Data and analysis code: Measuring intrinsic motivation with the Becker-DeGroot-Marschak mechanism during task engagement

Description

De-identified participant-level data and a self-contained analysis script for the study "Measuring intrinsic motivation with the Becker-DeGroot-Marschak mechanism during task engagement." Intrinsic motivation is measured as an incentive-compatible willingness to accept, elicited with the BDM mechanism during a LEGO-building task, at one of four randomly assigned completion checkpoints (20, 45, 70, 95 percent), to test the goal-gradient prediction. 1_data.csv holds one row per engaged participant. Of these 304, the 275 who reached their assigned checkpoint were analyzed, and 29 had insufficient progress. Running 2_analysis.py on 1_data.csv regenerates 3_results.html, which reproduces the reported statistics, tables and figure in the paper in the order they appear. Participants and sessions appear only as anonymized sequential labels (pid P001.., session S01..), and all directly identifying oTree fields are excluded.

Steps to reproduce

1. Install Python 3 with numpy, pandas, statsmodels, scipy, and matplotlib. Versions used: Python 3.14.4 and numpy 2.4.4, pandas 3.0.3, statsmodels 0.14.6, scipy 1.17.1, matplotlib 3.10.9 (any recent versions reproduce the same numbers). 2. Put 1_data.csv and 2_analysis.py in the same folder. 3. Run: python 2_analysis.py 4. This reads 1_data.csv and writes 3_results.html. Open it in a browser. Its blocks follow the paper (Sections 5.1, 5.2, 5.3, 6) with all numbers, tables and the figure inline. The main analysis uses the 275 participants with reached == 1, and the sample-selection funnel uses all 304. One bootstrap standard error uses a fixed seed (numpy default_rng(0), 500 resamples) and reproduces exactly on the deposited row order.

Institutions

• Nihon University

  Tokyo, Tokyo
• University of Tsukuba

  Ibaraki, Tsukuba
• The University of Tokyo

  Tokyo, Tokyo

Categories

Related Links

Licence