Gender differences in selection into self-competition

ABSTRACT

While previous studies demonstrated that, in many settings, women tend to be less willing than men to engage in interpersonal competition, this study focuses on selection into self-competition. Competing against own past performances can be an integral part of life, including job and sports. Using data obtained from a lab-in-the-field experiment, we find empirical evidence that women are, on average, more reluctant than men to compete against their own past performance. Our results suggest that this difference can be mainly explained by gender differences in risk preferences.

KEYWORDS:

• Competition
• gender differences
• experiment
• risk preferences

JEL CLASSIFICATION:

• C93
• C91
• C93
• J16

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Golf, for instance, seems to be strongly associated with self-competition (SC): ‘I began to see golf as a metaphor for living, for in golf you are basically on your own, competing against yourself and always trying to do better’. (Sean Connery, The Telegraph, August 25th, 2008).

² A prime example is an internet job advertisement of an Australian superstore that contained the following sentence: ‘So, if you have a winning attitude and love to compete against yourself on a daily basis when it comes to sales – what are you waiting for?’

³ On a 7-point scale from ‘does not apply at all’ to ‘fully applies’, participants evaluated the following statements: ‘In case of monetary investments, I am willing to take risks’. And ‘In general, I am willing to take risks’.

⁴ We control for mother’s, father’s, and both parents’ self-employment, because under the assumption that self-employed individuals are more competitive (Bönte and Piegeler 2013) this could indicate more competitive parents. We control for the order of treatments, because Buser (2016) finds men reacting to losing a two-person tournament by subsequently picking a more challenging target while women lower their performance. Despite not having feedback, individuals may have a guess about their performance, such that behaviour in one treatment might be affected by the preceding treatment. We also control for differences in performance in the piece-rate condition. In our data, however, men and women do not differ in performances in PR (as in other) conditions (male = 5.61, female = 5.45, diff = −0.16, S.E. 0.36, t = −0.45, p = 0.66), implying that performance in PR cannot explain gender difference in selection into SC.