This OSF repository contains materials associated with the study: Diversity promotes abstraction and cognitive flexibility in collective problem solving. Browsing the folders, you will find all stimulus items, analysis scripts, data sets, and models. Notice that the embedded Bayesian models are compiled using rstan on mac, and might therefore not run on windows. However, code is provided to rerun the models. This study is preregistered using the service aspredicted.org. In the preregistration pdf found on this page, you will find all the hypothesis tested and reported in the study. Notice, however, that the original plan was to analyze the data using mixed effect models (lme4 for RStudio). Meanwhile, we have decided to migrate to the more contemporary Bayesian framework as we believe that this allow us to model the analysis in a more appropriate way given the nature of the data. We also believe that the Bayesian framework is the future of statistics for the cognitive sciences. All results reported are analogous to those which can be obtained using a mixed effects regression approach (see also Fjaellingsdal, T. G., Vesper, C., Olesen, C., & Tylén, K. (2020). Abstraction and cognitive flexibility in collective problem solving: The role of diversity. In Proceedings from the annual meeting of the Cognitive Science Society 2020: https://cogsci.mindmodeling.org/2020/papers/0245/0245.pdf.