Political polarization is rapidly becoming one of the greatest public issues of our time. As the Pew Research Center has found, fewer Americans are taking centrist views while more are shifting toward the extremes. The consequences are dire: the Pew Research Center found that 73% of Americans say Republicans and Democrats “not only disagree over plans and policies, but also cannot agree on basic facts.” And animus has turned personal; one PRRI study found that 35% of Republicans and 45% of Democrats would be very unhappy if their child married someone from the opposing party.

What accounts for this polarization? Are Americans just more divided over policies, like tax rates, abortion, and immigration? Or are we living in a more tribal era in which political attitudes are driven by partisan loyalties and animus toward opponents—regardless of the policy?

Some may recall when the Jimmy Kimmel show asked Hillary Clinton voters in 2016 if they supported a plan to cut corporate and estate taxes while telling them the plan was Clinton’s idea. The interviewees dutifully agreed, only to be told later that it was actually proposed by Trump rather than Clinton. Kimmel concluded, “It seems to me most people pick a candidate and go along with whatever that candidate says.”

So which is it? Do voters just pick a candidate and align their opinions with that person? Or do voters select candidates who best represent their policy views? One of us (Michael Bernstein) collaborated with psychologists Nick Zambrotta (University of Massachusetts — Dartmouth), Lauren Micalizzi (Brown University), and Scott Martin (Brigham Young University – Idaho) to find out.

The researchers recruited 195 college students across three universities asking their opinion about several policies from both the Obama and Trump administrations. (See Appendix A for further details.) While the context and specifics of various plans will always differ between two administrations, the researchers selected broad policies that apply to both presidents. The researchers randomly assigned whether students were told that the policy came from President Obama or President Trump.[i] But the policy description itself was identical between conditions. By only varying the president that each policy is attributed to, the researchers were able to examine how support for the same policies differ according to the person enacting them. Here’s what they found:

Experiment 1: Immigration at the Border

In 2018, the Associated Press released a bombshell report of immigrant children being held in overcrowded “cages created by metal fencing.” But then the New York Post reported that some of the photographic evidence was actually taken in 2014, during the Obama administration (although reasons for detainment and the number of detained children differed). This offered a great opportunity for an experiment. Would Americans feel differently about holding immigrant children in overcrowded cells if they thought it happened under Obama or Trump?

Study participants were randomly divided into two groups in which one half read that “Trump” had “come under intense criticism for keeping many immigrants, including children, in overcrowded holding cells.” The other half read the same question but were told that “Obama” had done this. They were then asked if they agreed or disagreed that keeping immigrant children in overcrowded holding cells is “a health concern and human rights violation.”

As you can see, partisans react very differently depending on whether the Trump or Obama administration was said to have kept immigrant children in overcrowded holding cells. When this was presented as Obama’s policy, nearly half (44%) of Republicans “strongly agreed” it was a human rights violation and health concern. But only 13% of Republicans thought so when told this was Trump’s policy—a 31 percentage point drop. In either condition, Democrats were more alarmed at keeping immigrant children in holding cells. But while 63% of Democrats strongly agreed it was a violation when told it was Obama’s policy, this share rose to 95% if told it was Trump’s policy—a 32 percentage point increase. When examining how many participants indicated they either somewhat or strongly agree that immigrant children in overcrowded holding cells is a health concern and human rights violation, the same trend emerges. For Republicans, 70% believe it’s a human rights violation if told it was Obama’s policy but only 36% agreed if told it was Trump’s policy. Among Democrats, 84% indicated they agree if told it was Obama’s policy and 95% agreed under Trump.

Taken these results together, we see that Obama causes greater consensus on this issue: Republican and Democratic voters are 19 points apart in strong agreement that keeping immigrant children in overcrowded holding cells is a human rights violation. But Trump polarizes people—leading to a whopping 82-point difference between partisans.

These data imply that rising public support for increasing immigration, as Gallup found, may be partly accelerated in reaction to Trump. For those in favor of more liberalized immigration policies, Trump likely has done much for galvanizing support in favor of more immigration.

Experiment 2: Free Trade

While President Trump is known for his hardline trade stance, both Trump and Obama put tariffs on goods from China. Thus, the next experiment divided respondents into two groups and told one group that President Trump (and the other President Obama) had imposed tariffs on China because “they do not play by the same set of economic rules as the rest of the world.” Then they were asked if they supported the tariffs on China.

Again, partisans’ attitudes about free trade vary tremendously depending on whether they believed that Trump or Obama imposed the tariffs. Democratic voters were 4 times as likely to support tariffs on China when told Obama imposed them (58%) than Trump (14%)—this is a 44-point drop. Notice also, a majority of Democrats support tariffs if proposed by a Democratic president but turn toward freer trade under Trump. Republicans were more supportive of trade restrictions regardless of the president. But, more Republicans (67%) supported tariffs if Trump imposed them than if Obama (48%) did so—a 19-point difference.

