use http://www3.nd.edu/~rwilliam/statafiles/reg01.dta, clear correlate income educ jobexp race, means regress income educ jobexp race regress income educ jobexp race, level(99) set level 99 regress income educ jobexp race regress income educ jobexp race, beta regress income educ jobexp race regress, level(99) regress, beta vif collin educ jobexp race if !missing(income) test educ jobexp test educ=jobexp test educ=jobexp, coef nestreg: regress income (race) (educ jobexp) pcorr income educ jobexp race sw, pr(.05): regress income educ jobexp race sw, pe(.05): regress income educ jobexp race reg income educ jobexp if race==0 sort race by race: reg income educ jobexp bysort race: reg income educ jobexp * First, input the means, sds, and correlations matrix input Mean = (.44,.82,6.46,7.12) matrix input SD = (.5,.46,2.11,1.42) matrix input Corr = (1.00,-.59,-.46,-.26\-.59,1.00,.53,.33\-.46,.53,1.00,.72\-.26,.33,.72,1.00) * Now use corr2data to create a pseudo-simulation of the data corr2data sphrd pared aptd grades, n(10609) means(Mean) corr(Corr) sds(SD) . * Label the variables . label variable sphrd "Sephardim" label variable pared "Parental Education Scale" label variable aptd "Scholastic Aptitude" label variable grades "Grade Point Average" * Confirm that all is well corr sphrd pared aptd grades, means use http://www3.nd.edu/~rwilliam/statafiles/reg01.dta, clear regress income educ jobexp i.race reg income educ jobexp i.race c.educ#c.educ reg income educ jobexp i.race i.race#c.educ