Voters punish politicians for misinformation that portrays them in a favourable light, but not for inaccurate information that attacks their opponents

What impact does inaccurate information have on political campaigning? Outlining the results of three studies on the role of misinformation in politics, Michael D. Cobb finds that voters react to positive and negative information in very different ways. While negative misinformation, such as using false figures to attack a political opponent, tends to linger in the minds of voters even after it is retracted; there is no such effect when positive information about a politician is debunked. Indeed, voters appear to actively punish politicians in the aftermath of positive misinformation.

“Yeah, well, that’s just like, your opinion, man” says the Dude (Credit: Cea, CC by 2.0)

Someone once remarked, “a lie will go round the world while truth is pulling its boots on.” This assertion raises a good question: what happens when people encounter false information prior to receiving accurate information?

The effects of exposure to false information have important implications for politics. Imagine the following scenario. You learn about a politician reported to have been the deciding vote responsible for passing a bill that you strongly dislike. On the following day, you learn that a reporter mistakenly identified this politician as the key vote. In fact, that politician was sick the other day, and not even present when the bill was voted on. How do you feel about the politician afterwards?

Research that indirectly examines this kind of situation paints a pessimistic picture about the nature of misinformation. Surveys in the UK, for instance, reveal that misconceptions are prevalent about nearly every aspect of government. A recent survey carried out by the Royal Statistical Society and King’s College London, for example, indicated that around 26 per cent of people would rank foreign aid as one of the top three items of UK government spending, while in truth it accounts for only around 1.1 per cent of spending. Interestingly, most British citizens also say that the UK spends too much on foreign aid. Are these two findings causally related? As we all know, it is difficult to establish the causal order of relationships between two variables, misinformation and preferences, when analysing data from cross-sectional surveys. It is entirely possible that survey respondents decide they oppose spending on foreign aid prior to knowing much about it, true or false. When asked about spending on foreign aid, however, opponents might guess higher because that bolsters their policy preference, not because they were exposed to false information about spending that subsequently influenced their opinion.

To get around this problem, political scientists have conducted some clever studies. One of these simulated the effects of acquiring accurate information on aggregate level policy preferences. Scott Althaus imputed preferences for low-knowledge respondents by matching them with demographically identical respondents who were different only in that they were accurately informed about a policy. These simulations sometimes generated significantly different preference distributions, but rarely caused a minority to turn into a majority, or vice-versa.

Simulations are intriguing, but they can’t tell us whether misinformed individuals would actually accept accurate information as valid and replace or negate the misinformation. Jim Kuklinski and his colleagues addressed this problem by giving people accurate policy information before measuring their opinions. First, they demonstrated that most people were misinformed about welfare policy, often badly. Secondly, they found people who received accurate information had nearly identical preferences about welfare as those who held multiple misperceptions. The implication is that misinformation trumps accurate information.

As before, the prior interpretation of the results is not totally warranted. In this case, the study did not directly measure whether respondents who were already misinformed continued to believe the misinformation after receiving policy facts. Instead, one subset of the study received facts and their preferences were compared to those that did not get these same facts. Put differently, it is as equally possible that facts were believed but irrelevant for affecting preferences as it is that people refused to accept the validity of the accurate information.

Studies from psychology are arguably more definitive, and depressing. These studies have documented a phenomenon called, “belief perseverance”, where initial exposure to false information continues to exert influence on opinions even after it is credibly debunked. In a typical experiment of this sort, individuals are asked to preform some task. They also receive randomly assigned feedback about their performance on the task. So, for example, roughly half of study participants are told, at random, that they did very poorly or very well. Their actual performance is unmeasured.

When the study purportedly ends during the debriefing, they are told about how the feedback was determined and that it was fictitious. They are then asked to rate their own performance on the task, knowing full well the prior feedback was not diagnostic information. Nevertheless, individuals told they did poorly on the task believed they were less capable at it than those who received positive feedback. It appears that exposure to false information cannot be negated, even under ideal circumstances such as when the retraction of it is authoritative and people exposed to it acknowledge this.

Thinking again about how all of this matters for the real world of politics, the obvious implications are problematic. For one, if misinformation is “sticky,” especially negative misinformation, this creates the perverse incentive to level false accusations against opponents. To do so invites the very real possibility that, even with retractions, people will develop less favourable impressions of the victim based on information that was never valid and lacks any diagnostic utility. For another, it pushes rational debate to the sidelines, since falsehoods and innuendos can leave lasting impressions in a way that thoughtfully articulated policy disagreements can’t.

Even worse, it is harder to credibly debunk misinformation in politics. Some people, for example, will only be exposed to the misinformation. In addition, evidence suggests that partisans are prone to accepting falsehoods when they are unflattering about opponents. This means that the way in which people assess the validity of retractions is a highly subjective process. Truth becomes equivalent to opinions. As best explained by “the Dude,” Jeff Bridges’ character in the movie, The Big Lebowski, “Yeah, well, you know, that’s just, like, your opinion, man.”

Positive misinformation does not persist in the same way as negative information

My colleagues, Brendan Nyhan (Dartmouth) and Jason Reifler (University of Exeter), and I wondered if these findings applied equally to positive misinformation. Sometimes, that is, politicians inappropriately take credit for positive actions or enhance their resumes. In these instances, does exposure result in lasting favourable impressions, even after it turns out the positive information was untrue?

We recently published our findings in Political Psychology. Our results demonstrate that there is an asymmetry between false negative and positive information. While we replicated the classic belief perseverance effect in the case of exposure to negative misinformation, in which negative evaluations of a politician lingered even after study participants acknowledged that the initial information was bogus, the opposite occurred for positive misinformation. Instead of lingering positive evaluations, politicians were “punished” after false positive information about them was debunked.

Our explanation centers on how we believe people attempt to correct for exposure to misinformation. A well-known finding in political science and psychology is that “bad is stronger than good”. That is, negative information tends to be more memorable and persuasive. Yet, there is no evidence that people know this or take this distinction into account when attempting to adjust for exposure to false information. In the case of positive misinformation, people over-adjust when trying to correct for it; in the case of negative misinformation, they fail to sufficiently adjust their evaluations and return to their original baseline opinion.

Now, all of this assumes people are motivated to correct their evaluations, and that they fully understand the initial information was completely bogus. Other scholars have posited different explanations for belief perseverance effects, but it isn’t clear to us how those explanations, which focus on explaining the impact of negative misinformation, help us understand this asymmetry we uncovered and replicated in three separate studies.

In the end, these kinds of findings are worrisome. They suggest a non-rational process underlies impression formation, one that is perhaps inescapable. One that, if abused, can powerfully shape opinions and evaluations in ways that challenge classical normative positions about how opinion formation should proceed.

Note: This article gives the views of the author, and not the position of Democratic Audit, nor of the London School of Economics. It was originally posted on Europp – Euro Politics and Policy, and then on the LSE British Politics and Policy Blog.

Michael D. Cobb is Associate Professor of Political Science in the School of International and Public Affairs at North Carolina State University.