Changes in issue salience over the past few presidential elections

Racial issues, immigration, and guns rose as economic issues and abortion fell in importance

I’ve been focusing recently on different looks at how the political coalitions in presidential elections have changed as we moved from the Obama era to the Trump era. I used the Cooperative Election Studies to see how narrow demographic slices shifted among voting for Democrats, voting for Republicans, and sitting things out. I used the Views of the Electorate Research Survey to get a longitudinal look at within-individual changes.

And today I’ll start looking at the 2020 American National Elections Study (ANES). This dataset has some really strong features, including a wide range of demographic info, good coverage across various key political issues, reports of presidential election voting going back to 2012, and a very nice sample size (>8,000).

The picture below shows how strongly various specific issues related to vote choices in the 2012, 2016, and 2020 elections. Keep in mind this is all as measured in 2020. For the issue scales, I combined a number of ANES items relating to (1) whether they preferred more support for the poor and redistribution from the rich, or less (which I’ve labelled “Economic Issues” below), (2) the extent to which Blacks (relative to Whites) face unfair discrimination that needs to be addressed (“Racial Issues”), (3) gun regulation, e.g., banning assault weapons (“Guns”), (4) the extent to which the U.S. should be more/less welcoming of immigrants (“Immigration”), and (5) the extent to which abortion services should be legally available (“Abortion”).

I used these 2020 issue opinions to predict Democratic vs. Republican presidential votes (with non-voters and third-party votes in the middle) for the 2012, 2016, and 2020 elections. The results are below (showing standardized regression coefficients). You can see that differences in views on Economic Issues (i.e. redistribution) and on Abortion more strongly predicted people’s choices between Obama and Romney in 2012 than people’s choices in the Trump elections of 2016 and 2020. In contrast, their views on Racial Issues, Guns, and Immigration became bigger deals over these elections, with Racial Issues and Immigration having big jumps in salience in 2016.

What this is telling you is that not everyone holds perfectly aligned left-right opinions on these different issues, and that the bigger shifts in partisan voting (and non-voting) over recent elections have primarily come from people on opposite sides of a different political axis. On one end are people who combine more-conservative views on Economic Issues and Abortion with more-liberal views on Racial Issues, Immigration, and Guns, which describes a lot of the supporters of establishment Republicans – these folks were especially likely to switch their votes from R to D (or from R to Neither), and the Harris campaign has been making a big push this cycle to get even more of them to make the leap.

On the other end are people who combine more-liberal views on Economic Issues and Abortion with more-conservative views on Racial Issues, Immigration, and Guns, which describes many Whites who are lower in socioeconomic status and not very religious. These folks used to lean toward Democrats but had pretty low turn-out rates. They were among the core groups drawn to Trump’s White nationalist agenda.

Along with these issue changes, then, we also see demographic shifts. One of the big tensions in the current left-right issue clusters is that, while higher socioeconomic status is associated with conservatism on issues relating to economic redistribution from rich to poor, it’s lower socioeconomic status that is associated with conservatism on issues relating to group-based discrimination.

So, as redistribution issues have declined in salience while group-based issues have risen, we’ve seen the expected shift in the demographics of the party coalitions, as I’ve discussed in prior posts here and here. The Trump Republican coalition is predominantly less-educated straight White Christians, while the Democratic coalition consists of an alliance of racial minorities, LGBT, non-believers and other non-Christians, especially those with more education, which shows up pretty clearly when looking at the highest-impact demographic splits.

The decline in the salience of views on abortion that occurred from 2012 to 2020 is, of course, likely to reverse and rise again this year, given the 2022 Dobbs case overturning Roe. This will largely reinforce recent demographic trends, given that the abortion split primarily pits evangelical churchgoers and the less-educated against secular folks with more education.