Download Religious Commitment's Influence on Political Views of Evangelicals and Non-Evangelicals and more Study notes Political Science in PDF only on Docsity! GVPT 722: Simpler example of interpreting and illustrating interactive effects — Does religious commitment have a greater impact on the political evaluations of evangelical Protestants than on those of non-evangelical Christians? (1) Regression of feeling thermometer ratings of conservatives (recoded to range from 0 to 1 – contherm01) on a dummy variable for evangelicals (evang2), the religious commitment scale (commit – ranging from 0 to 1), and their interaction (evancomm) for Christian religious traditions only (“if denom2<5" means that regression is only for evangelical Protestants, mainline Protestants, black Protestants, and Catholics. Controls are included for region of residence (dummy variable for southerners (south)), gender (dummy variable for women (female)), race (dummy variable for whites (white)), education level (educ), income level (income01), and feeling thermometer ratings of liberals (recoded to range from 0 to 1 – libtherm01). . reg contherm01 evang2 commit evancomm south female white educ income01 libtherm01 if denom2<5 Source | SS df MS Number of obs = 614 -------------+------------------------------ F( 9, 604) = 8.28 Model | 2.4777554 9 .275306156 Prob > F = 0.0000 Residual | 20.0712139 604 .033230487 R-squared = 0.1099 -------------+------------------------------ Adj R-squared = 0.0966 Total | 22.5489693 613 .036784615 Root MSE = .18229 ------------------------------------------------------------------------------ contherm01 | Coef. Std. Err. t P>|t| [95% Conf. Interval] -------------+---------------------------------------------------------------- evang2 | -.1011775 .0525055 -1.93 0.054 -.204293 .0019381 commit | .0976955 .0374927 2.61 0.009 .0240637 .1713273 evancomm | .1926198 .0735004 2.62 0.009 .0482724 .3369671 south | .0324863 .015805 2.06 0.040 .0014469 .0635258 female | -.0117823 .0154475 -0.76 0.446 -.0421197 .018555 white | .0274139 .0181524 1.51 0.132 -.0082356 .0630633 educ | -.011768 .0308271 -0.38 0.703 -.0723093 .0487733 income01 | .0708926 .0313034 2.26 0.024 .0094159 .1323693 libtherm01 | -.1252726 .0388444 -3.22 0.001 -.201559 -.0489861 _cons | .5619846 .043144 13.03 0.000 .4772541 .646715 ------------------------------------------------------------------------------ (2) Predicting the effect of religious commitment on ratings of conservatives for non- evangelicals and for evangelicals (holding the evangelical dummy at either 0 or 1, allowing commitment to vary across its full range, and holding the control variables constant at their means): . summ south female white educ income01 libtherm01 Variable | Obs Mean Std. Dev. Min Max -------------+-------------------------------------------------------- south | 1212 .3440594 .4752567 0 1 female | 1212 .5330033 .4991155 0 1 white | 1212 .7227723 .4478147 0 1 educ | 1211 .5510597 .2680294 0 1 income01 | 1070 .6338573 .2727689 0 1 -------------+-------------------------------------------------------- libtherm01 | 1014 .5540237 .2070523 0 1 . gen nonevanslope=_b[_cons]+(_b[evang2]*0)+(_b[commit]*commit)+(_b[evancomm]*0*commit) +(_b[south]*.34)+(_b[female]*.53)+(_b[white ]*.72)+(_b[educ]*.55)+(_b[income01]*.63)+ (_b[libtherm01]*.55) (30 missing values generated) . gen evanslope=_b[_cons]+(_b[evang2]*1)+(_b[commit]*commit)+(_b[evancomm]*1*commit) +(_b[south]*.34)+(_b[female]*.53)+(_b[white]*. 72)+(_b[educ]*.55)+(_b[income01]*.63)+ (_b[libtherm01]*.55) (30 missing values generated)