COVID Precautions Broadly Used, Despite Rising Cases

COVID Precautions Broadly Used, Despite Rising Cases

By Tim Ludolph and Theo Reuter | Data Product Manager and Geospatial Data Scientist

As the country continues its struggle with the coronavirus, recently passing the grim milestones of 7M cases nationwide and over 200,000 deaths, two preventive measures loom in importance as we near the annual flu seasonface masks and social distancing. The CDC, WHO and other major health organizations are now unequivocal in their guidance that employing these two measures is critical to stemming the disease’s spread. 

Premise has examined the use of face masks and distancing in this country several times the past six months, including most recently in early September. In light of the recent milestones and the President and several White House officials testing positive for the virus after attending public events with no masks, we decided to examine the issue again to see how we are doing as a country.

Mask Use Remains Steady

The good news is that since our last look, it appears the number of contributors reporting they “always” wear masks when they leave the house has remained steady, hovering around 50% nationally. Totals may have risen or fallen slightly week to week, but the overall trajectory for those responding “always” has maintained or surpassed the 50% mark the past 11 weeks, as shown in the below graphic.

Weekly breakdown of responses to the question of wearing a mask when leaving the house

Additionally, the number of states where over 50% of contributors now say they “always” wear masks has risen to 24. When we looked at this issue back in July there were 27 states where 80% of respondents were wearing masks “sometimes” or “always.” There are now 42 passing that threshold now, which represents a significant jump in mask use nationwide. 

States with more than 50% of respondents “always” wearing masks (Sept.)

Looking deeper into these states, we find California has the most unique respondents “always” wearing masks, as well as the most submissions where this response is provided—by a wide margin, as seen in the below chart. Florida, Georgia, Illinois, New York and Texas form the next highest cluster, balanced relatively evenly between the number of contributors and submissions. 

Each of these states has spent the majority of its time with over 50% of their contributors saying they were “always” wearing masks, as represented by their bubble size. Five of the six cleared that threshold at least 11 of the last 22 weeks, with Texas representing the low end of that spectrum (11) and New York the high (22). 

Interestingly, though, Georgia—which has a comparable number of contributors and submissions as those states—has surpassed that mark far less frequently (as shown by its much smaller bubble). This means the majority of its contributors are still either “sometimes” or “never” wearing a mask, raising the question of which is the most important population to target for changing their behavior, which we will explore later in this post.

Breakdown of contributors by state who “always” wear a mask (May to Sept.)

Several of the remaining “over 50%” states—Colorado, Nevada, Virginia and Washington, for example—have fewer contributors and submissions where “always” was selected, relative to these other states. However the amount of time where that response was over 50% of their total is much higher—each of these states (and five others) have spent the same number of weeks at or above that threshold as Texas. (Hence the high number of large bubbles in the lower left.) 

This indicates a potential higher receptivity to “always” masking in these areas, despite the smaller sample size. If authorities can capitalize on this willingness and expand the overall numbers, these states may be better positioned to guard against the spread of COVID-19.

Looking at the flip side, the number of contributors who say they “never” wear masks has continued to drop the past 11 weeks, remaining below 15% and falling to its lowest level yet last week (11%). Unfortunately, among those not wearing masks, the second most frequent justification (28%) has consistently been they “do not believe face masks are effective for preventing the spread of COVID-19.” The top reason was “masks are uncomfortable” (41%). 

This is a critical data point that invites public officials and health groups to continue their efforts combating popular misperceptions and potentially lingering confusion from the beginning of the pandemic when there was a lack of clear guidance regarding masks’ effectiveness. 

Diving more into this population, California again had the most contributors and submissions saying they “never” wear masks, but it also had the lowest number of weeks where the percent of contributors saying so surpassed the national red line of 15% discussed earlier. (As reflected by their extremely small bubble.) Florida, Georgia, Illinois, New York and Texas again form the next highest cluster, relatively balanced between the number of contributors and submissions again, however, their bubble sizes highlight some differences from the above graphic for “always” wearing masks.

Breakdown of contributors by state who “never” wear a mask (May to Sept.)

Here we see Georgia’s bubble is far larger than before, which is a result of their spending 20 of the past 22 weeks above that 15% threshold. In contrast, Illinois, New York, and Texas have much smaller bubbles, having spent less than 9 weeks above that mark, despite comparable numbers of contributors and submissions. 

This means that while their populations of contributors “never” wearing masks is roughly the same, the amount of time they’ve been reporting doing so is higher in Florida and Georgia, suggesting they are more entrenched and may be harder to convince to change their behavior.

