Perceived Vulnerability to Disease, Resilience, and Mental Health Outcome of Korean Immigrants amid the COVID-19 Pandemic: A Machine Learning Approach

Abstract

This study examined the predictive ability of perceived vulnerability to disease (PVD), fear of COVID-19, and coping mechanisms on the Korean immigrants' psychological distress level amid the pandemic. Through purposive sampling, both foreign-born and US-born Korean immigrants residing in the US above the age of 18 years were invited to an online survey. Between May and June 2020, data collection took place, which yielded the final sample of 790 participants from 42 states. An artificial neural network (ANN) was used to verify variables that predict the level of psychological distress on the participants. The model with one hidden layer holding six hidden neurons showed the best performance. The error rate was approximately 27%, and the results from the sensitivity analysis, the receiver operating characteristics (ROC) curve, showed that the area under the curve (AUC) was 0.801. The most powerful predicting variables in the neural network were resilience, PVD, and social support. Implications for practice and policy are discussed.