54 60 0.60 0.56 -0.81 -0.70 0.69 0.56 0.29 61 0.70 0.53 -0.65 -0.40 0.17 0.54 0.25 62 0.06 0.49 -0.39 -0.10 0.25 0.66 0.14 63 0.49 0.31 1.07 1.59 1.44 0.58 -0.27 64 0.70 0.50 -0.47 0.01 0.68 0.54 0.22 65 0.17 0.50 -0.41 -0.25 0.18 0.64 0.13 66 0.49 0.54 -0.49 -0.68 0.31 0.58 0.19 67 1.02 0.51 0.07 -0.41 -0.41 0.48 -0.05 68 0.39 0.58 -1.00 -0.98 -0.11 0.60 0.29 69 0.81 0.51 -0.18 -0.27 0.60 0.52 0.09 70 0.81 0.65 -1.67 -1.73 1.42 0.52 0.51 71 1.12 0.50 0.20 -0.31 -1.07 0.46 -0.08 72 1.23 0.60 -1.20 -1.11 -0.31 0.44 0.39 73 0.81 0.50 -0.34 -0.02 -0.41 0.52 0.17 74 1.33 0.34 1.41 1.07 -0.88 0.42 -0.50 75 1.44 0.31 1.06 1.39 0.99 0.40 -0.29 76 1.23 0.54 -0.63 -0.54 0.68 0.44 0.24 77 0.91 0.35 0.65 1.40 1.51 0.50 -0.12 78 1.65 0.51 -0.07 -0.54 0.16 0.36 0.01 79 1.65 0.51 -0.57 -0.31 0.16 0.36 0.19 80 1.54 0.61 -1.23 -1.44 1.38 0.38 0.35 81 1.44 0.56 -0.92 -0.84 -0.12 0.40 0.30 82 1.65 0.31 0.77 1.38 -0.16 0.36 -0.18
(appendix continues^
![Page 140: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/140.jpg)
Appendix C (continued')
131
Logit Point Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
83 2.24 0.15 1.80 1.61 1.68 0.26 -0.45 84 1.23 0.34 0.99 1.21 1.08 0.44 -0.29 85 1.77 0.38 0.44 0.67 0.26 0.34 -0.07 86 2.00 0.51 -0.54 -0.58 -1.23 0.30 0.16 87 2.12 0.45 -0.18 -0.17 0.34 0.28 0.10 88 2.37 0.35 -0.12 0.64 -0.10 0.24 0.07 89 2.24 0.48 -0.32 -0.41 -0.48 0.26 0.09 90 2.51 0.31 0.31 0.60 0.29 0.22 -0.01 91 2.66 0.54 -0.07 -1.27 -0.02 0.20 -0.00 92 2.51 0.41 -0.21 0.05 -1.48 0.22 0.08 93 3.16 0.41 -0.49 -0.21 0.74 0.14 0.09 94 3.16 0.39 -0.21 -0.24 -1.19 0.14 0.06 95 2.51 0.42 0.33 -0.31 -1.48 0.22 -0.05 96 3.36 0.13 1.29 0.66 0.89 0.12 -0.20 97 4.64 0.21 0.00 0.24 -0.23 0.04 0.04 98 3.59 0.45 -0.78 -0.43 0.40 0.10 0.08 99 4.19 0.14 0.49 0.33 0.51 0.06 0.01
100 4.64 0.29 -0.26 0.10 -0.23 0.04 0.04 Mean -0.09 0.38 0.10 0.01 0.14 0.64 S.D. 1.96 0.15 0.84 0.72 0.80 Groups 2 Note: Raw score mean = 63.84 with a S.D. of 15.57. Mean person ability = 0.91 with a S.D. of 1.18. Test reliability (K.R. 20) = 0.94. Reliability of person separation = 0.94.
![Page 141: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/141.jpg)
132
Appendix C (continued)
Experiment 25: Summary of Item Fit Information for a Normally Distributed Item Difficulty Distribution With 25 Items. 100 Persons, and a 50% Chance of Guessing Correctly
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
1 -4.17 0.10 0.43 0.33 -0.57 0.99 0.02 2 -3.45 0.11 0.88 0.20 1.32 0.98 -0.03 3 -2.70 0.10 1.30 0.23 2.02 0.96 -0.12 4 -2.24 0.32 -0.39 -0.11 0.43 0.94 0.04 5 -1.50 0.25 2.69 0.07 -0.64 0.89 -0.74 6 -2.06 0.14 0.55 0.59 0.73 0.93 0.01 7 -1.62 0.19 1.18 0.35 2.13 0.90 -0.10 8 -1.28 0.36 -0.64 0.26 1.28 0.87 0.08 9 -1.08 0.41 -0.10 -0.46 -0.67 0.85 0.03
10 -0.66 0.35 0.12 0.49 -0.75 0.80 0.01 11 0.01 0.48 -0.28 -0.47 -0.71 0.70 0.09 12 -0.24 0.39 0.02 0.49 -0.03 0.74 0.03 13 0.01 0.47 -0.18 -0.42 -0.86 0.70 0.07 14 0.12 0.53 -1.07 -0.94 -1.35 0.68 0.24 15 0.29 0.47 -0.33 -0.10 1.20 0.65 0.08 16 0.51 0.51 -0.55 -0.66 0.18 0.61 0.14 17 0.73 0.57 -1.36 -1.63 1.19 0.57 0.31 18 1.19 0.49 -0.32 -0.37 -0.78 0.48 0.09 19 1.60 0.31 1.25 2.07 1.06 0.40 -0.26 20 1.50 0.54 -0.20 -1.38 0.21 0.42 0.01 21 2.04 0.40 0.62 0.51 1.85 0.32 -0.09 22 2.54 0.45 -0.49 -0.29 -0.29 0.24 0.10 23 2.99 0.31 0.16 0.82 -0.54 0.18 0.03 24 3.25 0.44 -0.66 -0.40 -0.32 0.15 0.09 25 4.25 0.22 0.30 0.43 -1.40 0.07 0.02
Mean 0.00 0.36 0.12 -0.02 0.19 0.64 S.D. 2.13 0.15 0.87 0.75 1.06 Groups 2 Note: Raw score mean = 16.02 and a S Mean person ability = 1.09 with a S.D. Test reliability (K.R. 20) = 0.76. Reliability of person separation = 0.77.
D. of 3.74. of 1.22.
![Page 142: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/142.jpg)
133
Appendix C (continued)
Experiment 26: Summary of Item Fit Information for a Normally Distributed Item Difficulty Distribution With 50 Items. 100 Persons, and a 50% Chance of Guessing Correctly
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
1 -4.01 0.17 -0.03 0.30 -0.47 0.99 0.02 2 -4.72 -9.99 -0.25 -1.20 0.04 1.00 -0.00 3 -3.29 -0.01 0.94 0.32 3.13 0.98 -0.01 4 -4.01 0.17 -0.03 0.30 -0.47 0.99 0.02 5 -2.86 0.04 0.65 0.33 0.29 0.97 0.01 6 -2.86 0.32 -0.79 -0.09 0.09 0.97 0.03 7 -2.31 0.14 0.62 0.24 1.14 0.95 -0.02 8 -2.86 0.13 0.31 0.20 0.29 0.97 0.02 9 -2.31 0.26 -0.43 0.08 0.46 0.95 0.04
10 -1.50 0.28 -0.30 0.20 -0.07 0.90 0.05 11 -1.92 0.20 0.18 0.18 -1.02 0.93 0.02 12 -1.92 0.21 0.04 0.22 -1.02 0.93 0.03 13 -1.27 0.28 0.23 0.12 -0.04 0.88 -0.00 14 -1.07 0.27 0.46 0.09 1.43 0.86 -0.01 15 -1.17 0.32 0.52 -0.18 -0.37 0.87 -0.05 16 -1.50 0.28 0.32 -0.09 0.60 0.90 -0.00 17 -0.89 0.46 -0.69 -0.89 1.02 0.84 0.07 18 -1.27 0.34 -0.62 0.03 1.38 0.88 0.07 19 -0.30 0.39 -0.46 -0.03 -1.13 0.76 0.09 20 -0.18 0.41 1.29 -0.49 -1.19 0.74 -0.37 21 -0.44 0.40 -0.56 -0.16 -0.55 0.78 0.09 22 -0.44 0.55 -1.59 -1.45 1.85 0.78 0.19 23 -0.30 0.36 0.54 -0.02 -1.13 0.76 -0.09 24 -0.06 0.27 1.45 0.92 0.20 0.72 -0.27 25 -0.12 0.38 0.16 0.14 -0.15 0.73 -0.02 26 0.06 0.33 1.17 0.43 0.42 0.70 -0.25 27 0.17 0.29 1.22 1.03 1.24 0.68 -0.23 28 -0.18 0.48 -1.05 -0.85 0.44 0.74 0.17 29 -0.18 0.42 -0.53 -0.24 0.44 0.74 0.10 30 0.49 0.37 -0.03 0.70 -1.54 0.62 0.02 31 0.44 0.32 0.90 1.11 1.25 0.63 -0.23 32 0.54 0.47 -0.67 -0.64 -0.72 0.61 0.18 33 0.89 0.41 -0.03 0.28 -1.09 0.54 0.04 34 0.89 0.58 -2.13 -2.28 1.10 0.54 0.52 35 0.89 0.43 -0.27 0.00 -1.09 0.54 0.09 36 1.08 0.37 0.86 0.69 -0.53 0.50 -0.24 37 1.48 0.31 1.02 1.43 -0.55 0.42 -0.25 38 1.13 0.55 -1.66 -1.74 1.77 0.49 0.42 39 1.73 0.33 1.48 0.78 1.03 0.37 -0.34
(appendix continues)
![Page 143: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/143.jpg)
Appendix C (continued)
134
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
40 1.53 0.59 -2.12 -2.15 0.77 0.41 0.48 41 1.58 0.40 0.29 0.17 -1.16 0.40 -0.04 42 1.78 0.48 -0.84 -0.62 -0.85 0.36 0.20 43 2.23 0.31 1.49 0.53 -0.99 0.28 -0.29 44 2.23 0.33 0.82 0.54 -0.63 0.28 -0.14 45 2.68 0.36 -0.15 0.17 -1.45 0.21 0.04 46 2.17 0.38 0.37 0.19 -0.07 0.29 -0.07 47 2.91 0.35 0.05 0.02 -0.18 0.18 0.02 48 3.45 0.28 -0.39 0.44 1.55 0.12 0.06 49 4.74 -0.07 1.57 0.47 2.81 0.04 -0.09 50 4.11 0.28 -0.17 -0.07 0.01 0.07 0.03
Mean -0.09 0.33 0.06 -0.01 0.13 0.66 S.D. 2.12 0.15 0.89 0.75 1.09 Groups 2 Note: Raw score mean = 32.79 with a S.D. of 6. Mean person ability = 1.09 with a S.D. of 1.04. Test reliability (K.R. 20) = 0.85. Reliability of person separation = 0.85.
64.
![Page 144: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/144.jpg)
135
Appendix C (continued)
Experiment 27: Summary of Item Fit Information for a Normally Distributed Item Difficulty Distribution With 100 Items. 100 Persons, and a 50% Chance of Guessing Correctly
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
1 -4.18 -0.04 1.12 0.35 1.84 0.99 -0.03 2 -4.18 0.14 0.07 0.31 -0.36 0.99 0.02 3 -4.18 0.02 0.74 0.35 -0.36 0.99 0.01 4 -4.89 -9.99 -0.26 -1.20 0.04 1.00 -0.00 5 -4.89 -9.99 -0.26 -1.20 0.04 1.00 -0.00 6 -3.04 0.05 0.73 0.30 1.86 0.97 -0.01 7 -4.18 0.09 0.30 0.33 -0.36 0.99 0.02 8 -3.46 0.16 -0.13 0.23 -0.01 0.98 0.03 9 -2.27 0.19 1.40 0.08 -0.89 0.94 -0.22
10 -2.48 0.26 -0.33 0.01 -1.51 0.95 0.04 11 -2.73 0.03 0.72 0.39 1.22 0.96 -0.00 12 -2.10 0.24 0.42 -0.07 -0.22 0.93 -0.02 13 -3.04 0.14 0.94 0.09 -0.08 0.97 -0.07 14 -1.94 0.13 1.01 0.29 0.76 0.92 -0.07 15 -2.27 0.17 0.87 0.11 -0.89 0.94 -0.08 16 -1.80 0.44 -1.11 -0.68 0.13 0.91 0.07 17 -2.10 0.25 -0.24 0.11 -0.47 0.93 0.04 18 -2.10 0.38 -0.29 -0.54 -0.22 0.93 0.02 19 -1.80 0.14 0.44 0.57 0.38 0.91 0.00 20 -1.68 0.38 -0.75 -0.37 0.37 0.90 0.06 21 -1.68 0.19 1.16 0.10 1.13 0.90 -0.12 22 -2.27 0.26 -0.10 -0.04 0.22 0.94 0.02 23 -1.16 0.41 -0.64 -0.56 0.43 0.85 0.08 24 -1.80 0.19 1.40 0.06 -1.23 0.91 -0.20 25 -1.35 0.11 1.18 0.79 1.09 0.87 -0.10 26 -1.16 0.33 -0.45 0.02 -0.67 0.85 0.07 27 -1.80 0.21 0.77 0.03 1.47 0.91 -0.06 28 -1.16 0.20 0.48 0.62 1.33 0.85 -0.01 29 -1.16 0.26 0.74 0.07 0.48 0.85 -0.06 30 -1.16 0.43 -0.68 -0.69 0.43 0.85 0.07 31 -0.99 0.27 0.01 0.45 -0.16 0.83 0.03 32 -0.99 0.33 0.17 -0.13 -1.41 0.83 -0.01 33 -1.68 0.28 1.12 -0.27 -1.14 0.90 -0.17 34 -0.77 0.34 -0.21 0.04 -0.15 0.80 0.06 35 -1.08 0.32 -0.02 0.02 0.17 0.84 0.02 36 -0.99 0.09 0.74 1.53 0.76 0.83 -0.04 37 -0.91 0.33 0.05 -0.02 -0.89 0.82 0.01 38 -0.56 0.20 1.40 1.08 0.63 0.77 -0.24 39 -1.08 0.37 -0.27 -0.19 0.65 0.84 0.04
(appendix continues)
![Page 145: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/145.jpg)
136
Appendix C (continued)
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
40 -0.77 0.32 -0.10 0.28 -1.50 0.80 0.04 41 -0.84 0.35 -0.01 -0.18 0.17 0.81 0.04 42 -0.63 0.41 -0.53 -0.47 0.13 0.78 0.10 43 -0.25 0.42 -0.20 -0.60 -0.59 0.72 0.06 44 -0.70 0.29 0.25 0.43 0.42 0.79 -0.01 45 -0.20 0.34 0.44 0.19 1.05 0.71 -0.05 46 -0.84 0.35 -0.05 -0.12 0.44 0.81 0.01 47 -0.03 0.31 0.09 0.96 0.23 0.68 0.01 48 -0.31 0.38 -0.31 -0.02 -0.63 0.73 0.09 49 -0.09 0.48 -1.14 -0.95 0.32 0.69 0.23 50 -0.20 0.38 -0.10 0.01 -0.40 0.71 0.04 51 0.18 0.42 -0.45 -0.34 -0.48 0.64 0.13 52 -0.09 0.37 -0.15 0.24 -1.28 0.69 0.07 53 -0.03 0.42 -0.67 -0.26 1.18 0.68 0.15 54 -0.43 0.38 -0.20 -0.16 -0.94 0.75 0.05 55 0.23 0.34 0.85 0.45 -0.68 0.63 -0.21 56 -0.31 0.29 0.34 0.86 -0.99 0.73 -0.06 57 0.63 0.37 0.16 0.41 -0.42 0.55 -0.00 58 0.13 0.39 0.15 -0.07 -0.88 0.65 0.00 59 0.53 0.44 0.25 -0.74 -0.39 0.57 -0.12 60 0.13 0.48 -0.78 -1.12 1.26 0.65 0.17 61 0.43 0.39 0.06 0.05 -0.33 0.59 -0.01 62 0.53 0.40 -0.29 0.13 0.28 0.57 0.13 63 0.13 0.38 0.16 0.10 0.67 0.65 -0.04 64 0.43 0.41 -0.21 -0.17 -1.37 0.59 0.08 65 0.87 0.36 0.33 0.63 -0.06 0.50 -0.05 66 0.72 0.37 0.46 0.49 -0.42 0.53 -0.14 67 0.82 0.50 -1.37 -1.34 -0.42 0.51 0.39 68 1.01 0.48 -0.54 -1.10 1.51 0.47 0.13 69 0.63 0.32 1.21 0.97 -0.42 0.55 -0.40 70 0.72 0.47 -0.79 -0.87 0.29 0.53 0.23 71 1.25 0.36 0.36 0.56 -1.39 0.42 -0.10 72 1.06 0.34 1.02 0.86 -1.06 0.46 -0.33 73 1.11 0.38 0.46 0.26 -1.19 0.45 -0.16 74 1.20 0.30 1.69 1.13 0.08 0.43 -0.57 75 1.30 0.41 0.01 -0.07 -0.01 0.41 -0.01 76 0.91 0.44 -0.36 -0.54 -0.38 0.49 0.11 77 1.66 0.39 -0.16 0.21 0.90 0.34 0.01 78 1.20 0.45 -0.72 -0.47 -0.13 0.43 0.23 79 1.66 0.45 -0.48 -0.58 0.90 0.34 0.12 80 1.94 0.39 -0.17 0.14 -0.30 0.29 0.05 81 1.82 0.56 -1.50 -1.86 1.51 0.31 0.24 82 2.24 0.40 -0.35 -0.10 0.30 0.24 0.08
(appendix continues)
![Page 146: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/146.jpg)
Appendix C (continued)
137
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
83 1.66 0.39 -0.08 0.17 -1.15 0.34 0.06 84 2.18 0.42 -0.29 -0.26 0.37 0.25 0.04 85 2.05 0.34 0.99 0.24 0.28 0.27 -0.20 86 2.18 0.39 0.16 -0.11 -0.53 0.25 -0.04 87 2.05 0.30 0.91 0.59 1.62 0.27 -0.11 88 2.51 0.39 -0.33 -0.14 0.56 0.20 0.06 89 2.37 0.17 1.39 1.31 0.83 0.22 -0.18 90 2.75 0.34 -0.09 0.10 -0.73 0.17 0.04 91 2.67 0.33 0.14 0.20 -0.42 0.18 0.00 92 2.31 0.42 -0.31 -0.33 -1.34 0.23 0.06 93 2.67 0.28 -0.06 0.80 -0.42 0.18 0.02 94 3.45 0.41 -0.91 -0.30 1.17 0.10 0.07 95 3.21 0.44 -0.80 -0.52 -0.99 0.12 0.07 96 3.33 0.51 -1.37 -0.91 1.30 0.11 0.09 97 4.52 0.19 1.06 0.00 -0.27 0.04 -0.10 98 4.27 0.27 -0.41 0.13 0.46 0.05 0.04 99 4.84 0.16 0.03 0.29 0.08 0.03 0.03
100 4.52 0.39 -0.92 -0.26 0.28 0.04 0.04 Mean -0.10 0.32 0.08 0.01 0.00 0.63 S.D. 2.11 0.13 0.69 0.58 0.83 Groups 2 Note: Raw score mean = 63.26 with a S.D. of 12. Mean person ability = 0.88 with a S.D. of 1.01. Test reliability (K.R. 20) = 0.92. Reliability of person separation = 0.92.
86.
![Page 147: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/147.jpg)
138
Appendix C (continued)
Experiment 7R- Summary of Item Fit Information for a Uniformly Distributed Item Difficulty Distribution With 25 Items. 25 Persons, and No Guessing
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
1 -1.07 0.42 -0.24 -0.05 0.48 0.84 0.11 2 -0.46 0.43 -0.15 0.15 -0.68 0.76 0.12 3 -1.95 0.33 0.01 0.11 -0.10 0.92 0.07 4 -0.46 0.26 0.67 0.88 0.21 0.76 -0.13 5 -0.75 0.40 0.21 -0.02 -1.49 0.80 0.02 6 -0.20 0.60 -0.65 -1.03 -0.09 0.72 0.25 7 -1.45 0.22 0.72 0.27 0.17 0.88 -0.06 8 0.93 0.33 0.84 1.10 0.52 0.52 -0.35 9 -0.20 0.44 0.22 -0.13 -0.46 0.72 0.02
10 -0.75 0.33 0.46 0.15 0.82 0.80 -0.02 11 -0.46 0.44 0.25 -0.22 -0.68 0.76 -0.02 12 -0.20 0.64 -0.77 -1.38 -0.09 0.72 0.28 13 0.04 0.52 -0.02 -0.44 -1.45 0.68 0.09 14 0.27 0.71 -1.11 -1.79 0.92 0.64 0.44 15 0.50 0.55 -0.36 -0.36 0.16 0.60 0.23 16 0.04 0.61 -0.68 -0.91 0.43 0.68 0.27 17 0.93 0.43 0.29 0.56 -1.18 0.52 -0.04 18 0.72 0.38 0.73 0.65 -0.66 0.56 -0.31 19 1.15 0.34 0.84 1.07 -0.32 0.48 -0.34 20 0.27 0.32 0.57 1.05 -0.15 0.64 -0.16 21 0.93 0.47 0.01 0.34 0.52 0.52 0.10 22 0.72 0.32 0.89 1.07 -0.12 0.56 -0.39 23 0.04 0.64 -0.76 -1.26 0.43 0.68 0.31 24 0.93 0.54 -0.28 -0.26 -1.18 0.52 0.20 25 0.50 0.36 0.53 0.89 -1.05 0.60 -0.17
Mean S.D. Groups
0.00 0.80
0.44 0.13
0.09 0.58
0.02 0.82
-0.20 0.70 2
0.68
Note: Raw score mean = 16.88 and a S Mean person ability = 1.06 with a S.D. Test reliability (K.R. 20) = 0.85. Reliability of person separation = 0.85.
