Does Unequal Access to Leadership Cause Gender Differences in Risk Aversion?

Abstract

In Western societies, the notion that women are more risk-averse than men has been long established, though certain people disagree. This gap is relevant when trying to explain other gender disparity phenomena observed in Western society. For instance, despite equal representation in certain fields, the number of men and women in leadership positions remains unbalanced. Researchers have sought to determine the cause behind the gender difference in risk aversion. The literature remains mixed, but this paper explores a comparatively new theory that exposure to leadership roles is a driving factor for higher levels of risk tolerance. After examining the current literature, this paper then proposes an experiment to investigate the theory. To isolate the effect of leadership positions, we emphasize a comparison of the gender gap in risk aversion between Western society and matriarchies, where women have more access to leadership. In fact, prior to the proposed experiment, we conducted an informal, preliminary survey in the indigenous Bri Bri community of Costa Rica, a culture with matriarchal values. We summarize our methods and results for this informal experiment, which indicate that there is a minimal–perhaps even negligible–gender gap in risk attitudes. Our findings indicate that, if conducted, our proposed experiment has the potential to advance the theory that exposure to leadership is the main cause of the gender gap in risk aversion.

Introduction

Understanding risk aversion is valuable not only in financial contexts but in nearly all other realms of society involving human decision-making. In the realm of gains, risk-averse individuals prefer a sure outcome with fewer gains over an even slightly uncertain outcome resulting in a higher or similar payout to avoid the uncertainty. Nowadays, a common notion exists that women are more risk-averse than men (see background). Despite equal representation in many fields, women hold only thirty-one percent of leadership positions globally¹. Since women are also observed to be more risk-averse than men, a natural question is whether there is a correlation between exposure to leadership positions and an individual’s appetite for risk. Investigating whether this correlation exists is relevant. If women are in fact more risk-averse because they’ve been less exposed to leadership positions, increased exposure to leadership positions would also close the gender gap in risk-taking. Given some careers require a higher risk tolerance to yield success than others, a lesser gender gap in risk aversion could amend gender differences in these fields. Furthermore, we will be able to better understand the effects of providing unequal opportunities in leadership for men and women. Although there are many theories as to why women are more risk-averse than men, we aim to first synthesize existing perspectives in the literature. We will then frame our approach of leadership exposure and examine the research conducted thus far about the theory. Since the effect of leadership exposure on risk tendencies has not been fully explored, we propose a new, formal experiment that could allow this notion to have more merit.

Background

There is extensive existing literature studying the differences in risk attitudes between men and women. Until 2014, literature supported the sentiment that women tend to be more risk-averse than men. However, this idea has now come under attack^2,3. Some critics have refuted the idea that risk aversion differs between genders, while others argue that perhaps it is less prominent than we previously believed.

Byrnes et al. analyzed 150 studies spanning 16 types of risk-taking behaviors, finding that although some behaviors elicited larger gender differences than others, overall, men were much more likely to take risks compared to women⁴. Eckel & Grossman found that women were more risk-averse than men in the context of a gambling game⁵. Charness & Gneezy discovered women were more financially risk-averse than men in their experiment involving an investment game⁶. Studies on risk aversion and gender differences often follow similar formats, which has led to critique. Some argue that sampling methods are highly influential in how people evaluate risks in a given scenario⁷. Other researchers have brought up the always-present confirmation bias and publication bias, both of which raise a reasonable concern that researchers may have inadvertently altered findings or misinterpreted data⁶. Despite potential issues with the research finding women to be more risk-averse than men, people have also made attempts to address the problems. For example, in the 2011 study by Charness & Gneezy mentioned previously, they primarily focused on studying risk aversion differences between men and women while tackling possible publication biases⁶. To do so, the specialty of their approach was in using large-scale data that had already been collected methodically without the intention of examining gender differences. Even after taking such measures, the findings by Charness & Gneezy still suggested women are more risk-averse than men.

In response to the many studies that have established women as more risk-averse than men, researchers have posed new questions about possible causes for this observed difference. Currently, the two leading explanations are societal factors and biological differences–both of which could also elicit psychological differences, a third explanation.

Some researchers argue that societal, external factors are the reason women are more risk-averse. For instance, in 2022, Morgenroth et al. conducted a study about taking risks in the workplace. According to the results of their study, women are more likely to face societal consequences for their risky behavior when compared to men⁸. Morgenroth et al. argue that inherently, women are not more risk-averse than men, but rather, women tend to face more backlash when they take risks. Therefore, due to these harsher consequences, women are less likely to take similar risks in the future. Other experimental economists also suggest that an individual’s past personal experiences affect later behavior, which may corroborate this theory.

