The Influence of Decision Contexts and Role Models on Female Risk Preferences

If you missed the last WAPPP Seminar, you wasted an opportunity to meet a former Jeopardy! contestant who is actually using Daily Doubles for more than just game show glory. Heidi Liu, Ph.D. Candidate in Public Policy at the Harvard Kennedy School and J.D. Candidate at the Harvard Law School, used a large dataset of Jeopardy! contestant betting patterns that allowed her to test whether female contestants exhibited different attitudes towards risk depending on the "gender context".

Heidi Liu, PhD and JD candidate at Harvard and former Jeopardy! contestant

How does this work? Heidi had a group of former contestants and another larger group of survey responders rate over 1,200 categories in the game show on the likelihood that men or women would correctly answer a question from that category. The categories where men were expected to perform better were coded as 'male categories' in the dataset, and in the same way, those where women were expected to do better were coded as 'female categories'. After noting that male and female contestants actually performed equally well on both types of categories, Lui presented her findings on the betting patterns of male and female contestants.

She found interesting differences between the genders. Men exhibit aggressive betting patterns across all categories. In contrast, women bet less aggressively overall. This result goes in accordance with robust evidence in the existing literature that categorizes risk-taking as an attribute of the masculine psychology. But the key contribution of Liu's research lies one level deeper. It turns out that women were found to be more risk averse when they were betting on 'male' categories, in which they are societally expected to underperform, than on the 'female' categories, where they were expected to do well. In other words, when dealing with a more masculine context, women tended to 'play it safe'.

Women are more likely to take risks when they're dealing with
more feminine contexts, regardless of ability

The Jeopardy! study is only one of the four studies that Liu presented to show how context can influence women's attitudes towards risk. She is building on previous research that identified increased risk-taking by men in physical tasks or financial decision making, but found no gender difference in risk attitudes in contexts such as prosocial heroism, as exhibited by some Holocaust survivors or even kidney donors, where women were just as likely to risk their lives or their health to save somebody else. Liu studied whether the gender content of certain contexts would be useful to predict women's attitudes towards risk. In one of the studies, she looked at women's decisions regarding a travel destination. When deciding on a destination for a professional conference, a male setting, women were much more likely to go with the safer option; in contrast when they were making a decision about a vacation destination, a more female stereotypical task, they tended to opt for the risky option. She reported similar findings for the other two studies. Women were more likely to choose risky options when making shopping decisions and less when making investment decisions, and the same happened with product choices.

Helping women become more comfortable with risk in male-dominated environments is important because they are contexts that greatly reward such behavior. Common attributes listed by venture capital firms, for example, as desirable in entrepreneurs often include words like 'aggressive', 'competitive', 'ambitious', 'relentless', 'with an appetite for risk', among many others that signal risk aversion as a negative. For this reason, Liu set up two additional studies that consider possible interventions to move women into risk-seeking territory. They focus on a critical aspect for any person's professional career: Role models.

Her first one exposed people to three different narratives of women who exhibited different degrees of counter-stereotypical behaviors and relatability. Only exposure to the story about a woman who was highly counter-stereotypical but also highly relatable had the effect of increasing risk-seeking among women. This means that the ideal role models may not necessarily be hyper-successful and unattainable, but rather, regular people who reach success on their own terms. The second study analyzes survey data from an Indian tech firm and finds that women who have a female supervisor are more likely to take risks at work. This goes in accordance with the results from the first study.

Heidi Lui is continuing her research on this topic as she works towards her PhD. Stay tuned to her work to continue learning about ways to get less women to play it safe and more women to just go for it.

Risky Business: How Environment Affects Women's Decision-Making

The concept of risk as a major factor in decision-making has been thoroughly examined since the 2008 financial crisis began. In this week’s WAPPP Seminar, “Risk in the background: How Men and Women Respond,” risk was examined using the lens of gender to better understand the causes and effects of the choices individuals make. Alexandra van Geen, a 2014-2015 WAPPP Fellow and Assistant Professor at Erasmus University's School of Economics in the Netherlands, discussed her research on the effect that gender has on risk-taking in different contexts.

