The overarching goal of this project is to develop methods to make social science research more rigorous, reproducible, and transparent. To make the work more concrete and to provide an exemplar, the project studies collective identity (CI) – roughly, an individual's mental and emotional connection with a broader community, category, or institution. We designed and constructed an online game platform where people play a novel group anagram game to foster the generation of CI within the participant group. The level of CI produced is measured by a proxy parameter called the DIFI (dynamic identity fusion index). We also model and simulate this game using various methods, including agent-based approaches. Finally, in support of rigorous methods for producing models of human behavior, various inference methods are studied to infer properties of human behavior models used in the social sciences. Inference methods are also used in the modeling of group anagram games: game data are used to devise models of human behavior and inference approaches are used to infer these models’ properties.