use https://www3.nd.edu/~rwilliam/statafiles/cruz, clear * 1A * Approach 1 - Generate dummies yourself tab ses, gen(ses) logit cruz male socialcons fiscalcons ses2 ses3 , nolog * Approach 2 - Let xi: generate the dummies xi: logit cruz i.male i.socialcons fiscalcons i.ses, nolog * Approach 3 - Yet another approach that somebody insists on trying logit cruz i.male i.socialcons fiscalcons i.ses , nolog * 1B * Possible margins & marginsplot commands. use https://www3.nd.edu/~rwilliam/statafiles/cruz, clear logit cruz i.male i.socialcons fiscalcons i.ses , nolog margins, at(fiscalcons = (-12(3)12)) marginsplot, noci name(g1) margins fiscalcons * 1C * 1. Basic margins commands use https://www3.nd.edu/~rwilliam/statafiles/cruz, clear logit cruz i.male i.socialcons fiscalcons i.ses , nolog margins male socialcons ses margins, dydx(*) * 2. Adjusted predictions for male social conservative fiscal moderates * at different SES levels use https://www3.nd.edu/~rwilliam/statafiles/cruz, clear logit cruz i.male i.socialcons fiscalcons i.ses , nolog quietly mtable, at(male=1 socialcons=1 ses=1 fiscalcons=0) rowname(Low SES) quietly mtable, at(male=1 socialcons=1 ses=2 fiscalcons=0) below rowname(Med SES) mtable, at(male=1 socialcons=1 ses=3 fiscalcons=0) below rowname(High SES) * 3. Examine effects of fiscalcons both by itself and with gender use https://www3.nd.edu/~rwilliam/statafiles/cruz, clear logit cruz i.male i.socialcons fiscalcons i.ses , nolog mcp fiscalcons, var1(20) plotopts(name(g2, replace)) mcp fiscalcons male, var1(20) plotopts(name(g3, replace))