Gender Differences in Preferences for Leadership Tasks

I run an online discrete choice experiment among employees in Germany to estimate their willingness-to-pay (WTP) for the two leadership tasks: team and project responsibility and two additional job attributes: commuting time and flexibility. The discrete choice experiment is part of a longer questionnaire that focuses on leadership roles.

2025-04-09

2025-06-01

Respondents’ choice between two hypothetical jobs.

I use respondents’ choice between two hypothetical jobs to estimate the WTP for two leadership related job tasks (team and project responsibility) and two additional job attributes (commuting time and flexibility).

My focus is on the heterogeneity in WTP for the leadership tasks by gender. The inclusion of commuting and flexibility serves as a benchmark, allowing for a comparison of the effect sizes of the leadership tasks with well-established job attributes.

My primary interest lies in the two leadership tasks. For these tasks, I explore whether common explanations for gender differences found in the literature can account for the patterns observed in the gender-specific WTP estimates.
Specifically, I examine interactions between gender and the following four constructs:
1. Respondents’ career aspirations
2. Respondents’ family preferences
3. Respondents’ competitiveness
4. Respondents’ share of female coworkers

Additionally, I conduct two types of exploratory analyses. First, I examine whether respondents sort into their current jobs based on job attributes, following an approach similar to Maestas et al. (2023). Second, I explore additional sources of heterogeneity in willingness to pay by respondents’ education, age, job type (e.g., sector, career opportunities), and other individual characteristics.
(For details on the constructs and survey questions, see the pre-analysis plan.)

The experimental design aims at identifying the workers’ WTP pay for the two job tasks: team responsibility and decision-making. Additionally, I plan to analyze workers' willingness to pay for other job attributes. The key features of the experimental design follows Maestas et al. (2023).

Each respondent participates in ten stated-preference experiments. In each experiment, survey respondents are asked to select between two jobs (A and B), each defined by a partially varying set of job attributes and wages.

For each respondent, I define a baseline job around which job attributes would vary. The baseline job is the respondent’s current job in order to generate hypothetical jobs that would appear realistic to the respondent. For this purpose, the respondents are asked about their current occupation, job attributes and wage prior to the discrete choice experiment. The five job attributes vary when selected and the wage always varies. In the following, I describe how I create the variation.

Starting from the respondent’s baseline job, I create hypothetical Job A and Job B by randomly selecting two non-wage attributes to vary across the two hypothetical jobs. Within each of the two randomly selected attributes, attribute values are chosen at random sequentially, first for Job A and then for Job B without replacement.

In every job choice, respondents see two jobs next to each other where two randomly selected job attributes and the monthly wage vary. Those three attributes are marked in red and all the other job attributes are the same as in their current job. The respondents were asked to select “Job A” or “Job B.”

(For details on the mapping between survey questions and discrete choice attributes, as well as the random wage generation, see the pre-analysis plan.)

Reference: Maestas, Nicole, Kathleen J. Mullen, David Powell, Till von Wachter, and Jeffrey B. Wenger. 2023. "The Value of Working Conditions in the United States and the Implications for the Structure of Wages." American Economic Review, 113 (7): 2007-47.

Not available

By a computer

At the individual level

No

-

3,000 individuals making 10 choices over 2 job profiles resulting in 30,000 choices over 60,000 job profiles

There are no treatment arms in my design, but the job attributes vary randomly in each individual choice experiment.