A Racist Disease (?) and Games with Models, 5/7/20 Data


We live in a world where some would insist that it is racist to say that it’s wrong to inject race into every public discussion, but doing so very often obscures more than it reveals.  As with claims about gender differences in pay, observed discrepancies often have more to do with things that are only incidental to racial groups.

Consider talk about racial disparities in COVID-19 infections.  Among the state Department of Health’s regular charts reporting daily results is one divided by race, and many have commented on the fact that a full 44% of all cases have been found among people identifying as Hispanic.  Even though they only represent around 13% of the population, Hispanics account for 44% of all positive cases of COVID-19.  Central Falls, which is around 66% Hispanic, leads the state in the number of cases of COVID-19 per capita, at 2,595 per 100,000 residents.

But focusing on race doesn’t tell us much about the disease, or even about discrimination.  Mostly it seems to be evidence that communicable diseases are especially dangerous to places that are densely populated.  In those terms, Central Falls is by far Rhode Island’s most densely populated city, at 16,147 people per square mile.  (The city is pretty much just a square mile.)

So, if we recalculated to rank our cities and towns by cases per person per square mile, Central Falls drops to 32nd in the state.  Providence becomes number 1 in the state by that method, but it is followed by Burrillville, North Kingstown, Smithfield, and Exeter.  That isn’t exactly a racial profile.  And it isn’t a close thing.  Burrillville has 0.25 cases per person per square mile to Central Falls’s 0.03.*  Given its density, Central Falls is doing very well.

“But,” the progressives will say, “the racist part is that we’ve isolated this one group in a densely populated area.”  Leave aside the fact that progressives have tended to want people to condense into urban areas, as well as questions of whether it is reasonable to blame the decisions of those communities on people who make different residential decisions and whether any of this is a function of race so much as the length of time immigrant groups have happened to be establishing their families in the area.

The key point, here, is that the most important bit of information we can know at the moment is how the disease is spreading and who is vulnerable.  Introducing race where population density provides explanation is a distraction, and distractions while dealing with a crisis can be deadly.

For further illustration of the point, return to the Dept. of Health’s race-based chart, and instead of looking at cases, click “hospitalizations.”  Whereas Hispanics account for 44% of all cases, they only account for 33% of all hospitalizations.  Even more striking: their percentage of fatalities is exactly in line with their percentage of the population: 14%.

In contrast, Non-Hispanic whites account for 79% of all deaths, despite being only 74% of the population.  Of course, this is because the disease is almost exclusively fatal to older people, and they are more likely to be white in Rhode Island, right now.  Would anybody claim that racism is to blame for this outcome, by concentrating our white people into nursing homes?

Turning to the regular COVID-19 report, trends are more or less as they had been.  Although the one-day increase in cases was significantly higher than it’s been in a number of days, the trend is still downward, as it is for hospitalizations and intensive care unit usage.  Meanwhile, the number of deaths is not improving as quickly as anybody would like.



  • Cases:
    • Projection for 5/7: 10,371
    • Actual for 5/7: 10,530
    • Projection for 5/8: 10,694
  • Hospitalizations:
    • Projection for 5/7: 307
    • Actual for 5/7: 318
    • Projection for 5/8: 300
  • Deaths:
    • Projection for 5/7: 376
    • Actual for 5/7: 388
    • Projection for 5/8: 394


* In the spreadsheet built for comparisons of density, I had a column multiplying my cases per person per population calculation by 100 to get it out of decimals as I reviewed the data.  In transcribing that number to text for this post, I didn’t account for this change.

  • Mario

    I have hospitalizations falling to 287 tomorrow, with 19 admissions and 42 discharges. Also another 13 or 14 deaths, as usual. I’m still trying to figure out how to estimate the number of new cases in a time of rising, but inconsistent, testing. I don’t even think I can assume a 9.5% positive rate, depending on if those Stop & Shop tests start getting tossed in.

    Anyway, I found a new way to try to estimate, so let’s see how it goes. 8.1% positive rate, which would be 250 cases, assuming we come just short of 3100 tests.

    As far as racism goes, I don’t think there is enough information available quite yet. As you showed, the disparity in hospitalizations and deaths is obviously a function of age and the racial aspect is just obscuring that. They’ll need to look to see if, after accounting for age, minority groups suffer worse from the disease, but I haven’t seen anything except speculation to suggest that that’s the case. There is good reason to think that their health is poorer, but that is an answerable question, so we should wait until it is answered. There is an open question of how exactly it is spreading, and I’m not sure that density gets to the heart of the problem. I’d bet public transportation is related, choice of occupation, shared laundry equipment, and maybe the air in apartment buildings is just not contained well from one dwelling to another. Racism aside, if problems are found in those areas, they’ll need to be addressed, because I think it’s clear that the oversaturation of the virus in areas like Central Falls is a primary contagion engine for the rest of the state.

  • ShannonEntropy

    The real reason the infection rate is so much higher in the Hispanic population is their much much higher rates of obesity, hypertension, diabetes & cardiovascular disease— and if you’ve got the first of those, yer almost certain to have the other three

    Obesity in particular is tied to higher infection rates and mortality:

    What percentage of that population is obese ?? Take a drive thru CF or SoPro some time; virtually everyone you’ll see is overweight and more than half are obese or morbidly obese. QED

  • Lou


    I suggest the personal attacks on Mr. Howard begin post haste. This attack on the KatzMath is totally uncalled for.

    • Justin Katz

      That’s silly. He made an argument. Of course, he did load it down with personal attacks, but if somebody makes an actual argument, I usually address that.

      • Lou

        So the “Freedom Covid Index” moves Providence (with over 36% of the state’s cases) from 5th to 1st and the conclusion we are to draw from that is…?

        • Justin Katz

          You’re mischaracterizing what I’m doing. I wasn’t creating an index. I was pointing out that density is more relevant than race and, in fact, a simple comparison of cases by density moves the state’s most-Hispanic city down the list, rather than securing its position at the top (which it is per capita). If race were a significant factor, one would think it would increase Central Falls’s standing if we adjust for population density.

          • Lou

            Again with the straw-man. Where is it suggested or implied that race is more relevant then density? It’s certainly not on the link to DOH you have provided.

          • Justin Katz

            It’s fair to think that DOH holds race to be more relevant than density because the agency displays a prominent chart showing the former but seeks no way to illustrate the latter. Dig a little, and they present cases per capita, but that’s not the same as population density, which is nowhere presented.

          • Lou

            You’re free to create your own narrative, but the fact is race is referenced after general information, geography, zip code, density, cities and towns, daily counts, age and sex.


            If that’s “prominent” to you, I suggest you are being disingenuous.

          • Justin Katz

            Everything on that page is more prominent than things that are only reported in the spreadsheet at the link at the bottom and certainly more prominent than a datapoint that isn’t mentioned at all.