Minimizing learning behavior in repeated real-effort tasks

In this paper, we discuss learning behavior and the heterogeneity of subjects' ability to perform in real-effort tasks. Afterwards, we present a novel variant of Erkal et al.'s (2011) encryption real-effort task which aims to minimize learning behavior in repeated settings. In the task, participants encrypt words into numbers. In our variant, we apply a double-randomization mechanism to minimize learning and heterogeneity. Existing experiments with repeated real-effort tasks find a performance increase of 12-14% between the first and second half. By contrast, our task mitigates learning behavior down to 2% between the first and second half. The data show that subjects show a small heterogeneity in performance.