Another theory is that women are more risk-averse than men biologically. Prior to 2009, research suggested that the testosterone hormone had effects on human behavior and cognition, including fear reduction and increased levels of aggressive behavior^9,10. Despite this, evidence about correlations between financial decisions and testosterone levels was scarce and contradictory. Things changed, however, when Sapienza et al. published research exploring this correlation¹¹. They considered the possibility that testosterone levels could cause gender differences in two ways: financial risk aversion and the likelihood that individuals choose career paths perceived to be riskier. In this case, they defined “risky” career paths as jobs in finance, whereas jobs in all other fields were classified as not risky. The studies measured participants’ salivary testosterone levels, and the researchers determined financial risk attitudes using a computer game. Two years after the initial study, around the students’ graduation time, the researchers categorized the students’ career path choices. Sapienza et al. concluded that testosterone levels in men and women impact risk aversion. However, shortly after publication, peers critiqued Sapienza et al.’s research. Most prominently, neuroscientist Daphna Joel published a letter in 2010 addressing the psychology and economics communities, pointing out several flaws in Sapienza et al.’s experimental procedures¹². Joel’s main critique was that Sapienza et al.’s data was not conclusive enough for the claims made. Specifically, Joel argued that the experimental procedure they used for determining testosterone levels was controversial and not concrete.

Both explanations could potentially explain a third argument for why men and women have different risk tendencies: psychological differences. Men and women evaluate risks differently, especially the perceived consequences of risks¹³. Additionally, women are found to be more sensitive to losses. Research suggests differences in confidence levels between genders, and that men display overconfidence in the payouts of their risky behaviors, whereas women are more aware of their position¹⁴.

Other explanations for gender differences in risk aversion have also emerged. Currently, most of these theories lack sufficient supporting evidence or have otherwise faced pushback. In general, since research to understand why we observe these differences is relatively new, researchers have yet to arrive at a consensus. Despite this fact, the more developed data about gender differences in risk existing in the first place strongly supports the conclusion that women are typically more risk-averse than men in Western societies. The more significant motive may now be to advance the research about causes for these differences and which theory is most truthful.

Theory

Though there is an expansive list of proposed explanations for why women are observed to be more risk-averse than men, we study the theory originally proposed by Liu & Zuo that exposure to leadership positions has the largest impact on the gender difference in risk-taking we observe today. Under the assumption that an individual’s risk preferences are not innate, the question emerges as to where men and women develop contrasting risk preferences throughout the course of their lives. In Western societies, women are underrepresented in leadership positions, even in fields in which they have equal representation. This disparity is especially apparent in fields with higher levels of risk involved, like in investment banking¹⁵. Because of this, young women may find a lack of female role models whom they can look up to–specifically women who have high risk tolerance. Studies conducted in similar contexts suggest women are more risk-averse than their male counterparts. A natural question is whether increased opportunities to acquire leadership positions lead to greater risk tolerance. One way to study this question is to look at matriarchies, in which women have more leadership roles. In matriarchal societies, women are given more political power and leadership authority. Although men also have leadership opportunities, they do not outnumber women in those roles. Specifically, we want to test whether gender affects risk aversion in matriarchies to the same extent as it does in Western societies. The challenge in investigating gender differences in risk aversion from a societal context is that a large portion of the information we have about women being more risk-averse than men comes from Western societies. Though perhaps it makes sense, it becomes more difficult to compare risk aversion in matriarchies to that of patriarchies due to the fact that today, the world is mostly made up of non-matriarchal societies. Although it is likely that exposure to leadership is not the only factor involved in shaping an individual’s risk tendencies, many existing factors cannot be controlled. For this reason, it is most productive that we focus on the factors that can be easily adjusted by influences such as policy change). Since it may be easier to change individuals’ intentions to acquire leadership positions rather than affect other possible factors such as differences in upbringing, it makes sense that leadership exposure is our focus.

