Getting the jab still the best thing if you can.
June 27, 2021 5:11 PM   Subscribe

While the vast majority of Covid deaths in the US is of unvaccinated people recent numbers out of UK show a 60% of recent UK Covid casualties had at least one dose of a vaccine (43% fully vaccinated). This is expected and no reason to not get fully vaccinated. The Guardian explains why.
posted by Mitheral (84 comments total) 23 users marked this as a favorite
 
This is so obvious that that it apparently requires extensive explaining makes me want to give up on the world. 100% of a population vaccinated with anything less than a 100% effective vaccine = 100% of infections will be in vaccinated people.
posted by praemunire at 5:25 PM on June 27 [29 favorites]


Going without shots is killing people, end of story. Every time people see conspiracy theories around vaccines or the disease or whatever else, every decent human being should drop what they are doing and do everything stamp that down, wherever that bullshit raises its ugly head.
posted by They sucked his brains out! at 5:39 PM on June 27 [14 favorites]


makes me want to give up on the world

...a plurality of people are fundamentally morons. It's depressing, I'm sorry to say it, and I get a lot of completely understandable pushback on it, but I genuinely think it's true. I'm not saying a plurality are irredeemable, but rather that all of us, literally all of us, are deeply stupid on particular topics and it's just a matter of which topic is at hand.

The only saving grace is that a reasonable number of people are able to recognize their own stupidity and let someone else (who is not quite so stupid) decide. I am a fucking idiot when it comes to aerodynamics (I barely comprehend the whole Navier-Stokes thing), so hey, that's a problem for the FAA et al.

...of course, then the problem arises: there are quite a lot of people, possibly even a plurality (I would say so, you might disagree), who are utterly convinced of their own rightness in all things. Those people are morons.

Once upon a time I thought perhaps The New Plague might make them correct their ways, but alas on that point I was myself deeply deeply stupid and wrong -- because they are not stupid, they are morons.
posted by aramaic at 5:41 PM on June 27 [37 favorites]


saw someone note at some point that, in the long run, everyone is going to wind up with COVID-19 antibodies one way or another, and it's probably better to do so via a vaccine than via a virus

it was certainly a useful framing device
posted by DoctorFedora at 5:41 PM on June 27 [31 favorites]


I'm not saying a plurality are irredeemable, but rather that all of us, literally all of us, are deeply stupid on particular topics and it's just a matter of which topic is at hand.

I fully accept this, very much including its application to self, but I guess I would hope after a year of pandemic that educated adults would mostly not be dumb on this basic of math.
posted by praemunire at 5:43 PM on June 27 [1 favorite]


60% of recent UK Covid casualties had at least one dose of a vaccine

Isn't this a side effect of the UK one dose strategy where the second dose of AZ was delayed by like 12 weeks so everyone got one dose at least even though they went way beyond the recommended 3-4 weeks. So there are a lot of people in the UK who've still only had one dose.

And I don't think the UK is vaccinating anyone under the age of 20 at this point.
posted by jmauro at 5:52 PM on June 27 [6 favorites]


'Moron' seems pretty harsh for something that is not intuitive at all given the way humans naturally process probability. We wouldn't need an official name and long wikipedia article on Base rate fallacy if it was something that was obvious or easy to teach to people. The discussions of representativeness in that article attempt to explain WHY humans are so bad at this particular type of math. People don't "do the math" that would make this obvious unless they suspect their intuition about probability is wrong, and if people aren't used to problems like this they won't be on the lookout.

I think the actual question is why isn't this taught better in schools. I definitely was not taught how to solve these kind of problems in my US schools, and I didn't really understand the confusion until I sat down and did the math myself. I remember being extremely confused by how my school tried to teach me probability
posted by JZig at 5:52 PM on June 27 [29 favorites]


pretty harsh for something that is not intuitive at all

You are completely correct, and also alas it doesn't matter at all -- because the stupid individual (ie., me and those like me) recognize we don't intuit the right answer, and therefore listen to those who do have the right answer.

...morons, of whom there are distressingly many, simply ignore what "those elites" say because their own intuition (or other morons on Facebook) cannot possibly be wrong.

I do not need to be convinced, because if (for example) people like Fauci says something about disease statistics and probabilities, I figure they know more than me. What with, y'know, their entire career being predicated on knowing that type of thing and being provably correct in the past. I am almost entirely unaware of how my car works; I pay my mechanic to understand it better than I. I peripherally understand HVAC systems, but I still pay someone else to fully understand them.

People who need to be "convinced" are morons, because they don't know how stupid they are.

I do not fully understand crystallography, to say nothing of protein structure, and I will probably never understand either one, but nevertheless my life will almost certainly depend upon them both being correctly understood (at least with reference to a specific set of proteins) at some point before my death. You, very likely, will also be depending upon the same thing whether you know it or not. Most of us will; it's just that too many of us fail to recognize that. We have whole systems built, with great effort over centuries, specifically intended to help all of us with this set of problems, but there are too many morons who refuse to admit the possibility of being wrong.
posted by aramaic at 6:13 PM on June 27 [35 favorites]


Guardian:
Why most people who now die with Covid in England have been vaccinated

AP Press:
Nearly all COVID deaths in US are now among unvaccinated

I'm sure this has a lot to do with timing of vaccinations and statistics, but it is confusing as all heck.
posted by eye of newt at 6:19 PM on June 27 [12 favorites]


