Let me take you on an adventure to see if our journalism about the elusive Obama-Trump voter (Obama in 2012, Trump in 2016) holds up to inspection.
To get a sense of what reporters, columnists and commentators said about the Trump-Obama voter, I searched the Nexis Uni and News Bank databases for March, April and May of this year. The profile was fairly consistent.
Conventional wisdom was that Obama-Trump voters are: swing voters, moderate, independent, white, working class, non-college educated, male, disaffected with politics and politicians, change seekers and concentrated in Rust Belt states.
During early 2019, an academic study of Obama-Trump voters came out, got some coverage and informed some of the chatter. The Democracy Fund Voter Study Group concluded that a majority of Obama-Trump voters still regard Donald Trump favorably, but “no other voting group has shifted away from the president more in the last two years.”
The study was a poll of 6,779 Americans, most previously surveyed as part of a longitudinal panel, specifically YouGov’s online panel, and surveyed between Nov. 17, 2018, and Jan. 7, 2019. In the group’s 2016 survey, some 85% held a favorable view of Trump, but that slipped to 66% this year.
I decided to test the broadcast commentary and press speculation about Obama-Trump voters against an even bigger set of YouGov data. The Cooperative Congressional Election Survey in late 2018 was a nationally stratified sample/survey of 452,755 people — with typically a thousand people or more in each congressional district.
The most notable result is the shrinking number of people who say they voted for Obama in 2012 and Trump in 2016. The CCES survey had only 2,131 voters fitting that description; that’s only one half of 1% of all respondents. We also may want to reconsider the pronoun “he” for Obama-Trump voters. Some 52.6% were female, not much different from the 53.6% of the rest of the sample.
Another surprise was ideological. On a five-point scale (1 = very liberal to 5 = very conservative) Obama-Trump voters were slightly more liberal, 3.26 versus 3.32, than other respondents. In terms of partisan identity, a streak of independence did show up in Obama-Trump voters. Compared to other respondents, they were 2.5 percentage points less likely to identify as Democrats and 2.9 percentage points less likely to identify as Republicans, but 7.7 percentage points more likely to call themselves Independents.
The mean age was 50.23, not much older than the other voters who averaged 49.61 years. The overall education level was the same between Obama-Trump voters and the rest of the respondents.
Obama-Trump voters were a notably white group, 81.7%, compared to the others in this sampling, 74.6%.
Economic stress was a mixed bag. Obama-Trump voters by a very slight margin were more likely to think the economy is getting worse, but for those same voters their average family income was higher than other respondents.
I defined Rust Belt residents as from the Great Lakes states, minus New York and adding West Virginia. Rust Belt voters made up a larger percentage of the Obama-Trump voters, 28.5% versus 23.4%, compared to other voters. Obama-Trump voters exceeded what one would expect (based on state percentage of the overall survey) in Illinois, Michigan, Minnesota, Ohio, Pennsylvania and Wisconsin, but trailed in Indiana and West Virginia.
In sum, journalism and punditry have gotten a few things right, but much more wrong, in characterizing the Obama-Trump voter. That voter is more likely to be white and from a Rust Belt state and to think of himself or herself as an independent, but that voter is not much different from the overall population in terms of age, education, political ideology and gender. The Obama-Trump voter may feel economically stressed, but actually has a slightly higher family income than other voters. The Obama-Trump voter does not vent stress via disapproval of his or her major elected officials more than any other person might.
Compared to the rest of the electorate, Obama-Trump voters spend less time with the news and thus may not be hearing our rather frequent mischaracterization of them.
Twice I have been elected to the Tennessee Democratic Party Executive Committee, and currently serve as a regional vice chair. I’m under no illusion that some magic issue or creative tactic will turn Blount County or any other areas quickly into bastions of Democratic support. On the other hand, testing assumptions about voters (in this case the rare and elusive Obama-Trump voter) should help Democrats to understand current voters, to appeal better to them and to expand the electorate in new directions.