elaboration elaboration extends our knowledge about an association to see if it continues or changes...
TRANSCRIPT
![Page 1: Elaboration Elaboration extends our knowledge about an association to see if it continues or changes under different situations, that is, when you introduce](https://reader035.vdocument.in/reader035/viewer/2022072008/56649d755503460f94a55f18/html5/thumbnails/1.jpg)
Elaboration
Elaboration extends our knowledge about an association to see if it continues or changes under different situations, that is, when you introduce an additional variable. This is sometimes referred to as a control variable because you are seeing if the original relationship changes or continues when you control for (hold the effects of) a new variable.
When you introduce one control variable the process is sometimes called first-order partialling. You can continue to add multiple variables, called second-order, third-order, and so on, for more elaborate models, but interpretation can get complex at that point especially if each of those variables has numerous values or categories. The original bivariate association is the zero-order relationship.
![Page 2: Elaboration Elaboration extends our knowledge about an association to see if it continues or changes under different situations, that is, when you introduce](https://reader035.vdocument.in/reader035/viewer/2022072008/56649d755503460f94a55f18/html5/thumbnails/2.jpg)
Voting in Election * Race of Respondent Crosstabulation
893 101 38 1032
71.2% 62.0% 50.7% 69.2%
337 56 27 420
26.9% 34.4% 36.0% 28.2%
19 5 10 34
1.5% 3.1% 13.3% 2.3%
5 1 6
.4% .6% .4%
1254 163 75 1492
100.0% 100.0% 100.0% 100.0%
Count
% within Race ofRespondent
Count
% within Race ofRespondent
Count
% within Race ofRespondent
Count
% within Race ofRespondent
Count
% within Race ofRespondent
voted
did not vote
not eligible
refused
Voting inElection
Total
white black other
Race of Respondent
Total
Chi-Square Tests
54.646a 6 .000
33.994 6 .000
27.663 1 .000
1492
Pearson Chi-Square
Likelihood Ratio
Linear-by-LinearAssociation
N of Valid Cases
Value dfAsymp. Sig.
(2-sided)
4 cells (33.3%) have expected count less than 5. Theminimum expected count is .30.
a.
![Page 3: Elaboration Elaboration extends our knowledge about an association to see if it continues or changes under different situations, that is, when you introduce](https://reader035.vdocument.in/reader035/viewer/2022072008/56649d755503460f94a55f18/html5/thumbnails/3.jpg)
Voting in Election * Race of Respondent * Married ? Crosstabulation
531 39 17 587
76.6% 67.2% 42.5% 74.2%
151 17 15 183
21.8% 29.3% 37.5% 23.1%
7 2 8 17
1.0% 3.4% 20.0% 2.1%
4 4
.6% .5%
693 58 40 791
100.0% 100.0% 100.0% 100.0%
362 61 21 444
64.5% 58.7% 60.0% 63.4%
186 39 12 237
33.2% 37.5% 34.3% 33.9%
12 3 2 17
2.1% 2.9% 5.7% 2.4%
1 1 2
.2% 1.0% .3%
561 104 35 700
100.0% 100.0% 100.0% 100.0%
Count
% within Race ofRespondent
Count
% within Race ofRespondent
Count
% within Race ofRespondent
Count
% within Race ofRespondent
Count
% within Race ofRespondent
Count
% within Race ofRespondent
Count
% within Race ofRespondent
Count
% within Race ofRespondent
Count
% within Race ofRespondent
Count
% within Race ofRespondent
voted
did not vote
not eligible
refused
Voting inElection
Total
voted
did not vote
not eligible
refused
Voting inElection
Total
Married ?yes
no
white black other
Race of Respondent
Total
![Page 4: Elaboration Elaboration extends our knowledge about an association to see if it continues or changes under different situations, that is, when you introduce](https://reader035.vdocument.in/reader035/viewer/2022072008/56649d755503460f94a55f18/html5/thumbnails/4.jpg)
Chi-Square Tests
75.921a 6 .000
39.848 6 .000
34.205 1 .000
791
4.865b 6 .561
3.926 6 .687
2.027 1 .155
700
Pearson Chi-Square
Likelihood Ratio
Linear-by-LinearAssociation
N of Valid Cases
Pearson Chi-Square
Likelihood Ratio
Linear-by-LinearAssociation
N of Valid Cases
Married ?yes
no
Value dfAsymp. Sig.
(2-sided)
5 cells (41.7%) have expected count less than 5. The minimumexpected count is .20.
a.
5 cells (41.7%) have expected count less than 5. The minimumexpected count is .10.
b.