UK is vaccinating over-18s at walk-in no-appointment vaccination centres all over now. Also, anyone who got their first dose at least eight weeks ago (not twelve as it was originally) can get their second now. Check the website to find a place near you; note that different centres serve different vaccines/ages/dates/times/etc. so read the details to find the one that's best for you. Don't delay!
posted by Chef Flamboyardee at 6:25 PM on June 27 [14 favorites]


Yes, thank you. I live in Indonesia and the breathless articles about how the doctors who have died recently have been fully vaccinated are giving me hives. All doctors were vaccinated early last year. Vaccination doesn't elide major comorbidities -- it just makes it less likely you'll get a case serious enough for them to matter. And if a health care system is in collapse, many people die who wouldn't otherwise die.
posted by frumiousb at 6:27 PM on June 27 [11 favorites]


it was certainly a useful framing device

No it's not, because a huge portion of the anti-vaxers will also argue that getting covid is no big deal, so rather than getting a "risky" vaccine, they prefer to just catch it.
posted by Candleman at 6:57 PM on June 27 [10 favorites]


This is so obvious that that it apparently requires extensive explaining makes me want to give up on the world

...a plurality of people are fundamentally morons. It's depressing, I'm sorry to say it, and I get a lot of completely understandable pushback on it, but I genuinely think it's true.

I teach math and I think that while it can be quickly explained to people how this works, it is not obvious, and the people who see "50% of deaths were among vaccinated people" and say "yikes, maybe the vaccine isn't so great after all," are not fools or "morons," they're just regular people like you and me who have not thought carefully about a somewhat tricky mathematical point, and are therefore wrong about something important.
posted by escabeche at 7:00 PM on June 27 [51 favorites]




No it's not, because a huge portion of the anti-vaxers will also argue that getting covid is no big deal, so rather than …

A framing device can be useful for a lot of things, not just convincing anti-vaxers.
posted by wemayfreeze at 7:45 PM on June 27 [4 favorites]


I also teach math, I understand Bayes' Theorem and base rate errors very well, and I honestly had a very hard time squaring "43% of the people who recently died of COVID in the UK had been fully vaccinated" with "The vaccine reduces the risk of mortality by a factor of 20". Conceptually, sure, I get why there's no contradiction, but that 43% figure still made me sit up.

What finally helped me in this article -- which I otherwise found pretty nebulous -- was the stat that in one of the older age groups, the vaccination rate was 93%; that's way higher than I would have guessed. (In the US, the CDC says 88% of adults over 65 have gotten at least one dose and 78% are fully vaccinated. Sucks to be us.)

We still need a lot more data for a complete picture, but just for illustration, let's suppose you have a certain population where 93% are vaccinated, and the risk of death is otherwise evenly distributed. (I'll also ignore the distinction between fully and not fully vaccinated.) If the vaccine did nothing, you'd expect 93 vaccinated people to die for every 7 unvaccinated people who die. Now reduce that 93 by a factor of 20, and the ratio becomes 4.65 : 7. Out of every 11.65 deaths, 4.65 are to vaccinated people. That's 40%, very close to the 43% figure in the headline. So there's a proof of concept, at least, that the math can check out.

And also a proof of concept for the disparity eye of newt mentions above. Change the 93% vaccination rate in the analysis above to 78%, and the results change dramatically: only 15% of deaths will be among the vaccinated.

I have to emphatically second escabeche: none of this is obvious.
posted by aws17576 at 8:16 PM on June 27 [52 favorites]


In case anyone is still wondering, the key point in the reasoning basically follows this structure:
1. Suppose everyone in your town of 10,000 people is vaccinated.

2. The vaccine is extremely effective, and prevents (for the sake of argument) 95% of people from catching Covid. That means that 500 people in your neighborhood STILL catch Covid (much fewer than if they had not been vaccinated).

3. The vaccine makes it very highly unlikely that even people who catch Covid will die. However, suppose that two of them do.

4. So…two people in your town died of Covid, even though they were vaccinated. Ergo: “100% of the people who died of Covid in your town were vaccinated.”
It’s basically the result of having a vaccine that isn’t 100% effective, but that’s far more effective than having no vaccine at all.

On edit: dammit, I should have just read your comment, asw17576, because you basically covered this.
posted by darkstar at 8:34 PM on June 27 [20 favorites]




One reason this might not be obvious to the observant reader is that until very recently we were told that while the vaccines were 95% effective at preventing Covid infection, they were 100% effective at preventing deaths in clinical trials.

From Reuters two months ago: "Firstly, both the Oxford/AstraZeneca and Pfizer/BioNTech vaccines have been found to give people 100% protection against severe COVID-19, hospitalisation and death in analysis of stage III clinical trial results (here , here)."

Now, suddenly the odds have been shifted to "We can perhaps assume there is at least 95% protection against Covid-19 death" with no explanation and no context. I know we can sort of guess why there might be a difference between clinical trial results and real-world statistics (or in this case the vague something "we can perhaps assume"), but really, an article in the science section that represents itself as an explainer should do a better job explaining.
posted by Umami Dearest at 9:01 PM on June 27 [14 favorites]


The great thing about this metric is that the percentage of vaccinated COVID deaths will be forever kept lower in the U.S. because there are so many unvaccinated people running around.
posted by Tell Me No Lies at 9:04 PM on June 27 [6 favorites]


Now, suddenly the odds have been shifted to "We can perhaps assume there is at least 95% protection against Covid-19 death" with no explanation and no context. I know we can sort of guess why there might be a difference between clinical trial results and real-world statistics (or in this case the vague something "we can perhaps assume"), but really, an article in the science section that represents itself as an explainer should do a better job explaining.

