Examination of proxy variables for approximating COVID-19 community risk

• In Event: Poster Session 05
  In Poster Session: PS 05 Topic - Solicited Content: COVID-19 Related

Abstract

As researchers race to examine how the COVID-19 pandemic has influenced children and their parents, best practices in modeling human development should be implemented to ensure validity of results. One such consideration is how to include the severity of COVID-risk that participants are facing within their community, as this varies across both geographic locations and time. Specifically, a quantification of risk will be examined as it relates to mothers’ COVID related stress from broad to narrow geographic locations (state vs. zipcode), along continuous and categorical measures (number of cases/deaths vs. low, medium, or high risk or level shut down), and based on general vs. specific dates related to the participants (date of interview vs. median date of interview window for entire wave).

Data will be drawn from the first two waves of a longitudinal study examining mothers’ stress and how this impacts their parenting practices and child behavior. English-speaking mothers (n = 298) with children under 3 years-of-age were recruited from Prolific and Facebook ads from 40 different states. Across the first wave of data these states varied in their level of shut-down, level of risk, number of cases experienced, and number of deaths recorded and, therefore, it is conceivable that mothers’ stress varied in part by how much imminent risk was perceived. Data was collected between April 10th and May 23rd (wave 1; n = 311) and August 8th and September 23rd (wave 2; n = 258). At both waves, mothers were asked to identify the state in which they lived; at wave 2, mothers were also asked to report on their zip code.

Wave 1 data is available for analysis and wave 2 data is in the process of being cleaned. Using participants’ reports of states and zipcodes, the validity of different proxy variables for COVID community risk will be correlated with a self-report measure of COVID stress that includes five subscales (stress related to relationships, ambiguity of situation, food insecurity, finances, and discrete stressors). Proxy variables include number of COVID cases and death by state and zip code at median date of data collection and exact date of data collection (eight distinct proxy variables) and classification of community restrictions in place by the state and local government from no restrictions to mandatory shut-down (5 point scale; 2 distinct proxy variables).

It is expected that different types of stressors will be more strongly correlated with more nuanced measures of how an individual is experiencing COVID risk within their community, but the analysis is exploratory in nature, as it is possible that there is a point of diminishing returns where the extra effort it takes to develop the proxy variables that are more tailored to participants may not be more highly correlated to COVID-stress. Researchers examining the impact of COVID on children and families need to include COVID community risk into their models as either covariates, predictors, or moderating variables to appropriately model risk and protective factors related to child development and parenting during the pandemic.

Authors

• Jody S. Nicholson, University of North Florida
  Presenting Author

• Sarah Cline, University of North Florida
  Non-Presenting Author

• Monique Villamor, University of North Florida
  Non-Presenting Author

• Shuqi Zhang, University of Texas at Austin
  Non-Presenting Author