“This is the operative statement. The others are inoperative.” This 1973 announcement by Richard Nixon’s press secretary, Ron Ziegler, effectively admitted to the mendacity of all previous statements issued by the White House on the Watergate scandal.

Most flip-flops by politicians are less monumental, although they can sometimes haunt an entire election campaign: During the 2012 presidential race, Republican candidate Mitt Romney was frequently accused of “flip-flopping” on key issues. For example, whereas Romney hailed the health care policy that he instituted as Governor of Massachusetts as a “model for the nation” in 2007, he argued that it would be “be wrong to adopt this as a nation” in 2010.

How often do voters detect when a candidate flipflops? How does recollecting that a candidate flip-flopped influence memory for the candidate’s current position?

A recent paper by Adam Putnam, Christopher Wahlheim, and Larry Jacoby in Memory & Cognition examined those questions in the laboratory.

Participants were presented with the statements of two fictitious candidates that participated in two (equally fictitious) debates, and who sometimes changed their positions between the debates. For example, during the first debate, one of the candidates might have stated that “Traditional marriage should be protected through a constitutional amendment”, whereas at the second occasion the candidate might have stated “Traditional marriage should be an issue for individual state legislatures.” (The opposing candidate likewise proffered two versions of their policy; for example, “Partners of the same sex should be recognized through marriage” and “Partners of the same sex should be recognized through civil unions.”)

Putnam and colleagues told participants that two candidates were running against each other for office and presented them with excerpts from two debates. Sometimes the candidates repeated themselves across debates, sometimes they addressed a topic at only the second debate, and sometimes they “flip-flopped” between positions. At test, participants attempted to recall each candidate’s position from the second debate and also indicated whether they thought the candidate had changed positions from the earlier debate.

Across three studies, Putnam and colleagues found that recall was worse overall for positions that had flip-flopped between debates, compared to positions that were consistently articulate by the candidate in both debates. Flip-flopping seems to make it harder for people to keep track of what a candidate ultimately stood for.

There were, however, several important qualifications to that overall result: First, the lower memory performance for the flip-flops was largely the result of people erroneously reporting a candidate’s first position as the current one—in other words, people didn’t update their memories accurately when they encountered the flip-flop. Indeed, they may not even have noticed the candidate’s changing opinion on the issue. In support, whenever people noticed the change—as revealed on a direct question about flip-flops—their memory for the changed position was as accurate as it was for control items that had only ever been mentioned during the second debate. Only when people failed to notice the change was their memory for flip-flop items particularly poor.

In a final twist, when participants were asked to monitor the second debate carefully and to indicate “online” whenever they noticed a flip-flop, their memory for changed positions exceeded that for control items when the change was recollected at the time of test. In other words, detecting a flip-flop can serve as a signal to encode the candidate’s changed position particularly well.

How do those memory processes translate into voter behavior?

This question cannot be answered on the basis of the study by Putnam and colleagues. The existing literature cannot fully settle the issue either: On the one hand, it is known that people generally value consistency because they see it as a marker of commitment to an issue. Broken commitments are frowned upon. Indeed, there is considerable evidence that politicians pounce on perceived flip-flops of opponents for their own gain. On the other hand, the impact of actual flip-flops on voters may be less than intuition and the rough-and-tumble of politics suggest: There have been mathematical demonstrations that even in the face of voter dislike of flip-flopping, it is advantageous for candidates to change their positions between primaries and the election, because the gain associated with moderating one’s position for the election—after running on a more extreme platform during the primaries within one’s party—outweighs the harm from flip-flopping. Accordingly, there is empirical evidence that people are willing to forgive inconsistency if a politician’s flip-flops bring a candidate in line with voters’ current stance.

Key to that forgiveness, however, is that people know what the candidate’s latest position is—and the article by Putnam and colleagues tells us that noticing and remembering the flip-flop is key to an accurate memory for the candidate’s current position.

Intriguingly, when opponents point to a candidate’s flip-flop, they may therefore merely pave a route towards forgiveness by those voters who can now embrace a candidate’s revised position.