The variants are a big part of the difference. All of the vaccines are weaker against Delta.
posted by frumiousb at 9:53 PM on June 27 [5 favorites]


> in the long run, everyone is going to wind up with COVID-19 antibodies one way or another

Not to be pedantic and morbid but do people that die from Covid necessarily develop antibodies before they’re dead? Dying or being maimed via long Covid seems to be a suboptimal outcomes regardless of if you develop antibodies
posted by fragmede at 10:02 PM on June 27 [1 favorite]


The variants are a big part of the difference.

Yes, I realize that. My point was that the explainer article could have done a better job of explaining it, rather than not even mentioning it. Also, made-up statistics preceded by "we can perhaps assume" don't give me great confidence.
posted by Umami Dearest at 10:06 PM on June 27


One reason this might not be obvious to the observant reader is that until very recently we were told that while the vaccines were 95% effective at preventing Covid infection, they were 100% effective at preventing deaths in clinical trials.

You are correct and I am so, so, so angry that even the drug companies seem unable to understand Poisson statistics and 95% confidence intervals. When you observe zero events in your "deaths" category you do not conclude that your vaccine is 100% effective at preventing deaths; instead you compute the statistical bound and say something like "at least 95% [or whatever] effective", and there is no "we can perhaps assume" about it at all. As you collect more statistics and detect some deaths, the bound will turn into a range of probabilities. FFS.
posted by heatherlogan at 10:11 PM on June 27 [15 favorites]


I don't get the anger about 100% versus 95%. Clinical trials are run with ten of thousands of people in presumably well controlled conditions. We're up near 3 billion jabs given in all kinds of conditions. Of course, there going to be differences.It is a true statement that some vaccines were 100% effective in presenting deaths in the trials. You can't expect that that means the vaccine will be 100% effective in the real world.
posted by rdr at 10:16 PM on June 27 [6 favorites]


The whole purpose of the analysis of the clinical trial data is to project (or estimate) how effective the vaccine will be in the real world.
posted by heatherlogan at 10:18 PM on June 27 [3 favorites]


And they did. 95 is pretty close to 100. You can't expect clinical trials to tell you everything. As an example the sample size is too small to say pick up on rare side effects. I think people saw the 100% effective and hoped the real world efficacy would be the same. Perhaps I'm missed it but I don't remember the drug companies or any serious person saying that the vaccine would be 100% effective against death in the real world.
posted by rdr at 10:26 PM on June 27 [5 favorites]


You are correct and I am so, so, so angry that even the drug companies seem unable to understand Poisson statistics and 95% confidence intervals [ . . . ] FFS

The line that makes you angry is from a Reuters column with no byline, not a drug company. A drug company that made a claim like the one you seem to think they made would be fined by the FDA.

FWIW the actual claim in Reuters (100% effective in the clinical trials) was a description of past results in the trials and was true. The tone is one of trying to convince readers to get vaccinated, so it's quite possible they made a conscious decision to not add a series of caveats that muddied the message.
posted by mark k at 12:11 AM on June 28 [8 favorites]




Just to throw another spanner in the statistics, the way the UK measures covid deaths is by counting those who die within 28 days of getting infected with covid. So if 100% of people are vaccinated with a vaccine that prevents 95% of infections and 100% of deaths from those infections, in the UK at least you would still officially get some percentage of vaccinated people "dying of covid."
posted by merlynkline at 12:18 AM on June 28 [3 favorites]


@merlynkline something similar in U.S.: headline like CDC says roughly 4,100 people have been hospitalized or died with Covid breakthrough infections after vaccination and then you find
more than 1,000 of those patients were asymptomatic or their hospitalizations weren't related to Covid-19, the CDC said.
Likewise in the "750 fully vaccinated people have died after contracting Covid", 19% were unrelated and probably everybody is head-faked by "after".
posted by away for regrooving at 12:37 AM on June 28 [1 favorite]


Isn't this a side effect of the UK one dose strategy where the second dose of AZ was delayed by like 12 weeks so everyone got one dose at least even though they went way beyond the recommended 3-4 weeks.

The reason that Astra-Zeneca doses are given 12 weeks apart is because they are 85% effective when spaced that way vs. only 55% if given less than 6 weeks apart. If you reduce the interval you kneecap the effectiveness.
posted by rednikki at 1:40 AM on June 28 [4 favorites]


While some people still have doubts and questions, in the scheme of things the UK has fairly consistently had the highest level of vaccine willingness in the world for the last six months. Which is a triumph of science communication, and media responsibility and regulation. We're 20 percentage points higher than France and the USA, for example.
posted by Klipspringer at 2:40 AM on June 28 [12 favorites]


> "The reason that Astra-Zeneca doses are given 12 weeks apart is because they are 85% effective when spaced that way vs. only 55% if given less than 6 weeks apart"

I'm first going to note that's a very confusing article, since it implies 76% efficacy with one dose and then ... 55% efficacy with two doses? Apples and oranges are being compared there somehow in a way the article isn't articulating very well.

That having been said, it absolutely seems to be true that waiting longer makes Astra-Zeneca more effective. However, that was not at all the reason the initial 12-week wait in the UK was introduced, since that was unknown at the time. It was a happy accident. The idea was to get first doses into as many (especially older and vulnerable) people as possible as quickly as possible, to offer them at least some protection.

But in any case ...

