Research Spotlight: New dataset explores how COVID-19 guidelines affected mental and physical wellbeing
The Library Communications Team periodically offers a Research Spotlight in Your Library. It’s a chance for researchers on campus to offer their knowledge and to introduce and expand on a project they have collaborated on with one of our library experts. The inaugural contribution highlights a collaboration with Data Services Librarian Megan O'Donnell and Assistant Professor Jacob D. Meyer.
When the pandemic hit in 2020, the ISU Wellbeing and Exercise (WellEx) laboratory research team realized that many things were going to be altered rather abruptly. In particular, as researchers who evaluate the links between behavior and mental health, we knew that any changes to the ways that people were able to move or exercise were likely to have associated effects on people’s wellbeing. The COVID-19 shelter in place, work from home, and other guidelines and restrictions offered a unique chance to study how societal behavior restrictions can affect a population’s mental health.
Therefore, we initiated an observational study of US adults designed to identify behavioral patterns and mental health changes early on in the pandemic (start of April, 2020) and track these patterns over the subsequent 14 months. To recruit for the study, we contacted all students, faculty, and staff at ISU, and reached out to professional organizations along with potential recruitment via snowball sampling methods (i.e., people could share the link with others) and a large source of participants came were ISU alumni who were contacted via working with the ISU Foundation. Altogether, over 3,000 adults completed the first survey within a week with each person who indicated they were interested in follow-ups receiving additional surveys regarding their current behaviors and mental health each week for 8 weeks and then each month for 12 additional months ending in June 2021 resulting in over 25,000 individual survey responses.
Sharing this data set publicly presented some challenges. The survey asked many demographic questions (age, race, ethnicity, gender, employment, state of residency, etc.) of the person answering the survey, which was important for understanding both health risks and the effects of COVID-19 restrictions. However, representation in the data set was uneven and the chance of re-identification for some combinations was high. To mitigate this risk some of the extensive demographic data was removed, combined, or recategorized. For example, because of how we recruited, the majority of participants reside in Iowa but there were had at least a handful of people from every state in the country and these people were at a higher risk of identification. We consulted with the library’s Research Data Services staff and the Office of Research Ethics to discuss how to mitigate risks like these while preserving the usefulness of the data. The end result is a data set with some additional deidentification-focused changes from the original, but with the changes and most of the original frequencies reported in the readme file accompanying the data.
Overall, much has been learned from this original dataset with more yet to be uncovered. The publications so far have shown multiple relationships between mental health and behavior over this period, including: that reductions in physical activity or increases in screen time were associated with worse mental health at the beginning of the pandemic, that those who worked from home were more sedentary and had more screen time than when working in-person, that maintaining or increasing outside time was important for lowering stress and positive mental wellbeing, that high sitting time across the early wave of the pandemic was associated with lasting depressive symptoms, etc. We are excited to have this data now publicly available and hope that other researchers will ask new and different questions from this longitudinal dataset and that others will be able to use it in additional ways (e.g., as an example dataset in university courses).
Access the COVID-19 and Wellbeing Survey data.
For questions about research data sharing contact email@example.com.