D. of 5.30. of 1.30.
![Page 148: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/148.jpg)
139
Appendix C (continued)
Experiment Summary of Item Fit Information for a Uniformly Distributed Item Difficulty Distribution With 50 Items. 25 Persons, and No Guessing
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
1 -2.37 0.07 0.54 0.39 -0.28 0.96 0.02 2 -0.74 -0.01 2.63 0.60 -1.38 0.84 -1.33 3 -1.59 -0.03 0.88 0.50 0.48 0.92 -0.06 4 -1.11 0.27 0.43 0.03 -0.39 0.88 -0.03 5 -0.74 0.21 0.55 0.26 -1.38 0.84 -0.06 6 -0.74 0.35 0.00 -0.07 -1.38 0.84 0.06 7 -0.74 0.18 0.26 0.54 -1.38 0.84 0.02 8 -0.43 0.12 0.49 0.86 -0.69 0.80 -0.05 9 -0.43 0.57 -0.86 -0.82 1.07 0.80 0.20
10 -0.74 0.10 0.72 0.61 0.90 0.84 -0.09 11 -0.17 0.43 -0.10 -0.31 -0.09 0.76 0.07 12 0.07 0.30 0.08 0.47 0.54 0.72 0.05 13 -1.11 0.31 0.14 0.05 -0.39 0.88 0.03 14 -1.11 0.42 -0.28 -0.21 -0.39 0.88 0.08 15 -0.74 0.36 -0.17 -0.03 -1.38 0.84 0.09 16 0.07 0.31 -0.14 0.58 0.39 0.72 0.08 17 -0.43 0.36 0.19 -0.14 -0.69 0.80 0.01 18 0.07 -0.04 1.73 1.56 1.77 0.72 -0.64 19 -0.17 0.32 0.17 0.14 1.12 0.76 0.03 20 0.90 0.16 1.31 1.46 0.76 0.56 -0.72 21 -0.17 0.45 -0.41 -0.29 -0.44 0.76 0.17 22 0.71 0.27 0.55 0.86 -1.05 0.60 -0.24 23 -0.43 0.35 -0.15 0.09 -0.69 0.80 0.08 24 -0.43 0.27 0.03 0.40 -0.69 0.80 0.05 25 -0.43 0.52 -0.67 -0.60 -0.69 0.80 0.18 26 -0.43 0.32 0.29 -0.05 0.24 0.80 -0.00 27 0.71 0.44 -0.14 -0.25 0.09 0.60 0.15 28 0.07 0.39 -0.20 0.05 -1.35 0.72 0.11 29 0.51 0.44 1.32 -0.81 -0.02 0.64 -0.89 30 0.51 0.52 -0.37 -0.84 1.24 0.64 0.19 31 -0.43 0.29 0.09 0.26 0.24 0.80 0.05 32 -0.17 0.25 1.06 0.17 -0.44 0.76 -0.33 33 0.51 0.54 -0.61 -0.96 -0.02 0.64 0.30 34 -0.74 0.11 1.05 0.49 -1.38 0.84 -0.23 35 1.29 0.40 0.14 0.13 -0.66 0.48 -0.02 36 0.07 0.32 -0.15 0.53 0.39 0.72 0.08 37 0.90 0.26 0.86 0.88 -0.64 0.56 -0.42 38 0.90 0.32 0.37 0.59 -0.64 0.56 -0.10 39 0.71 0.42 -0.33 0.01 0.09 0.60 0.23 40 0.51 0.60 -1.17 -1.20 -0.02 0.64 0.49
(appendix continues)
![Page 149: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/149.jpg)
Appendix C (continued)
140
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
41 0.90 0.31 0.42 0.72 0.76 0.56 -0.10 42 0.07 0.39 -0.26 0.05 0.39 0.72 0.12 43 0.30 0.34 -0.01 0.35 -0.08 0.68 0.07 44 0.07 0.63 -1.14 -1.31 0.39 0.72 0.35 45 2.10 0.62 -1.00 -1.18 0.43 0.32 0.37 46 1.48 0.49 -0.63 -0.39 -1.02 0.44 0.38 47 0.30 0.25 1.06 0.49 -0.67 0.68 -0.43 48 -0.17 0.56 -0.82 -0.80 1.36 0.76 0.24 49 1.68 0.57 -0.99 -0.91 0.00 0.40 0.46 50 1.29 0.61 -1.27 -1.36 -0.21 0.48 0.65
Mean S.D.
-0.00 0.86
0.34 0.17
0.11 0.78
0.03 0.68
-0.16 0.80
0.71
Groups Note: Raw score mean = 35.72 with a S.D. of 7 Mean person ability = 1.21 with a S.D. of 0.99. Test reliability (K.R. 20) = 0.87. Reliability of person separation = 0.85.
83.
![Page 150: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/150.jpg)
141
Appendix C (continued')
Experiment 30: Summary of Item Fit Information for a Uniformly Distributed Item Difficulty Distribution With 100 Items. 25 Persons, and No Guessing
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
1 -0.95 0.16 0.28 0.69 -0.07 0.80 0.00 2 -1.26 0.34 -0.09 0.02 -0.78 0.84 0.07 3 -0.95 0.50 -0.44 -0.59 -0.07 0.80 0.13 4 -0.03 0.46 -0.54 -0.45 1.11 0.64 0.31 5 -1.26 0.55 -0.70 -0.65 1.08 0.84 0.14 6 -1.26 -0.16 1.80 0.99 1.86 0.84 -0.51 7 -0.69 0.24 -0.01 0.64 0.48 0.76 0.05 8 -2.11 0.45 -0.47 -0.13 0.35 0.92 0.07 9 -0.45 0.46 -0.57 -0.36 -0.49 0.72 0.21
10 -0.95 -0.00 0.82 1.08 -0.42 0.80 -0.13 11 -1.63 0.07 0.47 0.56 -0.99 0.88 0.00 12 -0.45 0.33 -0.01 0.18 -0.21 0.72 0.06 13 -0.69 0.41 -0.28 -0.19 -1.57 0.76 0.14 14 -0.03 0.36 0.03 0.11 -0.16 0.64 0.09 15 -0.23 0.55 -0.92 -0.90 0.18 0.68 0.36 16 -0.45 0.46 -0.32 -0.50 -0.21 0.72 0.17 17 -0.45 0.38 -0.27 0.03 0.96 0.72 0.11 18 -0.45 0.40 -0.32 -0.10 0.96 0.72 0.17 19 -0.45 0.46 -0.30 -0.54 -0.21 0.72 0.16 20 0.36 0.40 -0.29 0.04 -0.22 0.56 0.24 21 -0.45 0.30 -0.05 0.38 -0.21 0.72 0.08 22 -0.69 0.54 -0.78 -0.73 1.74 0.76 0.23 23 -1.26 0.51 -0.37 -0.61 -0.78 0.84 0.10 24 -0.95 0.31 -0.11 0.21 -0.42 0.80 0.08 25 -0.03 0.37 -0.12 0.11 0.76 0.64 0.09 26 -1.63 0.36 -0.17 -0.04 -0.99 0.88 0.07 27 -0.03 0.61 -1.18 -1.47 0.76 0.64 0.51 28 -0.45 0.47 -0.52 -0.45 -0.21 0.72 0.22 29 -0.95 0.21 0.22 0.46 -0.42 0.80 0.02 30 -0.69 0.06 1.35 0.88 -1.57 0.76 -0.46 31 0.74 0.62 -1.56 -1.69 0.87 0.48 0.82 32 -0.23 0.57 -0.91 -1.14 -1.13 0.68 0.37 33 -0.23 0.52 -0.57 -0.87 0.18 0.68 0.26 34 -0.95 0.22 0.33 0.40 -0.42 0.80 -0.03 35 -0.69 0.62 -1.00 -1.15 0.48 0.76 0.27 36 -0.23 0.51 -0.75 -0.70 -1.13 0.68 0.32 37 -0.95 0.34 -0.04 0.04 -0.42 0.80 0.07 38 -0.45 0.36 -0.17 0.09 -0.49 0.72 0.11 39 -0.69 0.55 -0.66 -0.91 0.48 0.76 0.21 40 -0.45 -0.17 1.82 1.99 2.24 0.72 -0.71
(appendix continues')
![Page 151: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/151.jpg)
142
Appendix C (continued)
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
41 -0.23 0.35 -0.20 0.22 -1.13 0.68 0.12 42 0.55 0.36 0.59 0.06 0.17 0.52 -0.39 43 0.17 0.29 0.65 0.44 -1.08 0.60 -0.27 44 -0.45 0.53 -0.71 -0.76 0.96 0.72 0.26 45 0.17 0.32 0.59 0.22 -1.08 0.60 -0.24 46 -0.23 0.06 0.86 1.57 1.68 0.68 -0.29 47 -0.23 0.52 -0.63 -0.83 0.18 0.68 0.29 48 0.36 0.33 0.20 0.45 -0.22 0.56 -0.04 49 0.74 0.41 0.12 -0.26 0.87 0.48 -0.12 50 0.55 0.28 0.57 0.78 -1.26 0.52 -0.25 51 -0.03 0.48 -0.69 -0.51 -0.68 0.64 0.34 52 -0.23 0.21 1.64 0.40 0.53 0.68 -0.85 53 0.36 0.28 0.42 0.79 -0.22 0.56 -0.18 54 -0.45 0.54 -0.78 -0.84 -0.21 0.72 0.28 55 -0.23 0.59 -0.86 -1.37 0.18 0.68 0.35 56 0.17 0.60 -1.25 -1.50 0.09 0.60 0.61 57 0.55 0.55 -1.14 -1.12 -1.26 0.52 0.65 58 -0.23 0.05 2.73 0.86 -1.13 0.68 -1.65 59 -0.45 0.41 -0.32 -0.18 -0.49 0.72 0.15 60 0.93 0.52 -0.94 -0.71 0.40 0.44 0.53 61 -0.23 0.20 0.46 0.88 1.68 0.68 -0.09 62 -0.03 0.60 -1.09 -1.39 0.76 0.64 0.49 63 0.74 0.22 1.35 0.92 -0.39 0.48 -0.89 64 1.12 0.28 1.18 0.40 -0.71 0.40 -0.62 65 0.55 0.32 0.29 0.52 0.40 0.52 -0.10 66 0.74 -0.00 2.51 2.16 0.93 0.48 -1.77 67 0.17 0.33 -0.02 0.43 0.09 0.60 0.05 68 0.17 0.21 0.54 1.13 -1.08 0.60 -0.15 69 -0.45 0.57 -0.75 -1.05 0.96 0.72 0.27 70 -0.03 0.26 0.38 0.66 1.11 0.64 -0.10 71 1.12 0.44 -0.25 -0.25 -0.71 0.40 0.20 72 0.74 0.44 0.02 -0.45 -0.46 0.48 -0.07 73 -0.69 0.48 -0.37 -0.58 -1.57 0.76 0.15 74 0.74 0.34 -0.03 0.50 1.96 0.48 0.14 75 0.17 0.38 -0.26 0.15 0.50 0.60 0.22 76 -0.69 0.34 -0.18 0.19 0.48 0.76 0.10 77 0.36 0.49 -0.77 -0.67 -0.66 0.56 0.45 78 0.74 0.46 -0.25 -0.57 -0.46 0.48 0.15 79 1.12 0.17 0.98 1.32 0.76 0.40 -0.44 80 -0.03 0.20 0.97 0.85 -0.16 0.64 -0.44 81 0.74 0.55 -1.07 -1.15 0.87 0.48 0.60 82 0.93 0.56 -1.16 -1.11 0.40 0.44 0.62 83 1.53 0.52 -0.71 -0.58 0.40 0.32 0.29
(appendix continues)
![Page 152: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/152.jpg)
Appendix C (continued)
143
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
84 1.32 0.51 -0.68 -0.61 0.79 0.36 0.35 85 0.55 0.52 -0.71 -1.05 0.17 0.52 0.38 86 0.55 0.39 -0.24 0.10 -1.26 0.52 0.22 87 1.75 0.22 1.62 0.33 -0.55 0.28 -0.72 88 1.12 0.29 0.52 0.61 0.76 0.40 -0.15 89 0.93 0.19 1.54 1.07 0.23 0.44 -0.98 90 0.74 0.16 1.50 1.41 1.96 0.48 -0.93 91 0.55 0.29 0.58 0.66 0.17 0.52 -0.31 92 -0.03 -0.15 2.09 2.58 2.14 0.64 -1.06 93 0.36 0.20 0.98 1.13 -0.22 0.56 -0.47 94 1.53 0.36 -0.05 0.26 0.50 0.32 0.10 95 0.55 0.37 0.22 0.11 0.17 0.52 -0.11 96 1.12 0.33 0.41 0.34 -0.11 0.40 -0.21 97 0.74 0.31 0.77 0.36 0.87 0.48 -0.50 98 0.93 0.03 2.20 1.95 1.40 0.44 -1.38 99 0.17 0.46 -0.58 -0.46 0.09 0.60 0.36
100 1.53 0.44 -0.43 -0.14 0.40 0.32 0.18 Mean 0.00 0.36 0.05 0.03 0.10 0.62 S.D. 0.78 0.18 0.89 0.85 0.88 Groups 2 Note: Raw score mean = 62.36 with a S.D. of 17. Mean person ability = 0.66 with a S.D. of 0.95. Test reliability (K.R. 20) = 0.94. Reliability of person separation = 0.93.
01.
![Page 153: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/153.jpg)
144
Appendix C (continued)
Experiment 31: Summary of Item Fit Information for a Uniformly Distributed Item Difficulty Distribution With 25 Items. 50 Persons, and No Guessing
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
1 -0.94 0.27 -0.20 0.14 -0.35 0.86 0.07 2 -0.77 0.23 0.34 0.13 -0.83 0.84 -0.00 3 -1.36 0.31 -0.37 0.01 0.71 0.90 0.07 4 -0.21 0.41 -0.39 -0.43 1.04 0.76 0.11 5 -0.77 0.28 -0.18 0.13 -0.01 0.84 0.07 6 -0.09 0.41 -0.26 -0.55 -1.11 0.74 0.11 7 -0.94 0.20 0.43 0.24 -0.35 0.86 -0.02 8 -0.47 0.30 0.05 0.07 -0.85 0.80 0.02 9 -0.33 0.29 0.67 -0.08 -0.60 0.78 -0.13
10 0.25 0.33 -0.15 0.39 0.81 0.68 0.10 11 -0.09 0.26 0.11 0.65 -0.12 0.74 0.04 12 0.46 0.38 -0.20 -0.01 0.07 0.64 0.13 13 -0.47 0.35 -0.40 -0.06 -0.85 0.80 0.11 14 -0.09 0.38 -0.54 -0.08 1.27 0.74 0.15 15 0.25 0.39 -0.30 -0.23 -0.19 0.68 0.15 16 0.36 0.41 -0.34 -0.40 0.45 0.66 0.16 17 0.25 0.35 -0.07 0.09 -1.15 0.68 0.06 18 -0.21 0.36 -0.37 -0.06 1.04 0.76 0.12 19 0.25 0.07 1.38 2.15 1.64 0.68 -0.38 20 0.76 0.49 -0.85 -0.98 -0.01 0.58 0.35 21 0.56 0.36 -0.06 0.29 -0.37 0.62 0.09 22 0.66 0.36 0.16 0.21 -0.44 0.60 -0.01 23 0.85 0.22 1.54 1.46 0.39 0.56 -0.68 24 1.24 0.48 -0.42 -0.80 -0.43 0.48 0.14 25 0.85 0.49 -0.77 -0.99 0.38 0.56 0.34
Mean 0.00 0.33 -0.05 0.05 0.01 0.71 S.D. 0.66 0.10 0.57 0.67 0.76 Groups 2 Note: Raw score mean = 17.84 and a S. Mean person ability =1.18 with a S.D. Test reliability (K.R. 20) = 0.71. Reliability of person separation = 0.73.
D. of 3.90. of 1.02.
![Page 154: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/154.jpg)
145
Appendix C (continued)
Experiment 32: Summary of Item Fit Information for a Uniformly Distributed Item Difficulty Distribution With 50 Items. 50 Persons, and No Guessing
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
1 -0.45 0.24 0.81 0.97 1.66 0.76 -0.12 2 -0.73 0.51 -0.49 -0.80 -1.55 0.80 0.14 3 -0.89 0.41 -0.17 -0.10 -0.03 0.82 0.09 4 -1.69 0.37 0.15 -0.26 -0.83 0.90 0.03 5 -0.89 0.49 -0.41 -0.69 -0.03 0.82 0.11 6 -0.58 0.31 -0.00 0.80 0.55 0.78 0.01 7 -0.31 0.19 0.74 1.63 0.28 0.74 -0.18 8 -0.73 0.26 0.59 0.74 0.30 0.80 -0.06 9 -0.19 0.44 -0.11 -0.15 -1.06 0.72 0.10
10 -0.07 0.45 -0.36 -0.11 -0.53 0.70 0.16 11 -0.89 0.35 0.45 0.09 -0.80 0.82 -0.06 12 -0.45 0.41 -0.06 0.05 -0.40 0.76 0.08 13 -0.89 0.51 -0.60 -0.74 -0.03 0.82 0.14 14 -0.45 0.28 1.00 0.53 0.68 0.76 -0.21 15 -0.19 0.43 -0.28 0.00 -1.06 0.72 0.12 16 -0.73 0.24 0.91 0.73 0.30 0.80 -0.16 17 -0.31 0.36 0.02 0.56 -1.19 0.74 0.06 18 -0.73 0.40 0.29 -0.26 0.30 0.80 -0.01 19 -0.45 0.18 1.24 1.26 1.66 0.76 -0.27 20 -0.07 0.36 0.38 0.41 -0.62 0.70 -0.03 21 0.05 0.43 -0.05 -0.05 -0.10 0.68 0.07 22 -0.07 0.48 0.00 -0.63 -0.53 0.70 0.04 23 -0.45 0.34 0.39 0.34 1.66 0.76 -0.01 24 -0.89 0.44 -0.45 -0.09 1.24 0.82 0.09 25 -0.73 0.48 0.08 -0.74 0.28 0.80 0.01 26 0.38 0.57 -0.98 -1.47 0.01 0.62 0.37 27 0.58 0.50 -0.48 -0.64 -0.18 0.58 0.23 28 0.27 0.59 -1.07 -1.67 1.47 0.64 0.37 29 0.27 0.50 -0.05 -0.92 0.62 0.64 -0.07 30 -0.19 0.12 1.09 2.23 2.59 0.72 -0.23 31 0.48 0.48 0.04 -0.64 0.40 0.60 0.02 32 -0.31 0.46 -0.43 -0.21 -0.03 0.74 0.14 33 0.27 0.44 0.29 -0.30 0.62 0.64 -0.19 34 0.58 0.37 0.36 0.77 0.96 0.58 -0.04 35 0.48 0.43 0.07 0.09 -0.70 0.60 0.06 36 0.16 0.46 -0.40 -0.20 -0.98 0.66 0.19 37 0.16 0.63 -1.34 -2.05 1.98 0.66 0.41 38 0.79 0.36 1.28 0.48 -0.30 0.54 -0.76 39 1.09 0.27 1.13 1.77 1.40 0.48 -0.39 40 0.79 0.57 -1.09 -1.46 -0.30 0.54 0.45
(appendix continues)
![Page 155: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/155.jpg)
Appendix C (continued^
146
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit 41 0.69 0.55 -0.80 -1.19 0.26 42 0.16 0.44 -0.06 -0.07 -0.29 43 0.79 0.39 0.63 0.40 -0.30 44 0.79 0.48 -0.52 -0.32 -1.11 45 0.99 0.25 1.21 2.02 1.77 46 0.79 0.39 0.27 0.61 -1.11 47 1.09 0.40 0.02 0.57 0.60 48 1.19 0.59 -1.22 -1.81 0.68 49 0.89 0.53 -0.69 -1.10 0.19 50 0.58 0.37 0.38 0.77 0.05
Mean S.D. Groups
-0.00 0.67
0.41 0.11
0.03 0.66
-0.02 0.95
0.17 0.94 2
Mean item score
0.56 0.66 0.54 0.54 0.50 0.54 0.48 0.46 0.52 0.58
Logit Residual
Index
0.34 0.10
-0.25 0.27
-0.42 -0.03 0.09 0.47 0.31
-0.06
Note: Raw score mean = 33.90 with a S.D. of 9.85. Mean person ability = 1.00 with a S.D. of 1.22. Test reliability (K.R. 20) = 0.91. Reliability of person separation = 0.91.