> "Isn't this a side effect of the UK one dose strategy where the second dose of AZ was delayed by like 12 weeks so everyone got one dose at least"

Maybe a little but not really? Currently in the UK, 66.3% of the population has had one dose and 48.6% have had two, compared to, for example, 54% and 46% in the U.S. (that's total population, not adult population, so those may be lower numbers than you've seen elsewhere.) That's partly because the UK stepped up the pace of giving second doses in that past few months, as others have pointed out, but the UK has always been comparably high on both first and second doses worldwide.

It's true that a single shot seems to be somewhat less effective against the Delta variant, which is currently spreading rapidly in the UK, so that may be part of what's going on. Another possible factor is that Astra-Zeneca, which is broadly in use in the UK, seems to have a slightly higher time to reach full effectiveness after the shots -- possibly four weeks instead of two weeks, which means there's a somewhat greater window for a once-vaccinated person to catch Covid.

However, I'd say it's more likely that the main factor is that the UK has had a very strict policy of vaccinating older and more vulnerable populations first, gradually expanding the age range. Even if hospitalization and death rates drop dramatically among those populations post-vaccination -- which they have -- statistically you still might see as many deaths as in the younger, unvaccinated population.

That being said, even in the middle of a third coronavirus wave, hospitalizations and deaths are currently WAY down in the UK compared to the second wave. The vaccinations are very much working. Admittedly, there's always a bit of a lag between an increase in cases and an increase in deaths, but right now there's reason to be optimistic.
posted by kyrademon at 3:04 AM on June 28 [3 favorites]


> "Isn't this a side effect of the UK one dose strategy where the second dose of AZ was delayed by like 12 weeks so everyone got one dose at least"

This issue was more for the Pfizer vaccine - Pfizer recommended 3 weeks, with a max. of 6 weeks, but the UK pushed it further. I'm not sure of the details (and I think it's changed a few times=, but I believe they are now trying to maintain 8 weeks between doses (although a lot of the drop-in centres seem to be accepting people for a second dose at 22 days+ based on twitter.)

We're 20 percentage points higher than France and the USA, for example.

One theory regarding France is that they tend to be very negative when answering polls, but then do it anyway.
posted by scorbet at 3:23 AM on June 28


I live in an area of the US that's achieved a whopping 36% vaccination rate despite it being free, readily available and with all kinds of different incentives being offered. The area is populated by young military and trump supporting rural folks.

I will still be wearing a mask for a long time. Logic and mathematics are not convincing people that vaccination is more important than ideology.
posted by mightshould at 3:49 AM on June 28 [7 favorites]


Just to put this in a little more perspective, the UK is currently seeing slow but significant growth in case numbers as the Delta variant starts to sweep through the general population (having previously been somewhat contained).

As of yesterday's reporting (which it should be noted is skewed slightly by being a Sunday) we're up around 90% of what we were at the start of the year, but deaths have remained roughly at the same level (though they are too increasing, but very very slowly).

We're currently averaging around 20k cases per day, and 10~ deaths. The last time we had roughly this number in December 2020, we were seeing 500 - 600 deaths per day. This is the vaccine working.
posted by fight or flight at 4:01 AM on June 28 [6 favorites]


The initial claims that the Pfizer and/or Moderna vaccines offered 100% protection against severe Covid were made without statistical support. It was irresponsible, sensationalist headline-grabbing on the part of the manufacturers, and it shouldn’t have survived the review process. But it did. And it was extremely effective at grabbing headlines. Moderna was especially egregious about this.

The robust result out of those trials was that, out of their enormous pools of people who got vaccines or placebos, only about 200 got Covid, and only ten or eleven of the people who got sick were in the real-vaccine group (instead of half). That’s where we got the result that the vaccine is 95% effective against symptomatic infection. That result was fine.

The severity result was that about 15% of the people who got sick (that is, about thirty people) had to go to the hospital, and none of them were in the vaccine group. The manufacturers reported this as “we observed 100% protection against severe Covid.” But 15% of ten or eleven people is (2 ± 1.4) people (with a “one-sigma Poissonian error bar”). If you’re looking for 2 ± 1.4 people and you observe zero, that’s … not different. Not different enough to be interesting, anyway.

What Moderna should have reported — what the vaccine trials actually measured — was that the vaccine was so devastatingly effective at preventing disease entirely that the trials were too small to determine whether the vaccine made the disease less severe in people who did get it.

(I am remembering these numbers from November, without re-checking, because I’m sure there are more robust results now that vaccines have been in real people’s arms for half a year.)
posted by fantabulous timewaster at 4:40 AM on June 28 [11 favorites]


Perfection is the enemy of the good.
posted by warriorqueen at 4:57 AM on June 28 [1 favorite]


I am remembering these numbers from November, without re-checking

The German STIKO (group responsible for recommendations about vaccinations) has the info from the various trials together in a table on p34 here. (I was looking it up earlier.) It is in German, but it shouldn't be too hard to decipher in most cases. They include the 95% confidence interval for the various scenarios as well, and as expected for "serious cases", "hospitalisation" or "death", they have confidence intervals of "n.k" (not calculable), "k.a." (not given) or "-152.6 - 99.5".
posted by scorbet at 4:57 AM on June 28 [1 favorite]


"I teach math and I think that while it can be quickly explained to people how this works, it is not obvious, and the people who see "50% of deaths were among vaccinated people" and say "yikes, maybe the vaccine isn't so great after all," are not fools or "morons," "


A lot of probabilistic thinking is deeply counterintuitive. Anyone who's thought about the Monte Hall problem for a minute can attest to that. No argument there.