Literature Review

Although a handful of researchers have started to investigate a possible correlation between matriarchies and gender risk attitudes different from Western societies, literature on the topic remains scarce. Most prominent is a 2019 study by Liu & Zuo which researches the risk and decision-making of Chinese schoolchildren at a school where the students come from two unique cultures–one culture is a matriarchy while the other is a patriarchy¹⁶. Some students are Mosuo, a culture where women are in charge of important family decisions and grandmothers lead the household. Other students are Han, a culture with traditionally patriarchal values in which men are typically given preference in aspects such as decision-making and birthright. As the students intermingle with one another in their classes throughout elementary and middle school, the study investigates if and how culture may affect the children’s risk attitudes. Further, over the course of two years, they used a hypothetical lottery game to measure the risk attitudes of Mosuo and Han girls and boys from different grade levels. The study’s findings show that over time, the gender differences in risk behaviors of the Mosuo and Han children changed–both when comparing the children’s choices within their culture and across cultures. At younger ages, Mosuo boys were more risk-averse than Mosuo girls, while Han girls were more risk-averse than Han boys. However, by fourth and fifth grade, Mosuo girls become more risk averse than the Mosuo boys, and the gender gap between the Han girls and boys lessens, though Han girls are still more risk averse. Since the initial data point was measured prior to the children’s exposure to outside influences (non-familial), it suggests that cultures where women hold equal power yield populations with a smaller gender difference in risk aversion than cultures where women do not.

The findings of Liu & Zuo are insightful, and they merit further research into risk aversion in matriarchies. Though the study spanned two years, it did not follow the same group of schoolchildren. Instead, it studied the risk behaviors of different children in each grade at the school, which does not account for the possibility that over the years, as each unique class of students grew older, their risk attitudes were affected by extrinsic, changing factors (eg. evolving media, varying levels of social influence from outside the community). The study also works under the assumption that as the children’s risk attitudes change throughout their schooling, the change was directly caused by their peers of different cultural backgrounds. Finally, though overall there was a sizable number of student participants, the study acknowledges that certain class sizes could have been too small for accurate findings. Nonetheless, Liu et Zuo was one of the first in the field to investigate societal structures–matriarchies and patriarchies–in the context of gender differences in risk aversion.

Another notable contribution comes from Gneezy et al. in 2009, who published an investigation comparing competitive gender differences in matriarchies to those in patriarchies¹⁷. One of the first in the field to investigate the effect of culture and society at large on gender differences, the paper focuses on the matriarchal Khasi people and the patriarchal Maasai people of India. Though Gneezy et al. examine competition as the affected factor rather than risk aversion, their findings of opposite behavioral patterns between the Khasi and Maasai mirror that of Liu & Zuo’s initial findings that the more risk-averse gender switched between matriarchies and patriarchies.

Preliminary Experiment

In the preliminary phase of our research, we sought to identify whether exposure to leadership positions could be a potential driving factor in gender differences in risk behavior. So, to isolate the effect of leadership, we conducted an initial, informal survey in a community with matriarchal values, where men and women have equal access to leadership. If we received confirming results from our survey, it would have then been logical to explore a more comprehensive, formal experiment. The results of our informal experiment could also shape future considerations for a formal experiment. Considering our intentions and the informal nature of the experiment, the results we gathered are not to be taken definitively. Instead, we aimed simply to use the informal experiment as an indicator of whether future studies could be viable. Also, the study allowed us to pilot our experimental questions for the formal experiment.

We conducted our preliminary survey in the Bri Bri community of Costa Rica, which is an indigenous, matriarchal community. In particular, one significant Bri Bri value is that the men and women work collaboratively and in harmony. Due to the informal nature of the experiment, we were unable to garner an especially large and definitive sample size. Nevertheless, the results of our informal experiment are still intriguing. We asked 17 Bri Bri adults–eight men and nine women–a popular risk elicitation question to assess their financial risk behavior. The question, adapted from a review article on behavioral economics, asked participants which they would prefer if, hypothetically, they were given two options to win money¹⁸. We adapted the original question in two ways. First, we switched currencies by replacing US dollar amounts with Costa Rican colones, the most commonly used currency within the community. Second, even after currency conversion, we also changed the amounts used in the original question. These adjustments were necessary because of the cultural context in which we conducted the experiment. Because we wanted to propose two options of moderate financial amounts, our priority was ensuring that, after conversion, participants did in fact perceive our chosen currency amounts as such (see proposed experiment to understand our monetary choices). Despite this intention, we acknowledge the possibility that our arbitrary alterations of the survey question could have introduced unintended biases. For example, even after our considerations in choosing appropriate, moderate amounts, these choices may have been inaccurate. Inaccuracies in the monetary amounts we use could prevent individuals from maintaining their typical risk tendencies. Since the survey was informal, we determined these values subjectively based on what we deemed to be a “moderate” amount. One factor in our decision-making process was the prices of regularly purchased items in the community, which we will elaborate on in later sections. The first option was a ½ (50%) chance of winning 20,000 colones and a ½ (50%) chance of winning nothing. The second option was winning 10,000 colones. Although we asked participants orally in Spanish, they also received a paper copy of the question. The question tested the participants’ risk attitudes because although riskier, the first option resulted in an equal or greater expectancy than the second. However, due to the risk factor, risk-averse individuals would prefer the second, safer option.

