Skip to main content
. 2022 Mar 2;8(9):eabj9140. doi: 10.1126/sciadv.abj9140

Table 3. Predicting change toward Democratic vote choice incorporating closeness to perceived party positions on race, immigration, and China, 2016–2020 (fixed effects logit models).

Note: Entries are odds ratios corresponding to fixed effects logit analyses. Given that fixed effects logit automatically drops respondents who did not change preferences, a linear fixed effects analysis is also shown in table S2. The dependent variable equals to 1 for votes supporting the Democratic candidate and 0 otherwise.

Excluding nonvoters Including new voters
Odds ratio SE z value P value Odds ratio SE z value P value
Closeness to Democrats
on race
2.411 0.894 2.370 0.018 Closeness to
Democrats on
race
1.954 0.605 2.170 0.030
Closeness to Democrats
on immigration
1.186 0.280 0.720 0.471 Closeness to
Democrats on
immigration
1.087 0.228 0.400 0.690
Closeness to Democrats
on China
1.375 0.314 1.400 0.163 Closeness to
Democrats on
China
1.509 0.336 1.850 0.064
Concern about
COVID
× wave
1.441 0.452 1.160 0.245 Concern about
COVID
× wave
1.344 0.368 1.080 0.280
Personal finances 0.718 0.218 −1.090 0.275 Personal finances 0.765 0.217 −0.940 0.345
Unemployed 1.008 0.274 0.030 0.977 Unemployed 0.991 0.239 −0.040 0.969
Wave 0.975 0.977 −0.030 0.979 Wave 1.499 1.292 0.470 0.639
Observations (n) (174) Observations (n) (226)
Individuals (n) (87) Individuals (n) (113)