The source provides no methodology.
tl;dr The claim that "if age distribution were equal" is not complete. The original source adjusts COVID mortality by age and an ill-defined "metabolic health". They provide no methodology for either. Without knowing the details they could have done any amount ...
I suspect, given Steven Moore's track record of op-ed deceit, they applied these highly-correlated corrections one-at-a-time, or some similar swindle. Note that Vermont, also an elderly state, had below average covid mortality by any metric.
I would interpret the Britannica claim as a soft vague statement essentially saying that the Florida death rate is similar to the national average and not some kind of extreme outlyer. How you decide that is up to the person making the claim.
Great answer. one small detail you call out Kerpen for being a founder of the Committee to Unleash Prosperity. The strong implication is that that committee is opposed to Covid preventative measures and thus the source may be biased. However It might be best to spell out exactly what the committee is for those that don't know so as to make it more clear why it is relevant.
...this does imply Wikipedia should really look at updating their own sources. Unfortunately I don't know anything about Wikipedia process and don't really feel that well suited for trying to correct it. Any wikipedia pros want to try to get them pointed to provable metrics?
@AndrewLazarus It's a tough comparison though because population density matters in a pandemic. You can't compare Vermont to, say, Miami, even on a per-capita basis, because the environment is completely different. Isolating in Vermont is easy. Isolating in many areas of Florida, not so much.
They literally explained metabolic health as "by which we mean the pre-pandemic prevalence of obesity and diabetes", which seems like a sound definition given the context. So I'm not sure how you can claim that they didn't explain that.
@csstudent1418 If that statement passes for methodology, then sure. But nobody who wants to be taken seriously in research is going to pretend that's anything but a vague general statement.
For the record, at time of writing, the national per-capita COVID death toll (from NYT) is 330 deaths per 100K. Florida is at 387/100K, ranking 14th out of 56 (50 states + D.C. and territories), where 26 of them exceed the national average. Florida may have been middle of the pack at the time Brittanica made that statement, but, like many states with a weaker mitigations and reduced compliance with mitigations and vaccinations, they've been dying more for the last couple years, catching up to/surpassing states hit hard early.
As far as dating Brittanica's claim, it specifies "more than 75,000 deaths" while the current death toll (from stats including 2022-12-15, again, NYT numbers) is 83,200. According to Worldometer, Florida crossed the 75K deaths mark on 2022-06-02 (75,016 deaths; U.S. had 1,036,634 on same date). Using 2022 populations (22.1M for Florida, 338.3M for U.S.), that's 339.4 deaths/100K for Florida at the time, to 306.4/100K for U.S. as a whole. They were ~11% higher than average when claim could be made; they're ~17% higher as of 15 Dec.
@Jontia I started writing about how the claim is suspiciously similar to the "sacrifice the elderly" proposals, and yes, how COVID policy should be adjusted to your demographics. That got opinionated and I decided to go with simply questioning the source's methodology.
@csstudent1418 Can we reproduce their results with that information?
Factoring out obesity, but no other health factors is indeed a huge red flag. Literally red. Of the 19 states with >35% obesity rates, every single one has a legislature controlled by Republicans (excepting NE, which is officially non-partisan). 61% of the rest do not. Adjusting for obesity alone among all health conditions is a great way to statistically downplay whatever may be going on in Republican-controlled states.
...similarly of the 9 states with obesity rates <30% (10 if you count DC), every single one has a legislature controlled by Democrats. There is such a strong correlation there that if you factor out one, you are essentially factoring out the other. If you then try to reason about something else that is also strongly correlated with having a Republican-controlled legislature (eg: laxity of COVID prevention measures), you're doing the statistical equivalent of dividing by 0.
but this all teeters on the verge of devaluing lives. => uh... what? It's perfectly normal to say that losing 100 lives in the younger demographics isn't particularly concerning compared to losing 10k lives in the oldest demographics.
We must address whether ignoring a state's demographics is a relevant statistic => of course it is! If you ignore demographics you could end up with the absurd conclusion that college towns performed better than any other city during the pandemic, ignoring their disproportionate number of young people and the lack of elderly care homes.
@BenHocking we have excess mortality data now for 2020 and 2021, which are the only objective metric at the end of the day.
@JonathanReez it's a good measure, but excess mortality is also subject to tons of, well, subjectivity. Both in how you obtain the estimate and what you attribute the difference to.
Also: "It's perfectly normal to say that losing 100 lives in the younger demographics isn't particularly concerning compared to losing 10k lives in the oldest demographics." Depends on the specifics. If it's 10k lives that would have otherwise, absent covid, only lived a tiny amount longer, then it's a complicated tradeoff. Most people agree that a 30 year old dying is a lot more tragic than a 90 year old dying quickly of covid instead of sitting in a hospital for a year and dying 'naturally'.
