Sunday, November 29, 2020

Excess Deaths, Murder Statistics, and Sexual Offenders.

Long sections of my career were spent working with sexual offenders.  The behavior of staff is worth noting. There are rescuers: 

He's developmentally disabled and he mooned some schoolgirls from his bus.  There's no way he belongs on the sex offender list for life.

He was 19 and she was 16 and she has accused other men.

And there are punishers:

After he raped her he knocked her out and tried to set her on fire.

He molested all the girls in that family but only one had the courage to testify against him. 

Versions of these statements had made it into the chart, the hospital's official record of the patient's history, which can be brought into court and used as a reference for expert testimony.  Thankfully, that information cannot in and of itself be submitted as evidence.  Normal rules of evidence apply in court. This is a good thing, because all four of the above statements were false. For example, the man started forcibly raping the neighbor girl when she was 11, she first reported it when she was 16.  No one tried to set anyone on fire in that other case, or even knocked them out. He groped her while she was asleep and the house burned down two years later. People get activated around sex offenders.  They want certain things to be true. I shouldn't say "they."  I should say "we."

Usually the corrective can be fairly low key, with someone saying "I don't think the evidence for that is very good.  I've been doing the psychosocial history/talking with his attorney/going through the old records and I think this got added in.  It seems to come from a neighbor saying 'We always knew something was going on in that house.  I'll bet he molested all those girls.'" But sometimes it has to be a bit harsh, and though I was not a confrontative person by nature* I learned to be, because sometimes you are sitting at a table and have to say. "I have put a note in my official eval that this is not true and previous records claiming it should be ignored.  We have to stop saying this, both formally and here in the team room." Not easy when one of those claims is by your supervisor, who is sitting right there. I have experienced this in reverse as well, of making a statement and having another staff member saying  "That's just a rumor, started by her previous girlfriend while they were divorcing. There's no evidence for it."  It's pretty humiliating, but if you don't want to be part of keeping non- or low-level offenders locked up or dangerous people let out, you try and be a stand-up guy. When something isn't true you can't let that go.

We were overinfluenced by recent events in this.  When we get to the bottom and uncover that the criminal justice system has kept a guy locked up for ten years, came to us, and was gradually given more freedom over the next ten until the halfway house finally set him up in an apartment, and it turns out it was another guy all along, we are altogether too eager to believe the next person proclaiming innocence and let them out more quickly. Which incidentally, still isn't all that quickly. More evidence to never plead Not Guilty by Reason of Insanity. We own you for life, then.  Dealing with so many lying, manipulative bastards one yearns for someone to believe, to rescue. Then also, when someone we released six years ago commits a crime we just naturally snap into a more restrictive, even vengeful mode.


When discussing violent crime statistics and race, people want very badly for some numbers to be true and some false. It does not always break down neatly, as people will want the numbers to show their favorite theory, such as the presence of fathers, or early intervention, or having a better attorney is the primary driver of the numbers. It shows up in the explanations of the facts. The numbers are higher for young black men fighting because the police want to round them up and get them off the street and they have worse lawyers. But the next person will say Actually, the numbers for young black men are too low, because the police don't give a rat's ass what happens to them and just send them on their way so long as they aren't bothering white people.  Also, it's hard to get witnesses to testify so the police just shrug.  Bring whatever pre-judgement you want, you can undermine what the statistics are. Except, as Steve Sailer pointed out years ago, that all falls apart with homicide.  You have to have an actual body, you can't just say that the police are exaggerating or over-enforcing.  And if you have a body, you have to have an explanation, and if he bled out in the ER from gunshot wounds you can't just make that go away and say it was extreme obesity leading to a heart attack, no matter if he weighed 600 lbs. That was tried in mob cities decades ago, as in The Gang That Couldn't Shoot Straight. "He unfortunately died of a heart attack while he was being stabbed." You can make up ways to get around it in order to write a book or make a movie, but 99.9%, you have to have a body to start with, and all else flows from there.  Which is why the homicide statistics for victims is so important.