These data suggest that polls finding support for free trade at an “all time high,” may be the result of Democratic backlash against Trump’s restrictionist stance on trade.

Experiment 3: Government Lowering Drug Prices

How strong is conservative commitment to the free market in health care? The next experiment suggests perhaps not very if Republican leadership jettisons market-based solutions. Again people were randomly divided into two groups and asked if they supported President Trump’s (or President Obama’s) “plan of having the government intervene to lower prescription drug prices.”

Most Republican participants opposed government intervention if they thought the plan was proposed by Obama with only 24% in support. However, support rose dramatically to 72% in support if Trump proposed the plan. Regardless of the president, Democrats were far more supportive of government lowering drug prices. But somewhat fewer would support it if presented as Trump’s plan (86%) than Obama’s plan (97%).

Unlike most of the other experiments, Trump, rather than Obama, acts as a more unifying figure. When it comes to government reducing drug prices, Obama polarizes people. It may be that when Trump proposes a policy Democrats already strongly support, it retains Democratic approval and then adds Republican support. Whereas, if Obama proposed the policy, Republicans would not feel the pull of partisan loyalty to support the plan.

Experiment 4: Government Fighting the Opioid Epidemic

In the next experiment, respondents were asked if they supported increasing federal funding “to help curb the opioid epidemic” even though the “federal budget is tightly constrained.” Democrats overwhelmingly supported this spending regardless of whether they were told Obama proposed it (81%) or Trump did (84%). But Republican support rose moderately from 48% if it were Obama’s plan to 61% if they were told it was Trump’s plan.

Experiment 5: Iran Sanctions

In another experiment, respondents were asked if they supported imposing economic sanctions on Iran. However, one-half of the participants were told that Obama had imposed them, while the other half were told that Trump had done so.

Democrats were nearly twice as likely to support sanctions if Obama announced them (46%) than if Trump did so (24%). Republicans were more supportive in either condition, but somewhat more supportive of sanctions if they thought Trump imposed them (53%) than if they thought Obama did (42%). Taking these together, we find that Obama has a more unifying effect on public support for Iranian economic sanctions: Democratic and Republican voter support differs by about 4 points. But Trump polarizes people leading to a partisan gap of 29 points. If people believe Trump is behind the sanctions, Democratic support drops by nearly half and Republican support edges up by about 10 points.

Experiment 6: Drone Strikes

According to The Bureau of Investigative Journalism, both the Obama and Trump administrations launched drone strikes in Afghanistan, Pakistan, Somalia, and Yemen that had resulted in some civilians being killed. Thus, another experiment randomly divided respondents into two groups and asked if they supported a drone strike in the Middle East ordered by Obama (or Trump) that resulted in the death of suspected terrorists but also included civilian causalities.

Regardless of partisanship, few Americans support drone strikes that could result in civilian causalities. However, Democrats were about three times as likely to support a drone strike if they thought it was ordered by Obama (12%) than if they thought it was ordered by Trump (4%). Republicans were already more likely to support such a strike, but they were nearly twice as likely to support it coming from Trump (15%) as they would if they thought the order came from Obama (9%).

Experiment 7: Hate Crimes

The next experiment took a step back from public policy and examined public reaction to hate crimes. As it turns out, even interpretation of hate crimes is affected by party. Half of respondents were told that “after a hate crime resulted in the death of several U.S. citizens in 2016” that President Obama gave a speech that said the attack was “an attack on all Americans.”[ii] The other half of respondents were told that Trump had said this.[iii] (Both Obama and Trump described a hate crime that occurred during their respective administrations as an attack or assault “on all of us.”) Then respondents were asked if they agreed or disagreed that a “hate crime committed in the United States constitutes an attack on all Americans.”

While 36% of Republicans agreed if told that Obama had said this, this share rises to 70% if told Trump said it—a 34-point increase. In reverse, 76% of Democrats agreed if told Obama said this while less than half (44%) agreed if told Trump said it. Both Trump and Obama polarize Americans when it comes to thinking about hate crimes in the United States. Americans interpret the validity of the president’s points on this contentious topic largely through a partisan lens.

Implications

These data demonstrate that partisans allow the political leaders they like and dislike to color their views of public policy. Republican participants were more likely to favor the same described policy under Trump than Obama. Similarly, Democratic participants were more likely to favor the same described policy under Obama versus Trump. This is evidence that people are not objectively judging the merits of a policy. Instead, policy support depends not only on political ideology (e.g. Republicans are overall more supportive of drone strikes than Democrats) but also on the president who is promoting the policy.

This poses a serious concern for our ability to effectively communicate and reach consensus with one another about political issues. How can we forthrightly debate the merits of a presidential action when its support is, to some degree, a function of the person who occupies the Oval office?

If we want to find common ground, it is important to recognize tribalism, how we ourselves may be affected by it, and consider how it may shape the attitudes of our public.

Michael Bernstein, Ph.D. is an Assistant Professor at Brown University in the Department of Behavioral and Social Sciences

David Kemp contributed to this article.