William P. Jimenez, M.S., Old Dominion University
Purpose/Objectives
In this study, I scoped out the literature on workers’ psychological strain specific to the global COVID-19 pandemic and conducted a psychometric meta-analysis to explore how such strain is related to worker characteristics and other relevant work-related variables.
Background
Occupational health has seen particularly tremendous growth during the pandemic-so much so that there exist at least two dozen extant systematic reviews and meta-analyses involving COVID-19’s impact on healthcare workers (e.g., Dong et al., 2021; Dutta et al., 2020; Li et al., 2021; Marvaldi et al., 2021). These studies, however, focused on the prevalence of workers’ general mental-health outcomes. The present review and meta-analysis is the first in which coronavirus-specific psychological strain is examined in relation to worker characteristics and work-related variables.
Methods
I identified and reviewed literature (viz., Chandu et al., 2020; Cortez et al., 2020; Muller et al., 2021; Ransing et al., 2021) summarizing the measures researchers have recently developed to measure coronavirus-specific strain. Using Google Scholar, I conducted forward searches for studies implementing the measures and limited my queries to literature involving “employees,” “professionals,” or “workers.” This search yielded approximately 1,400 results. After excluding duplicates and irrelevant studies (e.g., conceptual articles, case studies, studies with non-worker samples), I tabulated the variables measured in the remaining studies to evaluate whether conducting a meta-analysis would be feasible. This assessment yielded the following insights:
– The two most popular coronavirus-specific strain measures that have been employed in work settings were the Fear of COVID-19 Scale (FCV-19S; Ahorsu et al., 2020) and the Coronavirus Anxiety Scale (CAS; S. Lee, 2020).
– Researchers have examined coronavirus-specific psychological strain in relation to many relevant variables, including workers’ demographics (e.g., age, gender, marital status), mental-health problems (e.g., general anxiety symptoms, general depressive symptoms, post-traumatic stress symptoms), general positive well-being (viz., resilience, quality of life), work-related well-being (viz., burnout, job satisfaction), and other relevant variables (e.g., trait mindfulness, sleep problems)-with at least three studies looking at each variable in relation to coronavirus-specific psychological strain.
I coded a total of 112 effect sizes across k = 37 independent samples with N = 12,391 workers. I proceeded to conduct an individual-correction (correcting measurement unreliability), random-effects psychometric meta-analysis (Schmidt & Hunter, 2015). If researchers did not provide reliability information, I imputed the corresponding database average. Although most effect sizes were Pearson correlations, some studies reported Spearman correlations, which I converted to the former (see Rupinski & Dunlap, 1996). Correlations involving gender or marital status were disattenuated as they involved dichotomous variables (see Schmidt & Hunter, 2015). Alternatively, if relevant information was provided for these variables (e.g., group means and standard deviations, group sample sizes), Pearson correlations were computed and disattenuated. Strain measure (viz., CAS vs. FCV-19S), publication type (viz., published vs. not published), and sample type (viz., healthcare vs. non-healthcare workers) were explored as categorical moderator variables (if a relationship had more than one subgroup comprising at least three samples, Z tests were conducted to formally assess subgroup differences; see Raju and Brand, 2003). Country was also coded, and the date (converted to a numeric value using “=DATEVALUE()” in Excel) a country reached a threshold of 20 for the Oxford COVID-19 Government Response Tracker’s (OxCGRT; https://www.bsg.ox.ac.uk/research/research-projects/covid-19-government-response-tracker) stringency index was explored as a continuous moderator variable (for another meta-analysis that incorporated this specific moderator variable, see Y. Lee et al., 2021).
Findings
See Table 1 for meta-analytic estimates. As most studies were published and involved healthcare workers, the only categorical moderator I was able to examine was coronavirus-specific strain measure. All overall strain-demographics relationships (except for the nonsignificant one involving professional tenure) were positive and small. Measure moderated relationships involving gender (women coded as higher) and marital status (married coded as higher) such that they were significantly stronger with the CAS. Overall relationships between strain and mental-health problems were positive and moderate to strong (remarkably so for post-traumatic stress symptoms: ρ = .87). Measure moderated the relationship between strain and general depressive symptoms such that the relationship was significantly stronger-remarkably so-with the CAS: ρ = .85. The relationship between strain and obsession with COVID-19 was strong and positive. Relationships with general well-being were negative and moderately strong. Relationships with burnout and job insecurity were positive and strong. The relationship involving job satisfaction was nonsignificant. Stringency index moderated the relationship with gender: meta-regression weight = 0.007, 95% CI [0.002, 0.013]. In other words, the sooner a country began to implement stricter “lockdown-style” policies, the smaller the gender difference in coronavirus-specific psychological strain.
Discussion
This study is the first psychometric meta-analysis of strain specific to COVID-19 in relation to important worker variables. All but two overall relationships (with professional tenure and job satisfaction) were significant. Additionally, some meaningful moderation occurred; however, given the widths of some credibility intervals, some of which included zero, more research identifying substantive moderator variables is warranted.
Conclusions
I intend to continually update this meta-analysis with additional relevant studies and meta-analytically compare coronavirus-specific psychological strain to other strain experiences workers experience.