I remember my training far less than I should.  We do not generalise in even the most obvious things sometimes.  But occasionally, I remember, and I have been trying to remember throughout the C19 crisis (and more recently, the election accusations from people who voted for Trump and counteraccusations from those who just don't like the guy and tell themselves it's about issues.) We have to be ten times more suspicious of what we hope is true, because that is where we are most likely misled.

All this in mind as I caught wind of a Johns Hopkins report saying there were no excess deaths from C19.  I consider Johns Hopkins to be reputable.  I was interested.

I waited until more info came out, as I do when I have kept my wits about me. We are in a period when studies come out showing that only 4% of the population has been infected, followed by claims that 40% has. Lots of finger pointing and claims of bad faith by the other side. Lyman Stone, who I have referenced before about excess deaths is not kind in discussing the retraction of the JHU study published in their student newspaper.  

Folks, that stupid JHU student newspaper piece has been retracted because it contained numerous blatantly false statements and elementary misreadings if the data. Just because you’re affiliated with JHU doesn’t mean you can’t be innumerate.

Please keep reading at his account. He does not shrink away from using the word "lie."

The reports have been circulating that the study was pulled because it offended the narrative or more mildly, that it was being misused by people who disagreed with the accepted narrative, and this is censorship because they should let science be free and open.  As far as I can tell, the only evidence for this is the beginning of the retraction announcement by the student editors, who then go on to admit that there was a lot wrong with the study itself, not just what people were doing with it. Retractionwatch, a publication I trust, has more of the story. Briand is in a graduate program for Economics, not medicine or disease. It looks like a great learning experience for those student editors.

People should have been alert from the first, because "no" excess deaths would mean that no one has died of CoVid, a position I don't think many take.  True, one could retreat to a position that it only hurried the deaths of 90% of its victims by 2-3 months, and the remainder could be cleaned up by jiggling around the flu statistics or something, but folks, those people died of something, and the people who were there watching them thought it looked like serious stuff, gasping for air while drowning in their own fluids. The CDC is saying at a midpoint 300,000 excess deaths. It's not enough to say "I don't trust 'em."  Find me better numbers.

I think I get why (some) conservatives cry that the balance between economic damage and health risk has been shoved out of whack, or that there has been inordinate focus on some unimportant safety measures. Those are indeed bad things, but they deserve to be argued on their own merits, not with made-up stuff. I even agree with a lot of that. I think a higher level of risk is justified.  I just don't get the drive on a fair number of prominent conservative sites, both the posters and the commenters, to insist that this is all overblown or even a hoax. Argue if you wish that 300,000 is still a small number.  Point out that politicians, prominently Democrats, have been hypocrites about what they allow for themselves versus what they demand of others.  Pound the table that people out there are being complete pricks (I ran into an irritating one myself yesterday). But stop seizing on stuff that tells you what you want to hear and waving it aloft without at least checking behind you to see if there is toilet paper trailing out of your belt. (Your opposition is also doing this.  It is infuriating.  But they have also been doing this your entire life. Don't imitate them.)

Actually, I do get it.  I've seen it on other topics my whole life, and I've even done it myself. I'm yelling at you, but I'm wincing because it's really me.

*I am on some things, but not as many as supposed. I feel guiltier about my cowardices than about my harshness.



Thos. said...

Well said, sir.

Texan99 said...

I'm prepared to believe it was false. I'd prefer to see it rebutted than hidden.

Excess deaths showed a spike--not spectacular, but detectible--last spring. Since then any continued excess mortality over the 5-year average has been pretty hard to see. The point is worth making even if the Johns Hopkins article was not a good treatment of the controversy; the rebuttal would be more convincing if it weren't accompanied by panicked censorship.

Douglas2 said...

Re: the JHU thing.