But the reports make it harder to figure out what's going on by emphasizing percentage rates without mentioning totals for context (or vice versa). When you see that 50% of deaths occur in people who are vaccinated, that's scary. When you see that it's 50% but that the total deaths have gone from 100 per day to 10 per day, your sense of what's happening is quite different! (I'm just picking arbitrary numbers for the example.)
posted by oddman at 5:19 AM on June 28 [2 favorites]


In Portugal, where we are (almost) all willingly taking our jabs, the number of new reported cases is back up again (over 1,000 a day in a population of 10m) but the number of deaths (averaging 2-3 per day with many days with 1 or none) much below the level the last time the daily infections were this high. So there is a conclusion to be drawn that the vaccination process is mitigating the effects of the virus. [And, also, that it's a really poor idea to open the flight corridor to the UK for a weekend because of a football game]
posted by chavenet at 5:47 AM on June 28 [8 favorites]


Another factor that can lead to a high number of cases/deaths in the vaccinated group is that this group heavily weighs those 50 and older. That group also comprises an overwhelming proportion of deaths to date (in Scotland, I believe the rationale for targetting people over 50 for initial vaccination is that they comprised over 95% of serious cases). That means that if we have somebody who has died from Covid 19 they are probably more likely than not to be an older person who has been the unlucky victim of "vaccine escape" rather than a younger, unvaccinated person who would have been at a very much lower level of underlying risk. Bayes at work indeed. - In fact the order of risk is probably older anti-vax people, then older vaccine escape victims - then everybody else.
posted by rongorongo at 6:13 AM on June 28


I remember being extremely confused by how my school tried to teach me probability
posted by JZig a


My 10th grade math teacher mostly talked to us about how to gamble on horse races , and at poker, and I attribute that with giving me a higher than normal understanding of how statistics can represent and misrepresent reality
posted by eustatic at 6:46 AM on June 28 [2 favorites]


From their official site: COVID-19 Vaccine AstraZeneca confirms 100% protection against severe disease, hospitalisation and death in the primary analysis of Phase III trials.

There's a fair amount going on between the lines. AZ "confirms" not claims. Also, if you read the trials themselves, they only had about 30,000 people vaccinated (IIRC) and only followed them about three months. That's not going to be a lot of cases (or time for reinfections) to draw on.

Of course, very few people read the actual trial results, including newspaper reporters.
posted by CheeseDigestsAll at 6:59 AM on June 28


But I think the Guardian headline is actually the main source of the educational problem.

Don't put the easily misunderstood statistics in your headlines, put the common sense interpretation in your headline.

The Guardian is leaning into people's misunderstandings, for clickbait. And this OP, regrettably, is doing the same thing. Therein lies the bulk of people's miseducation around probability.

I remember that the anti-vax movement was begun in England by an unethical doctor who was trying to sell his alternative vaccine; but he had a ton of help by the television push that scared people about the MMR vax.

The doctor was swiftly discredited, and his license was removed, and he had to move to Texas and sell books on the Alex Jones program to stay rich.

But the majority of UK residents were bombarded with television clips of kids crying as the got the MMR, so that s what people remember
posted by eustatic at 7:01 AM on June 28 [12 favorites]


Yeah, this "explainer" article doesn't help AT ALL for all the reasons stated above.
posted by lalochezia at 8:12 AM on June 28 [1 favorite]


Meanwhile, I saw Twitter footage of a rather large protest march in London Saturday, demanding an end to all COVID restrictions and declaring the usual kinds of barking mad things.

So, the best of luck and best wishes to all of our Euros who interact with these types of people on any regular basis.
posted by delfin at 8:16 AM on June 28


I'm glad to see oddman bring up the Monte Hall Problem because I think internalizing that lesson will change your brain for the better. It's deeply, incredibly painful to my brain for some reason to consider it, at least.
posted by OnTheLastCastle at 9:49 AM on June 28 [2 favorites]


I'm glad to see oddman bring up the Monte Hall Problem because I think internalizing that lesson will change your brain for the better. It's deeply, incredibly painful to my brain for some reason to consider it, at least.


I guess. I mean, I get it, the 'it' important factor being that the host never picks a correct door, which changes your odds from 1/3 to 1/2 so changing is useful I guess...but I think it's interesting how that implication of an extra 17% of odds is somehow valuable psychologically for a one-time event. It also assumes no discount rate of confidence for picking the correct door initially and then picking the incorrect one. Isn't that what the movie Top Gun was about? Statistics man.
posted by The_Vegetables at 9:58 AM on June 28


And this OP, regrettably, is doing the same thing.

I figured my headline title was pretty clear every one should get vaccinated. And my blurb says the figure people are already hearing [it's all over my social media] is normal and expected. I can see where it could be clickbaity if one hadn't heard the numbers before though.
posted by Mitheral at 10:06 AM on June 28


My point was more about how understanding our intuitive brains are wrong often is important. Percents are inherently meaningless when you're not aware of the underlying hard number.

For the Monte Hall Problem, it clicks a lot better if you imagine 100 doors. You have 1/100 chance of picking the correct one. He opens 98/100. (edit: doing too many things at once, you have a 99/100 chance of winning at this point) Scale improves our ability to understand the nature of the problem whereas only 3 doors is harder. Unfortunately when I see most people talk about teaching "critical thinking" it seems like (to me) they mean "Stop being so dumb, you big dummy." instead of "Grapple with the ways your brain shortcuts or can't correctly estimate things and be willing to put in the effort to understand even if it hurts."
posted by OnTheLastCastle at 10:14 AM on June 28 [2 favorites]


Human nature dictates that most people, myself included of course, believe what they would prefer to believe most of the time, but the age of the internet has turbocharged that tendency because no matter what you would prefer to believe you can find constant positive reinforcement online. It's so depressing.
posted by The Card Cheat at 10:16 AM on June 28 [3 favorites]




So, the best of luck and best wishes to all of our Euros who interact with these types of people on any regular basis.

