[Co-author of this article is Eric Tyberg, a retired IT executive and consultant residing in Lincoln, California. Originally from Falun, Wisconsin, he rose through the ranks at IBM and formed his own consulting business when IBM downsized.]
Per the Johns Hopkins University Coronavirus resource site, on April 1st Skagit County reported a cumulative total of 128 COVID-19 cases and Whatcom County reported 139. Between April 1st and June 1st, the cumulative number of total cases in Skagit County increased 304 percent ((389/128)*100) while in Whatcom County it increased 286 percent ((398/139)*100). Whatcom County has about 1.74 times more people (225,300) than Skagit (129,200), so it is not surprising that it has more cases, but as of April 1st, the number of cumulative total cases per 100,000 people in Skagit was 99.1, while in Whatcom it was considerably lower, 61.7. By June 1st, Skagit’s rate was 301.8 while Whatcom’s was only 176.7.
Testing rates, compromised health factors, the number of assisted care centers, and the presence of food processing plants affect the number of COVID-19 cases reported, but given that Skagit and Whatcom are roughly similar in these terms, what accounts for the higher rate in Skagit County? Four demographic factors worth examining as plausible ways to answer this question are: population density, age, race/ethnicity, and gender. If everything else is equal, the following assumptions should be true: (1) an area with high population density will have more cases than one with lower population density; (2) an “older” population will have more cases than a “younger” population; (3) a population with a higher proportion of minorities will have more cases than one with a lower proportion of minorities; and (4) a population with more men will have a higher rate than a population with fewer men.
Starting with population density, there are approximately 67 people per square mile in Skagit County and 90 per square mile in Whatcom. This factor suggests that we should see a higher rate of cases in Whatcom. In Skagit County, 2.47 percent of the population is aged 85 and over, while in Whatcom it is 2.11 percent. This indicates that we should see a higher rate in Skagit County, as we do. In terms of race/ethnicity, 72.62 percent of the population of Skagit County is White/non-Hispanic, while in Whatcom it is 78.69 percent. This difference also indicates that we should see a higher rate in Skagit County. In terms of gender, the proportions are virtually the same; in Skagit County 49.67 percent of the population is male, while in Whatcom it is 49.76 percent.
So, in terms of these four demographic factors there are two, age and race/ethnicity, that are consistent with the higher case rate in Skagit County, one (density) is not, and the counties are virtually even on the fourth, gender. Looking more closely at the age and race/ethnicity factors, we see that Skagit is not substantially different from Whatcom: less than a percent difference in the age index, and only about a six percent difference in terms of race/ethnicity. These do not seem sufficient to be driving the difference in rates we see between the two counties as of April 1st, and the much larger difference as of June 1st, with Skagit going from about 1.6 times more case rates as of April 1st to 1.71 by June 1st.
One additional difference between the two counties is the presence of Western Washington University in Whatcom County, which closed for spring quarter causing many students to return home. However, even if 10,000 students left Whatcom County, the remaining population is about 215,000, which yields case rates that are still lower in Whatcom on both dates, 64.7 and 185.7, respectively.
Could the difference be related to politics? Research suggests “…that Republicans are about three times more likely than Democrats to say their state is moving too slowly to reopen business and ease restrictions and Democrats are more likely than Republicans to report taking preventative measures like wearing masks in public.”
Given the assumptions and facts so far, it may be the case that as a Republican-leaning county, Skagit has views about containment measures that are more like the views in Franklin and Benton counties and less like the views generally held in more strongly Democrat-leaning Whatcom County. Unlike the Franklin County Commission and sheriff, both of whom attempted to defy containment orders issued by Governor Inslee, Skagit County officials did not attempt to defy the governor, but like Benton and Franklin counties, it has more of a Republican political orientation than does Whatcom, per the table to the left.
Although Hillary Clinton carried Skagit County in 2016, only 48 percent of its votes were for her, while in Whatcom County, 55 percent of its votes were for Clinton. Could the politicization of the pandemic be playing a role here? If so, “re-opening” areas may be putting Republican-leaning counties at greater risk than Democrat-leaning counties.
As an initial means of answering this question, we assembled 2016 presidential voting data by county along with the increase in cases since April 1st to three dates: May 1st, June 1st, and June 10th. We then summarized the results for Washington’s 39 counties as shown in the table to the right. As you can see, the increase in cases since April 1st is not only far higher for the 27 counties that voted for Trump (rose-colored background), but the gap is widening. Between April 1st and May 1st, the overall increase in the 27 Trump-voting counties was 431%, by June 1st it was 987%, and as of June 10th it was 1,228%; For the 12 Clinton-voting counties (blue-colored background), the respective increases are far lower, 209%, 270%, and 282%. Consistent with the research cited earlier these findings suggest: (1) that Republican-leaning counties in Washington were likely less observant about social distancing and other containment guidelines than Democrat-leaning counties; and (2) that these politically-related behaviors are becoming exacerbated as the re-opening process unfolds in Washington.
These results are sufficiently suggestive that we plan to do a more in-depth analysis of all 39 counties; similar to the analysis we have done here for Skagit and Whatcom counties, especially in terms of the increase in cases from April 1st as we move into the upcoming phases of re-opening.
Note:
Data Sources.
The COVID-19 data are from The Coronavirus Resource Center maintained by Johns Hopkins University.
The demographic data are for 2019 and taken from Small Area Demographic Estimates (SADE) by Age, Sex, Race and Hispanic Origin.
The land area data are taken from Wikipedia: Skagit and Whatcom.
Voting data are taken from Wikipedia.
