Mean Response
Regression or Logit Models
controlling for Sociodemographic variables
Women
Men
p-value
Coef
for “female”
Std Error
for “female”
p-value
for “female”
Table 5:
Women’s Rights,
Women’s Issues, and
Abortion Issues
W pred.prob M pred.prob Diff (W – M)
Therm: Women’s Libbers (0-97)
63.78 60.53 .003
2.53
1.17 .030
Therm: Feminists (0-97)
56.6 51.9 .000
4.09
.930 .000
Favorable:FeministMvmt (0-100)
(‘00)
52.54 46.29 .000
4.926
.339 .000
Favorable:FeministMvmt (0-10)
(‘04)
5.32 4.79 .000
.344
.033 .000
Women should not have an
equal role to men in running
things (1-7)
1.90 2.08 .007 -.181
.072 .012
Should do Less to stop job
discrimination against women
(1-4)
1.57 1.75 .000 -.104
.012 .000
Think Male Pres better at Most
Important Issue (1-3)
1.88 2.12 .000 -.221
.063 .001
Abortion
.513 .440 .000
.270
.063 .000
Oppose making RU-486
available (0/1)
.493
.426
.067
.352 .335 .000
.089
.022 .000
Fed. Govt should restrict
abortion (0/1)
.336
.316
.020
Scale of each variable in (parentheses) after the variable name.
[N= 529 to 23,460]
Cells are: Women’s and Men’s mean responses and the 2-sided difference of means t-test p-value.
: Coefficients, standard errors, and p-values for “female” in logit or regression models run with control variables
for age, race, education, income, and 7-point party ID. (Coefficients are bolded if statistically significant at p<.05)
: For Logit Models, the predicted probabilities of answering “1” for Women and Men (with all other variables set
to their means) and the difference between Women’s and Men’s predicted probabilities are included in italics.
(The difference is bolded when the 95% CIs do not overlap.)These are predicted using Clarify (King et al. 2000).
“Female” was not statistically significant for the following variables in this category:
- Abortion should never be allowed (t-test pvalue:.976) (regression coefficient pvalue:.449)
- Strongly favor laws restricting abortion (.196) (.694)
Crowder-Meyer
- 23