Loud though they are, this is an extremely small and almost irrelevant minority here. In the medium term, I actually think that despite the current rising wave of Delta cases here in the UK, the situation in September/October is going to be much uglier in parts of the US than here. That's because the overwhelming majority of people here who have not yet been vaccinated just haven't had their turn yet. You can see on pg4 here how high uptake is by age group. In the oldest cohorts, that is over 90%. Even in ethnic groups where uptake has been slower, a concerted campaign which started almost from the beginning has had substantial effects. Very soon there simply won't be substantial pockets of unvaccinated adults left anywhere. The question of what to do with children remains outstanding.

Compare that with parts of the American South where there are whole communities with low coverage. That's where we could see a really nasty flare-up that I worry about. (and of course... most of the world has had barely any doses.)
posted by atrazine at 10:54 AM on June 28 [4 favorites]


> "For the Monte Hall Problem, it clicks a lot better if you imagine 100 doors."

Not that this undercuts your overall point, but -- not for me, it doesn't. I understand the math and I get why it works, but when there's two doors left my brain intuitively tells me, "It's 50/50 and it makes no difference whether you switch or not." Even if a million doors have been opened beforehand.
posted by kyrademon at 11:25 AM on June 28 [1 favorite]


MetaFilter: extremely small and almost irrelevant
posted by Foosnark at 11:34 AM on June 28 [5 favorites]


(RE the Monty Hall problem, I was able to conceptually understand the math behind the explanation, and it made sense, but I didn’t really internalize it until I saw the graphics showing the three possible cases and the two possible choices and their outcomes for each. Then the lightbulb really lit up for me.)
posted by darkstar at 11:38 AM on June 28


"Monty Hall problem"

Y'all are killing me with this
posted by Windopaene at 11:38 AM on June 28 [4 favorites]


the Monte Hall Problem...I get it, the 'it' important factor being that the host never picks a correct door, which changes your odds from 1/3 to 1/2 so changing is useful I guess...

Actually that's not correct. The real answer is weirder. Changing doors increases your probability of winning to 2/3.

when there's two doors left my brain intuitively tells me, "It's 50/50 and it makes no difference whether you switch or not."

I like to think of this informationally. What information do you gain by seeing the host open a door?

At the beginning of the game your-door has a 1/3 chance of winning and not-your-door has a 2/3 chance of winning. If you had the option of saying "I'd rather pick BOTH of the not-your-doors" it's easy to see why your chance of winning would rise to 2/3. Also, you know for a fact that one of the not-your-doors is a loser.

When the host shows you a loser door you have learned... nothing. You already knew that one of the not-your-doors was a loser. Since you have gained no information your opinion should not change. You used to think that not-your-door had a 2/3 chance of winning, so you should still think that not-your-door has a 2/3 chance of winning. Reject your-door and switch to not-your-door.
posted by justsomebodythatyouusedtoknow at 11:41 AM on June 28 [3 favorites]


Monty Hall: imagine there are a million doors. The chance that you've picked the right one is one in a million. The chance that the right door is one of the other ones is 999,999 in a million. The host helpfully eliminates all of the other doors apart from one. The chance that this remaining other door is the right one is still 999,999 in a million; the chance that yours is the right door is still one in a million. Do you switch?
posted by rory at 12:08 PM on June 28 [6 favorites]


...a plurality of people are fundamentally morons.

Human nature dictates that most people, myself included of course, believe what they would prefer to believe most of the time, but the age of the internet has turbocharged that tendency because no matter what you would prefer to believe you can find constant positive reinforcement online. It's so depressing.

The Guardian is leaning into people's misunderstandings, for clickbait.

I keep going back and forth. Yes, people are stubborn, obstinate, and foolish, and I’ve made my share of disparaging comments about them, but modern times have placed us in the middle of an unacknowledged information war. The morons aren’t so much born as they’re made, whether it's Fox News, Facebook, or even The Guardian.
posted by Eikonaut at 12:24 PM on June 28 [2 favorites]


"Monty Hall problem" Y'all are killing me with this

Yeah, I realised I'd done it too just as the edit timer clicked off. Sorry to add to your deadness, Windopaene.
posted by rory at 12:25 PM on June 28


I have no regrets, Windopaene!! Thanks for giving me a laugh at such a bleak time, I can't believe I misspelled it.
posted by OnTheLastCastle at 1:16 PM on June 28 [1 favorite]



On edit: dammit, I should have just read your comment, asw17576, because you basically covered this.

but you broke yours into point form and thus it caught my eye while browsing down the page, so you get the free gift of a Mefi favourite.
posted by philip-random at 1:25 PM on June 28 [1 favorite]


And they did. 95 is pretty close to 100. You can't expect clinical trials to tell you everything. As an example the sample size is too small to say pick up on rare side effects.

