What drives preventative health behaviors one year into a pandemic? A replication and extension

Abstract

Objective: There is continued interest in understanding what leads people to engage in CDC-recommended COVID-19 prevention behaviors. We tested whether fear and COVID-19 worry would replicate as the primary drivers of six CDC recommended prevention behaviors. Methods and Measures: We recruited 741 adult participants during the second major peak of the COVID-19 pandemic in the United States (early 2021). Using very similar methods to the original study, participants completed a 10-day daily diary. Mixed effects models identified the strongest predictors of each individual prevention behavior as well as approach and avoidance behavior clusters. Results: At the between-person level, COVID-19 worry, COVID-19 perceived susceptibility, fear, and positive emotions all had positive zero-order associations with the prevention behaviors. However, with all predictors in the same model together, primarily COVID-19 worry remained significant for both the individual behaviors and behavior clusters. At the within-person level, only fear related to assessing oneself for COVID-19 and approach behaviors on the same day, but not the next day. Mediational analyses suggested COVID-19 worry, but not COVID-19 susceptibility, mediated the links between fear and approach/avoidance behaviors. Conclusion: Findings replicated worry about yourself or a loved one getting COVID-19 as the strongest predictor of prevention behaviors.

Keywords:

• Health behavior prevention behaviors COVID-19 health worry daily diary

CRediT model author contributions (https://credit.niso.org/)

D. J. Disabato: conceptualization, formal analysis, methodology, software, validation, visualization, writing original draft, writing review & editing. J. L. Foust: data curation, formal analysis, resources, software, visualization, writing original draft, writing review & editing. J. M. Taber: conceptualization, investigation, methodology, project administration, resources, supervision, writing original draft, writing review & editing. C. A. Thompson: conceptualization, data curation, funding acquisition, investigation, methodology, project administration, resources, supervision, writing review & editing. P. G. Sidney: conceptualization, methodology, project administration, supervision, writing review & editing. K. G. Coifman: conceptualization, methodology, project administration, resources, supervision, writing review & editing.

Disclosure statement

The authors report there are no competing interests to declare.

Data availability statement

The de-identified datasets and statistical analysis code associated with this manuscript are posted on the Open Science Framework (https://osf.io/waerb/).

Notes

1 The “difference” in the mean for Geography in the original study and Population Density in the present study stems from reversing the scoring in the present study for easier interpretation (i.e., most urban = 9 rather than most urban = 1).

2 A total of 24 participants endorsed the “Don’t know” or “Refused” response option.

3 The supplemental materials present the frequencies of the “Does Not Apply” responses.

4 The supplemental materials present the frequencies of the “Do Not Know” responses.

5 In addition to main effects, intervention condition could have interaction effects in the current study. To assess this possibility, we tested for intervention x between-person discrete emotion/risk beliefs interactions predicting COVID-19 prevention behaviors. A total of 60 interactions were tested in 60 separate models. All interaction effects were non-significant, except for intervention x disgust predicting wearing a face mask (b = .60, p = .047). Due to the large number of statistical tests and the p-value close to the .05 threshold, the significant interaction effect was likely due to chance and not interpreted. Table S4 in the supplemental materials reports the interaction term coefficients and their p-values.

6 Although we originally pre-registered conducting mediation analyses with the six individual behaviors using Multilevel Structural Equation Modeling, the models did not converge and we pre-registered an amendment removing the mediation analyses from our analysis plan (https://osf.io/ad92b/?view_only=ac8836d56f6d4a9ba1050f7999d197de). A reviewer suggested we conduct mediation analyses using the approach and avoidance behavior clusters. Because of the prior convergence issues and because almost all our significant results were at the between-person level, we simplified the mediation model to be only at the between-person level with the between-person components using linear regression.

Additional information

Funding

This research was supported in part by the U.S. Department of Education, Institute of Education Sciences, Grant R305U200004 to C. A. Thompson at Kent State University.