Two men and three women chose the first option, while six men and six women chose the second option. These results indicate ¼ (25%) of the men chose the riskier option while ? (33.33%) of the women chose the riskier option. However, the small sample size indicates that these differences are negligible, and had there been one more man surveyed, the percentage of men and women who chose the riskier option could be equal. Since a nearly equal number of men and women chose the riskier option, it indicates that the theory of exposure to leadership roles being the primary cause of gender differences in risk aversion could be true. At the same time, our small sample size limits the conclusions we can draw from the preliminary experiment. Because there were not many people surveyed, we do not have the power to identify significant differences between men and women in our experimental results. This strengthens the need for our proposed experiment. Also, even though our small sample size was not ideal for gathering the most definitive results, we had the special opportunity to converse one-on-one with each individual we surveyed. These one-on-one conversations provided us with further insight into why people chose the options that they did, as well as the context surrounding their decisions. For instance, after one of our participants chose the safer option, they specifically cited the possibility of losing money as the primary reason behind their choice. This aligns with the principle of risk aversion, where individuals prefer sure outcomes with smaller gains over even minimally risky outcomes with similar or better gains. In another instance, after a participant chose the riskier option, we learned of their role in the community as a shopkeeper. In this case, we can perhaps infer that their experience with financial transactions and their exposure to a leadership position may have influenced their greater risk tolerance.

Proposed Experiment

This paper proposes a future experiment that could further solidify the theory that unequal access to leadership positions drives the observed gender gap in risk aversion. To isolate the effect of access to leadership positions, we can examine the difference in risk attitudes between men and women in an environment where women have equal exposure to leadership roles. If we compare the magnitude of the gender gap in a matriarchy to Western society, we can better understand whether greater exposure to leadership increases an individual’s appetite for risk. We will specifically employ the diff-in-diff method for our comparisons, which we will elaborate on in later sections. We hypothesize that in a matriarchal society where men and women have equal exposure to leadership, there is a smaller difference in risk aversion levels between genders compared to Western societies.

Our experiment will follow the investment game framework of Gneezy and Potters 1997¹⁹. In this model, participants are given a sum of money x, and they choose what portion of x to invest in a risky option with a higher expected value. More risk-averse individuals tend to invest less of x into the risky option in favor of taking the money at face value. The common utility of this framework is convenient for our purposes, as an experiment conducted in a matriarchy based on this framework would be easy to compare to similarly framed studies already conducted in Western societies. Studies using the GP framework include a 2019 study by Horn and Kiss on gender differences in risk aversion and patience and a 2022 study by Holden and Tilahun on gender differences in investing^20,21.