When I read on blogs & twitter that a new "study" from some respected institution says just what would reinforce the views of the blog owner or twitter user, I'm primed to assume that it is not a 'study' at all.

As was the case here - - a news article about a webinar. No link to the webinar or presenter, but entering her name and subject allowed me to find it on youtube easily.

It's obviously a seminar intended as a primer for accessing and analyzing data sets. In fact we are at about the 20 minute mark when we download the .csv files from CDC for weekly numbers of all-cause deaths, and combine the (final) 2014-2018 set with the preliminary 2019-2020 set. A bunch of time up to that point has been spent on a step-by-step of navigating the CDC web-site, and then looking at weekly death data broken up by age groups.

Douglas2 said...

One of the Q&A questions is essentially "are the increased deaths just elderly who would have died soon anyway?"; and it was answered her first segment - pasting weekly total deaths by age group from the CDC site line by line into excel to build the stacked bar chart that shows very clearly that the proportion of deaths in any age group hardly changes at all through the course of the first wave. Hence the reported quote “The reason we have a higher number of reported COVID-19 deaths among older individuals than younger individuals is simply because every day in the U.S. older individuals die in higher numbers than younger individuals,”

All of these bars add to 100%, and this seems to me where the spurious charge that she 'compared percentages rather than total numbers' comes from in relation to the headline conclusions about 'excess deaths'.

Douglas2 said...

She emphasizes that the peak number of deaths in the peak week of the year is some 12000 higher than the max of peak week for recent previous years, talks repeatedly about excess deaths.

The anomaly in the data that she chooses to point out is that in the yearly 'peak' week for deaths -- which thru 2019 corresponds to the influenza death peak -- we always see that deaths from many of the top causes have their peak at the same time: Heart attack deaths peak, diabetes deaths peak, dementia deaths peak, etc. However with the 2020 April COVID-19 peak, not only is there no peak for these other causes, but they drop -- a result that no-one would predict.

This is where it gets weird -- but is it quite evident that this faculty member is a non-native speaker of English who struggles for words, so it is pretty easy to focus on a literal but unintended meaning when trying to follow her.

The number that these other causes have 'gone down' almost perfectly corresponds to the number that the deaths attributed to covid-19 exceeds the number of deaths attributed to heart disease (in the few weeks that they do). So:

(covid-deaths heart-disease deaths = missing deaths from non-covid causes.

This seems to me to be a strange statistic to focus on, and perhaps a good candidate for –– and kind of glosses over that the total number of deaths in those weeks is way beyond the range of statistically plausible predictions sans COVID.

At this point she says "don't shoot the messenger!", and sums up in saying that the number of deaths attributed to COVID, although awful, isn't as bad as we might think because it obviously encompasses many expected deaths from other causes. She also says that the average of weekly deaths is approx 60,000/wk for 2020 (actually 62,373 through week 32) which is "normal". In this I suspect she means not the colloquial meaning of 'normal', but within two standard deviations of the population-adjusted mean from previous years. It sure looks like that's not the case with the 2015-2020 data that's currently available from the CDC, but I couldn't say whether if we extend back through a few more bad flu years we might find similar averages. It will probably easier to compare the "death rate" per 1000 once it becomes known, rather than work out weekly averages

I can see that the Yanni Gu JHU newsletter highlighted quotes out of context that set off alarm bells, but the JHUNewsletter retraction misrepresents both what was stated in the article and what was shown in the webinar.

WmBriggs reports multiple dozens of his covid-skeptic readers asking for his analysis of the "study", and his reply is "don’t think there’s as much to it as some are hoping."

Did the Lyman Stone twitter-thread contain anything other than ad-hominem?

Assistant Village Idiot said...

Thank you Douglas. Stone is a PhD demographer at McGill and I have seen him show his work on other threads. His comment certainly suggests he did some looking at the data here, but he doesn't show his work that I can see. I followed lots of the threads of others and felt I was little wiser for the experience.

Assistant Village Idiot said...

