Imperfect recall may hamper social networks of women and racial minorities

There’s nothing more embarrassing than introducing yourself to somebody you’ve already met. From socialising at work to making friends at the gym, our lives depend on remembering not only people we’ve met, but relevant information about them. 

New research published in the Journal of Behavioural and Experimental Economics has taken a look at what kinds of information about people we remember. The team’s findings suggest that both women and those belonging to ethnic minority groups are more likely to be remembered when they comprise a small minority of a group – and more likely to be confused with each other when in larger numbers. 

The team’s first study took place after two real academic conferences. At the first, which took place in San Diego, women made up a third of presenters, and at the second, an Edinburgh-based event, women made up a fifth of presenters. Both conferences were single-session, meaning that participants would not have appeared more than once and complicated results. Most conference-goers attended all sessions of the conference.

Eighty-seven participants (33% female, 22% non-White, 26% over 40 years old) responded to an email from the research team a month after their respective conferences. They were asked to complete three online memory tasks. 

The first two tasks consisted of matching pictures of presenters to the titles of their papers. In the first task, participants attempted to match a paper title to one of four presenters' pictures; in the second task, the order was flipped, and participants attempted to match one of four titles to a single presenter’s picture. Each question had a time limit of 25 seconds to discourage participants from sneakily googling the answers.

The final task asked participants to provide the name and institution of a presenter shown in a picture – a far harder task, considering multiple choices were not provided and the 25 second time limit remained. The participants then provided their own demographic information, indicated which sessions they attended, and whether they knew the presenter before the conference. 

Analyses revealed that overall, the participants were better able to match faces to paper titles than recall their names. Participants were also more likely to accurately match a female presenter with their paper title compared to a male presenter, with the team estimating a difference of 14 percentage points between the genders, once they accounted for the effects of social proximity and academic achievement level. There was a similar effect for non-White presenters (significant to the 10 percent level). 

Surprisingly, presenters who had high levels of academic achievement were recalled to similar levels as female or non-White presenters – but not significantly so. And, perhaps unsurprisingly, sharing a field with the presenter had a strong and significant effect on recall, with participants much more likely to remember presenters within their area of interest and expertise. 

Participants were slightly better at remembering the institutions of non-White presenters if the institution was highly ranked, or in the same field. White participants were better at remembering White presenters, though the team notes that the control for the same race as the participant only includes observations about White participants due to low levels of non-White participants. 

Unlike male presenters, women were often confused with other women. Race also played a role in remembering women, with participants more likely to be confused between people of the same race and gender. White presenters escaped this fate, suggesting that the participants formed broad categories for women and non-White people. 

The team’s second study extended on these findings in a controlled experimental setting. Participants (N=387, 59% female, 15% non-White) were presented with 12 pictures of White and East Asian people from a picture database matched randomly to titles of academic papers. Each pairing was shown for 10 seconds. They were then asked to either match a person with a title, or a title with a person. This happened four times, with different levels of gender and racial balance. 

As was seen in the first investigation, the probability that a person was correctly matched to a title was higher when the correct answer involved a woman. However, this only happened in rounds in which they were in the minority. In these cases, women were 15 to 18 percent more likely to be correctly matched. East Asian people were more likely to be remembered only when women and East Asian people were in the minority.  

This second study also provided no evidence that people are likely to confuse White men with other White men. But, when the correct answer involved either a White woman or an East Asian man, gender and race were used to select an answer – people were more likely to incorrectly match a photo of someone of the same gender and race.

As the authors note, this paper is not without its limitations. Though it is very possible that these findings still reflect current trends, data were gathered around ten years ago, at conferences in April 2012 and May 2014, opening up the question of whether diversity awareness and other movements in academia may have altered these patterns of recall. The second, more tightly controlled study in particular is also abstracted from the realities of socialising, and may lack some ecological validity as a result. Further research could also explore whether this pattern exists outside of mainly White samples.

Overall, this paper suggests that people are better able to remember information about others when they are women or from a non-White racial group, but only when they are in the minority. When they are not in the minority, people from these groups become "blended together", remembered as homogenous groups. 

These findings have significant implications for the nature and size of social networks of different demographics, as well as suggest that this effect may to some extent impede certain demographics’ labour market success. Beyond being just personally frustrating for those belonging to these groups, this effect could also lead to material challenges. Many fields, including academia, rely heavily on recognition and social ties. Effects impacting how easily someone is recognised or remembered could therefore inhibit the social networks of members of these groups, who already face discrimination and systemic disadvantages. 

Read the paper in full: https://doi.org/10.1016/j.socec.2023.102008