Real-world effectiveness was always going to be less than 100%. The trials didn't include people with suppressed or otherwise inoperable immune systems. The vaccine- quite probably any vaccine- simply doesn't trigger antibodies in a small number of people, none of whom were in the trial for, I think, obvious reasons (we know vaccines won't work well for them).
posted by BungaDunga at 2:34 PM on June 28


That AstraZeneca headline also (deliberately?) suffers from deceptive grammar. It's as if I woke up in the morning, asked Siri from my the comfort of my air conditioned dwelling what the outside temperature was, and then later issued a technically accurate but entirely misleading statement, "I confirm in my bedroom it has reached 100 degrees."

Similarly, by "COVID-19 Vaccine AstraZeneca confirms 100% protection against severe disease, hospitalisation and death in the primary analysis of Phase III trials," obviously what they mean is that there was 100% observed protection among the participants of the Phase III trials, not that the Phase III trials confirmed that there was 100% protection generally. But someone uncritically perusing the statement will naturally assume the latter.
posted by xigxag at 3:06 PM on June 28 [1 favorite]


Everyone always overthinks the Monty Hall problem and should just play an interactive/simulation version. Yes, I'm suggesting that y'all oughta Monte Carlo Monty Hall.
posted by GCU Sweet and Full of Grace at 4:25 PM on June 28 [6 favorites]


To add to the naming confusion / hilarity, apparently Monty Hall's birth name was Monte.
posted by thefoxgod at 5:15 PM on June 28 [4 favorites]


Legit laughing at both of those comments!
posted by darkstar at 5:57 PM on June 28


Thanks, everyone, for this thread. It’s finally helped me unpick this knot that I’ve been worrying at for days.

I hate how this dumb pandemic makes me doubt everything I know/knew about immunology and vaccines. But I love how exposing my ignorance through these doubts lets me patch it up.
posted by disentir at 6:51 PM on June 28


Which is a triumph of science communication, and media responsibility and regulation

And, I would imagine, a triumph of the NHS! Because the US doesn't have such infrastructure to reliably connect people to medical providers, reaching certain traditionally underserved populations was always going to be difficult. I'm not referring to obstinate Trumpers. In my city, we seemingly plateaued at around a 60 percent vaccination rate because vaccination efforts ran into all the usual barriers driving healthcare inequity and lack of access. Issues like disparate education access, lack of access to reliable transportation, inequitable autonomy of employment circumstances, and historical mistrust of the medical establishment inevitably present themselves and act as barriers to vaccine uptake. The logistical problem of ensuring people get two shots for the MRNA vaccines within a certain timeframe complicate things even further.

That being said, I'm not sure if this is "bright side" or not, but I wonder if greater community spread in the US compared to Britain and Israel will help curb a third wave amongst unvaccinated populations in the US. From what I can tell, there is an overlap between populations with low vaccine uptake and populations that were already disproportionately affected by COVID. Obliviously, given rising cases in places like rural Missouri this protection isn't foolproof, but I still wonder, maybe its just a hope, if we won't escape the worst because of preexisting population immunity from people already infected with COVID.

Regarding lies, damn lies, and statistics, one of the aims of Common Core math was to help people better understand advanced mathematical concepts like statistics. Except, people went apeshit over Common Core. Still, from the perspective of someone who completed a doctoral program with more statistics than I ever wanted, I would have appreciated Common Core style math instruction in grade school. Your mileage may vary on that point, perhaps.

I think part of the problem, and I encountered this constantly in my education, is that probabilistic research seems to contain many measures that never seem to tell you quite what you want them to or think they should. For instance, I think most people just want to know the chances they'll catch COVID or get seriously ill from COVID if they've been vaccinated. You'd think the efficacy studies would tell you that, but they really don't. Instead, they convey an estimate of the reduction in chances of infection after vaccination of over the timeframe of the study. Generalizing those results forward in time or to the specifics of an individual's biology and life circumstances simply isn't possible.

For example, as people age, their immune systems inevitably began less able to mount a response to any disease. Because severe COVID becomes more common with advanced age, and we're proportionately vaccinating more of the elderly, its inevitable that certain people simply will be unable to mount an immune response. I think we have to keep that in mind when we look at incidences of "breakthrough" infections. I also think we need to keep in mind the risks we run in our daily lives we aren't even aware of, such as dying in an auto accident or from diseases like seasonal flu. Eventually, even if the vaccines aren't perfect, they still seem to drop the risk of COVID below that of the "background" risks we run in our day to day lives. At least so far. I think its likely we'd behoove ourselves to all get booster shots as the virus completes its evolution as it interacts with human immune systems. That's really what is happening with all these "variants".
posted by eagles123 at 8:00 PM on June 28 [1 favorite]


when there's two doors left my brain intuitively tells me, "It's 50/50 and it makes no difference whether you switch or not." Even if a million doors have been opened beforehand.

For me, the breakthrough is in understanding that it's really a two-step problem.

Scenario A: If you chose the correct door first, then the host opens an incorrect door, leaving you with an incorrect door, and you lose if you switch.

Scenario B: If you chose the incorrect door first, then the host opens the other incorrect doors, leaving the correct door, and you win if you switch.

Nothing to choose between them, right? Except there's not the same chance that you're in Scenario A as that you're in Scenario B. You need to look back a step to see how you got there. Because there are three doors, each equally likely to be correct, the chance that you got it right initially is 1 out of 3, and that you got it wrong is 2 out of 3. You're simply more likely to be in Scenario B to begin with, and thus you have a greater chance of winning by switching.

Or you can think of it as:

Scenario A: If you chose the correct door first, then the host opens an incorrect door, leaving you with an incorrect door, and you lose if you switch.

