Starting just after New Year, we have been tracking U.S. voters’ policy preferences. We see two very clear trends.
First, more people support Democratic over Republican positions.
[interstitial_link url=”https://www.washingtonpost.com/news/monkey-cage/wp/2016/01/22/think-you-can-predict-the-election-join-our-2016-forecasting-tournament/”]Think you can predict the election? Join our 2016 forecasting tournament.[/interstitial_link]
Second, voters are making up their minds. A few months ago, many more of them expressed “weak” or “indifferent” attitudes toward both Democratic and Republican policy positions — but increasingly they support a partisan position, either strongly or very strongly. That will almost certainly continue. To avoid cognitive dissonance, voters will make up their minds to agree with the party or politician they plan to vote for in November.
We’ll explain in more detail below.
Here’s how we investigated this
Specifically, we staked out 11 issues that span economic, social and foreign policy. We asked for opinions on abortion, anti-discrimination laws for sexual orientation, government efforts to reduce inequality, gun control, immigration, Iran, maternity leave, Medicare spending, government regulation, taxes for incomes over $400,000 and global warming. (You can see the exact wording of the questions here.)
[interstitial_link url=”https://www.washingtonpost.com/news/monkey-cage/wp/2016/03/06/trump-is-closer-to-what-most-republicans-believe-than-the-establishment-candidates-are/”]Trump is closer to what most Republicans believe than the ‘establishment’ candidates were[/interstitial_link]
We anchor each question at either the Democratic or the Republican position, and ask whether respondents agree or disagree. To be clear, the respondents see only options to support not support the position; we do not reveal to them our interpretation of which party’s position it is.
The technical stuff. We collected the data through Pollfish’s mobile polling system and then modeled and post-stratified to the expected voting population, as Andrew Gelman explained here at the Monkey Cage two years ago. This is not a traditional probability-based telephone survey, but a non-probability-based mobile survey. This method is much cheaper and faster, allowing us to track respondents weekly. And we are confident that modeling and post-stratification, combined with the large number of responses, allow us to track public opinion accurately with high time granularity.
[interstitial_link url=”https://www.washingtonpost.com/news/monkey-cage/wp/2014/04/09/tracking-public-opinion-with-biased-polls”]How to figure out public opinion accurately even when your sample isn’t representative[/interstitial_link]
To get even more technical: Note that we develop a dynamic statistical model that yields probabilities of sub-demographic groups supporting any issue, and is able to parse out noise from substantive movement. We then weight these probabilities based on the proportion of this sub-demographic in the likely voter space. We estimate the likely voter space leveraging Big Data on all registered American voters. The survey data was collected in 17 waves and includes 17,202 responses total.
[interstitial_link url=”https://www.washingtonpost.com/news/monkey-cage/wp/2016/05/03/will-it-be-clinton-or-trump-in-november-heres-whats-wrong-with-most-predictions/”]Will it be Clinton or Trump in November? Here’s what’s wrong with most predictions[/interstitial_link]
The figure below shows the average number of responses each week that were very strongly or strongly in favor of the Democratic position, weak or indifferent, or very strongly or strongly in favor of the Republican position.
Average percent of voting population holding strong or neutral issue positions
Data: Pollfish; Figure: David Rothschild
More voters hold positions aligned with the Democrats than with the Republicans
The general population is much more aligned with Democratic rather than Republican positions. For five issues, the Democratic position is much more popular than either the neutral or the Republican position. Those include increased taxes on high earners, legalizing abortion in cases of rape and incest, having anti-discrimination laws for sexual orientation, federally mandating that businesses offer maternity leave and increased gun control measures.
For two issues, the Democratic position and the neutral position are equally popular: whether the government should try to reduce income inequality and whether global warming exists.
American voters are decidedly neutral on two issues associated with Republicans: reducing Medicare costs by giving vouchers to subscribers and curtailing government regulations.
[interstitial_link url=”https://www.washingtonpost.com/news/monkey-cage/wp/2016/05/04/2016-should-be-winnable-for-republicans-but-can-they-win-with-trump/”]2016 should be winnable for Republicans. But can they win with Trump?[/interstitial_link]
But they agree with the Republicans on two issues: reducing immigration and considering military options to deal with Iran.
Voters are picking sides more firmly
Over time, voters are increasingly picking a side and committing to its policy positions – or to put it differently, voters are getting more polarized and constrained in their policy positions as the primary season gives way to the beginning of the general election. That’s especially true among voters who hold Democratic positions, but there’s a slight increase in support for Republican positions as well.
We expect this trend to continue through the election season.
Tobias Konitzer is a Ph.D. candidate in communication at Stanford University.
David Rothschild is an economist at Microsoft Research. Find him on Twitter @DavMicRot.