Add study: Decoy effect on survey participation

[tbhat-ops]

Added study_015: Decoy effect on survey participation (Stoffel et al., 2023)

This study implements the decoy effect (attraction effect), where introducing an inferior option shifts preferences toward a target option.
In this experiment, participants received an email invitation to complete a survey. In the control condition, they were offered a single (target) survey. In the decoy condition, they were offered a choice between the target survey (closed-ended questions, faster reward) and a decoy survey (open-ended questions, delayed reward).
The key outcome is whether participants choose the target survey. The study tests whether the presence of the decoy increases target selection, and whether this effect depends on presentation order (target-first vs decoy-first).

Key components:

• Between-subjects design: control vs decoy conditions
• Nested order manipulation: target-first vs decoy-first
• Primary outcome: target survey completion rate
• Includes replication of main findings:
  • Decoy increases target completion
  • Strong order effect within decoy condition

Implementation includes:

• Full study specification, ground truth, and metadata
• Materials for all conditions
• config.py for trial generation
• evaluator.py for reconstructing statistical tests and computing BAS

Notes:

• Focused on primary findings (F2 and F3) for evaluation
• Some secondary findings are included in ground truth but may not be fully reconstructed in evaluator

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
human-study-bench-hqiu	Ready	Preview, Comment	Apr 4, 2026 6:36pm