Choices About Competition: Differences by Gender and Hormonal Fluctuations, and the Role of Relative Performance Feedback

Abstract

Economic experiments have shown that when given the choice between piece rate and winner-take-all tournament style compensation, women are more reluctant than men to choose tournaments. These gender difference experiments have all relied on a similar framework where subjects were not informed of their relative abilities as compared to other potential competitors. I replicate these previous findings and then show that giving feedback about past relative performance moves high ability females towards more competitive compensation schemes, moves low ability men towards less competitive compensation schemes such as piece rate and group pay, and removes the gender difference in compensation choices. I then examine between and within-subjects differences in choices for females, across the menstrual cycle. I find that, before receiving relative performance feedback, women in the low-hormone phase of their cycle are less likely to enter tournaments than women in the high-hormone phase. Men are more likely to choose tournaments than females at either stage. There are no significant selection differences between any of these groups after they receive relative performance feedback.

Outrunning the gender gap—boys and girls compete equally

“…In this literature, however, only a few tasks have been used to measure competitiveness, and these tasks can arguably be considered as more male than female. Three studies find that gender differences in competitiveness vary with the task at hand (Gneezy and Rustichini 2004b, Günther et al 2009, Grosse and Reiner 2010, whereas another study find no difference in the gender gap between a maze task and a word task (Wozniak et al 2010). Meanwhile, work in social psychology suggests that individual perceptions about relative performance, such as (over)confidence, and especially stereotypes may have important implications for actual performance (Steele 1997, Shih et al 1999.…”

“…It has previously been shown that culture affects important economic decisions such as labor market participation and fertility (e.g., Fernández and Fogli 2006), and the institutional setting has been found to influence competitive behavior (e.g., Balafoutas and Sutter 2010, Gneezy et al 2009, Niederle and Yestrumskas 2008, Cotton et al 2009, Niederle et al 2009, Wozniak et al 2010. For example, the gender gap in selfselection has been shown to disappear with performance feedback (Wozniak et al 2010) and the difference in performance change vanishes with repetition of the competition (Cotton et al 2009). Women have also been found to compete more than men in a matrilineal society whereas men compete more than women in a patriarchal society (Gneezy et al 2009).…”

“…There is mixed evidence on the role of the task on the gender gap in competitiveness. Günther et al (2009) and Grosse and Reiner (2010) find a gender difference in performance change in a math task but not in a word task, whereas Wozniak et al (2010), using a maze task and the same word task, find no difference in the gender gap between the two tasks. 2 Thus, to explore competitiveness more generally than what has previously been done, in particular on children, we also look at what we consider more female tasks.…”

“…A study looking at competitiveness among men finds no relationship between self-selection into a tournament and current testosterone levels (Apicella et al 2010). Buser (2009) finds that women are less likely to self-select into a tournament when progesterone and estrogen levels are high whereas Wozniak et al (2010) find the opposite with women in the low-hormone phase being less competitive. 25 Meanwhile, 23 The difference between our results and those of Cárdenas et al (2010) in skipping rope is perhaps due to the larger sample size in the latter study (520 children).…”

Are females scared of competing with males? Results from a field experiment

“…Other works neutralize risk aversion by asking experiment's participants to make a choice characterized by the same risk as the tournament. In the same vein, beliefs are manipulated by changing the task performed by participants 4 (Niederle and Vesterlund, 2007;Wozniak et al, 2010) or, similarly to what we do in this paper, providing them information on their relative performance (Cason et al, 2010;Wozniak et al, 2010).…”