Modeling Decisions from Experience Among Frequent and Infrequent Switchers via Strategy-Based and Instance-Based Models

Abstract

In Decisions from Experience (DFE) research decision-makers search for information before making a final consequential choice in the sampling paradigm. Although DFE research involving the sampling paradigm has focused on accounting for information search and final choices using computational cognitive models. However, it remains to be seen how models implementing strategies (strategy-based models) and models relying upon memory retrievals (instance-based models) perform for final choices of participants with different switching behaviors. In this paper, we perform an individual-differences analysis and test the ability of strategy-based and instance-based models to explain final choices of participants with different switching behavior in the sampling paradigm. An instance-based model, which relies on recency and frequency memory processes, is calibrated to final choices of participants exhibiting frequent switching or infrequent switching between options. Also, we develop two strategy models: a summary strategy model and a round-wise strategy model. Both these models rely upon different switching behaviors and subsequent decision rules to derive choices. Results revealed that at the aggregate level, both the strategy-based and instance-based models explained consequential choices similarly when participants exhibited frequent switching. However, the instance-based model performed better than the strategy-based models when participants exhibited infrequent switching. Furthermore, at the individual level, the instance-based model was among the best models to fit to both frequent and infrequent groups. We highlight the implications of modeling experiential decisions using strategy-based and Instance-based models.