Adapting a learning perspective on cognitive control, we aimed to test whether the regulation of cognitive flexibility versus stability can be learned through rewards. In a line of task switching studies, we systematically reinforced either task switches or repetitions more to test whether this leads to more voluntary task switching (flexibility) or repeating (stability) on unrewarded free choice trials. In a first set of experiments, where these free choice trials were interspersed between cued reward trials, we found evidence for increased task switching following rewarded switches in an instructed but not uninstructed experiment version. The absence in the latter might be explained by design complexity overshadowing subtle manipulation effects. In a third experiment version, using a simpler, uninstructed design with free choice trials only, we again observed successful adaptation to the reward scheme – even on trials with masked reward feedback. In a fourth experiment, we used both cued and free choice trials but in a blocked manner. While we did find a reinforcement effect in the pilot study, it disappeared in a large (n=496) sample. Here, distinct susceptibility to the reward manipulation might be linked to individual differences assessed with psychiatric questionnaires, or other sample characteristics. Together, these studies help determine whether and when people can learn about the value of task selection strategies beyond the scope of a single task, and provide a self-regulating system to understand putative higher-order control processes.