The experiment will be conducted in the matriarchal Bri Bri community of Costa Rica. We chose the differences-in-differences method for our experimental design because our experiment involves two dimensions in which we are looking for differences (gender and age groups). Besides gender, we plan to incorporate age groups into our experiment as a factor that is affected by leadership exposure. Our experiment operates under the assumption that leadership exposure increases throughout one’s life, and older individuals have faced more leadership exposure than younger individuals. Specifically, we assume that the younger group we survey will vary less in their risk attitudes because they have experienced less exposure to leadership. In this case, it makes sense to employ the diff-in-diff method to examine the differences that occur over time from the younger group to the older group (as a result of leadership exposure). To satisfy the diff-in-diff framework we plan to employ for our later comparisons, we will survey two groups: younger participants aged 14-17, and older participants over the age of 25. For the younger group, our selected age range is high school students, so we assume that they have not yet been exposed to significant leadership positions–and, for the experiments conducted in Western society, that they have not yet been exposed to unequal access to leadership positions. In contrast, we assume that those in our older group have already been exposed to leadership opportunities. Our goal in surveying these two distinct groups is to observe participants’ risk attitudes before and after their exposure to leadership. But for efficiency and practicality, we are not surveying the same exact group of people. Instead, we know that in recent years, there have been minimal social changes, so we can assume that it is unlikely these two groups are fundamentally different from each other. We acknowledge that this assumption is perhaps optimistic, so we also propose the possibility of an additional, preliminary experiment to control for the effects of potential social changes. If we receive optimal results from such an experiment, we could advance to the experiment as planned. However, if our preliminary results suggest that there have been extreme social changes in recent years that affect the risk tendencies of our two age groups, we could modify our experiment to focus on the same people. After conducting a power analysis of our intended experiment, to achieve a power of 80%, the minimum sample size necessary is one hundred and thirty individuals²². The minimum detectable effect size (MDES) is 0.50 with a significance level of 0.05. The experiment will be conducted individually with each participant, in a location away from others. Instructions will be provided orally in Spanish, though a paper version will also be available for reference. Even with the written instructions, we acknowledge the possibility of encountering language barriers or missing cultural nuances. To minimize this risk, we will allocate time for each participant to ask any questions they may have. After receiving 10,000 colones (about $19), participants will be given the option to invest any portion of the money into a lottery game, and they will keep any money they do not bet (see next paragraph for the rationale behind using 10,000 colones). The lottery game entails a ½ chance of doubling the bet and a ½ chance of losing it. The outcome of the lottery game will be determined by a bag with ten balls: five red, and five black. Participants will be asked to choose either red or black, after which an experimenter will procure a ball from the bag at random. If the participant’s selected color matches the color of the ball drawn, they win the bet. If the color they choose does not match the color of the ball drawn, they will lose the bet. Before they make their bets, these probabilities for winning and losing money will be explained thoroughly to participants, after which they may ask any clarifying questions. The total sum of money participants take home will be a combination of any money they did not invest in the bet and any earnings they made.

To minimize the presence of biases in our experiment (like the ones we outlined in the background), we will find a third party to collect all experimental data to implement a double-blind protocol. Although the researchers will be informed of all the necessary information to conduct our experiment as intended, they will not know about our overarching goals for the study. Because of this, the researchers will be protected from potential biases such as confirmation bias and publication bias.

Since the Bri Bri people reside in Costa Rica, they primarily use Costa Rican colones, which was an important factor to consider during the design of our experiment. As we plan to compare the results of our experiment to similar experiments from Western society, the perceived value of the monetary amounts we use in our experiments must match. Our goal was to select an amount for the experiment that would facilitate different risk attitudes depending on the participant. If the amount is too large or small, all participants–regardless of their usual risk tendencies–will skew toward making similar decisions. Risk aversion is a localized property, and people tend to be more risk-averse over larger bets. In American society, for instance, most people would agree that given the choice, the average American would choose to keep five

hundred thousand dollars rather than bet the sum and risk losing it. We selected 10,000 colones (18.91 USD) as our amount, though in the context of the Bri Bri people, using indicators such as average monthly household income was not applicable in our decision-making. Since the Bri Bri people are indigenous and live mostly detached from Western society, they also possess different values. Though the community makes purchases, they do not spend nearly as much as those living on the outside. Their average household income is lower than in other parts of Costa Rica, but they also have fewer expenses, and they place more emphasis on the importance of nature than they do on money. With this knowledge, we still elected to use monetary values for our experiment for ease of comparison, though we paid extra attention to the amounts. We chose 10,000 colones because it is a commonly used amount denominated by a single Costa Rican bill. We also considered expenses, as one common expense in the community is a bag of rice, which costs 1,400 colones. Another rarer expense is a birthday cake from outside of the community, which costs 15,000 colones.

Although non-financial factors may also play a role in the participants’ risk attitudes, we elect to focus on financial risk aversion for our proposed experiment. Especially when compared to the ease of measuring financial risk tendencies, it is more difficult to examine an individual’s non-financial risk behaviors. Prior literature also provides us with reliable survey questions to measure financial risk attitudes.

