Skip to main content
. 2015 Aug 4;112(33):E4512–E4521. doi: 10.1073/pnas.1419828112

Table S4.

Treatment effect estimates for study 2 voting preferences

Predictor variable Presearch vote Postsearch vote
Coefficient SE Coefficient SE
Intercept −0.073 0.540 0.062 0.543
Sex
 Female 0 Referent 0 Referent
 Male 0.039 0.110 −0.135 0.119
 Other −0.430 0.922 −0.568 0.924
Race/ethnicity
 White 0 Referent 0 Referent
 Black 0.115 0.224 0.090 0.245
 Hispanic −0.435 0.235 −0.280 0.237
 Asian 0.366 0.238 0.668 0.291*
 Other 0.133 0.274 −0.072 0.291
Age group
 18–24 0 Referent 0 Referent
 25–44 −0.024 0.144 −0.083 0.157
 45–64 0.241 0.184 0.029 0.200
 65+ 0.258 0.411 0.685 0.519
Education level
 Less than ninth grade 0 Referent 0 Referent
 Ninth to 12th grade 0.024 0.548 0.732 0.550
 High school graduate 0.074 0.528 0.927 0.528
 Bachelors 0.094 0.529 0.842 0.530
 Advanced −0.050 0.543 0.549 0.544

The presearch and postsearch columns report the estimate and variance for both treatment groups using classical regression poststratification. Data for sex, race/ethnicity, age group, and education level came from the 2010 US Census. Data on the number of people who identify their sex as “other” came from a 2011 Gallup study.

*

P < 0.05.