Data and models from experiment where participants tapped to 91 different well-formed rhythms. There were also two tasks where participants selected their favourite rhythm from a set of well-formed and not-well-formed rhythms, and adjusted the long-short ratio of well-formed to their preferred value.