![Page 5: Elaboration Elaboration extends our knowledge about an association to see if it continues or changes under different situations, that is, when you introduce](https://reader035.vdocument.in/reader035/viewer/2022072008/56649d755503460f94a55f18/html5/thumbnails/5.jpg)
multiple correlation (R) is based on the Pearson r correlation coefficient and essentially looks at the combined effects of two or more independent variables on the dependent variable. These variables should be interval/ratio measures, dichotomies, or ordinal measures with equal appearing intervals, and assume a linear relationship between the independent and dependent variable.
Similar to r, R is a PRE when squared. However, unlike bivariate r, multiple R cannot be negative since it represents the combined impact of two or more independent variables, so direction is not given by the coefficient.
Multiple R2 tell us the proportion of the variation in the dependent variable that can be explained by the combined effect of the independent variables
Multiple Correlation
![Page 6: Elaboration Elaboration extends our knowledge about an association to see if it continues or changes under different situations, that is, when you introduce](https://reader035.vdocument.in/reader035/viewer/2022072008/56649d755503460f94a55f18/html5/thumbnails/6.jpg)
Uncovering which of the independent variables are contributing more or less to the explanations and predictions of the dependent variable is accomplished by a widely used technique called linear regression. It is based on the idea of a straight line which has the formula
Y= a + bX
Y is the value of the predicted dependent variable, sometimes called the criterion and in some formulas represented as Y' to indicate Y-predicted;X is the value of the independent variable or predictor; a is the constant or the value of Y when X is unknown, that is, zero; it is the point on the Y axis where the line crosses when X is zero; and b is the slope or angle of the line and, since not all the independent variables are contributing equally to explaining the dependent variable, b represents the unstandardized weight by which you adjust the value of X. For each unit of X, Y is predicted to go up or down by the amount of b.
Regression
![Page 7: Elaboration Elaboration extends our knowledge about an association to see if it continues or changes under different situations, that is, when you introduce](https://reader035.vdocument.in/reader035/viewer/2022072008/56649d755503460f94a55f18/html5/thumbnails/7.jpg)
the regression line which predicts the values of Y, the outcome variable, when you know the values of X, the independent variables. What linear regression analysis does, is calculate the constant (a), the coefficient weights for each independent variable (b), and the overall multiple correlation (R). Preferably, low intercorrelations exist among the independent variables in order to find out the unique impact of each of the predictors. This is the formula for a multiple regression line:
Y' = a + bX1 + bX2 + bX3 + bX4 …. + bXn
The information provided in the regression analysis includes the b coefficients for each of the independent variables and the overall multiple R correlation and its corresponding R2. Assuming the variables are measured using different units as they typically are (such as pounds of weight, inches of height, or scores on a test), then the b weights are transformed into standardized units for comparison purposes. These are called Beta () coefficients or weights and essentially are interpreted like correlation coefficients: Those furthest away from zero are the strongest and the plus or minus sign indicates direction.
![Page 8: Elaboration Elaboration extends our knowledge about an association to see if it continues or changes under different situations, that is, when you introduce](https://reader035.vdocument.in/reader035/viewer/2022072008/56649d755503460f94a55f18/html5/thumbnails/8.jpg)
Model Summary
.294a .086 .080 4.754Model1
R R SquareAdjustedR Square
Std. Error ofthe Estimate
Predictors: (Constant), Size of Place in 1000s, HoursPer Day Watching TV, Respondent's Sex, Highest Yearof School Completed
a.
ANOVAb
1299.940 4 324.985 14.381 .000a
13740.066 608 22.599
15040.007 612
Regression
Residual
Total
Model1
Sum ofSquares df Mean Square F Sig.
Predictors: (Constant), Size of Place in 1000s, Hours Per Day Watching TV,Respondent's Sex, Highest Year of School Completed
a.
Dependent Variable: Age When First Marriedb.
![Page 9: Elaboration Elaboration extends our knowledge about an association to see if it continues or changes under different situations, that is, when you introduce](https://reader035.vdocument.in/reader035/viewer/2022072008/56649d755503460f94a55f18/html5/thumbnails/9.jpg)
Coefficientsa
22.204 1.094 20.290 .000
.307 .062 .202 4.978 .000
2.360E-02 .088 .011 .267 .790
-2.112 .392 -.210 -5.384 .000
5.474E-05 .000 .011 .283 .777
(Constant)
Highest Year ofSchool Completed
Hours Per DayWatching TV
Respondent's Sex
Size of Place in 1000s
Model1
B Std. Error
UnstandardizedCoefficients
Beta
Standardized
Coefficients
t Sig.
Dependent Variable: Age When First Marrieda.
SNSex: 1=Male, 2=Female
In Words: Respondents who marry at younger ages tend to have less education and are female. Those who marry later tend to have more education and are male.