This study re-examines the link between agency and responsibility attribution—specifically, whether higher personal agency in decision-making leads to greater responsibility for outcomes. Agency-based attribution of responsibility underlies legal systems worldwide, from historical precedents such as Nuremberg trials to contemporary cases, with previous studies consistently finding that external constraints reduce both agency and responsibility. However, Malter et al. (2021) reported a striking asymmetry through their series of experiments: observers attribute greater responsibility to individuals who make free choices, while actors themselves experience heavier feelings of culpability when they have low agency. This mismatch between external attribution and internal experience of responsibility is theoretically important but has not yet been closely replicated.
We replicated and extended Malter et al.'s (2021) findings through three key innovations. First, we employed a unified 2×3 between-subjects framework manipulating Perspective (self vs. other) and Agency Level (following orders vs. taking advice vs. autonomous decisions), allowing direct statistical comparison of whether perspective determines the direction of the agency-responsibility relationship. Second, we refined measurement by focusing specifically on responsibility attribution as our central construct while also assessing perceived deservingness of punishment and accountability, providing a more nuanced account of attribution. Third, we introduced advice-taking as an intermediate agency condition to test whether observed patterns generalize across a broader spectrum of agency.
Using Prolific Academic, we recruited 1,600 participants (~267 per cell) to have 90% power to detect effects comparable to the original study's smallest reported effect (ηp² ≈ .011). Data collection is currently underway, with results available for presentation at ICAP 2026. Together, this replication and extension tests the robustness of the agency-responsibility attribution link across perspectives, measurement approaches, and levels of agency, offering both theoretical clarification and methodological advances.