Hacker News Reader

Comment by anm89 | Hacker News Reader

When I hear people with anti-science attitudes I think back to things like this where an attempt to get a catchy headline published, misleadingly states something that obviously isn't true. So is it a wonder that people are dismissive of the results of other scientific research like climate change research?

Overwork *CONTRIBUTES* to the death of more than 745k per year. It did not solely and directly cause those deaths. How would you even know if it did? But the claim of the title is that overwork killed them as if overwork popped out behind them with a gun and shot them. It's just obviously a nonsense claim. And once you put it this way, tons of things CONTRIBUTE to excessive deaths. Poor diet, exposure to many everyday chemicals, living in a city(air pollution).

In fact, what are the odds that overwork doesn't correlate with living in a higher air pollution environment? Anyway, the claim in the title is nonsense.

View on HN

This reminds me of a comment that I read years ago and still think of every now and then:

> A few months ago a person of the ones you mention in your first paragraph posted on FB, as a blow against religion, that religion was so unreasonable that parents had to train their kids since youth in order to believe. And I remember thinking at the time about all the years of training needed to get any non-superficial commanding of science.

https://philosophy.stackexchange.com/a/29763

Edit: I should add that I believe the title is justified in this case. The study[1] identifies overwork as a cause for developing deadly diseases. Without overwork, the given number of people wouldn't have died of these diseases. Unhealthy diet and other factors are counted as a consequence of overwork. It could be phrased as '3.6% of stroke deaths and y% of ischemic heart disease deaths could have been avoided by working less'.

[1]: https://www.sciencedirect.com/science/article/pii/S016041202...

> get any non-superficial commanding of science.

Oh you mean like everyone on Social Media who think they understand Science but never actually practiced it once in their lives?

"Science" gets thrown around a ton to refer to about 5 billion different things. There's the scientific philosophy; the results of an experiment; the consensus of a peergroup; the entirety of all disciplines that require consistently, repeatable evidence before thinking of making a claim; the entirety of all disciplines that think one-off but repeatable in theory evidence is a-ok; baking soda volcanoes...

Some other really weird ones: engineering, design, art, math

And my favorite: programming

I generally find conflating any two of these different things to be problematic enough that "science" is a useless word.

Your statement really resonate with me as a scientist (phd) and a father. I want my kid to learn science at an early age (not baking soda volcano).

But more how would you go about testing an hypothesis (design of experiments) how do you know if the result are statistically significant? if it’s statistically significant is it enough to say it’s true (light can behave like a particule and like a wave so is it both or neither ...)

You want your kids to learn the scientific method [1], not "science" ;)

[1] https://en.wikipedia.org/wiki/Scientific_method

exactly! I think most teacher just teach random fact about plants weather molecule ... and think they did a great job.

But they rarely teach the scientific method.

Don't knock the baking soda volcano. I made one with my son when he was about 7 or 8. There's a little engineering that goes into constructing the paper mache structure. A little art in painting it to look realistic. Some chemistry in understanding the release of carbon dioxide from acid-base reactions. Then there's the filming/presentation of the project, for a little media savvy.

+1 if you can get your hands on some ammonium dichromate for an even more impressive, but more dangerous demonstration.

These types of "experiments" are less about teaching science, and more about impassioning a love and interest for science.

"chemistry in understanding the release of carbon dioxide from acid-base reactions." This is exactly the part I think is less important.

My goal is not for my kid to learn a subset of chemistry theory from the explanation in a book or some fun experiment.

Instead my goal is for them to understand why the human race now have this knowledge. How did the first chemist came up with the theory, how did they verify the theory was correct ... why do we now believe this is how atoms work.

The meaning of the word has changed substantially, a long time ago science was "what is known" or "knowledge acquired by study". [1] There was a shift in the English language in the 19th century, when the Scientific Method was formalized, and science communities grew and became influential. The meaning of science was narrowed down to "these studies that use the scientific method", with philosophy, maths and others being left out more and more. This isn't true in other languages/science traditions though. In German for example, all fields share the word "Wissenschaft", which literally translates to science.