0.68
![Page 156: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/156.jpg)
147
Appendix C (continued)
Experiment 33: Summary of Item Fit Information for a Uniformly Distributed Item Difficulty Distribution With 100 Items. 50 Persons, and No Guessing
Item Logit Point. Unwt. Wt. Ability Mean Logit
Item item bis. total total between item Residual # diff. corr. fit fit fit score Index 1 -1.66 0.14 1.96 0.30 -1.29 0.90 -0.48 2 -0.73 0.46 -0.66 -0.31 0.52 0.80 0.13 3 -0.88 0.34 -0.17 0.30 0.23 0.82 0.06 4 -0.73 0.19 1.73 0.73 -0.06 0.80 -0.42 5 -1.43 0.24 0.01 0.52 -0.96 0.88 0.05 6 -1.05 0.38 -0.47 0.10 1.22 0.84 0.09 7 -0.45 0.43 -0.26 -0.12 -1.14 0.76 0.08 8 -0.45 0.37 0.04 0.25 -1.14 0.76 0.04 9 -0.45 0.34 0.19 0.39 -1.14 0.76 -0.01
10 -0.32 0.46 -0.34 -0.39 -1.09 0.74 0.09 11 -0.88 0.38 -0.01 -0.05 0.23 0.82 0.04 12 -0.88 0.17 0.66 1.05 0.38 0.82 -0.06 13 -1.43 0.28 0.29 0.14 0.33 0.88 0.01 14 -0.20 0.39 0.30 0.07 0.57 0.72 -0.02 15 -1.05 0.13 1.30 0.84 1.93 0.84 -0.19 16 -0.73 0.27 0.44 0.59 -0.94 0.80 -0.03 17 -1.23 0.22 0.49 0.57 -0.47 0.86 -0.02 18 -1.66 0.24 0.33 0.26 -1.29 0.90 0.00 19 -0.88 0.50 -0.84 -0.62 0.23 0.82 0.13 20 -0.88 0.30 0.65 0.18 0.38 0.82 -0.08 21 -1.05 0.41 -0.40 -0.13 -0.09 0.84 0.08 22 -0.20 0.43 -0.34 0.04 -0.53 0.72 0.13 23 -0.73 0.17 0.75 1.07 2.93 0.80 -0.05 24 -0.20 0.45 -0.39 -0.20 0.57 0.72 0.13 25 -0.32 0.52 -0.75 -0.83 -1.09 0.74 0.18 26 -1.05 0.38 -0.23 -0.05 -0.68 0.84 0.07 27 -0.88 0.22 0.46 0.79 -1.36 0.82 -0.03 28 -0.59 0.40 0.15 -0.11 -0.45 0.78 -0.00 29 -0.73 0.40 -0.10 -0.04 -0.94 0.80 0.04 30 -0.32 0.37 -0.10 0.39 -0.12 0.74 0.06 31 -0.20 0.42 -0.22 0.01 -0.53 0.72 0.08 32 -0.08 0.59 -1.31 -1.25 0.92 0.70 0.31 33 -0.45 0.28 0.47 0.76 0.32 0.76 -0.05 34 0.03 0.30 0.75 0.97 -0.30 0.68 -0.17 35 -0.45 0.39 -0.06 0.10 -1.14 0.76 0.06 36 -0.45 0.50 -0.89 -0.48 1.94 0.76 0.18 37 -0.08 0.56 -0.84 -1.16 -1.26 0.70 0.23 38 -0.45 0.28 1.07 0.48 0.32 0.76 -0.20 39 -0.73 0.41 -0.31 -0.12 -0.94 0.80 0.09 40 -0.08 0.37 -0.02 0.44 0.15 0.70 0.05
(appendix continues")
![Page 157: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/157.jpg)
148
Appendix C (continued!
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index 41 -0.59 0.45 -0.48 -0.22 0.79 0.78 0.10 42 -0.32 0.47 -0.47 -0.33 1.29 0.74 0.12 43 0.03 0.60 -1.14 -1.49 0.29 0.68 0.30 44 0.14 0.59 -1.33 -1.36 0.64 0.66 0.33 45 0.14 0.27 1.46 1.06 1.23 0.66 -0.44 46 0.36 0.54 -0.67 -0.96 1.29 0.62 0.19 47 0.36 0.55 -1.03 -0.97 -0.65 0.62 0.29 48 0.25 0.58 -1.18 -1.27 0.97 0.64 0.35 49 0.66 0.51 -0.53 -0.71 -0.23 0.56 0.18 50 -0.08 0.35 0.44 0.45 1.12 0.70 -0.04 51 -0.32 0.47 -0.32 -0.42 -1.09 0.74 0.10 52 0.46 0.63 -1.63 -1.82 0.81 0.60 0.50 53 0.03 0.44 -0.34 -0.03 -0.30 0.68 0.11 54 -0.08 0.40 -0.05 0.26 -1.26 0.70 0.03 55 -0.08 0.44 0.66 -0.43 -0.10 0.70 -0.21 56 -0.32 0.26 0.92 0.84 -0.12 0.74 -0.19 57 0.36 0.24 2.09 1.34 1.30 0.62 -0.77 58 -0.32 0.46 -0.57 -0.22 -1.09 0.74 0.13 59 -0.08 0.49 -0.53 -0.46 -1.26 0.70 0.17 60 0.03 0.38 0.32 0.36 -0.96 0.68 -0.07 61 0.46 0.60 -1.51 -1.39 2.30 0.60 0.45 62 -0.20 0.53 -0.68 -0.90 -0.53 0.72 0.18 63 0.25 0.52 -0.80 -0.65 -1.38 0.64 0.27 64 0.36 0.40 0.78 0.05 -0.65 0.62 -0.29 65 -0.08 0.33 0.45 0.66 0.15 0.70 -0.07 66 0.96 0.57 -1.33 -1.11 0.28 0.50 0.48 67 0.56 0.54 -1.04 -0.75 -0.80 0.58 0.37 68 0.66 0.38 1.13 0.38 -0.23 0.56 -0.44 69 0.66 0.50 -0.50 -0.60 -0.23 0.56 0.18 70 0.77 0.49 -0.21 -0.55 -0.87 0.54 0.08 71 -0.08 0.49 -0.78 -0.34 0.92 0.70 0.19 72 0.46 0.28 1.34 1.37 0.92 0.60 -0.47 73 0.56 0.33 0.68 1.04 2.06 0.58 -0.14 74 0.86 0.59 -1.40 -1.39 -0.24 0.52 0.49 75 0.03 0.63 -1.59 -1.65 1.99 0.68 0.37 76 0.46 0.39 0.13 0.61 0.00 0.60 0.02 77 0.25 0.46 0.32 -0.34 -1.38 0.64 -0.12 78 0.77 0.35 1.20 0.75 0.53 0.54 -0.47 79 1.36 0.47 -0.56 -0.22 0.33 0.42 0.24 80 0.36 0.27 0.88 1.52 0.45 0.62 -0.24 81 0.77 0.35 1.05 0.76 -0.48 0.54 -0.42 82 1.57 0.33 0.59 0.94 0.16 0.38 -0.13 83 0.66 0.37 0.66 0.63 -1.23 0.56 -0.23
(appendix continues)
![Page 158: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/158.jpg)
Appendix C (continued)
149
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
84 0.86 0.22 2.00 1.91 0.95 0.52 -0.79 85 0.25 0.65 -1.87 -1.91 0.97 0.64 0.49 86 0.66 0.40 0.53 0.29 -1.23 0.56 -0.17 87 0.86 0.34 0.95 0.93 -1.22 0.52 -0.35 88 0.96 0.38 0.72 0.59 0.51 0.50 -0.26 89 0.46 0.32 0.78 1.00 -1.40 0.60 -0.24 90 1.26 0.53 -1.02 -0.79 -0.04 0.44 0.37 91 1.06 0.42 0.46 0.10 0.01 0.48 -0.18 92 1.06 0.46 -0.27 -0.11 -1.33 0.48 0.12 93 0.56 0.38 0.68 0.57 -0.52 0.58 -0.23 94 0.77 0.47 -0.28 -0.31 -0.87 0.54 0.12 95 0.96 0.46 0.22 -0.24 -0.84 0.50 -0.11 96 0.77 0.28 1.60 1.29 -0.48 0.54 -0.63 97 1.06 0.61 -1.64 -1.64 0.76 0.48 0.56 98 0.36 0.45 -0.48 0.11 -0.64 0.62 0.18 99 0.46 0.29 1.10 1.25 -1.40 0.60 -0.33
100 1.47 0.38 0.97 0.37 0.72 0.40 -0.34 Mean -0.00 0.40 0.02 0.01 -0.07 0.67 S.D. 0.72 0.12 0.88 0.81 0.98 Groups 2 Note: Raw score mean = 67.44 with a S.D. of 18.91. Mean person ability = 0.96 with a S.D. of 1.08. Test reliability (K.R. 20) = 0.95. Reliability of person separation = 0.94.
![Page 159: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/159.jpg)
150
Appendix C (continued)
Experiment 34: Summary of Item Fit Information for a Uniformly Distributed Item Difficulty Distribution With 25 Items. 100 Persons, and No Guessing
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
1 -0.52 0.37 -0.29 0.06 -0.04 0.80 0.06 2 -1.02 0.30 -0.17 0.25 -0.70 0.86 0.05 3 -1.11 0.38 -0.69 -0.26 0.52 0.87 0.08 4 -1.33 0.32 0.48 -0.29 -1.36 0.89 -0.06 5 -0.68 0.47 -0.63 -1.01 -0.58 0.82 0.08 6 -0.84 0.50 -1.47 -0.94 1.82 0.84 0.13 7 -1.44 0.28 -0.13 0.05 -0.85 0.90 0.04 8 -0.06 0.23 2.44 1.29 1.16 0.73 -0.58 9 -0.60 0.35 0.30 -0.01 0.66 0.81 -0.03
10 0.00 0.44 -0.19 -0.48 0.12 0.72 0.05 11 -0.52 0.30 0.19 0.55 -1.35 0.80 0.01 12 0.18 0.44 -0.16 -0.44 0.13 0.69 0.06 13 -0.06 0.47 -1.02 -0.59 1.33 0.73 0.17 14 -0.38 0.38 0.37 -0.27 -0.89 0.78 -0.04 15 0.40 0.44 -0.47 -0.12 -0.29 0.65 0.13 16 0.40 0.37 0.93 0.48 -1.42 0.65 -0.22 17 0.00 0.20 1.26 2.06 2.66 0.72 -0.17 18 0.71 0.41 0.33 0.28 -1.02 0.59 -0.06 19 0.66 0.48 -0.69 -0.43 -0.32 0.60 0.18 20 0.66 0.41 -0.06 0.58 -0.45 0.60 0.03 21 0.97 0.61 -2.07 -2.57 1.79 0.54 0.49 22 1.12 0.38 0.82 0.78 -0.43 0.51 -0.22 23 1.07 0.38 0.56 1.04 1.19 0.52 -0.09 24 1.27 0.43 0.34 0.16 -0.83 0.48 -0.10 25 1.12 0.40 0.28 0.67 -0.43 0.51 -0.03
Mean 0.00 0.39 0.01 0.03 0.02 0.70 S.D. 0.82 0.09 0.90 0.88 1.11 Groups 2 Note: Raw score mean = 17.661 and a S.D. of 4.42. Mean person ability =1.18 with aS.D.of l . l l . Test reliability (K.R. 20) = 0.79. Reliability of person separation = 0.76.
![Page 160: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/160.jpg)
151
Appendix C (continued)
Experiment 35- Summary of Ttem Fit Information for a Uniformly Distributed Item Difficulty Distribution With 50 Items. 100 Persons, and No Guessing
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
1 -0.73 0.42 -0.22 -0.26 -1.29 0.79 0.06 2 -0.88 0.36 0.43 0.07 0.34 0.81 -0.03 3 -1.04 0.32 0.63 0.43 -1.30 0.83 -0.08 4 -0.45 0.31 1.11 0.77 1.17 0.75 -0.17 5 -1.22 0.33 -0.12 0.32 -0.75 0.85 0.04 6 -0.66 0.31 0.09 0.99 0.41 0.78 0.02 7 -0.96 0.40 -0.03 -0.16 -0.87 0.82 0.00 8 -0.09 0.34 2.57 0.76 0.37 0.69 -0.88 9 -0.26 0.40 0.17 0.17 -0.43 0.72 0.01
10 -0.88 0.38 0.25 -0.13 -0.78 0.81 -0.01 11 -1.32 0.12 2.74 0.89 3.22 0.86 -0.53 12 -0.20 0.37 0.14 0.60 -0.79 0.71 -0.00 13 -0.88 0.44 -0.64 -0.37 0.50 0.81 0.10 14 -0.52 0.37 -0.17 0.57 -0.15 0.76 0.06 15 -0.66 0.48 0.33 -1.03 -0.85 0.78 -0.07 16 -0.80 0.41 -0.16 -0.19 -1.19 0.80 0.04 17 -0.88 0.29 0.49 0.72 -0.78 0.81 -0.02 18 -0.66 0.43 0.38 -0.41 -0.85 0.78 -0.07 19 -0.33 0.43 -0.16 -0.01 -0.38 0.73 0.05 20 -0.33 0.49 0.43 -0.92 1.18 0.73 -0.15 21 -0.09 0.34 0.93 0.92 0.37 0.69 -0.17 22 0.14 0.45 -0.45 -0.12 -0.95 0.65 0.15 23 0.30 0.54 -0.68 -1.50 0.88 0.62 0.15 24 -0.20 0.48 -0.80 -0.61 -0.78 0.71 0.18 25 -0.03 0.50 -0.90 -0.69 -0.78 0.68 0.20 26 -0.39 0.47 -0.46 -0.64 0.26 0.74 0.09 27 -0.52 0.45 -0.49 -0.30 0.63 0.76 0.09 28 0.25 0.49 -0.74 -0.60 -0.06 0.63 0.16 29 0.51 0.47 -0.47 -0.32 -0.75 0.58 0.15 30 0.08 0.37 0.47 0.87 0.95 0.66 -0.08 31 0.35 0.48 -0.50 -0.54 -0.24 0.61 0.13 32 0.46 0.38 1.91 0.69 0.38 0.59 -0.69 33 0.30 0.46 -0.21 -0.30 0.88 0.62 0.07 34 0.56 0.36 1.12 1.11 -1.26 0.57 -0.28 35 0.41 0.56 -1.66 -1.63 -0.78 0.60 0.38 36 -0.20 0.54 -1.42 -1.18 2.10 0.71 0.23 37 0.76 0.45 0.02 0.08 -0.11 0.53 0.00 38 0.61 0.36 0.72 1.24 0.80 0.56 -0.12 39 0.35 0.43 -0.33 0.32 0.27 0.61 0.14 40 0.66 0.55 -1.48 -1.39 0.14 0.55 0.39
(appendix continues")
![Page 161: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/161.jpg)
152
Appendix C (continued)
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
41 0.91 0.41 0.36 0.56 -1.41 0.50 -0.06 42 0.76 0.49 -0.89 -0.38 -0.11 0.53 0.28 43 0.76 0.36 1.11 1.28 -0.11 0.53 -0.28 44 0.91 0.43 0.04 0.48 -0.45 0.50 0.05 45 0.86 0.38 0.65 1.07 -0.94 0.51 -0.13 46 1.01 0.42 0.98 0.11 0.97 0.48 -0.31 47 0.81 0.46 0.06 -0.24 -0.32 0.52 -0.02 48 0.96 0.51 -0.95 -0.92 -0.91 0.49 0.26 49 1.27 0.44 0.55 0.01 0.11 0.43 -0.14 50 1.17 0.51 -1.08 -0.85 0.71 0.45 0.30
Mean -0.00 0.42 0.07 -0.01 -0.07 0.66 S.D. 0.70 0.08 0.90 0.75 0.93 Groups 2 Note: Raw score mean = 33.23 with a S.D. of 9.92. Mean person ability = 0.92 with a S.D. of 1.18. Test reliability (K.R. 20) = 0.91. Reliability of person separation = 0.89.
![Page 162: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/162.jpg)
153
Appendix C (continued)
Experiment 36: Summary of Ttem Fit Information for a Uniformly Distributed Item Difficulty Distribution With 100 Items. 100 Persons, and No Guessing
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
1 -0.50 0.45 -0.36 -0.13 0.28 0.77 0.09 2 -1.36 0.42 -0.57 -0.03 0.24 0.87 0.07 3 -0.97 0.42 0.70 -0.45 0.09 0.83 -0.10 4 -1.25 0.42 0.24 -0.41 -0.94 0.86 -0.01 5 -1.71 0.46 -0.74 -0.53 -0.46 0.90 0.06 6 -1.36 0.41 -0.50 -0.05 0.24 0.87 0.08 7 -1.15 0.22 0.29 1.42 -0.60 0.85 0.00 8 -0.88 0.30 0.47 0.92 0.79 0.82 -0.05 9 -1.06 0.40 0.66 -0.25 0.38 0.84 -0.06
10 -1.36 0.40 -0.13 -0.21 0.14 0.87 0.05 11 -1.15 0.26 2.22 0.53 0.66 0.85 -0.41 12 -0.65 0.43 0.09 -0.20 -0.04 0.79 0.02 13 -0.65 0.47 -0.38 -0.53 -1.39 0.79 0.08 14 -0.43 0.35 0.70 0.66 -1.07 0.76 -0.09 15 -0.57 0.24 1.36 1.43 0.60 0.78 -0.18 16 -0.65 0.41 0.36 -0.09 0.86 0.79 -0.01 17 -0.72 0.35 0.50 0.57 0.25 0.80 -0.05 18 -0.37 0.34 0.85 0.86 0.62 0.75 -0.13 19 -0.30 0.37 1.80 0.26 -0.56 0.74 -0.41 20 -0.57 0.40 0.25 0.12 0.60 0.78 -0.01 21 -0.43 0.45 -0.20 -0.22 0.49 0.76 0.05 22 -0.72 0.47 -0.38 -0.46 -0.45 0.80 0.07 23 -0.17 0.30 1.10 1.32 1.29 0.72 -0.18 24 -0.80 0.43 -0.28 -0.19 0.52 0.81 0.07 25 -0.50 0.45 -0.18 -0.25 -0.67 0.77 0.06 26 -0.43 0.37 0.27 0.60 0.05 0.76 -0.01 27 -0.43 0.43 -0.19 0.06 0.05 0.76 0.06 28 -0.43 0.40 -0.06 0.39 0.05 0.76 0.02 29 -0.30 0.44 0.02 -0.16 -0.56 0.74 0.04 30 -0.05 0.50 -0.32 -0.80 1.57 0.70 0.02 31 -0.72 0.47 -0.70 -0.42 -0.45 0.80 0.12 32 -0.50 0.36 0.54 0.58 -0.67 0.77 -0.07 33 -0.57 0.45 -0.31 -0.26 -1.13 0.78 0.05 34 0.01 0.46 -0.57 -0.06 -1.42 0.69 0.15 35 -0.43 0.45 -0.56 -0.06 0.49 0.76 0.11 36 -0.30 0.51 -0.59 -0.85 0.09 0.74 0.12 37 -0.24 0.45 -0.27 -0.01 -0.59 0.73 0.08 38 -0.50 0.36 0.66 0.57 -0.67 0.77 -0.07 39 -0.37 0.47 -0.29 -0.41 -0.15 0.75 0.08 40 -0.05 0.55 -1.10 -1.29 0.23 0.70 0.21
(appendix continues)
![Page 163: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/163.jpg)
154
Appendix C (continued')
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
41 0.06 0.48 -0.57 -0.25 -0.06 0.68 0.15 42 -0.24 0.37 0.64 0.53 0.86 0.73 -0.06 43 -0.05 0.45 0.35 -0.35 -0.67 0.70 -0.08 44 -0.50 0.47 -0.44 -0.38 -0.74 0.77 0.06 45 0.06 0.50 -0.80 -0.60 0.29 0.68 0.20 46 -0.05 0.39 1.23 0.34 -0.91 0.70 -0.25 47 -0.05 0.49 -0.93 -0.39 -0.91 0.70 0.19 48 -0.11 0.44 -0.13 -0.07 -0.53 0.71 0.03 49 0.06 0.35 0.52 1.32 0.98 0.68 -0.03 50 -0.24 0.53 -1.01 -1.04 -0.59 0.73 0.19 51 -0.11 0.40 0.94 0.31 0.34 0.71 -0.19 52 0.01 0.26 1.40 1.97 2.36 0.69 -0.27 53 0.06 0.55 -1.13 -1.34 0.68 0.68 0.23 54 -0.11 0.52 -0.80 -0.85 -0.02 0.71 0.16 55 0.18 0.40 0.10 0.73 1.12 0.66 0.04 56 -0.24 0.35 0.54 0.87 0.09 0.73 -0.04 57 0.06 0.44 -0.32 0.30 -0.56 0.68 0.10 58 0.56 0.40 0.80 0.67 -0.13 0.59 -0.19 59 0.34 0.39 0.75 0.67 -0.27 0.63 -0.15 60 -0.05 0.49 -0.45 -0.65 -0.67 0.70 0.11 61 0.61 0.41 0.54 0.61 -0.48 0.58 -0.20 62 0.34 0.48 0.06 -0.52 -1.36 0.63 -0.07 63 0.56 0.49 -0.35 -0.58 0.16 0.59 0.11 64 -0.17 0.49 -0.79 -0.45 -1.45 0.72 0.12 65 -0.05 0.42 0.12 0.24 0.06 0.70 0.03 66 0.61 0.35 1.73 1.11 -0.48 0.58 -0.55 67 0.34 0.29 3.21 1.50 0.99 0.63 -1.06 68 0.23 0.45 -0.04 0.13 1.30 0.65 0.03 69 0.45 0.46 -0.52 0.13 0.31 0.61 0.14 70 0.29 0.44 0.20 -0.01 -0.60 0.64 -0.02 71 0.91 0.55 -1.37 -1.54 0.20 0.52 0.36 72 0.23 0.48 -0.13 -0.46 -1.00 0.65 0.05 73 0.34 0.38 1.37 0.75 0.34 0.63 -0.38 74 0.50 0.40 1.13 0.52 -0.63 0.60 -0.34 75 0.91 0.50 -0.84 -0.57 -1.20 0.52 0.26 76 0.23 0.39 2.20 0.43 -0.04 0.65 -0.77 77 0.61 0.43 0.31 0.32 -1.35 0.58 -0.06 78 0.66 0.47 -0.20 -0.22 0.03 0.57 0.05 79 0.61 0.40 0.32 0.87 -0.48 0.58 -0.03 80 0.56 0.44 -0.12 0.38 0.57 0.59 0.09 81 0.71 0.40 0.83 0.71 -1.51 0.56 -0.21 82 0.56 0.52 -1.17 -0.85 0.16 0.59 0.30 83 1.01 0.53 -1.21 -1.06 0.16 0.50 0.34
(appendix continues)
![Page 164: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/164.jpg)
Appendix C (continued)
155
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
84 0.66 0.53 -0.97 -1.19 1.28 0.57 0.25 85 0.29 0.51 -0.76 -0.82 0.90 0.64 0.19 86 0.34 0.37 1.03 0.85 -1.36 0.63 -0.29 87 1.32 0.52 -0.83 -1.05 0.86 0.44 0.23 88 0.96 0.54 -1.26 -1.27 -0.18 0.51 0.34 89 1.01 0.39 0.43 0.99 -1.14 0.50 -0.05 90 0.50 0.37 0.61 1.25 1.94 0.60 -0.05 91 0.91 0.53 -1.07 -1.23 0.20 0.52 0.29 92 1.17 0.47 -0.40 -0.18 -0.62 0.47 0.15 93 1.32 0.48 -0.45 -0.54 0.22 0.44 0.13 94 0.81 0.45 0.38 -0.10 -0.51 0.54 -0.12 95 0.40 0.43 0.21 0.35 -0.85 0.62 -0.01 96 1.22 0.57 -1.69 -1.89 0.81 0.46 0.43 97 1.22 0.55 -0.83 -1.78 1.38 0.46 0.18 98 1.07 0.49 0.31 -0.66 -0.62 0.49 -0.20 99 1.07 0.46 0.03 -0.23 -0.21 0.49 -0.02
100 0.71 0.44 1.39 -0.19 0.95 0.56 -0.63 Mean 0.00 0.43 0.07 -0.02 -0.04 0.68 S.D. 0.68 0.07 0.86 0.76 0.80 Groups 2 Note: Raw score mean = 68.05 with a S.D. of 20. Mean person ability = 1.00 with a S.D. of 1.20. Test reliability (K.R. 20) = 0.96. Reliability of person separation = 0.95.
