Exposure to Common Geographic COVID-19 Prevalence Maps and Public Knowledge, Risk Perceptions, and Behavioral Intentions

This survey study examines knowledge, risk perceptions, and behavioral intentions among survey respondents exposed to different types of COVID-19 prevalence maps.


Introduction
Several organizations have produced maps showing the prevalence of confirmed coronavirus disease 2019 (COVID-19) cases across the United States, but there is limited data on what map features are most effective at informing the public about infectious disease risk and motivating engagement with recommended health behaviors. 1 We assessed the association of 6 different COVID-19 maps with knowledge, risk perceptions, and behavioral intentions.

Methods
This survey study included US adults recruited between May 18 and 28, 2020, by Qualtrics Online Panels. This study was deemed exempt by the University of Iowa institutional review board, given the minimal risk to participants and collection of deidentified information. All respondents provided informed consent and were compensated for their participation. The survey was conducted online in English. This study follows the American Association for Public Opinion Research (AAPOR) reporting guideline.
After providing informed consent, respondents were randomized to see 1 of 6 maps ( Figure) or Author affiliations and article information are listed at the end of this article. Maps were collected on May 11, 2020. The New York Times map (D) is not included owing to licensing restrictions. The map used in the study was a heat map with a color scheme from light orange for areas with fewer cases moving to red for areas with more cases and gray representing areas with no cases reported. The map showed cases per capita (total reported cases per 100 000 people) at county level and is available on their website (https://www.nytimes.com/interactive/2020/us/coronavirus-us-cases.html). All other maps to not receive any information (no map) using an automated function within the Qualtrics software.

US Centers for Disease Control and Prevention
Respondents answered questions assessing their knowledge of confirmed cases of COVID-19 across the US (total cases and cases per capita), their perceived risk of COVID-19 (individual and societal), and their intentions to adhere to infection control guidelines. 2 Total cases and cases per capita knowledge were each assessed on scales of 4 items specifically about the total or per capita confirmed cases. Scores ranged from 0 to 1, with higher scores indicating greater knowledge about total or per capita numbers of confirmed COVID-19 cases. Individual risk perception was assessed on a scale of 9 items about perceived susceptibility and severity of getting COVID-19. Scores ranged from 1 to 7, with higher scores indicating greater perceived susceptibility and severity of getting COVID-19. Societal risk perception was assessed on a single item about whether the pandemic would be better or worse in 2 weeks. Scores ranged from 1, (indicating that the COVID-19 pandemic would be much worse in 2 weeks) to 7 (indicating the COVID-19 pandemic would be much better in 2 weeks). Intentions to adhere to COVID-19 guidelines were assessed on a scale of 15 guidelines (eg, "avoid gatherings of >10 people"). Scores ranged from 0 to 100, with higher scores indicating greater intent to adhere to the guidelines. Maps were available alongside questions for reference. Using planned contrasts, we compared these outcomes at 4 levels: map intervention (no map vs maps), visualization type (heat vs bubble), geographic level (state vs county), and case format (total vs per capita). Respondents self-reported demographic information, including age, gender, and race/ ethnicity.
All tests were 2-sided with P values adjusted using Holm-Bonferroni 3 correction for multiple comparisons. Significance was set at α = .05. Analyses were performed using R Studio statistical software version 1.1.463 (R Project for Statistical Computing).

Results
After excluding 2062 respondents who did not complete the survey, completed the survey in an unrealistically short time (ie, <9 minutes), or indicated that they did not provide high-quality answers (ie, respondents who answered "I will not provide my best answers" or "I can't promise either way" to the question "Do you commit to thoughtfully provide your best answers to each question in this survey?"), our final sample included 2676 respondents (completion rate, 57%).

Discussion
The findings of this survey study suggest that simply providing maps with COVID-19 case information was not necessarily associated with improved public knowledge, risk perception, or reported intent to adhere to health guidelines.
Limitations of this study include reliance on self-report and potential limited participation from individuals without internet access and lower English proficiency.
Based on the findings of our survey study, we encourage map developers to be mindful of the potential influence of reporting strategies on public knowledge and perception of the pandemic. We suggest developers present cases per capita using state-level heat maps rather than county-level