Comments by Readers
Michael Riordan
Jun 16, 2020You’re missing one major, singular, superspreader event, David: the March 10 Skagit Valley Chorale rehearsal, in which 52 out of 60 attendees got sick with flu-like symptoms. Most of those were diagnosed as due to Covid-19, while the remainder were assumed to have been infected. That initial major spike in cases then drove the per-capita Skagit County infections much higher on April 1 than Whatcom County. From news reports, there were also a few singers present from Island and Whatcom Counties, but the initial brunt was felt in Skagit. And fortunately for San Juan County, there were no singers from here present due to ferry schedule limits. Otherwise we’d probably have had a lot more cases than the 16 so far reported.
David A. Swanson
Jun 16, 2020Ineresting point. Do you believe that the consequences of this event would be driving the higher trend of case increase seen in Skagit through June 10th, three months later?
Michael Riordan
Jun 16, 2020Yes, indirectly. The initial wave ignited on March 10 should still have been propagating into April, but it would take serious computer analysis to determine whether that would continue into June. And it would depend on the relative effectiveness of public-health measures in the two counties. Beyond my pay grade.
Thomas R. Scott
Jun 16, 2020Some review:
A) Whatcom, while normally a border county, is currently a cul-de-sac given that the border is mostly closed. Conversely, Skagit is a nexus for Whatcom, Whidbey Island and all points south. And as of May 12, Skagit opened to eastern Washington via Highway 20. All but the southern areas (and eastern Washington) are cul-de-sacs.
B) Looking at a night satellite photo (c) 2012, one might extrapolate population densities based on the density of light emanating from the ground. Doing so shows the light density to be brightest in Bellingham but overall more diffuse across western Whatcom from Sumas and Lynden through less dense farmland to Bellingham and Ferndale. Skagit densities seem more compact with Burlington/Mt Vernon and Sedro Wooley closely nearby. The one outlier is Anacortes. However, the Anacortes population commonly travels to Burlington/Mt Vernon to purchase hard goods and many other needs. Another outlier is Oak Harbor of Island County. Oak Harbor also “shops” in Skagit for many needs.
C) Looking at topo maps or satellite photos during the day indicates that the Cascade Range “turns” a little upon entering Whatcom allowing more level land available compared to the more compressed Skagit level landmass. Further, more of Skagit’s flat land is within flood zones further compressing population density.
Some extrapolation and conjecture:
A) The point here is that Whatcom is more self-contained with virtually no through traffic, and thus less exposure to the “outside”. Conversely, Skagit is a hub for land travel. While Anacortes (Fidalgo Island) has a land bridge then a ferry south, few use that route when heading south (or just about anywhere else except for leisure travel). Arguably, to a slightly lesser degree Oak Harbor and vicinity on Whidbey Island tend to use the bridges to Fidalgo then east to the mainland and Burlington/Mt Vernon when travelling.
B) The point, is that the population appears more evenly spread and over a larger area in Whatcom than in Skagit.
C) Indicates more usable land for Whatcom than for Skagit. Note: That a portion of both counties are in communities outside the main population areas. The question is how much that may even out or “unskew” the average density numbers. Much of both counties is not used for living or conducting business. Measuring population density by simply taking the entire land mass respectively of the two counties without taking into account how much of that land is usable for living and/or conducting business is flawed. Given the geography of western Washington, it is very flawed.
When taken together, one may extrapolate much more exposure for the more constricted Skagit population in a broader communication with other communities around them.
Of course, much of this is conjecture as is the article. However, it stands to point out there are many other factors than simply the average population spread over mostly unused land masses and socio-political attitudes.
David A. Swanson
Jun 16, 2020Lest we forget, in addition to Highway 20 and I-5, there is that hotbed of traffic, Highway 2, likely jammed from Lake McMurray to Wickersham with all of those Snohomish and King county residents anxious to get to Skagit County. And of course, how could we forget the Mt. Vernon Amtrak and bus stations, which put their counterparts in Whatcom County to shame with all of their traffic. As for the Alaska Marine Highway Terminal in Fairhaven, it’s just a bagatelle in the limited transportation system in the Whatcom County cul-de-sac.
I wonder if the folks in Concrete, Marblemount, and Rockport turn out their lights earlier than they do in Anacortes, Burlington, La Conner, and Mt. Vernon. I am sure that the folks in Acme, Glacier, Nooksack and Sumas keep them on longer than they do in Bellingham. Of course, we could speculate that it was the folks in Newhalem that went over the Whatcom County line into Skagit County and turned out the lights in Concrete, Marblemount, and Rockport. And as long as we are speculating, it may be the case that some of the folks in Stehekin hiked over Cascade Pass to help the Newhalem folks with darkening Concrete, Marblemount, and Rockport. That Chelan County lot is a crafty bunch, especially the ones in the Stehekin Valley.
When I drove down to the Longhorn in Edison last week for its great chicken wings, first swinging by the Skagit County Airport to check on the Chuckanut South, one of my favorite brew pubs, I could see what you are talking about in regard to highway 20. It was just jammed with traffic. - all those folks from Oak Harbor coming up to shop, going against all of the folks from Snohomish and King counties who turned off the 2 to get on the 20 to Anacortes and LaConner. Since the 20 was so crowded (And as far as I could see when I got to its interection with Chuckanut Drive, so was the Bow Hill Road heading back to I-5). After seeing the traffic there, I turned north and came back to Bellingham on Chuckanut with no traffic in my direction, but the traffic running south to the Oyster Bar, Chuckanut Manor, the Longhorn, and points beyond was unbelieveable.
I sure enjoyed those chicken wings along with a couple of bottles of Rainer, but I have to learn how to eat and drink while wearing a mask. My t-shirt was covered with medium strength thai sauce, tartar sauce, grease, and Rainier stains. Good thing that the laundromats in the Whatcom County cul-de-sac weren’t overrun with Canadians.