13.
![Page 165: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/165.jpg)
156
Appendix C (continued)
Experiment 37: Summary of Item Fit Information for a Uniformly Distributed Item Difficulty Distribution With 25 Items. 25 Persons, and a 25% Chance of Guessing Correctly
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
1 -0.64 0.60 -0.41 -0.77 0.70 0.80 0.16 2 -1.43 0.21 0.46 1.07 0.10 0.88 0.05 3 -1.43 0.59 -0.26 -0.95 0.10 0.88 0.09 4 -1.43 0.24 1.54 0.33 0.46 0.88 -0.47 5 -1.00 0.38 0.23 0.46 -0.28 0.84 0.06 6 -0.04 0.56 -0.12 -0.48 -0.27 0.72 0.16 7 -0.64 0.42 0.03 0.55 0.70 0.80 0.08 8 0.23 0.31 0.78 1.57 1.93 0.68 -0.24 9 -0.04 0.63 -0.56 -1.02 -0.19 0.72 0.24
10 -0.64 0.41 0.63 0.16 -1.11 0.80 -0.05 11 -0.32 0.43 0.47 0.28 0.41 0.76 -0.01 12 -0.32 0.57 -0.34 -0.47 -0.87 0.76 0.18 13 -0.64 0.52 0.04 -0.40 -1.11 0.80 0.10 14 -1.98 0.14 0.70 0.87 -0.22 0.92 0.05 15 -0.04 0.63 -0.56 -1.02 -0.19 0.72 0.24 16 1.18 0.60 -0.30 -0.49 0.10 0.52 0.21 17 -0.04 0.58 -0.41 -0.52 -0.19 0.72 0.21 18 0.23 0.39 0.36 1.14 0.68 0.68 -0.09 19 0.23 0.69 -0.93 -1.59 1.67 0.68 0.34 20 1.89 0.62 -0.61 -0.44 -0.04 0.40 0.32 21 1.41 0.30 0.98 2.19 2.36 0.48 -0.28 22 0.23 0.51 -0.03 0.12 0.36 0.68 0.02 23 2.14 0.47 0.47 0.58 1.57 0.36 -0.07 24 1.65 0.66 -0.85 -0.82 0.44 0.44 0.40 25 1.41 0.51 0.46 0.25 -0.61 0.48 -0.17
Mean -0.00 0.48 0.07 0.02 0.26 0.70 S.D. 1.11 0.15 0.61 0.91 0.89 Groups 2 Note: Raw score mean = 17.40 and a S. Mean person ability = 1.32 with a S.D. Test reliability (K.R. 20) = 0.88. Reliability of person separation = 0.86.
D. of 5.52. of 1.60.
![Page 166: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/166.jpg)
157
Appendix C (continued)
Experiment 38- Summary of Ttem Fit Information for a Uniformly Distributed Item Difficulty Distribution With 50 Items. 25 Persons, and a 25% Chance of guessing Correctly
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
1 -0.46 0.44 -0.40 -0.26 -0.94 0.80 0.14 2 -1.14 0.21 0.17 0.37 -0.21 0.88 0.04 3 -1.62 0.35 -0.28 0.06 0.06 0.92 0.07 4 -1.14 0.51 -0.62 -0.45 0.39 0.88 0.11 5 -0.77 -0.01 2.42 0.68 -1.05 0.84 -1.12 6 -0.77 0.49 -0.58 -0.44 0.68 0.84 0.14 7 -2.40 0.05 0.64 0.40 -0.34 0.96 0.02 8 0.05 0.16 1.94 0.82 -0.93 0.72 -0.98 9 -0.19 0.04 1.28 1.24 2.50 0.76 -0.34
10 -0.77 0.37 -0.23 -0.02 0.68 0.84 0.10 11 -1.14 0.43 -0.40 -0.22 0.39 0.88 0.10 12 0.28 0.27 0.44 0.84 1.46 0.68 -0.06 13 -0.46 0.45 -0.41 -0.32 -0.94 0.80 0.14 14 -0.46 0.45 -0.45 -0.30 -0.94 0.80 0.15 15 -0.46 0.37 -0.35 0.14 0.96 0.80 0.12 16 0.05 0.46 -0.46 -0.24 -0.93 0.72 0.20 17 -0.77 0.42 -0.41 -0.17 0.68 0.84 0.12 18 -0.77 0.30 0.04 0.13 -1.05 0.84 0.07 19 0.05 0.34 0.58 0.11 -0.93 0.72 -0.16 20 0.05 0.52 -0.45 -0.78 0.20 0.72 0.19 21 0.28 0.33 -0.02 0.64 1.46 0.68 0.09 22 0.28 0.27 0.71 0.69 0.19 0.68 -0.22 23 0.49 0.57 -0.96 -0.94 -0.28 0.64 0.41 24 -0.19 0.35 1.04 -0.24 -0.30 0.76 -0.39 25 -0.46 0.50 -0.54 -0.58 -0.94 0.80 0.16 26 0.49 0.20 1.01 1.24 0.94 0.64 -0.36 27 0.28 0.14 1.40 1.20 1.46 0.68 -0.50 28 -0.19 0.52 -0.35 -0.86 -0.30 0.76 0.13 29 -0.46 0.33 -0.02 0.17 0.44 0.80 0.08 30 0.05 0.68 -1.36 -1.74 1.53 0.72 0.38 31 0.49 -0.11 2.54 2.60 2.03 0.64 -1.33 32 0.70 0.37 0.16 0.40 0.31 0.60 -0.07 33 0.49 0.31 0.15 0.85 -0.47 0.64 -0.03 34 0.49 0.40 -0.02 0.15 -0.47 0.64 0.04 35 0.05 0.53 -0.82 -0.63 1.53 0.72 0.24 36 0.05 0.38 -0.26 0.26 0.20 0.72 0.12 37 0.28 0.47 -0.25 -0.39 -1.02 0.68 0.15 38 0.05 0.28 0.54 0.53 0.77 0.72 -0.10 39 -0.19 0.21 0.43 0.76 -0.19 0.76 -0.05
(appendix continues)
![Page 167: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/167.jpg)
Appendix C (continued)
158
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
40 0.28 0.51 -0.65 -0.56 0.65 0.68 0.23 41 1.10 0.61 -1.13 -1.13 0.19 0.52 0.56 42 1.10 0.37 0.54 0.41 0.43 0.52 -0.30 43 0.70 0.60 -0.84 -1.33 0.31 0.60 0.39 44 1.10 0.60 -1.07 -1.07 0.19 0.52 0.54 45 1.10 0.39 0.15 0.47 0.43 0.52 0.02 46 0.90 0.31 1.16 0.56 -0.53 0.56 -0.66 47 0.28 0.33 0.29 0.40 0.19 0.68 -0.01 48 1.10 0.58 -0.98 -0.94 1.37 0.52 0.51 49 1.50 0.28 0.81 1.06 0.38 0.44 -0.31 50 0.70 0.49 -0.34 -0.42 0.31 0.60 0.20
Mean S.D.
-0.00 0.77
0.37 0.17
0.08 0.87
0.06 0.79
0.21 0.87
0.71
Groups Note: Raw score mean = 35.68 with a S.D. of 8. Mean person ability = 1.22 with a S.D. of 1.06. Test reliability (K.R. 20) = 0.88. Reliability of person separation = 0.86.
43.
![Page 168: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/168.jpg)
159
Appendix C (continued)
Experiment 39: Summary of Item Fit Information for a Uniformly Distributed Item Difficulty Distribution With 100 Items. 25 Persons, and and a 25% Chance of Guessing
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
1 -2.33 0.26 0.01 0.30 -0.25 0.96 0.04 2 -0.67 0.59 -0.91 -0.47 0.90 0.84 0.15 3 -0.67 0.23 0.87 0.27 0.70 0.84 -0.17 4 -1.05 0.30 0.63 0.02 -0.58 0.88 -0.10 5 -0.36 0.27 0.69 0.20 1.47 0.80 -0.10 6 0.16 0.57 -0.56 -0.78 -0.94 0.72 0.20 7 -0.09 0.51 -0.61 -0.26 0.21 0.76 0.17 8 -1.55 0.52 -0.62 -0.17 0.20 0.92 0.07 9 -1.05 0.15 0.34 0.57 -0.58 0.88 0.01
10 -0.09 0.22 0.60 0.67 0.83 0.76 -0.11 11 -0.36 0.39 0.27 -0.09 -0.03 0.80 -0.02 12 -0.09 0.47 0.04 -0.35 -0.87 0.76 0.02 13 -0.67 -0.07 1.72 0.80 2.16 0.84 -0.38 14 -0.36 0.18 0.65 0.63 -0.03 0.80 -0.09 15 -1.05 0.42 -0.32 -0.02 0.56 0.88 0.08 16 -0.67 0.73 -1.31 -1.10 0.90 0.84 0.18 17 -1.05 0.16 0.87 0.31 1.47 0.88 -0.11 18 -2.33 0.26 0.01 0.30 -0.25 0.96 0.04 19 -1.55 0.12 0.62 0.42 0.37 0.92 -0.03 20 -0.09 0.37 0.25 0.09 0.21 0.76 -0.04 21 -0.36 0.26 0.41 0.45 -0.39 0.80 -0.04 22 -2.33 0.08 0.50 0.40 -0.25 0.96 0.03 23 -2.33 0.26 0.01 0.30 -0.25 0.96 0.04 24 -0.67 0.43 -0.37 -0.02 -1.27 0.84 0.09 25 -0.36 0.05 1.37 0.85 1.47 0.80 -0.33 26 -0.67 0.64 -1.09 -0.65 0.90 0.84 0.16 27 -0.67 0.74 -1.33 -1.12 0.90 0.84 0.18 28 -1.05 0.24 0.55 0.22 -0.58 0.88 -0.06 29 -1.05 0.17 0.63 0.34 1.47 0.88 -0.04 30 -0.36 0.75 -1.43 -1.30 1.23 0.80 0.25 31 -0.67 0.35 -0.26 0.31 0.90 0.84 0.08 32 0.16 0.77 -1.67 -1.72 0.71 0.72 0.44 33 0.39 0.28 0.72 0.49 -0.62 0.68 -0.26 34 -0.67 0.45 -0.39 -0.07 -1.27 0.84 0.10 35 -1.05 0.38 -0.19 0.06 0.56 0.88 0.07 37 -0.67 0.52 -0.61 -0.28 0.90 0.84 0.12 38 -0.09 0.40 -0.22 0.16 0.21 0.76 0.09 39 0.39 0.40 0.21 -0.03 -0.62 0.68 -0.01 40 -0.67 0.44 -0.41 -0.03 0.90 0.84 0.10
(appendix continues)
![Page 169: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/169.jpg)
160
Appendix C (continued)
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
41 -0.67 0.34 0.30 0.08 -1.27 0.84 -0.03 42 0.39 0.33 0.28 0.43 -0.14 0.68 -0.10 43 -0.09 0.31 0.05 0.49 -0.87 0.76 0.04 44 0.16 0.08 1.37 1.13 1.45 0.72 -0.45 45 0.39 0.19 0.94 0.94 0.85 0.68 -0.34 46 -1.05 0.41 -0.32 0.01 0.56 0.88 0.08 47 -0.36 0.29 0.43 0.38 -0.39 0.80 -0.09 48 -0.36 0.24 1.20 0.17 -0.03 0.80 -0.30 49 -0.67 0.38 -0.27 0.14 0.90 0.84 0.09 50 -0.36 0.69 -1.21 -0.98 1.23 0.80 0.23 51 0.39 0.31 0.27 0.53 -0.14 0.68 -0.08 52 -0.09 0.17 1.33 0.60 0.83 0.76 -0.41 53 0.16 0.25 0.74 0.56 0.16 0.72 -0.20 54 -0.09 0.32 0.10 0.42 0.83 0.76 0.03 55 -1.05 0.54 -0.44 -0.45 0.56 0.88 0.08 56 0.60 0.38 0.19 0.12 -1.19 0.64 -0.09 57 -0.09 0.11 1.07 0.93 0.83 0.76 -0.25 58 -0.09 0.61 -0.84 -0.82 0.21 0.76 0.23 59 0.16 0.48 -0.44 -0.21 -0.94 0.72 0.19 60 0.60 0.59 -1.03 -0.87 -1.19 0.64 0.44 61 0.60 0.21 1.13 0.87 0.19 0.64 -0.53 62 1.19 0.24 1.13 0.87 0.67 0.52 -0.71 63 0.16 0.38 -0.12 0.21 -0.94 0.72 0.07 64 1.19 0.49 -0.22 -0.69 -0.12 0.52 0.06 65 0.99 0.24 0.78 0.96 1.30 0.56 -0.33 66 0.16 0.21 0.53 0.88 0.16 0.72 -0.13 67 0.39 0.62 -1.13 -0.98 -0.14 0.68 0.40 68 0.80 0.34 0.13 0.53 0.80 0.60 0.03 69 -0.09 0.13 0.79 0.97 -0.87 0.76 -0.16 70 0.60 0.45 -0.03 -0.24 -1.19 0.64 0.07 71 0.16 0.44 -0.10 -0.16 0.16 0.72 0.09 72 0.39 0.21 0.82 0.88 1.96 0.68 -0.23 73 1.37 0.48 -0.57 -0.57 -0.81 0.48 0.36 74 0.39 0.55 -0.84 -0.56 1.17 0.68 0.28 75 0.99 0.42 0.07 -0.09 -1.12 0.56 0.02 76 0.99 0.59 -1.04 -1.20 0.48 0.56 0.54 77 -0.09 0.45 -0.36 -0.08 0.21 0.76 0.12 78 0.80 0.38 0.40 0.07 0.80 0.60 -0.13 79 0.39 0.06 1.09 1.56 0.85 0.68 -0.43 80 -0.09 0.51 -0.50 -0.35 0.21 0.76 0.15 81 1.37 0.50 -0.66 -0.72 0.66 0.48 0.38 82 0.99 0.21 1.12 1.08 1.30 0.56 -0.54 83 0.80 0.35 0.28 0.38 -0.60 0.60 -0.06
(appendix continues)
![Page 170: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/170.jpg)
Appendix C (continued)
161
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
84 0.80 0.70 -1.79 -1.72 2.07 0.60 0.75 85 0.39 0.21 0.70 0.95 -0.62 0.68 -0.25 86 0.60 0.09 1.12 1.62 1.39 0.64 -0.46 87 0.39 0.49 -0.59 -0.25 1.17 0.68 0.21 88 1.56 0.44 -0.45 -0.33 0.18 0.44 0.28 89 1.75 0.71 -1.86 -2.47 1.00 0.40 0.80 90 0.80 0.73 -1.97 -1.95 2.07 0.60 0.82 91 1.75 0.30 0.09 0.55 -0.43 0.40 0.02 92 0.80 0.21 0.84 1.08 -0.60 0.60 -0.42 93 0.99 0.28 0.88 0.66 -1.12 0.56 -0.41 94 1.19 0.48 -0.58 -0.50 -0.12 0.52 0.36 95 1.56 0.24 1.35 0.67 0.18 0.44 -0.88 96 2.15 0.13 0.85 1.07 0.64 0.32 -0.27 97 0.39 0.63 -1.02 -1.13 1.17 0.68 0.36 98 0.80 0.31 0.45 0.53 0.80 0.60 -0.13 99 1.19 0.49 -0.61 -0.56 -0.12 0.52 0.37
100 0.80 0.41 -0.21 0.10 -0.60 0.60 0.16 Mean S.D. Groups
0.00 0.93
0.36 0.19
0.05 0.82
0.06 0.75
0.28 0.88 2
0.72
Note: Raw score mean = 72.28 with a S.D. of 15, Mean person ability = 1.25 with a S.D. of 0.92. Test reliability (K.R. 20) = 0.94. Reliability of person separation = 0.92.
98.
