Suppose you see something unusual – like a blue strawberry or a purple cat. You’ll be more involved in it and hopefully understand it. Psychologists have recognized this trend for years. Even babies stare at objects they find surprising for longer periods of time. We found that people also used more words to describe things that violated their expectations of others than those that met those expectations. We call this phenomenon the “unexpected refinement effect”.
According to research by my Northwestern University colleague Lauren Eskreis-Winkler and me, this effect can reveal hidden or subconscious biases in people from different racial or ethnic backgrounds. In the United States, society often associates black and Latino communities with those living in poor or at-risk neighborhoods. These associations became the seeds of implicitly biased thinking: no malice, people began to expect the worst for people from minority backgrounds.In a series of studies, we found that when people were asked to write about situations that ran counter to racial or ethnic-related expectations—for example, good things that happened to a racial or ethnic minority group—they use more words. The length of their responses suggests stereotypes that they may even think they don’t have but still influence their thinking.
To understand how stereotypes inspire unexpected expositions, we examined public records to compare reports associated with individuals of different races and ethnicities. For example, we reviewed 1,051 missing children posters produced by law enforcement agencies in California, Texas, Florida, and New York. State records identified the children as white, black or Hispanic.
When we analyzed the reports, we found that posters about white children were 30 percent longer than others. This difference is not a function of relative rarity: no group of children is statistically more likely to be missing. What happened? We suspect that the unexpectedly detailed instructions are at work. In the US, many people associate whites with better life outcomes, while blacks and Hispanics are associated with more negative experiences. Thus, these reports may reflect the authors’ conscious or unconscious stereotype that white children are less likely to disappear than black or Hispanic children. They thought the missing white child was an unusual and surprising event, which prompted them to write more about it.
Another set of public records showed a similar pattern. When forensic doctors wrote reports on unidentified bodies, those reports were 20 percent longer for whites compared to blacks or Hispanics. Again, we suspect that people are more surprised by unidentified whites than blacks or Hispanics, prompting them to write more.
To further explore this idea, we designed some experiments. Working with more than 1,200 people and several different scenarios, we provided participants with some basic details—such as photos and a brief description of the situation—and then asked them to write some form of report. Scenarios vary: For example, participants may know that the person being photographed is an award-winning teacher or a professor who was recently caught on drugs. In some cases, we also asked people how surprised they felt about a particular situation.
What we found repeatedly was that when negative events were associated with whites and positive events were associated with blacks, participants wrote more—indicating that they were surprised. For example, people reported a 25 percent increase in white teachers compared to black teachers fired for sexual harassment. Another set of articles about black teachers 30 percent more than white teachers, who won the award. Participants were even more surprised when white teachers made mistakes and black teachers won. The number of words they wrote revealed social stereotypes that white people had more positive life outcomes. We found the same surprise effect on exposition in a sample of all-Black participants who wrote more about a white college professor taking a class when accused of drug use than a black professor in the same scene.
Putting this evidence together, we start to see how elaboration can be a tool for identifying and documenting social stereotypes. People often do not admit to their negative expectations of social groups, even of themselves. We do not believe that most of the writers involved have a conscious desire to reduce blacks or Hispanics. Still, we would argue that the length of their writing reflects their endorsement of anti-stereotype events, such as good outcomes for members of a minority group or bad outcomes for members of a majority group.
This unexpected form of exposition can have major implications. In another set of experiments, we provided more than 400 people with reports of various lengths—such as unidentified bodies or missing children reports—and asked them to assign each a priority to help determine government spending to address the case. We found that longer reports get a higher priority from people who want the government to spend more to resolve these cases.
We even created two missing child and two unidentified body reports with nearly identical details but varying lengths. We found that 64% of respondents prefer to focus resources on reporting more lengthy cases. Long descriptions are more important to the reader. Because our earlier work suggested that longer reports of missing children and unidentified bodies often involved white victims, our findings raised the troubling possibility that people might prioritize these based on their longer reports case.
Writers and speakers, whether journalists, law enforcement or public officials, should be mindful when discussing negative incidents that have occurred to members of minority groups. When we rarely talk about these events, we should ask ourselves if we are less surprised and therefore assume that these events are somehow less interesting or lower priority.
As readers and listeners, we should recognize this underlying communication bias so that we can think critically about how stereotypes can in turn reinforce structural barriers. We can remember that what other people say is only part of their message: whatever they say is instructive.