Perceptions of Persistence and Gender

This study will vary the task used to examine persistence. There are two treatments: a male-stereotyped task and a female-stereotyped task.

2025-03-26

2026-12-31

Our primary outcomes include:
• Gender difference in persistence: the gap between men and women’s decision to do the task again, given positive and negative feedback
• First-order beliefs about gender differences in persistence, given positive and negative feedback
• Second-order beliefs about gender differences in persistence, given positive and negative feedback

Our secondary outcome includes gender difference in performance, which is the gap between the proportion of men and women performing above and below the threshold.

This is a laboratory experiment to be conducted in an experimental economics lab, recruiting participants from the university. The experiment consists of two parts.

In Part 1, participants complete a real-effort task (Emotion or Math, depending on their treatment) and state their prior belief about whether they exceed a performance threshold based on pilot data. Before receiving feedback, they state their posterior belief about exceeding the threshold and decide whether they would repeat the task under both possible feedback scenarios (above or below the threshold). After making these decisions using the strategy method, they receive noisy feedback on their performance, and their choice is implemented accordingly—either repeating the task or watching a video.

In Part 2, we elicit participants' beliefs about gender differences in task performance and persistence. They estimate whether men or women are more likely to pass the threshold and to repeat the task after receiving positive or negative feedback. Additionally, they predict how other participants in the session answered these same belief questions.

After completing Part 2, participants answer a survey covering demographics (e.g., age, nationality, and study level), their experience in the experiment, and measures to identify potential mechanisms underlying any observed gender differences in persistence and their perceptions, including risk preferences, grit, locus of control, gender norms attitudes, self-improvement and growth mindset, and Big Five personality traits.

Our main research questions are:
1. Are there gender differences in persistence, given positive feedback and negative feedback?
2. What are participants’ (first-order) beliefs about the gender differences in persistence, given positive feedback and negative feedback?
3. What are participants’ (second-order) beliefs regarding others’ beliefs about the gender differences in persistence, given positive feedback and negative feedback?
4. Are there misperceptions in these first-order and second-order beliefs, given positive feedback and negative feedback?
5. Are the answers to questions 1 to 4 different between the two treatments?

Not available

Participants will be recruited via email using ORSEE. They will choose between a list of available sessions and the session is randomized to a treatment before it is initialized.
Depending on the randomized treatment, participants complete either a female task (Emotion) or a male task (Math) in Part 1.

The randomization unit for this experiment is the individual participant.

No

300 participants

Based on power calculations, we aim to collect 300 datapoints in total (75 for each gender and for each treatment /task).

300 participants in total for a 2X2 factorial design:
- Men/Women x Math/Emotion task
- 75 participants per gender per task

With a sample size of 75 men and 75 women for each task or each feedback type (300 in total), at alpha=0.05 and power=0.80, using the following tests: • RQ1: Gender difference in persistence (between-subjects) - Type: Binary, Persisted vs. Did Not Persist - Fisher’s exact test (two samples): We can detect a 20-percentage points difference in the proportions of men and women doing the task again, assuming a baseline percentage of 60% of one gender doing the task again. • RQ2-4: Beliefs and misperceptions in beliefs about the gender difference in persistence - Type: Categorical/Ordinal, -1 Men>Women, 0 No gender difference, 1 Men<Women - Wilcoxon signed-rank test (one sample case): We can detect a minimum effect size d of 0.1661 standard deviations. • RQ2-4: Beliefs and misperceptions in beliefs about the gender difference in persistence, given positive vs. negative feedback (within-subject) - Type: Categorical/Ordinal, -1 Men>Women, 0 No gender difference, 1 Men<Women - Wilcoxon signed-rank test (matched pairs): We can detect a minimum effect size d of 0.1661 standard deviations. • RQ5: Treatment difference in the outcomes above (between-subjects) - Type: Categorical/Ordinal, -1 Men>Women, 0 No gender difference, 1 Men<Women - Mann-Whitney U Test/Wilcoxon rank-sum tests (two groups): We can detect a minimum effect size d of 0.3322 standard deviations.