I should have mentioned that a pdf of the study is still up online. I have not gone to look at it.

Sam L. said...

I am in the "over 70" cohort, but I live in a very rural area, so I am not worried.

Grim said...

I posted about this report, quite briefly, to invite discussion into the central claim. I don't think I made any conclusions about it in the post, but I am interested in trying to figure out what's going on with all of this. In the spirit of your narrative response here, let me lay out some background things too. I'll have to break it up into a few comments because there's a lot to say.

1) Personal history. I am not a mental health worker, but I have dealt with 'excess death' studies once before. This was the Lancet study of excess deaths from the Iraq War, published right before the 2004 election. In that case, the issue was a titanic confidence interval, which the media reported at the mid-point rounded up as if that were a real number. In fact, the study didn't really know how many people had died; it didn't have bodies to count, to use your metaphor. It had a guess about how many bodies there were, somewhere between 8,000 and 194,000.

But that wasn't the point, really. The point was political power. Here as there, reporting on deaths is meant to persuade us to yield power over our lives -- and to the same people, I notice.

Grim said...

2) Deaths and bodies. You use an interesting metaphor, a mafia killing of a guy who had a heart attack while being stabbed. There's a more real-life example readily available: George Floyd.

Is George Floyd dead? Yes, definitely. Did the police kill him? Well, arguably; he died in their custody, and might have gone on living if they hadn't tried to arrest him. Or maybe not; he had a lot of drugs in his system. Was his death because of the drugs, or the police, or were the police a kind of extra stressor on an already-taxed body that pushed him over the edge? Once we know what we think about that, what should we do about it?

That seems to me to be the thing this report was trying to get at. Lots of people may be dying "with" COVID; how many are dying "from" COVID? Is it the case that a lot of people who would have died anyway are being classified as COVID deaths? If so, the report makes sense on its own terms: we'd expect to see deaths from 'other causes' drop as COVID deaths rise. Yet then you can't make sense of the excess deaths.

Alternatively, is COVID an extra stressor that's killing people early, like cops kneeling on your neck adds stress to a body? If so, the 'excess deaths' may be deaths being stolen from 2021; then you can make sense of the excess deaths as well as the way that COVID rise and non-COVID decline mate up like puzzle pieces. IF so, 2021 may be a good year for statistics. Like NYC we can enjoy a lull in death rates because we already killed off the vulnerable.

The excess deaths being reported are also widely off from the rate we're told COVID would kill people. If that's true, deaths from non-COVID causes should be UP, not down, which doesn't make sense with these reports. I also posted just yesterday a report from Japan showing suicide rates spiking to levels not seen in years. Yet, as a commenter points out, it's only a few hundred deaths.

So which theory is true, or is there another theory that makes sense of it? I don't know.

Grim said...

3) Pragmatics and power. There's a real need to know what's true, because there's a group of power hungry would-be tyrants using this and everything else they can to grasp for power. These people would need to be resisted even if the facts were much worse than they are, though, which reduces pressure to 'make the facts fit.'

But there's also pragmatic difficulties. My mother spent Thanksgiving completely alone, eating dinner by herself in her little house. I spent the time on the phone with her (in fact, I'm not sure how she managed to eat her Thanksgiving dinner, since I can't remember her ever stopping talking long enough to chew). She had plenty of family nearby; within a mile, her daughter and grand-daughter and sons-in-law were all gathered. But she felt like the virus was too dangerous, and she had to shut herself in to be safe.

In Canada, I read this weekend, a lady had herself euthanized (as is legal there) because she didn't want to face another lockdown. She couldn't stand the idea of being lonely again. If she'd been allowed to be social, she might have gotten sick and died; so to protect her, doctors killed her at her request. No risk that way.

These considerations add weight to the concern about whether COVID is being over-sold. I don't know if it is, but I do know that the picture we are being presented is one-sided in a way that encourages these negative externalities.

Aggie said...

