# Prediction of U.S. daily mask wearing and social distancing using psychologically valid agents during three waves of COVID-19

## TL;DR

This paper introduces a model that predicts mask wearing and social distancing behaviors in the U.S. during the first three waves of the pandemic using psychologically valid agents.

## Contribution

The novelty lies in integrating computational cognitive theory with individual behavior change to model public health behaviors.

## Key findings

- The model achieved an R2 of 0.93 for predicting daily mask wearing.
- It also achieved an R2 of 0.62 for predicting social distancing behaviors.
- The model was effective during the first three waves of the pandemic before vaccines were available.

## Abstract

We present Regional Psychologically Valid Agents (R-PVAs) as a modeling approach to predicting transmission-reducing behaviors and epidemiology. The approach builds upon computational cognitive theory and formalizes aspects of theories of individual-level behavior change. We present R-PVA models of social distancing and mask wearing in response to dynamics in the physical and information environments in the 50 U.S. states. The models achieve strong goodness-of-fits for predicting day-to-day mask-wearing (R2 = 0.93) and social distancing (R2 = 0.62) for the first three waves of COVID-19, prior to the rollout of vaccines.