Individual and group interactions weave society together, and their influence on relationships changes as time unfolds. Events occurring at each level are described using different concepts at each level of research, preventing us from predicting the future, let alone comprehensively understanding what is happening in front of us. Gaming simulations help us understand this phenomenon. Unlike mathematical models or multiagent simulations, experiential simulation gaming allows us to experience what is happening in the modeled system. However, it is also difficult to draw generalizations because they involve numerous interactions among many participants. This issue is addressed in this chapter. Psychological experimentation is an experimental paradigm based on specific theory. Under controlled conditions, the interaction steps are maintained pre-post or, at most, within a few steps, and programmed bots are commonly used as interaction partners. The goal is to extract generalization rules by eliminating parts of complex phenomena. To combine such psychological experiments with gaming simulations, it is necessary to discover the structure that exists in gaming and to link the constructs that form that structure to psychological research. This chapter presents an educational program that combines a trust game experiment with a simulated society game to provide experiential learning about environmental issues. This study attempts to model a phenomenon that cannot be understood by psychological constructs using machine learning. Psychological experiments attempt to understand the complex in a simple manner, whereas machine learning attempts to understand the complexity as it is. This chapter proposes a method to advance the understanding of latent factors in gaming simulations by combining these two.