Using machine learning to identify important predictors of COVID-19 infection prevention behaviors during the early phase of the pandemic

PsyCorona Collaboration, Caspar J. Van Lissa*, Wolfgang Stroebe, Michelle R. vanDellen, N. Pontus Leander, Maximilian Agostini, Tim Draws, Andrii Grygoryshyn, Ben Gützgow, Jannis Kreienkamp, Clara S. Vetter, Georgios Abakoumkin, Jamilah Hanum Abdul Khaiyom, Vjolica Ahmedi, Handan Akkas, Carlos A. Almenara, Mohsin Atta, Sabahat Cigdem Bagci, Sima Basel, Edona Berisha KidaAllan B.I. Bernardo, Nicholas R. Buttrick, Phatthanakit Chobthamkit, Hoon Seok Choi, Mioara Cristea, Sára Csaba, Kaja Damnjanović, Ivan Danyliuk, Arobindu Dash, Daniela Di Santo, Karen M. Douglas, Violeta Enea, Daiane Gracieli Faller, Gavan J. Fitzsimons, Alexandra Gheorghiu, Ángel Gómez, Ali Hamaidia, Qing Han, Bertus F. Jeronimus, Yasin Koc, Joshua Krause, Maja Kutlaca, Anton P. Martinez, Kira O. McCabe, Solomiia Myroniuk, Boglárka Nyúl, Anne Margit Reitsema, Michelle K. Ryan, Edyta M. Sasin, Samiah Sultana, Kees Van Veen

*Bijbehorende auteur voor dit werk

OnderzoeksoutputAcademicpeer review

2 Citaten (Scopus)
2 Downloads (Pure)

Samenvatting

Before vaccines for coronavirus disease 2019 (COVID-19) became available, a set of infection-prevention behaviors constituted the primary means to mitigate the virus spread. Our study aimed to identify important predictors of this set of behaviors. Whereas social and health psychological theories suggest a limited set of predictors, machine-learning analyses can identify correlates from a larger pool of candidate predictors. We used random forests to rank 115 candidate correlates of infection-prevention behavior in 56,072 participants across 28 countries, administered in March to May 2020. The machine-learning model predicted 52% of the variance in infection-prevention behavior in a separate test sample—exceeding the performance of psychological models of health behavior. Results indicated the two most important predictors related to individual-level injunctive norms. Illustrating how data-driven methods can complement theory, some of the most important predictors were not derived from theories of health behavior—and some theoretically derived predictors were relatively unimportant.

Originele taal-2English
Artikelnummer100482
Aantal pagina's14
TijdschriftPatterns (New York, N.Y.)
Volume3
Nummer van het tijdschrift4
Vroegere onlinedatum2022
DOI's
StatusPublished - 8-apr-2022

Citeer dit