Repetition, not number of sources, increases both susceptibility to misinformation and confidence in the accuracy of eyewitnesses

Abstract

Are claims more credible when made by multiple sources, or is it the repetition of claims that matters? Some research suggests that claims have more credibility when independent sources make them. Yet, other research suggests that simply repeating information makes it more accessible and encourages reliance on automatic processes—factors known to change people's judgments. In Experiment 1, people took part in a “misinformation” study: people first watched a video of a crime and later read eyewitness reports attributed to one or three different eyewitnesses who made misleading claims in either one report or repeated the same misleading claims across all three reports. In Experiment 2, people who had not seen any videos read those same reports and indicated how confident they were that each claim happened in the original event. People were more misled by—and more confident about—claims that were repeated, regardless of how many eyewitnesses made them. We hypothesize that people interpreted the familiarity of repeated claims as markers of accuracy. These findings fit with research showing that repeating information makes it seem more true, and highlight the power of a single repeated voice.

Introduction

People who witness the same event will often remember it differently: a student remembers her teacher saying that the father of psychology is Freud, while another remembers it being Wundt; one person at the parade remembers hearing shots coming from the grassy knoll while another remembers hearing shots from the book depository; an eyewitness testifies in court that the burglar fled the scene in an electrical company van—another remembers the van with a different company name. “I saw the burglar drive off in an RJ's electricians van,” Aidan mistakenly reports to Emily, who remembers it as an AJ's electricians van. These differences can even lead people to be misled about what really happened: the father of psychology becomes Freud; the shots come from the grassy knoll; and Emily's memory contains RJ's electricians (French et al., 2008, Gabbert et al., 2003).

We also know that repeated misleading claims do more damage to people's memories than claims made only once (Mitchell and Zaragoza, 1996, Zaragoza and Mitchell, 1996). What we do not know are the answers to two questions: Does one person who repeats misleading claims do more damage to people's memories than that same person making the claim only once? And when those misleading claims are repeated, does it matter how many people make those claims? For instance, suppose an eyewitness (call him Aidan) tells other eyewitnesses, “The burglar drove an RJ's electrician's van.” Would Aidan's claim do more damage to another eyewitness's memory if Aidan states the claim once, or repeats it three times? By contrast, suppose that Aidan says, “The burglar drove an RJ's electrician's van.” Later, another eyewitness (Ben) also says the burglar drove an RJ's electrician's van, and still later a third eyewitness (Cheryl) says the same thing. Would Aidan, Ben and Cheryl's converging claims damage another eyewitness's memory more than if Aidan had simply repeated the same claim three times? Put another way, do claims do more damage when made by multiple sources, or is it the repetition of claims that matters? That is the question we ask here.

On the one hand, the idea that a claim does more damage when uttered by independent sources is intuitively appealing. And research supports it: when several people tell us something, we put more trust in the overlapping portion of their accounts; similarly, we put more trust in our own memories of an event when others remember it the same way (Harris and Hahn, 2009, Ross et al., 1998). That trust is justified: For instance, when multiple eyewitnesses identify the same perpetrator, they are generally more accurate than when one eyewitness does (Clark & Wells, 2008). These findings are also reminiscent of work by Bråton, Strømsø, and colleagues showing that students better comprehend multiple passages when they remember the source of each (Bråten et al., 2011, Strømsø et al., 2010).

On the other hand, simply repeating information can change a person's judgments. People have come to 1) decide that repeated non-famous names are famous; 2) like a stock better after repeatedly viewing the same day's news reports about it from multiple TV channels; 3) be more confident that they visited a novel campus after repeatedly seeing photos of it, and 4) believe that one person's opinion better represented the population's opinion when it was repeated than when it was not (Brown and Marsh, 2008, Jacoby et al., 1989, Unkelbach et al., 2007, Weaver et al., 2007). Considered as a whole, this research suggests that the number of times a claim is made might be more important than the number of people who say it.

But why does repetition lead people to make these errors? One possibility is that previous encounters make information more accessible and more fluently processed; this accessibility then leads to feelings of familiarity, which are often interpreted as truth (Alter and Oppenheimer, 2009, Arkes et al., 1991, Dechêne et al., 2010, Kelley and Lindsay, 1993, Lindsay, 2008, Unkelbach, 2007, Unkelbach and Stahl, 2009). As a result, repeated information feels truer than unrepeated information.

When it comes to repetition, some research suggests that source may play an important role in its effects, suggesting that the number of times a claim is made and the number of people who say it will interact. For instance, Zaragoza and Mitchell (1996) asked subjects to read misleading questions about a film, and found that subjects who were misled multiple times were more likely to incorporate that misinformation into their memory than subjects that were misled only once. But when they repeated misinformation using multiple presentation styles (written, video, audio), subjects were even more misled than when they repeated misinformation using only one presentation style (e.g., audio; Mitchell & Zaragoza, 1996). Awareness matters as well—Bacon (1979) found that people who realized information was repeated found it truer than people who did not realize it. Other research has shown that trustworthiness matters: repeated information from a trustworthy source becomes more believable than repeated information from an untrustworthy source (Begg et al., 1992, Unkelbach and Stahl, 2009). Taken together, this research suggests that if one person repeats a claim it might make that claim more misleading than were it not repeated—but if several people all make the same claim, it might trump repetition alone, making that claim more misleading still.

We explore these issues in two experiments. In Experiment 1, we asked if repeating misleading claims changes the way people report details about a witnessed event, regardless of how many eyewitnesses repeat those claims. To answer this question, people took part in an experiment adapted from the well known misinformation paradigm: they watched an event, then read a misleading description of the event, and finally were tested on their memory for what they remembered seeing. Typically, many people report seeing the misleading details in the event (Loftus et al., 1978, Mitchell and Zaragoza, 1996, Takarangi et al., 2006).

People first watched a video of an electrician who stole items while doing repairs at a client's house. Later, they read three eyewitness police reports—ostensibly written over three consecutive days—about the activities of the electrician. Sometimes, all three reports misled people about what happened in the video; other times only one of the three reports misled people. To manipulate source, we told half the people that three different eyewitnesses made these reports; we told the other half that the same eyewitness made all three reports. For example, people read three eyewitness reports from Day 1, Day 2 and Day 3: for half of the people, Eyewitness 5 made the Day 1 report; Eyewitness 9 made the Day 2 report; and Eyewitness 16 made the Day 3 report. The other half read the same reports—but all three reports were attributed to Eyewitness 9. Later, we asked everyone to take a surprise memory test to tell us what they saw in the event. In Experiment 2, we examined how these factors affect the kind of situation in which police officers, judges, and jurors find themselves: considering eyewitness reports about an event they never saw.