Masculine Norms and Infectious Disease: The Case of COVID-19

During the novel coronavirus pandemic, early data suggested that men were slightly more likely to contract COVID-19 than women, less likely to seek medical attention for the disease, and far more likely to die as a result of COVID-19. While several studies have explored this gender gap, none has attempted to isolate the psychological processes underpinning this phenomenon. In this research note, I suggest that sexism partly explains these differences. Using data from a large (N = 100,689) survey of American adults conducted between March and June 2020 by the Democracy Fund and the University of California, Los Angeles (Nationscape), I find that sexist beliefs, a component of masculine norms, are consistently the strongest predictor of coronavirus-related emotions, behaviors, policy attitudes, and ultimately contracting COVID-19. This study highlights how gender ideology can impact health and impede government public health efforts.


B Questions, Descriptive Statistics, and Scale Construction B.1 Nationscape Data
Nationscape is a large, weekly online survey conducted by Lucid for the Democracy Fund and researchers at UCLA that was designed to collect weekly snapshots of the American electorate throughout the 2019-2020 primary and general elections (Tausanovitch and Vavreck 2020). This cross-section survey is in the field every day of the week and includes weekly collections of N∼6,250 responses. While the sample is opt-in, a representativeness assessment of the data finds that the samples are comparable to those collected by well-known pollsters like Pew and YouGov (Tausanovitch et al. 2019). More information on the survey can be found at https://www.voterstudygroup.org/nationscape.

B.2 IV: Sexism Scale
Respondents indicated whether they agreed or disagreed (5-pt Likert) with the following statements. The first two are part of an established old-fashioned sexism battery (Swim et al. 1995) and the last two, which tap into perceptions of attitudes toward gendered social hierarchies, were created by researchers at Democracy Fund Voter Study Group for their panel study.
• "I would be more comfortable having a man as a boss than a woman" (reversed, mean=2.82) • "Women are just as capable of thinking logically as men" (mean=1.66) • "Increased opportunities for women have significantly improved the quality of life in the United States" (mean=2.03) • "Women who complain about harassment often cause more problems than they solve" (reversed, mean=2.70) Responses for each range from 1=strongly disagree to 5=strongly disagree. While these measures do not directly measure adherence to masculine norms, sexism is a central component of the masculinity belief system and is highly correlated with conforming to masculine norms (Smiler 2006). Questions were added together into a sexism scale (α=0.56) that was re-scaled to range between 0 and 1 (mean=0.32, sd=0.18).

B.3 DV: Behaviors, Attitudes, and Sickness
The dependent variables are a series of items measuring: (1) concern about coronavirus; (2) self-reported precautionary behaviors; (3) attitudes toward pandemic-related state and local policies; and (4) whether the respondent had or has contracted COVID-19. Full question wording below:

Concern
• "How concerned are you about coronavirus here in the United States?" (1=Very concerned; 2=Somewhat concerned; 3=Not very concerned; 4=Not at all concerned; mean=1.57)

Precautionary Behavior
Have you done any of the following things in response to the spread of coronavirus?

Sick with COVID-19
Have any of the following people been sick with coronavirus?

C Regression Tables
Ordered probit regression coefficients. Standard errors in parentheses. Two-tailed t-tests. All models use survey weights.  Logistic regression coefficients. Standard errors in parentheses. Two-tailed t-tests. All models use survey weights.  Ordered probit regression coefficients. Standard errors in parentheses. Two-tailed t-tests. All models use survey weights. Logistic regression coefficients. Standard errors in parentheses. Two-tailed t-tests. All models use survey weights. Regression type noted above columns. Standard errors in parentheses. Two-tailed t-tests. All models use survey weights.  550.023 Akaike Inf. Crit. 25,976.810 12,293.090 29,283.090 17,140.040 Note: * p<0.1; * * p<0.05; * * * p<0.01 Regression type noted above columns. Standard errors in parentheses. Two-tailed t-tests. All models use survey weights. Regression type noted above columns. Standard errors in parentheses. Two-tailed t-tests. All models use survey weights. Regression type noted above columns. Standard errors in parentheses. Two-tailed t-tests. All models use survey weights.