![Page 171: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/171.jpg)
162
Appendix C (continued)
Experiment 40- Summary of Item Fit Information for a Uniformly Distributed Item Difficulty Distribution With 25 Items. 50 Persons, and a 25% Chance of Guessing Correctly
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
1 -2.04 0.26 0.03 0.22 0.13 0.94 0.04 2 -0.62 0.48 -0.41 -0.61 -1.04 0.82 0.10 3 -1.42 0.32 -0.38 0.36 0.57 0.90 0.07 4 -0.32 0.37 0.12 0.30 -0.33 0.78 0.03 5 -0.18 0.18 1.39 1.34 1.50 0.76 -0.28 6 -0.79 0.09 1.76 1.17 -0.46 0.84 -0.33 7 -0.32 0.23 1.93 0.85 1.87 0.78 -0.53 8 -1.42 0.43 -0.53 -0.32 0.57 0.90 0.07 9 -0.62 0.61 -1.23 -1.33 1.32 0.82 0.17
10 -0.62 0.31 2.12 -0.12 -1.04 0.82 -0.62 11 -0.05 0.56 -1.04 -0.84 1.20 0.74 0.21 12 0.08 0.55 -0.95 -0.87 -0.74 0.72 0.23 13 0.43 0.26 1.10 1.63 0.52 0.66 -0.28 14 0.08 0.32 0.62 0.92 -0.39 0.72 -0.13 15 0.32 0.52 -0.75 -0.55 0.14 0.68 0.18 16 0.20 0.44 -0.17 0.05 -0.95 0.70 0.10 17 -0.32 0.42 -0.47 0.27 0.69 0.78 0.11 18 1.07 0.62 -1.64 -1.57 0.99 0.54 0.53 19 0.97 0.44 0.05 0.28 -0.90 0.56 0.04 20 0.76 0.42 0.02 0.53 -1.03 0.60 0.08 21 0.43 0.50 0.43 -0.63 -0.61 0.66 -0.15 22 0.43 0.49 0.59 -0.63 -0.61 0.66 -0.20 23 1.48 0.48 -0.04 -0.20 -1.01 0.46 0.05 24 1.07 0.41 0.52 0.69 0.72 0.54 -0.19 25 1.38 0.57 -1.18 -1.01 -0.40 0.48 0.42
Mean S.D. Groups
-0.00 0.89
0.41 0.14
0.07 1.01
-0.00 0.84
0.03 0.91 2
0.71
Note: Raw score mean = 17.86 and a S.D. of 4.66. Mean person ability = 1.28 with a S.D. of 1.19. Test reliability (K.R. 20) = 0.82. Reliability of person separation = 0.76.
![Page 172: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/172.jpg)
163
Appendix C (continued)
Experiment 41: Summary of Item Fit Information for a Uniformly Distributed Item Difficulty Distribution With 50 Items. 50 Persons, and a 25% Chance of Guessing Correctly
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
1 -2.27 0.27 0.32 -0.11 -0.20 0.94 0.00 2 -1.07 0.21 0.25 0.38 0.45 0.84 0.03 3 -1.25 0.21 0.04 0.51 -0.14 0.86 0.05 4 -1.07 0.29 0.08 -0.00 0.45 0.84 0.05 5 -0.77 0.28 0.23 0.16 -0.50 0.80 0.02 6 -0.50 0.38 -0.04 -0.46 0.28 0.76 0.07 7 -1.25 0.32 0.10 -0.08 -0.14 0.86 0.00 8 -1.45 0.41 -0.35 -0.47 -0.56 0.88 0.08 9 -0.77 0.13 0.58 0.94 1.74 0.80 -0.07
10 -0.91 0.29 0.08 0.11 -0.99 0.82 0.02 11 -0.63 0.37 -0.19 -0.28 -0.07 0.78 0.10 12 -0.63 0.31 0.18 -0.05 1.36 0.78 0.03 13 -0.38 0.34 -0.12 0.04 -0.57 0.74 0.06 14 -0.15 0.29 0.28 0.31 -1.04 0.70 -0.08 15 -0.15 0.25 0.36 0.69 -0.26 0.70 -0.03 16 -0.38 0.38 -0.02 -0.42 -0.57 0.74 0.07 17 0.16 0.30 0.38 0.48 -0.52 0.64 -0.08 18 -0.63 0.29 0.18 0.19 -1.10 0.78 0.02 19 0.06 0.47 -0.66 -0.98 0.89 0.66 0.28 20 0.16 0.26 0.34 0.93 0.52 0.64 -0.03 21 -0.15 0.39 -0.14 -0.38 -0.26 0.70 0.12 22 -0.04 0.37 -0.27 -0.01 0.57 0.68 0.16 23 -0.50 0.31 0.01 0.16 -1.14 0.76 0.06 24 -0.50 0.42 -0.33 -0.62 -1.14 0.76 0.14 25 0.46 0.46 -0.30 -1.29 0.54 0.58 0.17 26 0.16 0.30 0.55 0.29 -0.52 0.64 -0.17 27 0.36 0.47 -0.66 -1.09 0.15 0.60 0.34 28 0.46 0.18 0.89 1.88 2.46 0.58 -0.33 29 0.65 0.37 -0.11 0.14 -0.81 0.54 0.14 30 0.06 0.31 0.07 0.41 0.90 0.66 0.07 31 -0.50 0.23 0.43 0.53 -1.14 0.76 -0.08 32 -0.04 0.31 0.07 0.38 1.26 0.68 0.07 33 0.84 0.24 0.84 1.40 -0.76 0.50 -0.33 34 0.16 0.42 -0.39 -0.52 -0.79 0.64 0.23 35 0.06 0.44 -0.51 -0.71 -0.05 0.66 0.24 36 0.84 0.36 0.05 0.25 -0.61 0.50 0.06 37 0.65 0.37 0.05 -0.06 0.45 0.54 0.06 38 0.93 0.36 0.18 0.20 -1.41 0.48 -0.01 39 0.84 0.20 1.14 1.81 0.30 0.50 -0.50
(appendix continues)
![Page 173: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/173.jpg)
Appendix C (continued)
164
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
40 0.06 0.31 0.08 0.42 -1.58 0.66 0.06 41 1.22 0.42 -0.15 -0.40 0.51 0.42 0.11 42 0.65 0.54 -1.10 -2.00 0.45 0.54 0.53 43 1.12 0.40 -0.04 -0.29 0.00 0.44 0.07 44 0.84 0.48 -0.56 -1.26 1.22 0.50 0.29 45 0.84 0.29 0.34 1.12 0.30 0.50 -0.04 46 0.93 0.43 -0.30 -0.59 -0.05 0.48 0.19 47 0.84 0.32 0.40 0.60 -0.61 0.50 -0.12 48 0.65 0.41 -0.37 -0.31 -0.81 0.54 0.27 49 0.84 0.27 1.28 0.88 0.30 0.50 -0.78 50 1.22 0.54 -1.25 -1.62 1.32 0.42 0.50
Mean S.D.
0.00 0.79
0.34 0.09
0.04 0.48
0.02 0.78
-0.04 0.86
0.66
Groups Note: Raw score mean = 32.82 with a S.D. of 8. Mean person ability = 0.87 with a S.D. of 1.07. Test reliability (K.R. 20) = 0.87. Reliability of person separation = 0.90.
28.
![Page 174: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/174.jpg)
165
Appendix C (continued)
Exper iment 47.: Summary of Ttem Fit Information for a Uniformly Distributed Item Difficulty Distribution With 100 Items. 50 Persons, and a 25% Chance of Guessing Correctly
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
1 -1.46 0.34 0.02 0.15 0.44 0.90 0.05 2 -0.81 0.48 -0.59 -0.18 1.00 0.84 0.09 3 -1.00 0.52 -0.76 -0.46 0.81 0.86 0.11 4 -0.81 0.44 0.43 -0.47 -0.11 0.84 -0.04 5 -1.46 0.26 0.71 0.32 1.30 0.90 -0.05 6 -1.00 0.29 0.28 0.62 -1.00 0.86 0.00 7 -0.81 0.51 -0.76 -0.33 1.00 0.84 0.12 8 -0.63 0.42 0.48 -0.33 0.86 0.82 -0.05 9 -0.81 0.51 -0.72 -0.37 1.00 0.84 0.09
10 -0.18 0.47 -0.29 -0.19 -0.55 0.76 0.12 11 -0.32 0.50 -0.47 -0.42 0.45 0.78 0.12 12 -0.32 0.45 -0.19 -0.15 -1.07 0.78 0.09 13 -0.63 0.34 0.15 0.59 -0.15 0.82 -0.02 14 -1.00 0.46 -0.50 -0.22 0.81 0.86 0.10 15 -0.05 0.58 -0.95 -1.02 1.89 0.74 0.23 16 -0.47 0.62 -1.14 -1.32 1.35 0.80 0.19 17 -0.32 0.51 -0.57 -0.44 0.45 0.78 0.14 18 -0.81 0.44 0.07 -0.33 -0.11 0.84 0.04 19 -0.81 0.36 0.26 0.24 -0.11 0.84 0.01 20 -0.32 0.56 -0.81 -0.95 0.45 0.78 0.18 21 -1.00 0.55 -0.31 -1.05 -1.00 0.86 0.06 22 0.20 0.41 0.28 0.19 -0.47 0.70 -0.12 23 -0.18 0.33 1.00 0.45 0.81 0.76 -0.19 24 -0.63 0.38 0.09 0.19 -0.60 0.82 0.05 25 -1.22 0.47 -0.26 -0.43 -1.30 0.88 0.07 26 -1.00 0.19 0.73 1.00 1.71 0.86 -0.06 27 -0.81 0.40 -0.09 0.09 -0.52 0.84 0.06 28 -0.63 0.42 -0.06 -0.04 -0.15 0.82 0.06 29 0.20 0.34 0.85 0.67 -0.47 0.70 -0.29 30 -0.47 0.49 -0.37 -0.47 0.17 0.80 0.11 31 -1.00 0.40 -0.22 0.12 0.81 0.86 0.07 32 -0.63 0.44 -0.20 -0.09 -0.15 0.82 0.07 33 -0.05 0.56 -0.64 -0.99 -0.15 0.74 0.18 34 -0.63 0.60 -0.76 -1.34 -0.15 0.82 0.13 35 -0.32 0.36 0.35 0.37 1.19 0.78 0.00 36 0.32 0.43 0.47 0.09 -0.21 0.68 -0.11 37 -1.00 0.41 0.19 -0.25 0.35 0.86 0.03 38 0.20 0.34 3.99 0.26 -0.47 0.70 -2.71 39 -0.05 0.40 0.07 0.36 -0.97 0.74 0.03
(appendix continues)
![Page 175: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/175.jpg)
166
Appendix C (continued)
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
40 0.44 0.21 1.57 1.96 1.68 0.66 -0.50 41 -0.05 0.38 0.27 0.42 0.42 0.74 -0.01 42 -0.47 0.15 1.48 1.31 2.53 0.80 -0.31 43 -0.81 0.44 -0.16 -0.18 -0.52 0.84 0.07 44 -0.32 0.50 -0.11 -0.60 -1.07 0.78 0.06 45 -0.05 0.45 -0.18 0.02 -0.97 0.74 0.10 46 -0.32 0.47 -0.44 -0.12 0.45 0.78 0.13 47 0.08 0.21 2.16 1.22 1.05 0.72 -0.75 48 -0.47 0.34 0.52 0.35 0.45 0.80 -0.03 49 -0.47 0.41 0.13 -0.01 0.45 0.80 0.04 50 0.55 0.38 0.32 0.65 0.45 0.64 -0.08 51 -0.05 0.33 0.61 0.82 -0.97 0.74 -0.12 52 0.32 0.41 0.30 0.32 -1.04 0.68 -0.14 53 0.08 0.46 -0.35 0.08 0.21 0.72 0.13 54 -0.05 0.51 -0.49 -0.49 -0.15 0.74 0.17 55 -0.18 0.30 1.27 0.80 -0.55 0.76 -0.38 56 0.55 0.21 1.62 1.96 2.09 0.64 -0.53 57 0.20 0.46 0.03 -0.16 -0.47 0.70 0.06 58 -0.63 0.39 -0.27 0.42 1.17 0.82 0.09 59 0.32 0.37 0.20 0.80 1.22 0.68 0.04 60 -0.32 0.50 -0.57 -0.39 0.45 0.78 0.13 61 0.20 0.51 -0.60 -0.42 0.53 0.70 0.19 62 0.20 0.32 1.13 0.75 -0.47 0.70 -0.30 63 -0.32 0.31 0.77 0.67 -1.07 0.78 -0.19 64 -0.05 0.26 0.75 1.20 2.28 0.74 -0.08 65 0.97 0.57 -1.11 -1.22 0.20 0.56 0.42 66 -0.18 0.39 0.30 0.24 -0.43 0.76 -0.02 67 0.08 0.46 -0.29 -0.09 0.00 0.72 0.13 68 -0.32 0.25 0.62 1.18 1.19 0.78 -0.05 69 0.44 0.43 -0.14 0.28 -1.16 0.66 0.12 70 0.66 0.45 0.49 -0.21 -0.08 0.62 -0.31 71 0.66 0.38 0.21 0.87 -1.43 0.62 0.00 72 0.76 0.51 -0.50 -0.59 -0.82 0.60 0.24 73 0.20 0.55 -0.73 -0.88 0.53 0.70 0.22 74 -0.18 0.54 -0.86 -0.59 1.71 0.76 0.19 75 0.44 0.39 0.04 0.67 -0.17 0.66 0.07 76 0.32 0.37 1.47 0.20 0.27 0.68 -0.50 77 0.32 0.44 -0.07 0.16 0.84 0.68 0.04 78 0.76 0.42 -0.07 0.44 -0.47 0.60 0.11 79 0.44 0.48 -0.31 -0.25 -1.16 0.66 0.17 80 0.20 0.37 0.40 0.58 0.67 0.70 -0.03 81 0.66 0.44 0.44 -0.02 -0.08 0.62 -0.14 82 0.97 0.49 -0.14 -0.49 0.20 0.56 0.03
(appendix continues)
![Page 176: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/176.jpg)
167
Appendix C ('continued')
Logit Point. Unwt. Item item bis. total
# diff. corn fit
Wt. Ability Mean Logit total between item Residual fit fit score Index
83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99
100
1.18 1.28 0.66 1.18 0.44 1.28 0.76 1.08 0.32 1.08 1.28 1.18 0.76 0.20 0.66 0.55 1.79 1.48
Mean S.D. Groups
0.00 0.70
0.41 0.46 0.36 0.58 0.37 0.63 0.37 0.38 0.59 0.41 0.49 0.49 0.48 0.35 0.48 0.45 0.54 0.34
0.43 -0.20 0.28
-1.08 0.22
-1.64 0.52 0.24
-0.93 0.23
-0.53 -0.47 -0.40 0.55
-0.46 -0.13 -0.94 0.94
0.35 -0.02 1.00
-1.50 0.86
-2.20 0.79 0.90
-1.41 0.53
-0.41 -0.38 -0.20 0.72
-0.19 0.11
-1.09 1.05
-1.09 -0.97 0.96 1.02
-0.17 1.40 0.33 0.20
-0.21 -0.99 -0.40 -1.09 -0.47 0.67
-0.08 -0.68 0.40 0.16
0.42 0.10
0.06 0.78
0.04 0.73
0.14 0.88 2
0.52 0.50 0.62 0.52 0.66 0.50 0.60 0.54 0.68 0.54 0.50 0.52 0.60 0.70 0.62 0.64 0.40 0.46 0.72
-0.13 0.15 0.00 0.40 0.03 0.57
-0.18 0.02 0.29
-0.00 0.26 0.23 0.20
-0.18 0.20 0.13 0.34
-0.32
Note: Raw score mean = 72.06 with a S.D. of 19 Mean person ability = 1.27 with a S.D. of 1.24. Test reliability (K.R. 20) = 0.96. Reliability of person separation = 0.94.
42.
![Page 177: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/177.jpg)
168
Appendix C (continued)
pvpg>t-impnt A-\- Summary of Ttem Fit Inform?*™" fr>r a TTniformlv Distributed Item Difficulty Distribution With 9.5 Items. 100 Persons, and a ,hance of Guessing Correctly
Logit Point. Unwt. Item item bis. total
# diff. corr. fit
Wt. Ability Mean Logit total between item Residual fit fit score Index
1 2 3 4 5 6 7 8 9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
-2.05 -2.26 -1.19 -1.56 -0.89 -0.80 -0.71 -0.55 -0.06 -0.26 -0.40 -0.06 -0.06 0.48 0.19 0.19 0.25 0.97 0.70 0.65 1.02 1.33 1.22 1.80 2.02
Mean S.D. Groups
-0.00 1.11
0.33 0.12 0.34 0.21 0.42 0.37 0.27 0.38 0.48 0.29 0.41 0.46 0.44 0.52 0.39 0.51 0.51 0.60 0.36 0.48 0.49 0.50 0.34 0.33 0.54
-0.40 1.22 0.57 1.35
-0.55 -0.42 0.70 0.15
-0.37 0.95
-0.12 -0.42 0.27
-0.91 0.24
-1.00 -0.94 -1.78 1.39 0.02 0.05
-0.73 1.38 3.16
-0.97
-0.17 0.38
-0.20 0.22
-0.38 0.19 0.75
-0.00 -0.54 1.08
-0.01 -0.42 -0.50 -0.89 0.63
-0.60 -0.70 -1.93 1.13
-0.38 -0.47 -0.18 1.92 1.32
-0.97
0.43 1.55
-1.01 0.06 0.54
-0.74 0.20 0.65
-0.09 1.27
-0.34 -1.29 -0.33 1.12 0.02 0.02 0.27 1.85 0.02
-0.48 -1.47 -1.31 2.08
-0.11 0.45
0.40 0.11
0.11 1.06
-0.03 0.83
0.13 0.94 2
0.94 0.95 0.88 0.91 0.85 0.84 0.83 0.81 0.74 0.77 0.79 0.74 0.74 0.65 0.70 0.70 0.69 0.56 0.61 0.62 0.55 0.49 0.51 0.40 0.36 0.71
0.05 -0.10 -0.06 -0.16 0.08 0.08
-0.04 0.01 0.08
-0.11 0.05 0.10
-0.05 0.20
-0.01 0.19 0.20 0.41
-0.35 -0.07 -0.02 0.23
-0.33 -1.10 0.21
Note: Raw score mean = 17.63 and a S Mean person ability = 1.28 with a S.D. Test reliability (K.R. 20) = 0.81. Reliability of person separation = 0.78.
D. of 4.53. of 1.22.
![Page 178: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/178.jpg)
169
Appendix C (continued)
Experiment 44: Summarv ot item 111 i niffioiiitv Distribution With 50 Items,
niormaiioii iui a uiumiihy jyiauiuw^ 100 Persons, and a 25% Chance of Guessing
Correctly
Logit Item item
# diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability Mean Logit between item Residual
fit score Index
1 2 3 4 5 6 7 8 9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39
-0.69 -0.77 -1.27 -1.00 -1.09 -0.62 -1.00 -0.69 -0.77 -0.69 -0.30 -0.56 -0.62 -0.84 -0.13 -0.77 -0.24 -0.36 -0.18 -0.77 -0.13 -0.36 -0.24 -0.07 0.20 0.25 0.25
-0.02 0.09 0.25 0.09 0.45 0.25 0.69 0.30 0.79 0.69 0.45 0.50
0.34 0.30 0.29 0.25 0.27 0.42 0.39 0.22 0.45 0.31 0.29 0.29 0.28 0.26 0.42 0.36 0.43 0.38 0.37 0.23 0.32 0.34 0.31 0.44 0.24 0.44 0.41 0.45 0.42 0.40 0.37 0.35 0.34 0.38 0.41 0.43 0.42 0.47 0.41
-0.24 0.40 0.78 1.57 0.28
-0.83 -0.72 0.91
-0.76 0.36 0.14 0.62 0.19
-0.02 -0.48 -0.51 -0.51 1.09
-0.01 0.60 0.11 0.42
-0.00 -0.71 1.18
-0.55 -0.61 -0.96 -0.62 -0.46 -0.27 0.28 0.72 0.20
-0.12 -0.36 -0.42 -0.93 -0.37
0.06 0.06
-0.21 0.38 0.24
-0.47 -0.33 0.68
-0.81 0.11 0.74 0.36 0.46 0.54
-0.41 -0.03 -0.64 -0.41 -0.02 0.66 0.66 0.06 0.61
-0.64 1.52
-0.64 -0.13 -0.64 -0.32 0.09 0.36 0.62 0.41 0.34
-0.31 -0.40 -0.13 -0.95 -0.04
0.40 0.67
-0.71 0.31
-1.33 0.44 0.31 0.40 0.01
-0.83 -1.19 -0.17 0.12 0.93
-0.16 0.01
-0.78 -0.37 -0.14 -0.26 -0.16 -0.37 -0.79 -0.81 0.19 0.80
-0.69 -0.47 -0.78 -1.02 -0.78 0.82
-0.69 -1.05 -1.50 -0.62 0.72
-0.22 -0.11
0.80 0.81 0.87 0.84 0.85 0.79 0.84 0.80 0.81 0.80 0.74 0.78 0.79 0.82 0.71 0.81 0.73 0.75 0.72 0.81 0.71 0.75 0.73 0.70 0.65 0.64 0.64 0.69 0.67 0.64 0.67 0.60 0.64 0.55 0.63 0.53 0.55 0.60 0.59
0.05 -0.02 -0.09 -0.34 -0.02 0.12 0.08
-0.11 0.09
-0.03 0.01
-0.06 -0.01 0.03 0.10 0.08 0.09
-0.29 0.03
-0.06 -0.02 -0.06 0.03 0.15
-0.25 0.14 0.16 0.20 0.15 0.13 0.09
-0.02 -0.20 -0.01 0.04 0.12 0.16 0.25 0.15
![Page 179: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/179.jpg)
170
Appendix C (continued)
Logti Point Unwt. Wt Ability Mean Item item bis. total total between item
# diff. corr. fit fit fit s c o r e
Logit Residual
Index
40 41 42 43 44 45 46 47 48 49 50
0.64 0.35 0.93 0.45 1.08 0.69 0.79 0.84 0.84 1.22 1.12
0.27 0.49 0.22 0.42 0.51 0.42 0.53 0.31 0.36 0.33 0.50
1.88 -0.83 2.36
-0.63 -1.35 -0.50 -1.71 1.43 0.21 0.42
-0.78
1.43 -1.27 2.23
-0.14 -1.51 -0.16 -1.68 1.06 0.60 1.05
-1.45
0.98 0.63 1.11
-1.19 0.33
-1.05 1.03
-0.18 -1.08 -0.09 1.21
-0.00 0.82
0.03 0.77
Mean 0.00 0.37 S.D. 0.66 0.08 Groups Note: Raw score mean = 33.82 with a S.D. of 8.52. Mean person ability = 0.94 with a S.D. of 0.97. Test reliability (K.R. 20) = 0.88. Reliability of person separation = 0.86.