Discussion

We will compare the results of our experiment conducted in the Bri Bri community to similar studies conducted in Western societies using a diff-in-diff framework. We make the assumption that for all experiments conducted in both settings, the younger group of people has yet to be exposed to significant leadership positions, regardless of gender. In contrast, we assume that for experiments from both settings, certain people in the older groups have been exposed to leadership positions. However, in the Bri Bri community, we identify this exposure as more equal between genders compared to Western society. On this basis, we can define the “treatment” being unequal exposure to leadership positions, which occurs only in the Western society experiments, while the results of our experiment in the Bri Bri community are from a “control” group. We will primarily compare our results to the results of Horn and Kiss from 2019 and Holden and Tilahun from 2022. Both studies use the GP framework to explore gender differences in risk aversion, but Horn and Kiss also research gender differences in patience, while Holden and Tilahun investigate gender differences in financial risk aversion and investment behavior²³. Although we plan to compare the results of our study with the results of the other two studies, it is important to note that our study does not intend to replicate the other two. Rather, we will add to the findings of the other studies and provide a new angle that approaches risk tendencies from a leadership lens. The participants of both studies are adults, which aligns with our older, post-treatment group. We may need to conduct further research later to find relevant studies for the younger-aged group. We will compare the raw data for gender difference percentages in risk aversion across our study and the other studies before and after treatment. We will first quantify the difference between our study and the other studies before treatment, which in our case, is the younger group. If the difference between our study and the other studies remains the same after treatment, we can conclude that exposure to leadership positions is not a primary cause of the observed gender difference in risk behavior in Western society. However, if the difference after treatment changes significantly, then we can infer gender differences in risk aversion are due to lack of equal exposure to leadership roles.

Conclusion

Today, women are commonly viewed as more risk-averse than men. There is also a significant earnings gap between genders, and even in fields where the gender breakdown is equal, women are still underrepresented in leadership positions. We study potential causes behind the observed gender gap, as risk tendencies especially pertain to careers that favor higher levels of risk tolerance. There are many existing hypotheses, but we specifically examine a relatively new theory that access to leadership positions is the driving factor behind people’s risk attitudes. The 2019 study by Liu & Zuo is especially relevant in this exploration, and it also paves the way for further examinations. More studies could further validate the findings of Liu & Zuo, and perhaps we could replicate their findings in a new environment. The large-scale experiment we propose, which isolates the effect of leadership positions, will provide valuable insight into whether access to leadership roles is a major factor in people’s risk behavior. If our hypothesis is confirmed, perhaps people in Western society will place more emphasis on equal access to leadership roles, and with that, the landscape around gender differences and risk tendencies may evolve too.

