Perceived control is a core component of the sense of agency, yet it remains unclear what people actually judge when explicitly rating their agency. Previous studies have assessed subjective ratings of “control over,” “responsibility for,” or “predictability of” an outcome, often treating these concepts as interchangeable measures of agency. However, it is uncertain whether such ratings reflect a single underlying construct.

To clarify this, we developed a novel paradigm that systematically manipulated control and predictability. Participants chose between two pie charts (free-choice) or were assigned one (forced-choice). Each chart contained colored fields of varying sizes, with each color representing a possible outcome and field size indicating its probability. Participants were instructed that one field would be selected randomly and that their goal was to maximize the chance of obtaining a specific color. Control and predictability were manipulated through variations in choice options (free vs. forced choice), number of possible outcomes (colored fields), and outcome probabilities (field sizes). After each selection, participants rated their perceived control, predictability, and responsibility.

Linear mixed-model analyses revealed that these ratings were only moderately correlated, indicating partially distinct underlying constructs and substantial interindividual variability. Responsibility ratings showed the greatest variability and were least explained by objective task parameters. These findings suggest that explicit control ratings capture heterogeneous facets of agency, underscoring the need for clearer conceptual and methodological differentiation.