-0.16 0.72 2
0.56 0.62 0.50 0.60 0.47 0.55 0.53 0.52 0.52 0.44 0.46 0.68
-0.59 0.20
-0.73 0.19 0.37 0.18 0.46
-0.44 -0.01 -0.05 0.16
![Page 180: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/180.jpg)
171
Appendix C (continued)
Experiment 45- Summary of Ttem Fit Information for a Uniformly Distributed Item Difficulty Distribution With 100 Items. 100 Persons, and a 25% Chance of Guessing Correctly
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
1 -0.84 0.45 -0.43 -0.74 -0.25 2 -1.17 0.33 0.12 0.03 -1.17 3 -1.17 0.35 -0.03 -0.05 -1.17 4 -1.36 0.34 -0.48 0.03 0.54 5 -1.09 0.40 -0.48 -0.32 -1.32 6 -1.47 0.26 0.05 0.43 -0.04 7 -0.70 0.26 0.95 0.89 -0.58 8 -1.09 0.37 -0.60 0.06 1.04 9 -0.84 0.44 -0.41 -0.57 -0.25
10 -0.84 0.39 0.15 -0.17 0.69 11 -0.24 0.29 0.75 1.14 1.50 12 -1.09 0.33 -0.30 0.32 0.09 13 -0.63 0.44 -0.58 -0.42 0.44 14 -0.49 0.39 0.44 -0.17 0.42 15 -0.43 0.36 0.44 0.23 0.15 16 -0.49 0.42 -0.36 -0.21 0.03 17 -0.37 0.48 -0.63 -0.79 0.49 18 -0.63 0.33 0.33 0.45 -0.51 19 -0.24 0.31 1.04 0.71 0.13 20 -0.77 0.46 -0.58 -0.80 0.00 21 -0.37 0.38 -0.02 0.23 -0.36 22 -0.63 0.31 1.18 0.45 0.94 23 -0.63 0.34 0.12 0.34 -0.96 24 -0.63 0.37 0.02 0.06 0.13 25 -0.49 0.32 1.21 0.28 -1.05 26 -0.70 0.34 0.38 0.31 -0.84 27 -0.24 0.35 -0.04 0.69 -0.81 28 -0.84 0.32 0.20 0.37 0.04 29 -0.07 0.41 -0.11 -0.04 -0.77 30 -0.56 0.26 1.18 0.86 0.68 31 -0.43 0.29 0.79 0.81 0.91 32 -0.24 0.25 0.71 1.61 2.08 33 -0.24 0.29 0.54 1.21 2.08 34 -0.63 0.37 -0.44 0.31 0.44 35 -0.37 0.37 0.15 0.20 -0.36 36 -0.49 0.38 -0.29 0.24 0.03 37 -0.24 0.39 0.28 0.04 -0.81 38 -0.01 0.49 -0.72 -0.93 -0.42 39 -0.19 0.54 -1.29 -1.43 1.69
Mean item score
Logit Residual
Index
0.81 0.85 0.85 0.87 0.84 0.88 0.79 0.84 0.81 0.81 0.72 0.84 0.78 0.76 0.75 0.76 0.74 0.78 0.72 0.80 0.74 0.78 0.78 0.78 0.76 0.79 0.72 0.81 0.69 0.77 0.75 0.72 0.72 0.78 0.74 0.76 0.72 0.68 0.71
0.06 0.01 0.02 0.07 0.07 0.03
-0.13 0.08 0.05
-0.05 -0.10 0.05 0.09
-0.07 -0.04 0.05 0.10
-0.03 -0.17 0.08 0.05
-0.21 -0.00 0.03
-0.24 -0.05 0.04
-0.00 0.05
-0.16 -0.12 -0.07 -0.04 0.06
-0.00 0.07
-0.04 0.16 0.17
(appendix continues)
![Page 181: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/181.jpg)
172
Appendix C (continued)
Logit Point. Item item bis.
# diff. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
0.70 0.18 1.35 0.74 -0.07 0.28 0.40 0.13 0.72 -0.02
-1.71 -1.36 1.20 0.78 0.20 -0.70 -1.03 1.09 0.71 0.12 0.35 0.43 -1.27 0.71 -0.07 1.29 -0.57 0.49 0.74 -0.39
-0.43 -0.23 -1.27 0.71 0.09 0.89 0.53 -0.35 0.65 -0.17
-1.03 -0.58 0.62 0.70 0.19 1.54 1.66 2.21 0.63 -0.34 0.71 0.92 0.61 0.71 -0.08 0.91 0.94 1.73 0.73 -0.12 0.79 1.25 0.13 0.72 -0.13
-0.21 0.57 -1.27 0.71 0.08 -0.27 -0.25 0.16 0.72 0.03 -1.14 -1.52 1.04 0.68 0.22 -0.38 -0.71 0.62 0.70 0.08 -0.18 0.09 -0.94 0.61 0.06 0.24 0.17 -0.71 0.64 -0.02
-0.03 -0.51 -0.76 0.72 0.02 -0.82 -1.23 -0.06 0.59 0.19 -1.59 -1.85 1.65 0.57 0.38 0.28 0.80 -0.38 0.60 -0.04
-0.86 -0.51 -1.29 0.58 0.25 -1.40 -1.40 1.10 0.62 0.30 -0.57 -0.30 -0.05 0.64 0.14 0.43 0.61 0.54 0.63 -0.06 0.73 1.40 0.25 0.62 -0.14
-1.61 -1.61 1.65 0.62 0.35 0.36 0.50 0.64 0.61 -0.06
-0.59 0.12 -0.77 0.59 0.18 -0.29 -1.25 -0.57 0.63 0.05 0.38 0.78 -1.34 0.53 -0.05 0.50 -0.98 0.23 0.63 -0.24
-1.33 -1.70 . 0.15 0.56 0.33 -1.11 -0.93 0.23 0.63 0.24 1.62 2.49 1.92 0.53 -0.36
-0.26 -0.36 0.15 0.56 0.09 -1.65 -1.69 -0.34 0.51 0.42 -0.54 -0.26 0.52 0.57 0.17 0.02 0.33 -0.13 0.54 0.04 1.65 0.74 -0.79 0.50 -0.59 0.09 0.03 -1.34 0.53 -0.01
40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82
-0.37 -0.24 -0.63 -0.19 -0.19 -0.37 -0.19 0.15
-0.13 0.25
-0.19 -0.31 -0.24 -0.19 -0.24 -0.01 -0.13 0.36 0.20
-0.24 0.46 0.56 0.41 0.51 0.31 0.20 0.25 0.31 0.31 0.36 0.46 0.25 0.76 0.25 0.61 0.25 0.76 0.61 0.86 0.56 0.71 0.91 0.76
0.35 0.36 0.55 0.50 0.36 0.43 0.43 0.36 0.48 0.29 0.32 0.30 0.28 0.37 0.43 0.54 0.47 0.43 0.41 0.44 0.52 0.56 0.38 0.48 0.54 0.46 0.38 0.32 0.55 0.39 0.44 0.51 0.39 0.49 0.55 0.51 0.26 0.46 0.56 0.46 0.42 0.37 0.43
(appendix continues)
![Page 182: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/182.jpg)
173
Appendix C (continued)
Logit Point. Unwt. Wt. Ability Mean Item item bis. total total between item
# diff. corr. fit fit fit score
Logit Residual
Index
83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99
100
0.81 0.91 1.11 0.91 0.81 1.06 0.81 0.86 0.56 0.76 0.86 1.01 0.86 0.51 0.86 1.01 0.96 0.71
0.45 0.35 0.50 0.30 0.57 0.48 0.48 0.29 0.50 0.47 0.47 0.43 0.36 0.46 0.40 0.45 0.26 0.48
0.58 0.93
-0.78 2.09
-1.58 -0.40 -0.91 2.13
-0.92 -0.62 -0.56 0.32 0.95
-0.31 0.21
-0.38 1.73
-0.69
-0.34 1.26
-0.92 1.49
-2.02 -0.57 -0.33 1.70
-0.83 -0.44 -0.46 0.03 0.99
-0.33 0.63 0.04 2.40
-0.43
0.04 -0.08 -0.73 -0.08 0.04
-0.32 -0.10 0.26
-0.17 -1.34 0.97
-0.78 -0.48 -0.53 0.26
-0.09 2.40
-0.72
0.52 0.50 0.46 0.50 0.52 0.47 0.52 0.51 0.57 0.53 0.51 0.48 0.51 0.58 0.51 0.48 0.49 0.54 0.67
-0.28 -0.21 0.19
-0.64 0.38 0.10 0.28
-0.67 0.25 0.19 0.17
-0.10 -0.24 0.11
-0.00 0.14
-0.42 0.21
Mean S.D. Groups
-0.00 0.65
0.40 0.08
0.00 0.84
0.02 0.90
0.08 0.90 2
Note: Raw score mean — 66.82 with a S.D. of 18 Mean person ability = 0.91 with a S.D. of 1.08. Test reliability (K.R. 20) = 0.95. Reliability of person separation = 0.94.
92.
![Page 183: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/183.jpg)
174
Appendix C (continued)
Experiment 46- Summary of Item Fit Information for a Uniformly Distributed Item Difficulty Distribution With 25 Items. 25 Persons, and a 50% Chance of Guessing Correctly
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
1 -0.69 0.52 -0.16 0.29 0.75 0.80 0.05 2 -1.54 0.57 -0.03 -0.45 0.07 0.88 0.08 3 -0.35 0.38 1.84 0.62 -0.74 0.76 -0.81 4 -1.07 0.64 -0.20 -0.79 -0.25 0.84 0.11 5 -0.69 0.54 1.89 -0.70 -1.17 0.80 -0.87 6 -1.07 0.34 0.99 0.62 1.75 0.84 -0.14 7 -0.35 0.78 -1.24 -1.66 1.09 0.76 0.28 8 -1.54 0.44 0.48 0.13 0.55 0.88 0.01 9 0.22 0.32 1.33 1.29 0.50 0.68 -0.62
10 0.22 0.24 1.41 1.80 1.78 0.68 -0.57 11 0.47 0.48 0.42 0.44 1.07 0.64 -0.23 12 -0.05 0.50 0.03 0.34 1.16 0.72 0.09 13 0.22 0.56 0.17 -0.22 -1.53 0.68 0.04 14 0.71 0.48 0.01 0.64 0.50 0.60 0.14 15 0.71 0.34 0.78 1.59 0.50 0.60 -0.23 16 -1.07 0.53 -0.16 0.16 0.40 0.84 0.06 17 0.71 0.61 -0.71 -0.47 -1.35 0.60 0.35 18 0.22 0.52 1.18 -0.14 -1.53 0.68 -0.50 19 -0.35 0.72 -0.93 -1.10 1.09 0.76 0.25 20 0.22 0.50 -0.02 0.50 -1.53 0.68 0.09 21 1.38 0.58 -0.52 -0.50 -0.15 0.48 0.32 22 1.38 0.52 0.36 -0.16 -0.15 0.48 -0.12 23 0.47 0.64 -0.80 -0.67 -0.39 0.64 0.33 24 0.47 0.60 -0.64 -0.33 -0.39 0.64 0.30 25 1.38 0.65 -0.93 -1.24 -0.15 0.48 0.45
Mean 0.00 0.52 0.18 -0.00 0.08 0.70 S.D. 0.86 0.13 0.88 0.85 1.00 Groups 2 Note: Raw score mean = 17.44 and a S.D. of 5.97. Mean person ability = 1.20 with a S.D. of 1.52. Test reliability (K.R. 20) = 0.91. Reliability of person separation = 0.90.
![Page 184: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/184.jpg)
175
Appendix C (continued)
Exneriment 47- Summary of Item Fit Information for a Uniformly Distributed Item Difficulty Distribution With 50 Items. 25 Persons, and a 50% Chance of Guessing Correctly
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
1 -1.24 0.49 -0.35 -0.07 0.13 0.88 0.08 2 -0.50 0.53 -0.56 -0.13 0.70 0.80 0.13 3 -1.24 0.72 -1.02 -1.17 0.13 0.88 0.12 4 -0.83 0.54 -0.23 -0.48 -0.34 0.84 0.09 5 -0.50 0.63 -0.58 -0.84 -1.14 0.80 0.16 6 -0.20 0.67 -0.96 -1.03 0.97 0.76 0.25 7 -0.50 0.14 2.35 0.79 1.00 0.80 -1.04 8 0.06 0.09 1.73 1.63 -0.23 0.72 -0.73 9 -0.83 0.52 -0.25 -0.31 0.41 0.84 0.10
10 -1.77 -0.20 1.80 0.90 3.41 0.92 -0.31 11 -0.83 0.12 0.72 1.13 -0.34 0.84 -0.06 12 -0.50 0.59 -0.19 -0.74 -1.14 0.80 0.09 13 -0.83 0.26 0.70 0.55 1.63 0.84 -0.07 14 -0.50 0.64 -0.68 -0.87 -1.14 0.80 0.17 15 -0.50 0.45 0.36 -0.11 -1.14 0.80 -0.02 16 -0.20 0.66 -0.93 -0.98 0.97 0.76 0.25 17 -1.77 0.40 0.01 0.01 -0.18 0.92 0.05 18 0.31 0.48 -0.24 0.02 0.24 0.68 0.09 19 0.31 0.70 -1.13 -1.54 0.24 0.68 0.38 20 -0.20 0.16 0.85 1.33 1.84 0.76 -0.21 21 0.31 0.50 -0.07 -0.18 -0.99 0.68 0.12 22 0.31 0.48 -0.23 0.01 -0.99 0.68 0.17 23 -0.50 0.20 0.87 0.93 2.44 0.80 -0.13 24 -0.20 0.40 0.83 0.02 0.39 0.76 -0.21 25 0.06 0.33 0.47 0.79 -0.26 0.72 -0.14 26 0.31 0.63 -0.88 -0.96 0.24 0.68 0.33 27 -0.83 0.48 -0.43 0.10 0.41 0.84 0.10 28 0.96 -0.01 2.75 2.71 1.49 0.56 -1.74 29 0.53 0.51 -0.46 -0.08 -0.94 0.64 0.24 30 -0.50 0.42 -0.17 0.38 0.70 0.80 0.07 31 0.31 0.62 -0.97 -0.82 1.55 0.68 0.34 32 0.31 0.45 -0.11 0.26 -0.99 0.68 0.12 33 0.75 0.19 1.27 1.67 0.89 0.60 -0.61 34 0.75 0.45 -0.04 0.18 -0.54 0.60 0.13 35 0.06 0.51 -0.22 -0.14 -0.26 0.72 0.13 36 0.31 0.28 0.83 0.97 0.73 0.68 -0.26 37 0.06 0.35 0.91 0.41 1.28 0.72 -0.24 38 -0.50 0.55 -0.60 -0.20 0.70 0.80 0.15 39 0.53 0.77 -1.83 -2.13 1.86 0.64 0.61
(appendix continues)
![Page 185: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/185.jpg)
Appendix C (continued)
176
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
40 0.75 0.41 0.06 0.48 -0.54 0.60 0.07 41 0.31 0.30 0.93 0.83 0.24 0.68 -0.38 42 0.96 0.36 0.50 0.77 -1.47 0.56 -0.15 43 0.75 0.53 -0.59 -0.27 -0.21 0.60 0.32 44 0.31 0.69 -1.24 -1.30 0.24 0.68 0.41 45 0.75 0.34 0.91 0.76 0.89 0.60 -0.34 46 1.16 0.61 -1.10 -1.21 -0.43 0.52 0.55 47 1.56 0.26 1.61 1.05 -0.39 0.44 -1.03 48 0.96 0.66 -1.31 -1.48 1.55 0.56 0.59 49 0.75 0.40 0.20 0.57 0.89 0.60 0.04 50 0.96 0.63 -1.17 -1.24 0.38 0.56 0.54
Mean S.D.
-0.00 0.75
0.44 0.20
0.04 0.99
0.02 0.96
0.30 1.03
0.72
Groups Note: Raw score mean = 35.80 with a S.D. of 9. Mean person ability = 1.22 with a S.D. of 1.17. Test reliability (K.R. 20) = 0.92. Reliability of person separation = 0.88.
97.
