PRIMA (Vorstudie)

In 2018, Forbes magazine compiled a list of who were considered the 75 ‘Most Powerful People’ in the world. The ranking was based on the following criteria: It was determined how many other people one individual has got influence over, the amount of financial resources they control, and whether their influence encompasses more than one sphere of society. The resulting list comprises politicians, academics, media-moguls, business-people, and investors. What every individual has in common is that their careers were accompanied by fierce competition - be it in the form of electoral campaigns, fighting over market shares, taking over rivalling firms, competing for a promotion within an organisation, or striving for recognition as a leading researcher in one’s field. Most of the listed people are male: Among the 75 named individuals, there are only five women. The Forbes list makes no claim to accuracy, but does illustrate two circumstances very clearly: Firstly, societal influence is gained through competition. And secondly, women are vastly underrepresented in the most powerful positions across various spheres of society.

In their influential experimental paper, Niederle & Vesterlund (2007) explored the link between gender and the willingness to enter a competition. They found that women have a significantly lower preference for competition than men. This finding was confirmed by a large number of similar follow-up studies (see: Markowsky & Beblo, 2022).

The lack of gender balance in powerful positions calls for extensive interventions from policy makers: As it stands, the representation of women’s and men’s perspectives and interests is far from equal. Furthermore, the imbalance in high-ranking and well-compensated roles has negative implications for individual social security (American Academy of Actuaries, 2017) and team productivity (Desvaux et al., 2017).

With PRIMA, we aim to give advice on policy matters by focusing on competition as a mediating factor in the relationship between gender and societal power.

We aim to answer the following questions:

• What mechanisms in competitive domains prevent equal representation, and how can they be removed?
• What are recommendations for hiring- and promotion procedures in organisations as well as for entrance examinations at higher-education facilities that do not entail gender bias?
• What kind of environment facilitates a gender-neutral and merit-based selection of candidates?

Our data of interest is collected in experiments on competitive behaviour, where the experimenters use real-effort tasks to elicit performance. The trait of competitiveness is reflected by the participant’s decision between a competitive and a non-competitive payment scheme for the same task. This type of experiment, to date, already inspired two meta-analyses at our chair. Their study pool and findings serve as a basis for the PRIMA project:

When do we observe a gender gap in competition entry? A meta-analysis of the experimental literature (Markowsky & Beblo, 2022) explores the gender-specific rates for tournament entry across 120 studies. We find:

• Men on average choose competition 13 percentage points more often than women.
• The results are driven by the standard subject pool of university students performing mathematical tasks in a computer laboratory.
• Risk preferences and confidence are systematically related to the gender gap.
• A large portion of the gap remains unexplained.

 In a follow-up study titled Gender and performance under competitive pressure: A quantitative meta-analysis of experimental studies (Gardiner & Markowsky, work in progress), we compare the performances of men and women between payment schemes, synthesising results from 78 papers. Main insights are:

• Men’s average scores are around 4 percent higher than women’s, irrespective of competitive pressure.
• We find no differences in women's and men's reaction to competitive pressure in terms of performance.
• The results are driven by the standard subject pool of university students performing mathematical tasks in a computer laboratory.

The experimental data of interest offers a great deal of potential, which is impossible to tap into with the “conventional” approach of synthesising reported summary statistics and regression results that were used in the abovementioned meta-analyses. For the PRIMA project, we aim to merge all data sets at our disposal, creating a meta-data set on the individual observation level. Our previous meta-analyses combine for 150 experimental studies and observations from more than 60.000 individuals – giving an idea about the potential sample size and statistical power of our meta-data set. The benefit of this approach, compared to the standard meta-analysis method, where one has to rely on published information, is the complete freedom as to exactly what knowledge is extracted from the data. For our research, this circumstance is reflected as follows:

• We will not be limited to using the data in the way that was of primary interest to the original study. Especially performance values, even though measured, are not the main outcome in most experiments in our pool and are often neglected by the authors when presenting results. PRIMA offers the potential to include all these values in our analyses on performance.
• We can explore the residual gender gap in tournament entry that remains after controlling for trait moderators such as personality traits or biological differences without relying on published regression tables. Such trait moderators are often elicited but their coefficients are not often included in the (published) regression tables.
• A “conventional” meta-analysis is based on synthesising mean differences and presupposes independent groups that are represented by the respective means. Since the data at hand includes repeated performance observations for the same individual (for each task stage) and therefore produces means from dependent groups, the standard procedure is flawed and might induce bias. Using primary data will allow a within-subject analysis and therefore circumvent this issue.

Using the list of experiments included in Markowsky & Beblo (2022), and Gardiner & Markowsky (work in progress) as a starting point, we are continuously collecting data sets. To date, we have created a preliminary individual meta-data set with data from around 25 experiments, amounting to ca. 15,000 participants.

The project is currently on hold as we await approval for financial support, which will enable us to begin data harmonisation and data analysis.

If you have elicited experimental data on behaviour in competition and are willing to contribute to the PRIMA project, we kindly invite you to send us your data via christopher.gardiner"AT"uni-hamburg.de. We would also appreciate any documentation that makes it easier for us to understand your data and homogenise the variables according to our meta-data set (such as code, variable codebook, …). We want to emphasise that our focus lies on aggregated results. At no point of the project would we disclose data or code from individual papers.

If you have any further questions, feel free to email us(christopher.gardiner"AT"uni-hamburg.de).  

• American Academy of Actuaries (2017), ‘Women and social security’, Washington, DC, American Academy of Actuaries. May 2017.
• Forbes Magazine (2018), ‘The World's Most Powerful People’, Retrieved March 3, 2023. 
• Desvaux, G., Devillard, S., de Zelicourt, A., Kossoff, C., Labaye, E., Sancier-Sultan, S. (2017) ‘Women matter: ten years of insights on gender diversity’, McKinsey & Company. October 2017.
• Markowsky, E. & Beblo, M. (2022), ‘When do we observe a gender gap in competition entry? a meta-analysis of the experimental literature’, Journal of Economic Behavior and Organization 198, 139–163.
• Niederle, M. & Vesterlund, L. (2007), ‘Do women shy away from competition? Do men compete too much?’, Quarterly Journal of Economics 122, 1067 – 1101.