Scenario B: If you chose incorrect door #1 first, then the host opens incorrect door #2, leaving the correct door, and you win if you switch.

Scenario C: If you chose incorrect door #2 first, then the host opens incorrect door #1, leaving the correct door, and you win if you switch.

Each of the three scenarios here is of equal probability, and so you can see that in 2 out of 3 cases, you win by switching.
posted by praemunire at 8:30 PM on June 28 [5 favorites]


Just alerting folks to the existence of a MeTa thread related to the casual use of the term 'moron' on MetaFilter as a whole, precipitated by this thread.
posted by lemur at 9:00 PM on June 28 [5 favorites]


I’ve found the best way to explain the Monte Hall problem is to change it by adding some doors. Suppose you could pick from one million doors. Then the host eliminates all but two doors. The door you picked and one other door. Behind one of those remaining doors is a prize and the other door has nothing. Should you keep your original guess or switch. Ok suppose it was a thousand doors, etc.
posted by interogative mood at 9:31 PM on June 28


I've found the best way to explain the Monte Hall problem…
Somebody check on Windopaene.
posted by Strutter Cane - United Planets Stilt Patrol at 12:17 AM on June 29 [1 favorite]


Monte Python.
posted by spitbull at 1:07 AM on June 29 [1 favorite]


To add to the naming confusion / hilarity, apparently Monty Hall's birth name was Monte.

Wait, what? It is Monty? Man. Windopaene's comment was ambiguously worded enough that it made me think I'd misspelled it. Now I have to reprogram my brayne.
posted by rory at 1:28 AM on June 29 [2 favorites]


when there's two doors left my brain intuitively tells me, "It's 50/50 and it makes no difference whether you switch or not." Even if a million doors have been opened beforehand.

It's best to think of it as a 1 step problem, with the original 1/3 winner vs 2/3 loser ratio, and with a door opened, they swap.

statistically:
if you stick with original = 1/3 chance of being a winner
if you randomly switch after a door is opened, it does become 50/50 chance of winning
if you always switch, it's 2/3 chance of winning.
posted by The_Vegetables at 7:02 AM on June 29


Guys, um ... thanks for the repeated explanations to me of why the Monty Hall problem works the way it does, and I do mean thanks because I understand that they were well meant, but I already understand the math, and seriously, at least as far as I am concerned, you can stop now.
posted by kyrademon at 8:05 AM on June 29


Behind two doors are additional explanations of the Monty Hall problem, and behind a third door is the original topic of the post...
posted by eponym at 8:10 AM on June 29 [11 favorites]


I just want to check in and follow up on a previous comment in a Covid related vaccine thread about waiting to get vaccinated.

I've now had my first jab of the Moderna vaccine and so far everything is ok. The only side effects I've had from the first dose was a fairly sore arm and some mild lymph node activity.

Because I'm really sensitive to my body I did notice some other effects that I'm pretty sure were not psychosomatic, some mild tingling and flushing as the vaccine did it's thing.

Curiously if anything I didn't feel fatigued by the dose at all, but if anything I felt energized and all keyed up like I had too much coffee or did some stimulants or something. Like it got me all amped up for about 12 hours in a really strange and curious way.

I also have a care plan in place for the second dose in case I have complications from the second dose like the very rare blood clots or heart inflammation. I will have some aspirin and blood thinners on hand just in case things get weird or I have a bad reaction.

As I mentioned previously someone in my household ended up with diagnosed blood clot problems after their J&J single shot dose and they had to go on blood thinners, but they recently got the all clear on that and they're doing fine.

I am also comforted by the new studies that the mRNA variants are starting to show in studies that they are providing long term if not life long immunity, even to the new mutations as known so far.

And one of the really cool things to come out of this massive global effort in mRNA vaccines is that it's opening the door to things like trials for a vaccine for HIV. The recent research into mRNA vaccines is just the tip of the proverbial iceberg - they've been studying mRNA for years if not decades now long before the pandemic started.
posted by loquacious at 11:05 AM on June 29 [1 favorite]


loquacious, just wanted to mention, as far as I understand, the clotting problems are only seen with the adenovirus vector vaccines, so just the J&J and the Astra Zeneca. I got the J&J (or the Janssen, as they call it here) and since I have an unusual blood vessel in my brain I did my homework and called my GP to discuss. For those of us who got the J&J or the Astra Zeneca the thing to watch out for is severe headaches and stroke-like symptoms, and you would want to call the doctor immediately, and definitely not take anything for it until you've been seen by a doctor, since the clotting issues they cause are unusual in that they are made worse by some blood thinners.

But yeah, in your case, with Moderna, it shouldn't be a concern. I'm glad to hear your household member got through the clotting issues ok. It's certainly something I've been thinking about over the past couple of weeks as I move out of the risk window.
posted by antinomia at 2:29 PM on June 29


I got the Pfizer vaccine fairly early on because of my job. The first shot I just got a slightly sore arm, nothing major. The second shot I was fine for about 24 hours, and then I felt like a train ran over me. I went to sleep for a night, and then I woke up feeling fine. I was actually kind of thankful for the symptoms because they confirmed the first shot trained my body to recognize the infection. Even better, it seems as though the immune response generated by the MRNA vaccines is pretty robust and long lasting, even against the variants.

MRNA Vaccines Provide Long Lasting Immunity to COVID
posted by eagles123 at 8:39 PM on June 29 [1 favorite]


« Older Nothing Really Matters   |   The story of Chuck E. Cheese Newer »


You are not currently logged in. Log in or create a new account to post comments.