![Page 186: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/186.jpg)
Ill
Appendix C (continued)
E x p e r i m e n t dR- Summary of Ttem Fit Information for a Uniformly Distributed Item Difficulty Distribution With 100 Items. 25 Persons- and a 50% Chance of Guessing Correctly
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
1 -2.39 0.26 -0.01 0.27 -0.20 0.96 0.04 2 -1.14 0.19 0.63 0.20 -0.67 0.88 -0.08 3 -0.21 0.33 -0.17 0.19 0.18 0.76 0.09 4 -3.13 -9.99 -0.41 -1.04 0.05 1.00 -0.00 5 -0.78 0.58 -0.85 -0.64 0.92 0.84 0.15 6 -1.14 0.41 -0.34 -0.13 0.59 0.88 0.08 7 -0.78 0.23 0.29 0.25 0.65 0.84 -0.00 8 -3.13 -9.99 -0.41 -1.04 0.05 1.00 -0.00 9 -1.63 0.54 -0.72 -0.27 0.24 0.92 0.08
10 -0.48 0.54 -0.66 -0.65 -0.38 0.80 0.17 11 -2.39 0.40 -0.30 0.12 -0.20 0.96 0.04 12 -0.48 0.27 -0.04 0.37 -0.38 0.80 0.06 13 -1.14 0.59 -0.92 -0.54 0.59 0.88 0.12 14 -0.21 0.21 0.15 0.66 -0.85 0.76 0.01 15 0.02 0.47 -0.51 -0.41 0.21 0.72 0.22 16 -0.48 0.55 -0.55 -0.76 -0.38 0.80 0.14 17 -0.78 0.30 -0.10 0.21 -1.18 0.84 0.07 18 -0.21 0.05 0.90 1.06 0.18 0.76 -0.24 19 -1.63 0.16 0.32 0.30 0.28 0.92 0.02 20 -0.48 0.57 -0.84 -0.69 1.23 0.80 0.19 21 0.02 0.36 -0.15 0.10 0.21 0.72 0.11 22 -0.48 0.23 0.06 0.48 -0.38 0.80 0.05 23 -0.21 0.38 -0.27 -0.01 -0.85 0.76 0.12 24 -1.14 0.15 0.97 0.22 -0.67 0.88 -0.19 25 -0.48 0.50 -0.67 -0.41 1.23 0.80 0.17 26 -0.21 0.49 -0.57 -0.48 0.18 0.76 0.19 27 -0.48 0.09 0.82 0.68 -0.06 0.80 -0.15 28 -1.14 0.27 -0.02 0.19 0.59 0.88 0.06 29 -0.48 0.33 -0.09 0.09 -0.06 0.80 0.08 30 -0.48 0.49 -0.61 -0.37 1.23 0.80 0.16 31 0.02 0.52 -0.78 -0.56 0.65 0.72 0.26 32 -0.78 0.31 0.25 -0.01 0.65 0.84 -0.00 33 -0.21 -0.20 2.32 1.51 0.83 0.76 -0.95 34 -0.48 0.18 0.67 0.38 1.45 0.80 -0.11 35 -0.21 0.22 0.70 0.34 0.83 0.76 -0.16 36 -0.78 -0.07 1.28 0.85 2.11 0.84 -0.26 37 0.45 0.11 1.01 1.33 1.50 0.64 -0.37 38 -1.14 0.36 -0.26 0.02 0.59 0.88 0.07 39 -0.78 0.52 -0.54 -0.51 -1.18 0.84 0.12
(appendix continues')
![Page 187: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/187.jpg)
178
Appendix C (continued)
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
40 -0.21 0.61 -0.91 -1.04 0.18 0.76 0.25 41 -0.21 0.31 -0.04 0.23 0.18 0.76 0.06 42 0.24 0.31 0.87 0.06 -0.48 0.68 -0.38 43 0.02 0.41 -0.17 -0.21 -1.04 0.72 0.12 44 -0.21 0.49 -0.66 -0.43 1.53 0.76 0.20 45 -0.21 0.59 -0.84 -1.00 0.18 0.76 0.24 46 -0.78 0.44 -0.36 -0.25 -1.18 0.84 0.10 47 0.45 0.61 -1.26 -1.34 -1.58 0.64 0.54 48 -0.21 0.14 0.34 0.89 -0.85 0.76 -0.03 49 0.24 0.36 0.93 -0.28 -0.27 0.68 -0.48 50 0.45 0.46 -0.60 -0.37 1.49 0.64 0.26 51 -0.21 0.24 0.27 0.43 0.83 0.76 -0.01 52 0.45 0.33 0.00 0.35 0.34 0.64 0.06 53 1.02 0.35 -0.05 0.37 0.95 0.52 0.14 54 0.24 0.21 0.81 0.64 -0.27 0.68 -0.27 55 0.02 0.18 0.62 0.70 0.21 0.72 -0.14 56 0.45 0.52 -0.78 -0.82 1.49 0.64 0.35 57 0.02 0.34 -0.20 0.23 -1.04 0.72 0.11 58 0.84 0.35 -0.01 0.33 0.38 0.56 0.10 59 -1.14 0.46 -0.41 -0.28 0.59 0.88 0.08 60 -0.21 0.28 0.60 0.02 0.83 0.76 -0.12 61 0.45 0.30 0.21 0.46 -1.58 0.64 -0.05 62 0.65 0.64 -1.52 -1.63 1.89 0.60 0.70 63 0.45 0.23 0.41 0.82 0.34 0.64 -0.10 64 0.84 0.52 -0.94 -0.89 0.19 0.56 0.53 65 -0.21 0.35 -0.00 -0.02 -0.85 0.76 0.06 66 0.24 0.45 -0.56 -0.25 1.08 0.68 0.22 67 1.21 -0.06 2.35 2.47 2.40 0.48 -1.52 68 -0.48 0.45 -0.19 -0.39 -0.38 0.80 0.08 69 -0.21 0.50 -0.65 -0.52 1.53 0.76 0.20 70 0.65 0.50 -0.74 -0.73 -0.51 0.60 0.39 71 0.45 0.35 -0.14 0.24 -1.58 0.64 0.10 72 0.65 0.04 1.28 1.94 0.96 0.60 -0.67 73 0.84 0.45 -0.34 -0.46 0.19 0.56 0.21 74 0.65 0.26 0.95 0.49 -0.34 0.60 -0.53 75 0.24 0.55 -0.85 -0.90 -0.48 0.68 0.35 76 0.65 -0.09 2.30 2.36 1.99 0.60 -1.40 77 0.45 0.27 0.33 0.57 0.34 0.64 -0.09 78 0.84 0.21 0.83 1.06 -1.31 0.56 -0.45 79 0.24 0.53 -0.81 -0.77 -0.48 0.68 0.33 80 0.24 -0.05 1.91 1.62 0.93 0.68 -0.88 81 0.65 0.62 -1.43 -1.48 0.85 0.60 0.67 82 1.21 0.31 0.66 0.32 0.34 0.48 -0.42
(appendix continues)
![Page 188: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/188.jpg)
Appendix C (continued)
179
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
83 0.45 0.28 1.21 0.21 -1.58 0.64 -0.64 84 0.02 0.44 -0.43 -0.29 -1.04 0.72 0.19 85 0.65 0.31 0.89 0.16 -0.34 0.60 -0.50 86 0.84 0.54 -1.04 -0.96 -1.31 0.56 0.58 87 0.24 0.53 -0.72 -0.80 1.08 0.68 0.29 88 0.84 0.32 0.37 0.38 0.19 0.56 -0.21 89 1.21 0.50 -0.68 -0.82 -1.39 0.48 0.41 90 1.40 0.30 0.80 0.32 -0.43 0.44 -0.51 91 1.02 0.18 1.07 1.25 0.95 0.52 -0.59 92 0.24 0.38 -0.25 0.05 -0.27 0.68 0.15 93 0.84 0.18 0.97 1.19 0.38 0.56 -0.48 94 1.02 0.65 -1.75 -1.88 1.83 0.52 0.91 95 1.02 0.17 1.38 1.19 -0.56 0.52 -0.86 96 0.45 -0.03 2.20 1.66 1.50 0.64 -1.14 97 0.24 0.27 0.20 0.55 0.93 0.68 -0.01 98 1.40 0.42 -0.17 -0.29 -0.43 0.44 0.14 99 0.84 0.32 0.36 0.33 -1.31 0.56 -0.14
100 0.84 0.32 0.20 0.45 -1.31 0.56 -0.03 Mean -0.06 0.34 0.05 0.07 0.15 0.71 S.D. 0.88 0.19 0.84 0.80 0.94 Groups 2 Note: Raw score mean = 71.28 with a S.D. of 15. Mean person ability =1.12 with a S.D. of 0.89. Test reliability (K.R. 20) = 0.93. Reliability of person separation = 0.91.
00.
![Page 189: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/189.jpg)
180
Appendix C (continued)
Experiment 49: Summary of Item Fit Information for a Uniformly Distributed Item Difficulty Distribution With 25 Items. 50 Persons, and a 50% Chance of Guessing Correctly
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
1 -1.51 0.39 -0.25 0.05 0.56 0.88 0.07 2 -0.48 0.45 -0.23 -0.05 -0.93 0.76 0.10 3 -0.35 0.55 -0.53 -0.97 -0.47 0.74 0.14 4 -0.48 0.55 -0.94 -0.78 0.52 0.76 0.19 5 -0.62 0.40 1.44 -0.05 0.26 0.78 -0.49 6 -0.10 0.59 -1.20 -1.07 1.23 0.70 0.27 7 -0.77 0.26 0.34 1.14 -0.03 0.80 -0.00 8 -0.35 0.25 0.73 1.50 0.74 0.74 -0.16 9 0.14 0.61 -1.42 -1.45 0.82 0.66 0.37
10 -0.10 0.33 1.01 0.77 1.03 0.70 -0.19 11 0.02 0.33 0.48 1.18 1.58 0.68 -0.03 12 -0.35 0.56 -1.01 -0.82 1.74 0.74 0.22 13 0.02 0.38 0.42 0.72 -0.47 0.68 -0.05 14 0.47 0.55 -0.88 -0.73 0.81 0.60 0.31 15 0.02 0.46 -0.04 -0.11 -0.47 0.68 -0.02 16 0.36 0.42 0.91 0.35 -0.58 0.62 -0.31 17 0.25 0.57 -1.01 -0.97 -1.26 0.64 0.30 18 -0.10 0.49 -0.32 -0.32 -0.03 0.70 0.13 19 0.02 0.51 0.02 -0.65 -0.67 0.68 0.03 20 0.02 0.43 -0.26 0.43 -0.47 0.68 0.12 21 0.89 0.43 0.13 0.66 0.40 0.52 0.03 22 0.57 0.43 0.67 0.45 -0.32 0.58 -0.23 23 0.57 0.25 2.39 1.83 0.69 0.58 -0.98 24 0.57 0.45 0.07 0.33 0.69 0.58 0.06 25 1.30 0.51 -0.47 -0.32 0.41 0.44 0.21
Mean 0.00 0.45 0.00 0.05 0.23 0.68 S.D. 0.58 0.11 0.89 0.87 0.79 Groups 2 Note: Raw score mean = 16.92 and a S.D. of 5.27. Mean person ability = .99 with a S.D. of 1.25. Test reliability (K.R. 20) = 0.84. Reliability of person separation = 0.80.
![Page 190: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/190.jpg)
181
Appendix C (continued)
Experiment 50: Summary of Item Fit Information for a Uniformly Distributed Item Difficulty Distribution With 50 Items. 50 Persons, and a 50% Chance of Guessing Correctly
Logit Point. Unwt. Wt. Ability Mean Logit Item item bis. total total between item Residual
# diff. corr. fit fit fit score Index
1 -1.37 0.24 0.22 0.28 0.75 0.88 0.05 2 -0.82 0.13 0.64 1.39 0.89 0.82 -0.03 3 -1.17 0.46 -0.54 -0.75 0.82 0.86 0.10 4 -1.17 0.08 1.49 0.89 1.68 0.86 -0.33 5 -0.82 0.24 0.79 0.30 1.99 0.82 -0.08 6 -0.53 0.37 0.29 -0.08 -1.24 0.78 -0.02 7 -0.99 0.30 0.24 0.24 -0.53 0.84 0.02 8 -1.87 0.22 0.22 0.26 -0.18 0.92 0.04 9 -0.67 0.33 0.08 0.27 0.11 0.80 0.04
10 -0.82 0.49 -0.67 -0.98 1.14 0.82 0.14 11 -0.53 0.50 -0.67 -0.93 0.38 0.78 0.17 12 -0.53 0.28 0.23 0.75 0.13 0.78 0.04 13 -0.02 0.53 -0.59 -1.33 -0.95 0.70 0.22 14 -0.26 0.10 1.39 2.04 3.16 0.74 -0.31 15 -0.39 0.37 0.18 0.02 -0.70 0.76 0.03 16 -0.26 0.55 -0.68 -1.61 0.87 0.74 0.19 17 -0.26 0.15 0.86 1.96 2.39 0.74 -0.13 18 -0.26 0.30 0.11 1.03 -0.75 0.74 0.05 19 -0.39 0.44 0.15 -0.72 -0.70 0.76 0.00 20 -0.14 0.48 -0.35 -0.63 1.10 0.72 0.13 21 -0.02 0.39 -0.01 0.29 0.86 0.70 0.08 22 -0.02 0.30 0.83 0.85 -0.23 0.70 -0.20 23 -0.26 0.53 -0.76 -1.24 0.87 0.74 0.21 24 -0.53 0.48 -0.38 -0.96 -1.24 0.78 0.14 25 0.21 0.48 -0.52 -0.31 0.94 0.66 0.19 26 -0.26 0.42 -0.29 -0.06 -0.75 0.74 0.13 27 -0.26 0.29 0.67 0.74 0.55 0.74 -0.10 28 -0.02 0.32 2.11 0.48 -0.95 0.70 -1.15 29 0.54 0.52 -0.62 -0.63 0.03 0.60 0.26 30 0.75 0.45 0.01 0.15 0.06 0.56 0.07 31 0.32 0.59 -1.03 -1.66 0.31 0.64 0.35 32 0.21 0.44 -0.21 0.06 -1.36 0.66 0.14 33 1.27 0.38 0.64 1.17 -0.39 0.46 -0.15 34 0.32 0.45 -0.10 -0.15 -0.93 0.64 0.09 35 -0.14 0.52 -0.61 -1.10 0.05 0.72 0.21 36 0.96 0.59 -1.09 -1.34 0.70 0.52 0.42 37 0.43 0.31 0.59 1.45 0.33 0.62 -0.15 38 0.54 0.64 -1.37 -2.37 1.76 0.60 0.45 39 1.06 0.58 -0.98 -1.21 0.23 0.50 0.38
(appendix continues)
![Page 191: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/191.jpg)
Appendix C (continued)
182
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
40 0.54 0.51 -0.50 -0.70 0.03 0.60 0.21 41 0.54 0.39 0.19 0.76 -0.09 0.60 0.01 42 0.64 0.50 1.10 -0.87 0.42 0.58 -0.65 43 0.64 0.45 -0.04 0.04 0.42 0.58 0.09 44 0.64 0.50 -0.35 -0.56 -0.68 0.58 0.16 45 0.54 0.26 0.93 2.06 2.39 0.60 -0.27 46 1.06 0.46 0.07 0.17 0.23 0.50 0.04 47 0.96 0.46 0.08 0.16 0.14 0.52 0.02 48 0.96 0.36 0.85 1.20 -1.11 0.52 -0.28 49 0.64 0.46 -0.23 0.08 -0.68 0.58 0.17 50 1.06 0.53 -0.51 -0.53 -0.96 0.50 0.25
Mean S.D.
0.00 0.72
0.40 0.13
0.04 0.72
-0.03 1.00
0.23 1.02
0.69
Groups Note: Raw score mean = 34.30 with a S.D. of 9. Mean person ability =1.13 with a S.D. of 1.33. Test reliability (K.R. 20) = 0.91. Reliability of person separation = 0.91.
80.
![Page 192: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/192.jpg)
183
Appendix C (continued)
Experiment 51 • Summary of Ttem Fit Information for a Uniformly Distributed Item Difficulty Distribution With 100 Items. 50 Persons, and a 50% Chance of Guessing Correctly
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
1 -0.47 0.37 0.14 0.22 0.13 0.80 0.01 2 -1.42 0.44 -0.64 -0.33 0.46 0.90 0.08 3 -0.98 0.46 -0.46 -0.40 -0.97 0.86 0.07 4 -0.47 0.31 0.19 0.58 -1.14 0.80 0.03 5 -1.42 0.31 0.04 0.12 -0.71 0.90 0.04 6 -0.98 0.33 -0.14 0.28 0.82 0.86 0.06 7 -0.47 0.19 1.45 0.75 1.60 0.80 -0.24 8 -1.18 0.43 -0.53 -0.26 0.64 0.88 0.08 9 -1.69 0.24 0.33 0.20 -0.16 0.92 0.01
10 -0.47 0.20 1.21 0.92 1.60 0.80 -0.21 11 -0.98 0.54 -1.15 -0.62 0.82 0.86 0.12 12 -0.63 0.18 0.99 0.90 0.87 0.82 -0.11 13 -0.63 0.25 1.21 0.45 0.87 0.82 -0.20 14 -0.98 0.36 -0.24 0.16 0.82 0.86 0.07 15 -1.18 0.45 -0.83 -0.17 0.64 0.88 0.09 16 -0.79 0.37 -0.29 0.16 -0.52 0.84 0.07 17 -0.19 0.44 -0.26 -0.07 0.66 0.76 0.06 18 -0.47 0.27 1.31 0.52 -1.14 0.80 -0.31 19 0.07 0.53 -0.58 -0.70 1.14 0.72 0.13 20 -0.47 0.64 -1.47 -1.49 1.32 0.80 0.20 21 -0.47 0.56 -1.08 -0.82 1.32 0.80 0.17 22 -0.98 0.42 -0.18 -0.35 0.32 0.86 0.06 23 -0.47 0.38 -0.29 0.32 -1.14 0.80 0.07 24 -0.79 0.44 -0.34 -0.34 -0.52 0.84 0.08 25 -0.79 0.43 -0.42 -0.16 -0.52 0.84 0.08 26 -2.03 0.21 0.14 0.33 0.05 0.94 0.03 27 -0.47 0.51 -0.88 -0.53 0.13 0.80 0.15 28 -0.47 0.20 1.51 0.87 -1.14 0.80 -0.32 29 -0.19 0.44 -0.03 -0.16 -0.64 0.76 0.03 30 -0.79 0.48 -0.86 -0.32 0.99 0.84 0.12 31 -0.19 0.42 0.23 -0.16 -0.34 0.76 -0.00 32 -0.06 0.56 -1.08 -0.82 1.83 0.74 0.23 33 -0.19 0.43 -0.10 -0.08 -0.34 0.76 0.05 34 -0.33 0.42 0.90 -0.37 -1.17 0.78 -0.23 35 -0.33 0.40 0.87 -0.13 -1.17 0.78 -0.24 36 0.30 0.35 0.97 0.59 0.42 0.68 -0.29 37 0.19 0.37 0.65 0.38 0.79 0.70 -0.11 38 -0.19 0.41 0.87 -0.19 0.87 0.76 -0.22 39 -0.79 0.27 0.23 0.58 -0.12 0.84 0.02
(appendix continues)
![Page 193: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/193.jpg)
184
Appendix C (continued)
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
40 0.07 0.27 1.55 0.81 1.14 0.72 -0.39 41 0.30 0.61 -1.27 -1.35 1.59 0.68 0.28 42 -0.79 0.27 0.15 0.66 -0.52 0.84 0.02 43 -0.33 0.31 0.31 0.68 0.09 0.78 -0.02 44 -0.47 0.35 0.15 0.30 0.48 0.80 0.01 45 0.19 0.52 -0.61 -0.63 -0.85 0.70 0.17 46 -0.06 0.19 1.59 1.36 0.50 0.74 -0.40 47 -0.19 0.40 0.17 0.10 -0.34 0.76 0.01 48 0.41 0.60 -1.43 -1.17 1.82 0.66 0.35 49 0.30 0.32 1.17 0.68 1.34 0.68 -0.26 50 0.19 0.49 -0.17 -0.47 0.42 0.70 0.01 51 -0.06 0.41 -0.01 0.14 -0.83 0.74 0.05 52 0.63 0.46 0.22 -0.21 -0.32 0.62 -0.03 53 0.41 0.39 0.87 0.24 0.02 0.66 -0.24 54 0.07 0.51 -0.87 -0.37 0.10 0.72 0.20 55 0.30 0.20 1.41 1.72 2.14 0.68 -0.39 56 0.07 0.32 0.23 0.96 -1.52 0.72 -0.03 57 -0.06 0.18 1.41 1.38 2.35 0.74 -0.29 58 0.41 0.44 0.57 -0.10 -1.54 0.66 -0.20 59 -0.06 0.33 0.48 0.60 0.50 0.74 -0.05 60 0.07 0.64 -1.49 -1.63 1.14 0.72 0.29 61 0.52 0.60 -1.04 -1.40 0.37 0.64 0.31 62 0.07 0.43 0.16 -0.06 0.10 0.72 -0.05 63 -0.33 0.31 0.24 0.71 -1.17 0.78 -0.01 64 0.52 0.51 -0.59 -0.47 -0.43 0.64 0.21 65 0.63 0.30 0.95 1.27 1.14 0.62 -0.30 66 0.30 0.37 0.21 0.62 0.42 0.68 -0.03 67 0.74 0.45 -0.26 0.09 -1.22 0.60 0.13 68 0.07 0.35 0.21 0.65 0.11 0.72 -0.04 69 -0.19 0.10 1.67 1.71 0.87 0.76 -0.33 70 0.63 0.63 -1.69 -1.59 0.70 0.62 0.48 71 0.52 0.50 -0.47 -0.40 1.27 0.64 0.13 72 0.63 0.52 -0.80 -0.51 0.24 0.62 0.27 73 0.07 0.46 0.11 -0.39 0.11 0.72 0.02 74 0.52 0.55 -0.97 -0.85 -0.81 0.64 0.30 75 0.30 0.39 0.42 0.40 -0.38 0.68 -0.11 76 0.41 0.40 0.41 0.37 0.02 0.66 -0.07 77 0.30 0.63 -1.55 -1.44 0.71 0.68 0.36 78 0.74 0.58 -0.71 -1.36 1.02 0.60 0.19 79 0.07 0.62 -1.36 -1.37 0.10 0.72 0.28 80 0.63 0.48 -0.34 -0.20 0.70 0.62 0.14 81 0.52 0.54 -0.78 -0.79 0.37 0.64 0.24 82 0.94 0.43 0.35 0.12 -1.51 0.56 -0.17
(appendix continues)
![Page 194: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/194.jpg)
185
Appendix C (continued)
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
83 0.52 0.56 -1.10 -0.90 1.27 0.64 0.32 84 1.04 0.44 -0.15 0.20 -0.55 0.54 0.11 85 1.14 0.62 -1.63 -1.73 -0.21 0.52 0.53 86 0.52 0.49 -0.36 -0.39 0.37 0.64 0.10 87 0.30 0.56 -1.01 -0.90 1.59 0.68 0.26 88 0.41 0.39 0.25 0.41 1.00 0.66 0.01 89 1.24 0.50 -0.51 -0.49 -0.87 0.50 0.21 90 0.07 0.20 1.66 1.32 0.11 0.72 -0.45 91 0.52 0.42 -0.09 0.40 -0.43 0.64 0.08 92 0.84 0.43 0.04 0.25 0.39 0.58 0.02 93 1.14 0.31 1.27 1.27 1.70 0.52 -0.46 94 0.94 0.29 1.23 1.44 1.66 0.56 -0.44 95 0.94 0.40 0.34 0.55 -1.51 0.56 -0.08 96 1.85 0.50 -0.57 -0.61 1.31 0.38 0.19 97 0.52 0.54 -0.92 -0.72 0.37 0.64 0.28 98 1.14 0.30 1.30 1.37 0.04 0.52 -0.46 99 0.84 0.39 0.33 0.59 0.39 0.58 -0.06
100 0.84 0.23 1.72 1.82 1.24 0.58 -0.58 Mean S.D. Groups
-0.00 0.70
0.41 0.12
0.02 0.88
0.03 0.80
0.26 0.92 2
0.72
Mean person ability = 1.23 with a S.D. of 1.09. Test reliability (K.R. 20) = 0.95. Reliability of person separation = 0.93.
51.
