Sensitivity Analysis for Computational Models in Psychology: A Tutorial
Year: 2026
Type: preprint
Abstract: There has been an increasing interest in formal models in psychology, allowing researchers to betterunderstand and intervene on psychological systems. However, the methods and tools for understandingand analysing these models are rarely used in psychological research. This is not necessarily due to a lackof methodology, but rather due to them not being familiar in this field. This paper introduces one suchmethod from computational science that can be used to determine the influence of model parameterson an output variable: (global) Sensitivity Analysis (SA). We first discuss the main goals of an SA andpotential applications, such as determining which parameters are the most relevant for causing a changein a chosen model output, such as suicidal thoughts or frequency of panic attacks. We then presenttwo methods that can be used to conduct an SA: the variance-based Sobol SA and the density-basedPAWN. Both of them are demonstrated using an existing computational model in a tutorial-like fashion,provided by (annotated) code in R. The results are connected to empirical literature and implicationsfor model evaluations are discussed. Lastly, we aggregate the entire process into a workflow and codepipeline that can be easily applied to other models. Although this paper only demonstrates the basicapplication of SA, further extensions and purposes, such as model calibration, are discussed.
Language: English
Subfield: Statistics, Probability and Uncertainty
Field: Decision Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7160846582