Misreading the data: Moral convictions influence how we interpret evidence of anti-women bias

How do our moral beliefs shape the way we interpret evidence of societal issues like gender discrimination? A recent study in the European Journal of Social Psychology found that individuals with strong commitments to gender equality are more likely to trust rigorous studies showing bias against women. However, the study also points to a darker side: the same moral conviction can lead to biased reasoning, causing people to infer discrimination even when the evidence says otherwise.

“We hear a lot of worries about gender discrimination against women in the fields of Science, Engineering, Technology, and Mathematics (the so-called STEM). And it is clear that having more women in these fields would be a good thing,” explained study authors Antoine Marie and Hualin Xiao, a postdoctoral fellow at Ecole Normale Supérieure and a postdoctoral researcher at the Université Clermont Auvergne, respectively.

“Surprisingly, however, recent evidence is mixed with some studies suggesting this hiring discrimination has largely diminished, if not reversed, in the U.S. and European countries, while others suggest the persistence of implicit gender bias against women in hiring.”

“Also, we noticed that in everyday conversations, people often conflate the fact there are smaller numbers of women than men in some jobs with the idea of women being discriminated against during hiring processes just because they are women. This is a wrong interpretation: women may also apply less to some jobs because they are less interested in them – and the same for men. This belief is in part due to the fact that many people believe that men and women are a blank slate at birth, which we don’t necessarily believe to be true.”

“There is some evidence from cross-cultural psychology that men and women are interested in different things in life and have different life goals, on average,” the researchers continued. “Notably, for instance, women are typically less interested on average to go through hyper-competitive career paths than men. They are also more constrained to choose between having children or staying in academia because of a shorter fertility window and greater maternal duties.”

Free daily newsletter

“Women also tend to be less interested in jobs that make them analyze problems related to the non-organic world (e.g., engineering, coding, vs. human biology or developmental psychology, for instance). This is perfectly fine, people should be able to do whatever they want, but the inference is often that if there are gender differences in interests, this is bad news, because they will be used to prevent women from accessing male-typical jobs: politics, engineering, informatics, non-organic sciences, etc.”

To examine the influence of moral commitment to gender equality on perceptions of gender bias in hiring, the researchers conducted a series of experiments involving over 3,500 participants from the United States and the United Kingdom. Participants were recruited through online platforms such as Prolific and Amazon Mechanical Turk, and they were presented with summaries of scientific studies that either demonstrated or refuted the existence of gender discrimination in hiring practices.

In the first series of experiments, the researchers aimed to investigate whether participants’ moral commitment to gender equality influenced their evaluations of a study showing evidence of gender discrimination. The study used in these experiments was based on the well-known Moss-Racusin et al. (2012) research, which found that male applicants were rated more favorably than equally qualified female applicants for a lab manager position. Participants in experiments were asked to evaluate the quality of the study, the accuracy of its findings, and the reliability of its methods.

In addition to evaluating the study, participants also answered questions designed to measure their moral commitment to gender equality. This was done using a three-item scale, which asked participants to rate how much they agreed with statements like, “Achieving gender equality is a moral imperative” and “The fight for gender equality is central to my identity.”

The researchers found a clear relationship between participants’ moral commitment to gender equality and their evaluations of scientific evidence. Participants who expressed stronger moral convictions about gender equality were more likely to positively evaluate studies that provided rigorous evidence of gender discrimination against women in STEM hiring processes. These individuals rated the studies as more accurate, reliable, and of higher quality compared to participants with weaker moral commitments.

Interestingly, the researchers found that differences in how men and women evaluated these studies were largely explained by variations in their moral commitment to gender equality. While earlier studies had suggested that men were generally more skeptical about claims of gender discrimination, this research showed that moral commitment, not gender alone, was the key factor.

In another experiment, the researchers expanded the scope of their investigation by introducing a new variable: the direction of the hiring bias. Rather than presenting only evidence of discrimination against women, participants were randomly assigned to read summaries of studies showing either a bias against women or a bias in favor of women. The summaries were based on the same type of research designs as the earlier experiments, but the findings were manipulated to either confirm or contradict the common belief that women are disadvantaged in hiring.

