In the Democratic presidential debate Tuesday evening, CNN’s Anderson Cooper prefaced his first question to Hillary Clinton by noting that “even some Democrats believe you change your positions based on political expediency,” pointing out Clinton’s evolving positions on same-sex marriage, immigration, and trade. He concluded by asking: “Will you say anything to get elected?” Later Martin O’Malley picked up on this theme, saying, “right before this debate, Secretary Clinton’s campaign put out a lot of reversals on positions on Keystone and many other things.”

Implicit in both Cooper’s question and O’Malley’s accusation is the notion that the American public values consistency among its politicians, and will punish them for changing positions. However, the political reality is often complex: available information changes over time, and political issues themselves evolve, something Clinton hinted at in her response to O’Malley.

Still, with the memory of John Kerry’s 2004 campaign relatively fresh in the minds of many Democrats, accusations of “flip-flopping” are likely to raise alarm bells in any candidate’s war room. But just how damaging is flip-flopping, really?

In a forthcoming paper, we tested the effect of flip-flopping in two experiments. In each experiment, participants were presented with a brief description of a hypothetical politician, including basic background information and his positions on several policy issues. However, for some respondents, the story also indicated that the candidate had changed positions on one of the issues.

In one experiment we randomly varied four issues on which the politician changed his position. The first two issues were whether abortion should be permitted in all cases and whether to raise the Social Security retirement age. The other two issues were arguably more complex: whether to send ground troops to fight the Islamic State and whether to issue more permits for nuclear power plants. Of course, very few issues are easy to fully understand, but we expected most people to feel more confident in their opinions about the first two issues than the latter two.

What did we find? People were less favorable to a candidate who had changed positions on one of the “easier” issues. But they were not less favorable to a candidate who had changed positions on either of the more complex issues, ISIS and nuclear power plant construction.

To be sure, across all of these issues, people evaluated the politician more favorably when they agreed with his current position. Thus, it’s not that people did not care about his position on more complex issues. Instead, they were simply not inclined to punish him for flip-flopping on these issues.

So what does this tell us about Hillary Clinton? First, our findings suggest that on complex issues — like whether to approve the Trans-Pacific Partnership or the Keystone XL pipeline — her shifting positions are unlikely to matter much to voters. Although voters may be willing to express a preference on these issues, they also appear to recognize that they are complex, and are apparently willing to give politicians a pass for changing their mind.

But what about Clinton’s alleged change of position on same-sex marriage? Our findings suggest that this is the type of “easier” issue where voters are likely to penalize politicians for changing positions. But two things mitigate this possibility for Clinton.

First, the second experiment we conducted found that people are less concerned about flip-flopping when it occurred in the more distant past. And Clinton’s shift on same-sex marriage was more than two years ago.

Second, we found that voters care more whether the candidate takes the “right” position than whether the candidate has always held that position. If enough people support a particular position, it may be wise for a politician to bear the costs of being “inconsistent” and switch to that position. Given that a majority of Americans supports same-sex marriage, Clinton’s switch is arguably the right move.

It’s important to note that our study cannot address whether Clinton’s opponents could successfully attack her as simply opportunistic. However, they do suggest that few voters are likely to be concerned about her shifts on the TPP or Keystone pipeline. Americans are not inherently averse to politicians whose positions “evolve” over time.
David Doherty is Assistant Professor of Political Science at Loyola University, Chicago. Conor Dowling is Assistant Professor of Political Science at the University of Mississippi. Michael G. Miller is Assistant Professor of Political Science at Barnard College, Columbia University.