Talk Elections

Gallup this week released 2017 job approval ratings for President Trump in all 50 states. There’s a lot of interesting stuff in there. But one data point seemed to catch people’s attention more than others: Trump managed only a 39 percent approval rating in Texas; 54 percent of Texans said they disapproved of Trump.

[...]

But should we take these figures at face value? It’s possible that O’Rourke and Cruz may end up in a close race, but I’d be cautious about using the Gallup numbers to argue that Cruz might lose. The reason is simple: The group of people Gallup is polling is the entire adult population, rather than registered or likely voters. And in Texas, there’s a big partisan gap between the adult population and the electorate.

Nationally, voters, as a group, typically lean a bit more Republican than the adult population.1 But Texas is a special case: The electorate tends to be way more GOP-leaning than adults overall are.

You can see this at work in polling that Marist College did of likely voters and all adults in nine states during October and November of the 2016 campaign.2 When Marist asked whether respondents approved or disapproved of then-President Barack Obama, it found, on average, a 5-percentage-point drop in Obama’s net approval rating when the population changed from adults to likely voters. In Texas, the decline was 13 percentage points — by far the largest of the states polled. In other words, the likely electorate in Texas was 13 points more GOP-leaning than the state’s overall adult population.

Why the difference? In short, the answer can be found in the ethnic and racial makeup of the adult and voter populations. Whites are more Republican-leaning than nonwhites, so the more white an electorate is the more Republican it will likely be. According to the U.S. Census Bureau’s Current Population Survey, non-Hispanic whites made up 47 percent of all Texas adults in 2016 but accounted for 61 percent of Texans who voted that year. The only other states with a gap that big were California and Arizona, which also had 14-point differences and have large Hispanic populations.3 That’s double the average gap (7 points).

But Texas is different from Arizona and California in an important way — it has a larger partisan gap between whites and nonwhites, which exacerbates the effect of the ethnically/racially-driven differences between all adults and voters. White voters in Texas are far more Republican than white voters in Arizona and California are.4 So, that Texas’s electorate is so much whiter than its adult population has a huge effect.