The results confirmed that participants, on the whole, were more skeptical of studies showing a bias favoring women than those showing a bias against women. Contrary to the researchers’ expectations, however, moral commitment to gender equality was not significantly associated with skepticism toward studies showing a hiring bias favoring women. Participants with strong moral commitments did not show increased distrust in evidence of women being favored compared to the overall sample.

“I was personally surprised that there was no association between being more morally committed to gender equality and being more skeptical of evidence of gender bias favoring women,” Marie told PsyPost. “I would have expected more motivated skepticism towards this result, because it is somewhat contrary to people’s expectations and might be seen as bad news for the gender equalitarians’ struggle to mobilize people against sex-based hiring bias.”

In yet another experiment, the researchers sought to further explore the mechanisms behind the relationship between moral commitment and research evaluation by introducing a two-step process involving both prediction and evaluation. In the first step, participants were asked to predict the outcome of a study before being shown the results.

Specifically, they were presented with a study summary about gender discrimination in hiring, but unlike in previous experiments, the study was described as being about to be conducted rather than already completed. Participants were asked to predict what they thought the results would be, based on their expectations of gender bias in academia.

After making their predictions, participants were shown the actual results of the study, which either confirmed or contradicted their predictions. The study summaries were similar to those used in earlier experiments, showing either a hiring bias against women or a bias in favor of women.

As expected, participants with stronger moral commitments to gender equality predicted that the study would show bias against women. When the results aligned with their predictions, these participants rated the study as more credible and reliable.

Finally, in a set of experiments, the researchers presented participants with fallacious scientific studies that drew incorrect conclusions from their data. In these experiments, participants were shown a study summary that concluded there was discrimination against women, but the data within the study actually showed the opposite—women were more likely to be hired than men.

In one experiment, the researchers included a control group where participants read about discrimination against left-handed individuals instead of gender discrimination. In the other experiment, the control condition focused on discrimination based on height, allowing for a more direct comparison with the gender condition, as both height and gender involve population divisions close to 50/50.

The results showed that participants with higher moral commitment to gender equality were indeed more likely to accept the study’s faulty conclusion. Despite the data contradicting the idea of discrimination against women, these participants were more inclined to endorse the study’s findings, showing that their moral beliefs could lead them to overlook evidence that did not support their views.

Additionally, participants in these experiments were more likely to accept the flawed gender discrimination conclusion than the fallacious conclusions in the control conditions. This indicates that moral commitment to a highly charged issue like gender equality can lead people to reason in biased ways, accepting conclusions that align with their beliefs, even when those conclusions are unsupported by the data.

But as with all research, there are some caveats to consider.

“We only collected data in Western, English-speaking countries where there is already a good deal of concern for gender equality and for promoting gender parity,” Marie and Xiao noted. “Most research in quantitative social science is based on Western samples. Collecting data in non-western countries would be desirable. There, men may care much less than women about the lack of females in intellectual professions – and may even see this as good news, especially in Middle Eastern and Asian patriarchal societies.”

“Our studies were correlational in nature because it would not be possible to assign a dose of high vs. low moral commitment to people and see how it affects their assessment of research on gender bias in hiring. Thus, it is not possible to completely disentangle the role of stronger prior beliefs about the size of gender discrimination from stronger moral convictions that equality must to be advanced. Whenever one measures moral attitudes, one captures a mix of factual or statistical beliefs, and some more hot moral cognition, and it’s hard to see in what proportion.”

Nevertheless, the findings offer valuable insights into how moral commitment influences the way people assess evidence of gender discrimination. While strong moral convictions can drive positive social change by fostering support for equality, they may also impair objective judgment, leading to biased evaluations of scientific evidence.

“We are interested in spreading the word about the negative effects of strong moral commitments,” the researchers said. “Moral convictions that gender equality must be promoted can motivate people to undertake beneficial action (e.g. making sure they are not biased against women in hiring processes; avoiding subtle things like mansplaining, etc.). At the same time, such moral convictions can also make them refuse nuanced and evidence-based discussions of the issue; or reject robust evidence that gender-based hiring discrimination might in some cases have gone down, etc. Morality binds and blinds.”

The study, “Moral commitment to gender equality increases (mis)perceptions of gender bias in hiring,” was authored by Hualin Xiao, Antoine Marie, and Brent Strickland.