References

1. Gilligan, Chris. “States With the Highest Percentage of Female Top Executives.” U.S. News & World Report. Last modified March 6, 2023. https://www.usnews.com/news/best-states/articles/2023-03-06/states-with-the-highest-percentage-of-women-in-business-leadership-roles#:~:text=In%202021%2C%20women%20made%20up,31.7%25%20of%20top%20executive%20roles. [↩]
2. Nelson, Julie A., Are Women Really More Risk-Averse than Men? (September 21, 2012). INET Research Note #012, Available at SSRN: https://ssrn.com/abstract=2158950 or http://dx.doi.org/10.2139/ssrn.2158950 [↩]
3. Nelson, Julie A. “Are Women Really More Risk-Averse than Men? A Re-Analysis of the Literature Using Expanded Methods.” Journal of Economic Surveys 29, no. 3 (2014): 566-85. https://doi.org/10.1111/joes.12069. [↩]
4. Byrnes, James P., David C. Miller, and William D. Schafer. “Gender Differences in Risk Taking: A Meta-analysis.” Psychological Bulletin 125, no. 3 (1999): 367-83. https://doi.org/10.1037/0033-2909.125.3.367. [↩]
5. Eckel, Catherine C., and Philip J. Grossman. “Forecasting Risk Attitudes: An Experimental Study Using Actual and Forecast Gamble Choices.” Journal of Economic Behavior & Organization 68, no. 1 (2008): 1-17. https://doi.org/10.1016/j.jebo.2008.04.006. [↩]
6. Charness, Gary, and Uri Gneezy. “Strong Evidence for Gender Differences in Risk Taking.” Journal of Economic Behavior & Organization 83, no. 1 (2012): 50-58. https://doi.org/10.1016/j.jebo.2011.06.007. [↩] [↩] [↩]
7. Pedroni, Andreas, Renato Frey, Adrian Bruhin, et al. “The Risk Elicitation Puzzle.” Nature Human Behaviour 1, no. 11 (2017): 803-09. https://doi.org/10.1038/s41562-017-0219-x. [↩]
8. Morgenroth, Thekla, Michelle K. Ryan, and Cordelia Fine. “The Gendered Consequences of Risk-Taking at Work: Are Women Averse to Risk or to Poor Consequences?” Psychology of Women Quarterly 46, no. 3 (2022): 257-77. https://doi.org/10.1177/03616843221084048. [↩]
9. Archer, John. “Testosterone and Human Aggression: An Evaluation of the Challenge Hypothesis.” Neuroscience & Biobehavioral Reviews 30, no. 3 (2006): 319-45. https://doi.org/10.1016/j.neubiorev.2004.12.007. [↩]
10. Hermans, Erno J., Peter Putman, Johanna M. Baas, Hans P. Koppeschaar, and Jack van Honk. “A Single Administration of Testosterone Reduces Fear-Potentiated Startle in Humans.” Biological Psychiatry 59, no. 9 (2006): 872-74. https://doi.org/10.1016/j.biopsych.2005.11.015. [↩]
11. Sapienza, Paola, Luigi Zingales, and Dario Maestripieri. “Gender Differences in Financial Risk Aversion and Career Choices Are Affected by Testosterone.” Proceedings of the National Academy of Sciences 106, no. 36 (2009): 15268-73. https://doi.org/10.1073/pnas.0907352106. [↩]
12. Joel, Daphna, and Ricardo Tarrasch. “The Risk of a Wrong Conclusion: On Testosterone and Gender Differences in Risk Aversion and Career Choices.” Proceedings of the National Academy of Sciences 107, no. 5 (2010). https://doi.org/10.1073/pnas.0913844107. [↩]
13. Croson, Rachel, and Uri Gneezy. “Gender Differences in Preferences.” Journal of Economic Literature 47, no. 2 (2009): 448-74. https://doi.org/10.1257/jel.47.2.448. [↩]
14. Jacobsen, Ben, John B. Lee, and Wessel A. Marquering. “Are Men More Optimistic?” SSRN Electronic Journal, February 1, 2008. https://ssrn.com/abstract=1030478. [↩]
15. Boorstin, Julia. “Survey: It’s Still Tough to be a Woman on Wall Street — But Men Don’t Always Notice.” CNBC Make It. Last modified June 26, 2018. Accessed April 10, 2024. https://www.cnbc.com/2018/06/25/surveyon-wall-street-workplace-biases-persist—but-men-dont-see-t.html. [↩]
16. Liu, Elaine M., and Sharon Xuejing Zuo. “Measuring the Impact of Interaction between Children of a Matrilineal and a Patriarchal Culture on Gender Differences in Risk Aversion.” Proceedings of the National Academy of Sciences 116, no. 14 (2019): 6713-19. https://doi.org/10.1073/pnas.1808336116. [↩]
17. Gneezy, Uri, Kenneth L. Leonard, and John A. List. “Gender Differences in Competition: Evidence from a Matrilineal and a Patriarchal Society.” Econometrica 77, no. 5 (2009): 1637-64. http://www.jstor.org/stable/25621372. [↩]
18. Kahneman, Daniel, and Amos Tversky. “Prospect Theory: An Analysis of Decision under Risk.” Econometrica 47, no. 2 (1979): 263. https://doi.org/10.2307/1914185. [↩]
19. Gneezy, Uri, and Jan Potters. “An Experiment on Risk Taking and Evaluation Periods.” The Quarterly Journal of Economics 112, no. 2 (1997): 631-45. http://www.jstor.org/stable/2951248. [↩]
20. Horn, Daniel, and Hubert Janos Kiss. “Gender Differences in Risk Aversion and Patience: Evidence from a Representative Survey.” SSRN Electronic Journal, 2019. https://doi.org/10.2139/ssrn.3313749. [↩]
21. Holden, Stein T., and Mesfin Tilahun. “Are Risk Preferences Explaining Gender Differences in Investment Behavior?” Journal of Behavioral and Experimental Economics 101 (December 2022): 101949. https://doi.org/10.1016/j.socec.2022.101949. [↩]
22. Dong, Nianbo, and Rebecca Maynard. “PowerUp!: A Tool for Calculating Minimum Detectable Effect Sizes and Minimum Required Sample Sizes for Experimental and Quasi-Experimental Design Studies.” Journal of Research on Educational Effectiveness 6, no. 1 (2013): 24-67. https://doi.org/10.1080/19345747.2012.673143. [↩]
23. Horn and Kiss, “Gender Differences”; Holden and Tilahun, “Are Risk.” [↩]