ch 12 – inference for proportions yms 12.1 inference for a population proportion
TRANSCRIPT
CH 12 – INFERENCE FOR PROPORTIONS
YMS 12.1Inference for a Population Proportion
Ch 9 Sampling Distributions
is an unbiased estimator of population proportion p
Standard deviation of is if the population is at least 10 times n
Sampling distribution of is approximately normal if np and n(1-p) are at least 10
Use z-scores to standardize
p
p
p
p p
n
( )1
Conditions for Inference
To be representative: Data are from an SRS from the population of interest
To accurately calculate standard deviation: Population is at least 10 times n
To use normal calculations: Counts of successes/failures must be at least 10
Standard Error Replacing p with in standard deviation
formula
Test of Significance Ho: p = po
Verify that npo and n(1-po) are at least 10 Formula
Confidence Interval Verify that n and n(1- ) are at least 10 Form
p p
p
zp p
p pn
o
o o
( )1
*( )
p zp p
n
1
Choosing the Sample Size
Margin of error
Use a guess for p* Based on previous data Use the conservative estimate of 0.5 (unless
you believe is closer to 0 or 1 because then p* = 0.5 will give you a much larger sample size than necessary)
p
zp p
nm*
* ( * )1
Which to use in formulasand conditions?
Hypothesis Tests Use po because that is the distribution
you’re comparing your result to Confidence Intervals
Use because you don’t have any other values (remember you’re using the CI to estimate the true proportion p)
p698 #12.14-12.15, 12.17
p
YMS 12.2Comparing Two Proportions
Sampling Distribution of When samples are large, the sampling
distribution is approximately normal.
Mean
Variance
p p1 2
p p p p p p1 2 1 2 1 2
( ) ( )p p p p
p p
n
p p
n1 2 1 2
2 2 2 1 1
1
2 2
2
1 1
Confidence Intervals for Comparing Two Proportions Same form as for two means and
standard error is replacing p with
Conditions are still SRS, a population at least 10 times n, but now n1 , n1(1- ), n2 , and n2(1- ) are all greater than 5
p706 #12.22-12.23
p
p1 p1p 2
p 2
Pooled Sample Proportion
Because both samples actually come from one huge population, we combine the sample results to estimate the unknown population proportion p
FormulaX X
n n1 2
1 2
Significance Tests for Two Props
Replace and with pooled in standard error formula and the conditions for count of successes and failures
Other conditions remain the same!
Test Stat
p707 #12.24-12.26
A Civil Action – text, video and article
p1 p 2 p
1 2
1 2
ˆ ˆ
1 1ˆ ˆ(1 )
p pz
p pn n