![Page 195: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/195.jpg)
186
Appendix C (continued)
Experiment 52: Summary of Item Fit Information for a Uniformly Distributed Item Difficulty Distribution With 25 Items. 100 Persons, and a 50% Chance of Guessing Correctly
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
1 -0.69 0.36 0.21 -0.01 0.72 0.83 -0.00 2 -0.78 0.31 0.33 0.30 0.02 0.84 -0.00 3 -0.61 0.17 2.19 1.04 2.02 0.82 -0.36 4 -0.61 0.38 -0.40 0.12 -0.67 0.82 0.06 5 -1.28 0.29 0.64 0.07 -1.15 0.89 -0.05 6 -0.32 0.31 0.95 0.42 2.37 0.78 -0.10 7 -0.61 0.36 -0.33 0.22 -0.62 0.82 0.05 8 -0.46 0.40 0.01 -0.22 -1.59 0.80 0.02 9 -0.18 0.43 -0.71 -0.07 0.74 0.76 0.12
10 -0.32 0.31 1.18 0.54 -0.65 0.78 -0.21 11 0.19 0.23 1.41 1.58 2.78 0.70 -0.26 12 -0.25 0.50 -0.58 -0.94 -1.22 0.77 0.08 13 -0.32 0.33 0.45 0.41 1.05 0.78 -0.04 14 0.24 0.43 0.09 -0.25 0.19 0.69 0.02 15 -0.05 0.51 -1.25 -0.77 1.74 0.74 0.18 16 0.01 0.58 -1.79 -1.52 1.89 0.73 0.25 17 0.41 0.28 1.16 1.40 0.79 0.66 -0.22 18 0.52 0.37 0.93 0.41 0.25 0.64 -0.23 19 0.77 0.36 0.59 0.85 0.84 0.59 -0.08 20 0.57 0.41 0.32 0.08 0.09 0.63 -0.06 21 0.41 0.56 -1.63 -1.47 0.73 0.66 0.32 22 0.52 0.54 -1.45 -1.42 1.15 0.64 0.31 23 0.67 0.46 -0.09 -0.53 2.24 0.61 -0.04 24 0.97 0.44 -0.27 -0.30 0.89 0.55 0.07 25 1.21 0.48 -0.89 -0.88 0.47 0.50 0.25
Mean S.D. Groups
0.00 0.62
0.39 0.10
0.04 1.00
-0.04 0.83
0.60 1.17 2
0.72
Mean person ability = 1.20 with a S.D. of 1.03. Test reliability (K.R. 20) = 0.79. Reliability of person separation = 0.71.
![Page 196: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/196.jpg)
187
Appendix C (continued)
Expe r imen t 53- Summary of Ttem Fit Information for a Uniformly Distributed Item Difficulty Distribution With 50 Items. 100 Persons, and a 50% Chance of Guessing Correctly
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
1 -0.90 0.29 -0.02 0.17 -0.82 0.83 0.04 2 -1.25 0.24 -0.19 0.39 -0.40 0.87 0.05 3 -1.07 0.24 0.30 0.34 -0.13 0.85 -0.00 4 -0.99 0.26 0.68 0.11 -0.83 0.84 -0.07 5 -0.90 0.32 -0.18 -0.04 0.30 0.83 0.04 6 -0.90 0.40 -0.54 -0.65 -0.50 0.83 0.07 7 -0.90 0.25 0.41 0.38 -0.82 0.83 -0.03 8 -0.90 0.39 -0.67 -0.52 -0.50 0.83 0.09 9 -1.16 0.33 -0.33 -0.16 -0.13 0.86 0.05
10 -0.61 0.51 -1.49 -1.30 1.89 0.79 0.17 11 -0.29 0.44 -0.92 -0.57 -0.07 0.74 0.16 12 -0.54 0.39 -0.37 -0.38 -0.19 0.78 0.07 13 -0.11 0.34 0.04 0.42 -1.18 0.71 0.02 14 -0.99 0.33 -0.31 -0.03 -0.83 0.84 0.05 15 -0.61 0.28 0.24 0.53 -1.01 0.79 -0.01 16 -0.35 0.48 -1.23 -0.98 1.20 0.75 0.18 17 -0.06 0.42 -0.45 -0.43 -0.86 0.70 0.11 18 -0.41 0.37 -0.24 -0.08 0.29 0.76 0.05 19 -0.06 0.51 -1.33 -1.30 0.13 0.70 0.24 20 -0.23 0.31 0.19 0.65 0.18 0.73 0.00 21 0.11 0.28 1.15 0.91 1.29 0.67 -0.20 22 -0.29 0.52 -1.51 -1.40 1.97 0.74 0.22 23 0.21 0.39 -0.15 0.09 -0.08 0.65 0.06 24 -0.17 0.27 0.78 0.86 1.25 0.72 -0.10 25 0.05 0.24 0.86 1.54 0.26 0.68 -0.15 26 0.26 0.47 -1.14 -0.79 0.19 0.64 0.24 27 0.21 0.45 0.04 -0.83 1.23 0.65 -0.08 28 -0.11 0.28 0.54 0.91 0.33 0.71 -0.06 29 0.26 0.44 -0.54 -0.46 -0.61 0.64 0.14 30 0.21 0.45 -0.60 -0.67 -0.08 0.65 0.15 31 0.32 0.42 0.25 -0.56 0.28 0.63 -0.06 32 0.21 0.38 -0.26 0.29 -0.08 0.65 0.07 33 0.42 0.39 0.05 0.06 -0.46 0.61 0.01 34 0.26 0.30 0.91 0.97 -1.12 0.64 -0.16 35 0.42 0.14 3.23 2.68 3.82 0.61 -0.99 36 0.52 0.44 -0.37 -0.42 -0.62 0.59 0.11 37 0.42 0.50 -1.41 -1.18 0.92 0.61 0.33 38 0.52 0.41 0.46 -0.20 0.80 0.59 -0.19 39 0.52 0.48 -0.08 -1.15 0.16 0.59 -0.04
(appendix continues)
![Page 197: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/197.jpg)
Appendix C (continued)
188
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
40 0.52 0.37 0.16 0.49 -0.62 0.59 -0.03 41 0.81 0.27 1.39 1.85 1.76 0.53 -0.36 42 0.66 0.45 -0.59 -0.39 -0.38 0.56 0.19 43 0.91 0.38 0.11 0.58 1.27 0.51 0.02 44 0.76 0.27 1.78 1.78 1.99 0.54 -0.53 45 0.81 0.33 1.72 0.81 0.04 0.53 -0.57 46 1.10 0.48 -0.81 -0.79 -1.11 0.47 0.23 47 0.66 0.36 0.63 0.62 -0.38 0.56 -0.18 48 0.91 0.27 1.55 1.91 1.78 0.51 -0.39 49 1.05 0.47 -0.94 -0.64 -1.21 0.48 0.28 50 0.71 0.49 -1.05 -1.01 -0.05 0.55 0.29
Mean S.D.
0.00 0.65
0.37 0.09
-0.01 0.94
0.05 0.92
0.17 1.04
0.68
Groups Note: Raw score mean = 33.96 with a S.D. of 8. Mean person ability = 0.97 with a S.D. of 1.00. Test reliability (K.R. 20) = 0.88. Reliability of person separation = 0.86.
63.
![Page 198: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/198.jpg)
189
Appendix C (continued)
Exper iment 54- Summary of Ttem Fit Information for a Uniformly Distributed Item Difficulty Distribution With 100 Items. 100 Persons, and a 50% Chance of Guessing Correctly
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
1 -1.21 0.20 0.51 0.59 -0.93 0.89 -0.02 2 -0.55 0.31 0.63 0.14 0.67 0.82 -0.06 3 -1.10 0.34 0.27 -0.13 -0.54 0.88 -0.02 4 -1.46 0.20 0.58 0.48 -0.92 0.91 -0.04 5 -0.90 0.48 -0.96 -0.74 0.04 0.86 0.09 6 -0.72 0.45 -1.04 -0.45 1.34 0.84 0.11 7 -0.99 0.42 -0.62 -0.46 -0.23 0.87 0.07 8 -0.63 0.30 0.45 0.20 0.95 0.83 -0.02 9 -0.72 0.48 -1.28 -0.57 1.34 0.84 0.12
10 -0.63 0.37 -0.27 -0.07 0.06 0.83 0.06 11 -0.48 0.49 -1.05 -0.79 0.27 0.81 0.12 12 -0.80 0.42 -0.77 -0.38 1.18 0.85 0.09 13 -0.55 0.42 -0.66 -0.31 -1.23 0.82 0.09 14 -0.72 0.15 1.05 1.09 0.38 0.84 -0.10 15 -1.33 0.27 0.49 0.34 0.23 0.90 -0.07 16 -0.63 0.20 2.16 0.56 0.06 0.83 -0.36 17 -0.40 0.28 1.08 0.33 0.90 0.80 -0.13 18 -0.99 0.35 -0.63 0.11 0.84 0.87 0.08 19 -1.33 0.07 1.20 0.95 1.32 0.90 -0.08 20 -0.40 0.33 0.03 0.30 -0.43 0.80 0.03 21 -0.55 0.43 -0.53 -0.46 0.02 0.82 0.07 22 -0.40 0.45 -0.80 -0.58 -0.43 0.80 0.11 23 -0.80 0.36 -0.11 -0.11 -0.94 0.85 0.03 24 -0.90 0.44 -0.52 -0.56 0.04 0.86 0.06 25 -0.33 0.42 -0.44 -0.34 -1.44 0.79 0.08 26 -0.40 0.32 0.25 0.24 0.90 0.80 0.01 27 -0.48 0.27 0.63 0.53 1.17 0.81 -0.08 28 -0.40 0.29 0.33 0.57 0.90 0.80 -0.01 29 -0.06 0.45 -0.50 -0.64 -0.94 0.75 0.09 30 -0.33 0.36 -0.13 0.08 -1.44 0.79 0.05 31 0.17 0.48 -0.92 -0.86 -0.51 0.71 0.18 32 -0.26 0.33 0.01 0.39 -0.60 0.78 0.04 33 -0.55 0.39 -0.11 -0.18 0.02 0.82 0.02 34 -0.40 0.30 0.47 0.36 0.90 0.80 -0.03 35 -0.72 0.26 0.27 0.60 -0.74 0.84 -0.02 36 -0.63 0.30 0.78 0.21 -1.21 0.83 -0.10 37 -0.33 0.37 0.15 -0.08 -0.24 0.79 0.01 38 -0.19 0.56 -1.63 -1.46 1.74 0.77 0.20 39 0.17 0.39 -0.26 0.11 -1.17 0.71 0.08
(appendix continues-)
![Page 199: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/199.jpg)
190
Appendix C (continued)
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
40 -0.33 0.30 1.38 0.12 0.63 0.79 -0.21 41 -0.40 0.29 0.81 0.41 0.08 0.80 -0.09 42 0.06 0.38 -0.09 -0.02 1.73 0.73 0.06 43 -0.13 0.42 -0.14 -0.40 -0.22 0.76 0.04 44 0.12 0.46 -0.91 -0.60 0.77 0.72 0.15 45 0.34 0.26 2.46 0.88 -1.42 0.68 -0.73 46 0.29 0.48 -0.74 -0.89 0.79 0.69 0.14 47 0.06 0.38 -0.27 0.13 1.21 0.73 -0.03 48 0.12 0.40 -0.42 -0.01 -0.89 0.72 0.11 49 0.00 0.39 -0.35 -0.01 -0.03 0.74 0.09 50 0.12 0.36 -0.15 0.43 -0.89 0.72 0.05 51 0.40 0.42 -0.37 -0.27 -1.04 0.67 0.09 52 -0.19 0.39 0.07 -0.19 0.37 0.77 0.00 53 0.29 0.34 0.80 0.34 0.10 0.69 -0.15 54 -0.90 0.34 -0.28 0.12 0.04 0.86 0.05 55 0.12 0.28 0.63 0.98 1.49 0.72 -0.07 56 -0.06 0.43 -0.63 -0.43 -0.94 0.75 0.11 57 -0.48 0.22 0.42 1.04 0.38 0.81 -0.04 58 0.06 0.34 0.15 0.47 0.46 0.73 0.02 59 0.45 0.34 0.15 0.79 -1.11 0.66 0.01 60 0.17 0.41 -0.23 -0.15 0.99 0.71 0.05 61 0.34 0.42 0.14 -0.49 -0.29 0.68 0.01 62 0.12 0.40 -0.23 -0.11 -0.72 0.72 0.06 63 0.23 0.38 -0.21 0.29 0.56 0.70 0.06 64 0.23 0.39 -0.01 0.04 -0.48 0.70 0.03 65 0.61 0.41 0.13 -0.32 0.95 0.63 -0.03 66 0.50 0.28 0.94 1.26 0.92 0.65 -0.16 67 0.29 0.42 -0.56 -0.19 -0.78 0.69 0.13 68 0.40 0.38 0.82 -0.03 -0.66 0.67 -0.19 69 -0.26 0.39 -0.41 -0.02 -0.60 0.78 0.08 70 0.45 0.49 -1.21 -0.85 0.19 0.66 0.26 71 0.45 0.43 -0.49 -0.24 0.19 0.66 0.12 72 0.45 0.40 -0.40 0.07 -0.14 0.66 0.14 73 0.55 0.41 0.27 -0.38 0.05 0.64 -0.05 74 0.76 0.35 0.61 0.60 -0.65 0.60 -0.13 75 0.66 0.35 0.37 0.53 -0.97 0.62 -0.05 76 0.45 0.42 -0.08 -0.32 -0.62 0.66 0.03 77 1.00 0.46 -0.36 -0.88 0.03 0.55 0.08 78 0.34 0.39 -0.18 0.13 0.46 0.68 0.08 79 0.55 0.34 0.42 0.69 -0.02 0.64 -0.04 80 0.86 0.38 -0.00 0.38 0.19 0.58 0.04 81 0.71 0.46 -0.76 -0.75 -0.54 0.61 0.21 82 0.40 0.40 -0.12 -0.08 -0.66 0.67 0.06
(appendix continues)
![Page 200: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/200.jpg)
191
Appendix C (continued)
Item #
Logit item diff.
Point, bis. corr.
Unwt. total fit
Wt. total fit
Ability between
fit
Mean item score
Logit Residual
Index
83 0.76 0.41 -0.17 -0.17 -0.19 0.60 0.06 84 0.61 0.43 -0.73 -0.25 -0.42 0.63 0.18 85 0.91 0.45 -0.76 -0.62 -0.73 0.57 0.20 86 0.76 0.26 1.05 1.74 2.78 0.60 -0.20 87 0.95 0.41 -0.25 -0.00 -1.33 0.56 0.07 88 0.76 0.37 0.37 0.34 -0.65 0.60 -0.05 89 0.76 0.36 2.10 -0.01 -0.19 0.60 -0.75 90 0.40 0.33 0.51 0.66 0.17 0.67 -0.08 91 0.61 0.35 0.43 0.61 0.37 0.63 -0.07 92 0.55 0.45 -0.98 -0.44 0.05 0.64 0.23 93 0.66 0.30 0.79 1.16 1.30 0.62 -0.16 94 1.10 0.48 -1.13 -1.02 0.65 0.53 0.34 95 0.76 0.36 1.67 0.27 0.14 0.60 -0.60 96 0.61 0.45 -0.36 -0.68 1.51 0.63 0.09 97 0.71 0.40 -0.01 -0.06 -0.54 0.61 0.05 98 0.81 0.29 1.02 1.42 0.49 0.59 -0.21 99 0.86 0.37 0.32 0.39 -1.25 0.58 -0.07
100 1.19 0.49 -1.05 -1.24 1.21 0.51 0.30 Mean S.D. Groups
0.00 0.63
0.37 0.08
0.02 0.75
0.03 0.59
0.04 0.86 2
0.73
Note: Raw score mean = 72.78 with a S.D. of 16.61. Mean person ability = 1.23 with a S.D. of 1.00. Test reliability (K.R. 20) = 0.94. Reliability of person separation = 0.92.
![Page 201: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/201.jpg)
REFERENCES
Allen, M. J., & Yen, W. W. (1979). Introduction to measurement theory. Belmont, CA: Wadsworth.
Andersen, E. B. (1973). Goodness of fit test for the Rasch model. Psvchometrika, 38,123-140.
Andrich, D. (1988). Rasch models for measurement. Sage University paper series on Quantitative Applications in the Social Sciences (Series No. 07-068). Beverly Hills: Sage.
Bamdorff-Nielsen, O. (1978). Information and exponential families in statistical theory. New York: Wiley.
Crocker, L.,&Algina, J. (1986). Introduction to classical and mordern test theory. New York: Holt, Rinehart, and Winston.
Ferguson, G. A. (1981). Statistical analysis in psychology and education. New York: McGraw-Hill.
Glaser, R. (1949). A methodological analysis of the inconsistency of responses to test items. Educational and Psychological Measurement 9, 721-739.
Glaser, R. (1952). The reliability of inconsistency. Educational and Psychological Measurement. 11. 60-64.
Gustafsson, J-E. (1980). Testing and obtaining fit of data to the Rasch model. The British Journal of Mathematical and Statistical Psychology. 33. 205-233.
Habermann, S. (1977). Maximum likelihood estimates in exponential response models. The Annals of Statistics. 5. 815-841.
Lord, F. M., & Novick, M. R. (1968). Statistical theories of mental test scores. Reading, MA.: Addison-Wesley.
Luppescu, S. (1992). SIMTEST 2.1 [Computer program for simulating test data]. Chicago: University of Chicago Press.
192
![Page 202: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/202.jpg)
193
Mosier, C.I. (1941). Psychophysics and mental test theory: II. The constant process. Psychological Review. 47. 235-249.
Neter, J., Wasserman, W., & Whitmore, G. A. (1978). Applied statistics. Boston: Allyn and Bacon.
Rasch, G. (1961). On general laws and the meaning of measurement in psychology. Proceedings of the Fourth Berkeley Symposium on Mathematical statistics and probability. Berkeley: University of California Press, 4, 321-333.
Rasch, G. (1980). Probabilistic models for some intelligence and attainment tests (Rev. ed.). Chicago: University of Chicago Press.
Smith, R. M. (1985). A comparison of Rasch person analysis and robust estimators. Educational and Psychological Measurement. 45. 433-444.
Smith, R. M. (1986). Person fit in the Rasch model. Educational and Psychological Measurement. 46. 359-372.
Smith, R. M. (1988a). A comparison of the power of Rasch total and between item fit statistics to detect measurement disturbances. A paper presented at the annual meeting of the American Educational Research Association, New Orleans.
Smith, R. M. (1988b). The distributional properties of Rasch standardized residuals. Educational and Psychological Measurement 48. 657-667.
Smith, R. M. (1991a). The distributional properties of Rasch item fit statistics. Educational and Psychological Measurement. 51. 541-565.
Smith, R.M. (1991b). TP ARM: Item and person analysis with the Rasch model. Chicago: MESA Press.
Smith, R. M., & Hedges, L.V. (1982). A comparison of the likelihood ratio % and Pearsonian %2 tests of fit in the Rasch model. Educational Research and Perspectives. 9. 44-54.
Stevens, S. S. (1946). On the theory of scales of measurement. Science. 103. 677-680.
Thorndike, R. L. (1949). Personnel selection: Test and measurement techniques. New York: Teachers College Press.
![Page 203: MEASUREMENT DISTURBANCE EFFECTS ON RASCH FIT …/67531/metadc279376/... · Logit Residual Index (LRI) were examined through simulated data sets of varying sample sizes, test lengths,](https://reader033.vdocument.in/reader033/viewer/2022041921/5e6be316154c713196372e36/html5/thumbnails/203.jpg)
194
Thurstone, L. L., & Chave, E. J. (1929). Measurement of attitudes. Chicago:
University of Chicago Press.
Torgerson, W. S. (1958). Theory and methods of scaling. New York: Wiley.
Walker, H. M.; & Lev, J. (1953). Statistical inference. New York: Holt, Rinehart,
and Winston.
Wollenberg, A. L. van den. (1982). Two new test statistics for the Rasch model.
Psvchometrika. 47.123-140.
Wright, B. D. (1977). Solving measurement problems with the Rasch model. Tnnrnal of Educational Measurement. 14, 97-116.
Wright, B. D., & Douglas, G. A. (1977). Best procedures for sample-free item analysis. Annlied Psychological Measurement. 1. 281-294.
Wright, B. D., & Linacre, J. M. (1991). mnSTFPS: Rasch analysis for all two-facet
models. Chicago: MESA Press.
Wright, B. D., & Masters, G. N. (1982). Rating scale analysis. Chicago: MESA
Press.
Wright, B. W., & Mead, R. J. (1978). RTC ATCalibrating items with the Rasch model (Research Memorandum 23A). Chicago: University of Chicago, MESA Statistical Laboratory, Department of Education.
Wright, B. D., & Panchapakesan, N. (1969). A procedure for sample-free item analysis. Educational and Psychological Measurement. 29. 23-48.
Wright, B. D., & Stone, M. (1979). Best test design. Chicago: MESA Press.