While acknowledging that previous research states women are generally more risk averse than men, Professor van Geen pointed out that this assumption is based on experiments involving isolated risks. Decisions to take risks rarely occur in isolated situations, however, due to the fact that there is both background risk (current risk) and realized risk (past risk) to consider. Instead, van Geen sought to measure the effect that gender has on risk in more realistic decision-making situations.

In a study conducted at Harvard’s Decision Science Laboratory, 160 students split evenly by gender were placed in an experiment where a third party threw a die to determine each student's payoff. To simulate what van Geen refers to as “background risk,” each participant had a 50% chance of receiving either $2 or $30. In addition, there were two fixed sum groups, one that received the low payout of $2 and one that received the high payout of $30, automatically.

Some of van Geen’s findings confirmed what many already understand to be true about gender and risk: as a baseline, women are more risk averse and less risk-seeking than men. However, van Geen found that women take more risks after receiving the fixed sum or with the presence of background risk, i.e. the roll of the dice in the experiment.

This reduction in risk aversion can be explained by women's sensitivity to the income effect, as in, the potential to earn money without risking a loss. In fact, the effect of income and potential income from past wins (via winning in the dice roll) eliminated gender differences in risk-taking altogether. One potential explanation for this is that women are more sensitive to income because of their higher baseline risk aversion. Additionally, men may narrowly frame a decision and not consider past or present risk in the same way that women do.

The income effect does not explain all the behavior around risk, however, van Geen argued. The effects of the experience of winning in and of itself was also examined. Winning produces several outcomes: (1) emotion, (2) subject expectations (i.e. belief an individual will keep winning once they have, or "gambler's fallacy"), and (3) an elevated level of hormones, namely testosterone. This study found that men increased their risk-taking after winning, despite the fact that there was no apparent income effect on men, indicating that it was the experience of winning itself that altered future behavior.

From these experiments, van Geen concluded that women experience an income effect that reduces risk aversion, while men don't experience an income effect at all. This effect is transient, however, and women eventually return to their baseline risk preferences. The presence of background risk also decreases women’s risk aversion but not men’s, which can also be explained by the effects of income and potential income. Lastly, men increase risk-seeking behavior after winning, while there is no effect for women.

Risk can have serious impact on decision-making, which in turn affects policy, van Geen argued, closing with a few potential policy implications of her research. Policy changes could encourage women to make riskier microfinance investments that lead to better outcomes, while behavioral nudges could discourage male day traders from being overconfident and seek too much risk after a win. Understanding how context affects gender's relationship with risk may help us close these micro-gender gaps, which could then chip away at larger inequities.

Fill In the Blank - WAPPP Seminar Series

The Woman: Katie Baldiga, Doctoral Candidate, Harvard University
The Talk: Gender Differences in Willingness to Guess and the Implications for Test Scores
The Question: Do Men and Women Perform Differently on SATs?

The short answer, according to Katie Baldiga of Harvard University, is yes.

In a recent study, Baldiga discovered that when men and women were given multiple choice tests, men were more likely to guess answers they didn't know, while women were likely to skip questions they didn't know, resulting in an on-average reduction of 5 points from their raw score. In SAT terms, this is a 70 point drop.

This carries implications for how men and women perform in testing environments, says Baldiga, because the data suggests that such tests reward risk-takers rather than those who are risk averse. Ruling out lack of knowledge or confidence as factors, Baldiga was able to find a link between women's risk aversion, their desire to compete, and how many questions they skipped. The results have interesting implications: women answer only what they think they know, while men tend to answer regardless of the consequences.

The results carry important implications for future standardized test design. Might we, for instance, design a multiple-choice format that doesn't penalize skipped questions? Is there a way of controlling for risk in standardized tests?

If we do so, we may even the playing field between those who take risks and those who don't. Or, as Victoria Budson mentioned, we may even the playing field between those who guess right and few times, or those who guess wrong, but guess often -- and are rewarded for it.

*Photo courtesy of Education Week

Effie-Michelle Metallidis is an MPP1 candidate at the Harvard Kennedy School and a participant in WAPPP's From Harvard Square to the Oval Office program.