Social Distancing Also Strong 

Shifting to the other major precaution, social distancing, we find results similar to the mask usage dataset. The number who “always” or “sometimes” socially distance has risen since we last wrote, maintaining a level of over 80% the past 22 weeks. (If you add in those who “somewhat” distance it’s been over 90% for that same stretch!) In contrast, those who “always” or “sometimes” wear masks have only been over 80% in the past 15 weeks.

Weekly breakdown of responses to the question of social distancing when leaving the house

Of the two recommendations, distancing has remained the clear favorite and the more readily adopted precaution since we began collecting data on contributors’ COVID-related perspectives in April. When you look at how different age groups approach the issues, there is some minor disparity. Both 16-25 and 26-35 year olds “always” distance more than they “always” mask, while those 36-45 and over 45 prefer the reverse, “always” wearing masks more than they “always” distance.

These responses raise the question of how we most effectively get out of this pandemic? With more than 80% of contributors saying that they “always” or “sometimes” employ both face masks and distancing—and doing so for months—you might expect that to be enough to slow the virus’ spread. And yet, as recently as this week, 21 states again have rising COVID case counts

So how do we break out of this cycle—should authorities spend more time trying to convince individuals who “sometimes” wear masks and/or distance to start “always” doing so, or is the better expenditure of energy getting those who “never” do to start doing so “sometimes?”  To delve into this, we decided to map the states and counties that flipped since we last wrote—either moving from a majority of contributors “sometimes” wearing masks/distancing to the majority “always” doing so, or shifting from “never” responses outnumbering “sometimes” to the reverse. 

We’re essentially looking for one of two things—is it more important to move those in the middle up, to become the most cautious group possible “always” wearing masks and/or distancing? Or is it more critical to bring the bottom up, converting those who “never” wear masks/distance into people who “sometimes” do? By overlaying these changes with ongoing case count data, we can begin to see if there is any correlation and potentially determine which state change, if either, is most crucial. Here are our initial results:

Number of states changing status from for the question of wearing face masks (July to Sept.)

Looking at the state level we can see 10 that changed positively from July to September. Six states leaped from the middle up (i.e., with “sometimes” falling behind “always” as the most popular response, shown in bright green). At the same time, four represented bottom-up improvements (i.e., “never” becoming less popular than “sometimes,” shown in dark green). All but two of these states currently have steady or falling case counts (Idaho and Montana have rising counts as of last week). 

Unfortunately, three states showed trends in the negative direction in that span (i.e., the number of “always” or “sometimes” responses shrank while “never” grew, colored in red). Of these, only Rhode Island had a rising case count as of last week. 

These results look positive on the surface but don’t show enough granularity to tell what’s happening on the ground and how widespread these changes actually are. To do that we need to look at the county level, similar to previous posts.

As you can see in the graphic below, the number of counties that changed status from July to September is relatively low nationwide—the overwhelming majority of counties’ responses stayed the same in those three months. For those that did change, the number of counties going from the middle up (bright green) is unfortunately almost as numerous as those relapsing (red). Far less prevalent are those going from the bottom up, with only a couple scattered across the country (dark green).

Counties changing status for the question of wearing face masks (July to Sept.)

Comparing these counties against the Johns Hopkins case count data shows an unclear correlation thus far (dark purple indicates high case numbers, while lighter/white colors indicate low counts). Some of the green counties do appear to coincide with improving case counts from July to September (those in Washington and Florida, for example), while others do not (those in California). The counties backsliding on these behaviors are equally inconclusive, which could mean these state changes occurred too recently to see a correlating spike in case counts, or there are other variables at play more influential to that ultimate count.

The same holds for the comparison of counties flipping on the question of social distancing—an almost equal number of middle up counties and relapsers, with far fewer bottom-up counties nationwide. Like the masking question, some correlate with decreases in case counts (Arizona and Washington, for example), and some do not (those in Utah). 

Counties changing status for the question of social distancing (July to Sept.)

At this point it’s too soon to derive much from the data—there are simply too few counties changing state for us to make any meaningful assessments. However, Premise will continue to research this behavior in the coming months and attempt to determine what is the most effective strategy—bringing the bulging middle up, or dragging the smaller numbers of “never” at the bottom toward the middle? Keep checking our website for ongoing updates and insights into the virus’s impact or contact us at

About Tim Ludolph and Theo Reuter

Tim Ludolph is Premise’s data product manager, ensuring our collection, analysis and visualization efforts provide insights from our global contributor network that have maximum impact on our customers’ daily operations. He has over fourteen years experience working with numerous U.S. government agencies and is a master’s graduate of American University’s School of International Service.

Theo Reuter is a Geospatial Data Scientist who is passionate about data and how it can be used to tell stories. Recently graduated from the University of Maryland College Park with an MS in GIS, he has been hard at work at Premise since January 2020 providing mapping, scripting and spatial data analysis, leveraging the spatial nature of the data Premise collects.