Homework #9—Logistic regression—Page 1
Soc 63993, Advanced Social Statistics II
Homework No. 9
Logistic Regression
I.
As we saw in the class handout on the PSI teaching example, 8 of the 14 students who
were in PSI got A’s compared to only 3 of the 18 students who were in a conventional classroom.
Verify that those numbers are consistent with the following results that we get when GRADE is
(logistically) regressed on PSI only. Recall that GRADE = 1 if grade is an A, 0 otherwise, PSI =
1 if in psi, 0 otherwise. [HINT: Compute the log odds for those in psi and those not in psi, and
then take it from there.]
. use http://www.nd.edu/~rwilliam/xsoc63993/statafiles/logist.dta
. logit grade psi, nolog
Logistic regression
Number of obs
=
32
LR chi2(1)
=
5.84
Prob > chi2
=
0.0156
Log likelihood = 17.670815
Pseudo R2
=
0.1418

grade 
Coef.
Std. Err.
z
P>z
[95% Conf. Interval]
+
psi 
1.89712
.831665
2.28
0.023
.2670865
3.527153
_cons 
1.609438
.6324555
2.54
0.011
2.849028
.3698478

II.
Download
lrb.dta
from the course web page. We use a sample of Southern Baptists from
the GSS in this homework. General Social Surveys from 1973 to 1991 are used to make one big
sample. All married Southern Baptists between the ages of 20 to 25 (all 61 of them!) are in the
data file. The dependent variable is
happymar
, respondent’s marital happiness (1 = Very
Happy, 0 = Otherwise).
church
, Church attendance (1 = Often attends, 0 = other),
female
(1
= female, 0 = male), and
educ
, Years of education, are the DVs.
Use Stata to run the logistic regression of happymar on church, female and educ. Then answer
the following questions.
1.
What assumptions of OLS would be violated if OLS was used to approach this problem?
2.
Interpret the logistic regression coefficients. What do the parameters tell you about the
determinants of marital happiness? What can you say about the size and magnitude of
effects?
3.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
This is the end of the preview.
Sign up
to
access the rest of the document.
 Spring '11
 RichardWilliams
 Statistics, Regression Analysis, Logit, Logistic function, Multinomial logit

Click to edit the document details