Thanks for this, AVI it's a superbly crafted argument. We live in an age of disinformation and too much of that sometimes makes it difficult to suss out what is creditable. The very nature of the internet makes it a 'volume' business, and therefore quality must be sifted for when chasing stories.

Whatever the mortality data says, whatever the policy makers decree, whatever the politicians use for their power plays, the stunning issue for me in the pandemic is the lack of materials being created to assist the individual in understanding their personal risk. This should be a straight-forward exercise, a simple programming and internet site-building task to create a place where a person can go to help them articulate and understand their exposure. How hard could it be to create such assays? I haven't seen one yet. Shouldn't one be able to supply basic demographic information, basic health information, geographical location and perhaps a short behavioral survey to get a return risk evaluation? I don't think that's unreasonable. Given the widely ranging COVID conditions across the nation, I don't think it's without merit, either.

Instead, we all individuals are being told to err on the side of extreme caution, all of us, and in some places this comes with threats of isolation in one form or another - even withholding of basic needs, water and electricity. At the bottom of this is the understanding that we should not be considering ourselves as individuals, only as members of a collective group. And therefore, for many the premise of mandate is shot down before even launching. It certainly is as far as I'm concerned.

My mom is 92 and has been in Assisted Living isolation since last March. We recently got her (and ourselves) a Facebook Portal so that we can at least look at each other in approximate life-size, real-time. It has made a big difference (these things are great, incidentally). It could also make a huge difference if the policy information being broadcast had indications of common sense, of data support, or sensibility, and yes, of some kind of sensitivity to the human condition. All of that aside - it is disheartening when our public policy leaders abandon the commitment to integrity and start issuing policies that favor demographic groups on the one hand and penalize others on the other hand. That cannot stand; Logic and human history tell that tale amply. If nothing else, the loss of the Public's trust in the face of perceived tyranny, in a time of hardship and need, never leads to constructive outcomes.

Donna B. said...

News outlets are looking for the worst news and they'll twist and spin, selectively quote, and then ask a question that presumes the answer they prefer. Often there's not a link to the actual study they are reporting on, just another report about it. Then someone on Facebook or Twitter links that article and adds another layer of "interpretation". Because of that, I have a little bit of sympathy for those who have decided it must be a political hoax, since that conclusion is supported by all the ways the data is twisted -- it can't all be true.

And then there's the matter of the death certificate and which information on it is used how. Then, there's the matter of how conscientious the person filling out the death certificate is that particular day -- which was a problem long before C19. There was a link on Instapundit about that problem and how it's now complicated by C19 and the effort to get all the C19 data possible.

@Grim - I imagine your mother thought long and hard before deciding to stay home. I know I did before I decided not to stay home. The deciding factor for me was the grandchildren I only get to see 3 times a year without a pandemic. I wasn't willing to give one of those up.

@Aggie - I've come to despise the phrase "out of an abundance of caution".

Christopher B said...

The CDC is saying at a midpoint 300,000 excess deaths. It's not enough to say "I don't trust 'em." Find me better numbers.

The JHU News-Letter says this "According to the CDC, there have been almost 300,000 excess deaths due to COVID-19." The phrase "300,000 excess deaths" links here -

The headline on that page is "United States COVID-19 Cases and Deaths by State" with a summary total at the top. Total deaths in the US since 21 January to 30 November 2020 are 266,051.

No where on this page can I find the word 'excess'. It appears to be simply a tabulation of the total number of COVID deaths to date by state.

I did find this CDC paper published in October this year that discusses the number of excess deaths from all causes to date

A summary blurb of the report states the following

"As of October 15, 216,025 deaths from COVID-19 have been reported in the United States; however, this might underestimate the total impact of the pandemic on mortality.

Overall, an estimated 299,028 excess deaths occurred from late January through October 3, 2020, with 198,081 (66%) excess deaths attributed to COVID-19. The largest percentage increases were seen among adults aged 25–44 years and among Hispanic or Latino persons."