* From the "BJPS replication file.do" file from Jardina and Piston 2021, * located here (https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/A3XIFC&version=2.0), * run in the version 2 "Qualtrics_2016_BJPS_raw.dta" dataset line 446 * through line 526. Line 526 should match Table 1 of Jardina and Piston 2021. * Run the next set of commands to check for non-linearity: sum aofmanwb if white==1 reg trumptherm0to1 aofmanwb bwstereoindex republican conservative female age2 income2 educ south if white==1 margins, atmeans at(aofmanwb=(0(0.1)1)) marginsplot recode aofmanwb (0/0.499=1 "Dehumanize Whites") (0.5=2 "Equal ratings") (0.501/1=3 "Dehumanize Blacks"), gen(aofmanwb3) tab aofmanwb3 reg trumptherm0to1 ib2.aofmanwb3 bwstereoindex republican conservative female age2 income2 educ south if white==1 margins, atmeans at(aofmanwb3=(1(1)3)) marginsplot, level(83.4) tab aofmanwb if white==1 recode aofmanwb (0/0.2500=1) (0.2501/0.4999=2) (0.5=3) (0.5001/0.7499=4) (0.7500/1=5), gen(aofmanwb5) tab aofmanwb5 reg trumptherm0to1 ib3.aofmanwb5 bwstereoindex republican conservative female age2 income2 educ south if white==1 margins, atmeans at(aofmanwb5=(1(1)5)) marginsplot, level(83.4) recode aofmanwb (0/0.4850=1) (0.4851/0.4999=2) (0.5=3) (0.5001/0.5149=4) (0.5150/1=5), gen(aofmanwb5b) tab aofmanwb5b reg trumptherm0to1 ib3.aofmanwb5 bwstereoindex republican conservative female age2 income2 educ south if white==1 margins, atmeans at(aofmanwb5=(1(1)5)) marginsplot, level(83.4) * Comparisons of version 1 and version 2 // Run the section below in version 2 codebook race reg aofmanwb ib4.ideology if race==1 & latino==2 tab aofmanwb3 if race==1 & latino==2 tab aofmanwb5 if race==1 & latino==2 tab aofmanwb5b if race==1 & latino==2 tab female male, mi tab race latino, mi tab1 race latino male education party1 ideology income aof* if state_1=="Missouri":state_1 & birthyear_1=="1949":birthyear_1, mi clear all *** * In version 1 of the dataset ("Qualtrics_BJPS.dta") run line 451 through 462 of the Jardina Piston replication do file. * Then run: sum aofmanwb tab aofmanwb recode aofmanwb (0/0.499=1 "Dehumanize Whites") (0.5=2 "Equal ratings") (0.501/1=3 "Dehumanize Blacks"), gen(aofmanwb3) recode aofmanwb (0/0.2500=1) (0.2501/0.4999=2) (0.5=3) (0.5001/0.7499=4) (0.7500/1=5), gen(aofmanwb5) recode aofmanwb (0/0.4850=1) (0.4851/0.4999=2) (0.5=3) (0.5001/0.5149=4) (0.5150/1=5), gen(aofmanwb5b) // Compare to version 2 of the dataset: codebook race reg aofmanwb ib4.ideology if race==1 & latino==2 tab aofmanwb3 if race==1 & latino==2 tab aofmanwb5 if race==1 & latino==2 tab aofmanwb5b if race==1 & latino==2 // Summary statistics sum aofmanwb if race==1 & latino==2, de sum aofmanwb if race==2 & latino==2, de sum aofmanwb if race>=3 & race<=5 , de prop aofmanwb3 if race==1 & latino==2, level(83.4) // 1 and 3 are strict dehumanization prop aofmanwb3 if race==2 & latino==2, level(83.4) // 1 and 3 are strict dehumanization prop aofmanwb3 if race>=3 & race<=5 , level(83.4) // 1 and 3 are strict dehumanization prop aofmanwb5b if race==1 & latino==2, level(83.4) // 1 and 5 are non-strict dehumanization prop aofmanwb5b if race==2 & latino==2, level(83.4) // 1 and 5 are non-strict dehumanization prop aofmanwb5b if race>=3 & race<=5 , level(83.4) // 1 and 5 are non-strict dehumanization prop aofmanwb3 if latino==2, over(race) level (83.4) prop aofmanwb5b if race==1 & latino==2, level(83.4) prop aofmanwb5b if race==1 & latino==2, level(83.4) tab race latino, mi codebook male sum thermind_2 if race==1 & latino==2 tab1 race latino male education party1 ideology income aof* if state_1=="Missouri":state_1 & birthyear_1=="1949":birthyear_1, mi clear all * Back to version 2 of the dataset tab1 race latino male education party1 ideology income aof* if state_1=="Missouri":state_1 & birthyear_1=="1949":birthyear_1, mi