@SnakeDoc Without knowing how they arrived at the "adjusted COVID mortality rates", they might as well have been written on a napkin. The claim relies on those rates.
@JonathanReez Age is already accounted for by looking at morality rates by age.
@Schwern Not really, because none of the various ways they reached this conclusion seem totally preposterous. We're debating what the absolute best method of determining outcome is, while ignoring all sorts of indicators the outcome for Florida was clearly not worse than any average. The rest is just an academic challenge... but the facts remain - Florida, by all measures thus far, did not fare worse than the average, and may have actually fared better than the average state's outcome.
@SnakeDoc Yup, those numbers on the napkin could be right. They could also be wrong. We have no way to know. The only other source in the claim shows Florida doing worse in every age group except a bit better with 75+. If you want to check the claim with a verifiable source, write up an answer. Also, have a look at my update about the dangers of averages.
@Schwern The entire update is basically an excerpt from the book "How to Lie using Statistics". It's nothing new. However, you seem to be casually throwing away the real data which does not align with your conclusion that the claim was baseless and unverifiable. We can look at per-capita and averages and have a very good idea. Is it perfect? No - but we can finagle the statistics all day until it tells the story you want instead of reality.
@SnakeDoc 1) It's nothing new, but it's sadly still relevant 70 years later. 2) "you seem to be casually throwing away the real data" What is the real data? 3) "we can finagle the statistics all day until it tells the story you want instead of reality" That's my point. Especially when we can't see the finagling.
@Schwern I'm not sure why you believe there's no data. We have fatality stats from Florida, and their population is known - therefore we can extrapolate per capita rates rather easily. On that metric alone, Florida fared no worse than the average, if not better. Refuting the claim wholesale simply because there are various methods of calculating that fatality rate (age, demographics, prior health, economic status, etc) does seem to be throwing out the bathtub and the water. The claim is "Plausible" - to quote the Mythbusters, not "Busted".
@SnakeDoc If you want to try to reproduce the claim based on verifiable data, I encourage you to write up an answer! That's not the route I chose, but its good to have multiple approaches. Please go ahead and then it can be voted and commented on.
@SnakeDoc, florida fired the person compiling the statistics and replaced them with a Destantis crony which doesn't give any confidence they've reported anything correctly. The initial wave of infections in the US was not in Florida. By the time Florida was hit hard we had already learned a ton about how to treat it and prevent worse case outcomes. The whole idea that Florida fared better is a pointless comparison being made entirely for political reasons
@Jontia: adjusting the factors to account for your local conditions is one of the ways you determine what the LOCAL policies should be. Treating Appalachian coal miners the same way you do SF Bay Area techies isn’t going to do them much good and vice-versa.
The chart you posted is great, but I’m not sure it really makes your particular point. Can we really say that government policy is what saved 3 lives per capita between 15-24 and not simply randomness? This seems particularly likely since the CDC only reports 2875 deaths in that age group nationwide. I would be extremely hesitant to claim government policy or any other factor is definitely responsible for such a small difference between states. It’s also baffling that Florida would do better with the elderly than California if this was the explanation.
I do also want to reiterate what I posted in an early comment that I guess was deleted: government policy doesn’t tell us much about what people were actually doing. If more at-risk Floridians were staying home and masking regardless of what the governor said than Californians, it would hardly prove the policy worked.
@Shamshiel: Here on the west coast, we had people bragging about defying the mandates. Also, locally (and judging by TV, nationwide), the masks many were wearing ranged in efficiency from 7% to 60%. I even had hospital "greeters" make me remove my N95 and put on one of those useless things that leaks on all sides.
@Shamshiel Well said, we're talking about such small fractions here, it all seemed like a practice in delaying the inevitable. Meanwhile apparently caring about economics was considered devaluing lives. Funny enough here in Ontario the "top health official" admitted two years of over protecting children may have contributed to the recent rise in RSV/flu hospitalizations.
Your chart in fact supports the claim. Fewer people died in Florida, and surprisingly it was the most vulnerable age group that was spared. Big surprise to me.
By the way, the one measure that implicitly incorporates all the mortality factors like age distribution, life style differences and different health conditions is over- (or under-) mortality. Simply compare the actual mortality during Covid with the one to be expected in each month, and you get an idea. That still does not account for climate differences. A warm state where even in winter lots of activities are outdoors -- or can be moved outdoors -- will be less affected by Covid than one in a colder climate. I would suspect that Covid mortality shows a significant north south gradient.
Discussion on answer by Schwern: Did …
Discussion on answer by Schwern: Did …