Of all Americans who completed the survey, 7% had been tested for COVID-19 by June 2020.

Margin of error = ±1.2%

Women, college graduates, Democrats more likely to self-isolate to reduce coronavirus risks

Women, older Americans, Democrats and people with more education are more likely to try to isolate themselves from contact with other people to reduce COVID-19 transmission risks, according to a new Tufts University national survey.

The survey also identified notable differences in whether people have received testing for COVID-19 based on geographic regions, age group, educational level, political affiliation, income, race/ethnicity and gender. Those who have been tested for COVID-19 most commonly live in the Northeast, are affiliated with the Democratic party, and are African American, according to the research.

“The results of our survey indicate that there are significant demographic and geographic differences in how people respond to COVID-19 pandemic risks, and that these disparities in protective responses need to be taken into account by public health and public policy officials,” said Tom Stopka, an epidemiologist at Tufts School of Medicine, and a co-lead on the study. “As public health officials continue to increase access to testing across the U.S. in light of persistent surges in COVID-19 infections in many states, they need to consider how to increase testing in geographic hotspots and the highest-risk groups to better understand infection patterns and inform data-driven public health and clinical responses,” Stopka added.

The survey was designed and analyzed by Tufts University’s Research Group on Equity in Health, Wealth and Civic Engagement. The research group previously released data showing that only 57 percent of Americans plan to get vaccinated for COVID-19. The group will soon release additional research about the economic impacts of the pandemic.

Self-Isolation Rates Differ

Overall, 71% of adult Americans say they have tried to separate themselves from others to avoid COVID-19.

Women are about nine percentage points more likely to self-isolate (75% versus 66% for men). People with a bachelor’s degree are almost 20 percentage points more likely to self-isolate than those with a high school diploma (79% versus 59.3%, respectively). Democrats (77.5%) and Independents (73%) are about 10 points more likely to self-isolate than Republicans (66%). Almost eight out of ten Americans who are 60 or older report trying to self-isolate, while two-thirds of respondents in all other age categories report similar attempts.                        

The disparities among those who are more likely to self-isolate may reflect opportunities and perceived risks, according to the researchers. For instance, retired people and people who work at computers can isolate more easily than people who provide in-person services. Also, people in older age groups may be paying close attention to the elevated risks that they face based on national and international epidemiological data and news.

Below age 60, all age groups are self-isolating at the same rate. That may counter speculation that the young are flouting restrictions

Peter Levine, associate dean of academic affairs at Tufts’ Tisch College

Testing rates vary by geographic region

Overall, 7% of those polled say that they have personally been tested for COVID-19, and 17% report that someone in their family has been tested.

These testing rates vary by region. In the Northeast, 10% have been tested, and 21% have a family member who has been tested, compared to 5% and 14% in the Midwest, respectively, with the other regions in between. Democrats are significantly more likely to have been tested (9%) or to have a tested family member (22%) than Republicans (6% and 13.5%, respectively). African Americans are the most likely to be tested (10%) or to have had a test in the family (26%), compared to 5% and 14% for Whites, respectively. Increased testing may reflect the higher rate of infection among African Americans that has been widely reported.

The survey was fielded online by Ipsos using its KnowledgePanel. The sample was nationally representative, and the number of complete responses was 1,267.