This paper investigates how the self-selection of tasks affects worker performance. Specifically, it investigates the impact of aligning tasks with workers’ preferences and the effect of providing workers with greater autonomy in choosing their tasks. To answer these questions, we conducted an online experiment in which participants engaged in one of two real-effort tasks. We exogenously varied whether participants were either assigned their preferred or non-preferred task, or if they had the opportunity to actively self-select their task. The results show that participants who self-selected a task or were randomly assigned their preferred task increased their output by about 33%-43% of the standard deviation compared to those assigned their non-preferred task. This increase in output is linked to both enhanced productivity and extended time spent working on the task. In essence, our results underscore that workers’ performance depends crucially on whether they work on their preferred task, with less emphasis on whether the task was self-selected or assigned.