[1] https://www.etymonline.com/word/science

Funny story about this. When I was a kid we had the whole "design a science experiment" in our science class. My dad was an electrical engineer, so I did an experiment with connecting up a clock through resistors to drain batteries more quickly than normal to test which batteries lasted longest (IIRC, I think the Walgreens brand ended up being best).

When I went to my teacher with the experiment, she laughed at me and told me it was absurd, because batteries wouldn't drain that quickly; she was really quite rude and aggressive about it. I'm still not sure what she was thinking; I had to contradict her multiple times and say that my Dad was an electrical engineer several times; she was still skeptical.

I learned while young that people in positions of authority, even when nominally "credentialed" in science, are often completely incompetent, not curious, not open-minded, and are instead arrogant and patronizing. She thought she knew what the answer was and she hadn't seen it before, so she just ridiculed me. This was in one of the best school systems in the country, in one of the best schools in that district.

> When I went to my teacher with the experiment, she laughed at me and told me it was absurd, because batteries wouldn't drain that quickly; she was really quite rude and aggressive about it. I'm still not sure what she was thinking; I had to contradict her multiple times and say that my Dad was an electrical engineer several times; she was still skeptical.

This would only be surprising if your teacher was a Physics teacher. If you want to be extra charitable a Science teacher. Most people don’t care about it understand the scientific method and it’s a profoundly unnatural way of thinking, much like doing formal logic.

> This would only be surprising if your teacher was a Physics teacher.

I'm not sure. Physics teachers often have to deal with absolute laws of physics, e.g. "you can't make a perpetual motion machine".

You can't ever demonstrate that something doesn't exist (philosophy of the black swan/inviting disproof by counter-example), so it's challenging for teachers to communicate this as a fact to curious students: the result is sometimes teachers appearing scornful.

One slightly-misunderstood law of conservation and you're suddenly a physics teacher who is scorning the truth.

I learnt a similar lesson about teachers early on at school and extended it to all classes of authorities through life. See also Are Experts Real?[1]

[1]: https://fantasticanachronism.com/2021/01/11/are-experts-real...

> people in positions of authority, even when nominally "credentialed" in science, are often completely incompetent

That's genuinely a baffling lesson to take from that incident. I mean, your father was the credentialed one, with an engineering degree, and the teacher was the unqualified one who was just a science teacher. Surely the lesson should be the exact opposite?

> Overwork CONTRIBUTES to the death of more than 745k per year.

I understood the headline figure to be suggesting that 750k people per year around the world are dying early, i.e. 10 years earlier than they would otherwise.

(Unlike you, I find it unsurprising and likely true to the point of not being headline-worthy: this is not very many people against the working age population)

The original article doesn't quite refer to that definition, but does say that 23.3 million Disability Adjusted Life-Years are lost per year, so it does seem to be working on an approximately equivalent basis.

Saying that everything around us "contributes" to death is pointless hyper-factualism. It's useful to have a scientific attempt to quantify how much each factor contributes.

At the same time, I don't think anybody would have issues with claims like "smoking kills X people a year", even though it's also a relationship that isn't 100% direct. The two claims do seem qualitatively different to me, but I'm not sure why. What's the concrete difference between "overwork kills X people a year" and "smoking kills X people a year" that makes one a sound claim and the other sound like nonsense?

I think smoking is a much more concrete relationship though, because you’re inhaling something which directly and physically damages your body. The tie to death is also more obvious because there aren’t other factors which could make a difference. If you smoke a lot, you’ll get lung cancer —- eating a good diet or exercising won’t stop the bad effects. Whereas with overwork, maybe if you have excellent routines like getting good sleep, exercise a lot, meditate, hydrate, etc, you’ll be able to combat the stress level. So you’re probably not quite as doomed as you might be with smoking.

Like, with smoking, it is the actual thing you consume which kills you. But with overwork, it’s not that you’re working $x hours too many, but that you’re super stressed out. (Even achieving low stress levels with normal work hours can be hard!) I think phrasing it as stress makes the tie more obvious to me because high stress levels are more obviously damaging.

I disagree. You can see clear effects from prolonged stress on the body. And in both cases, people still get lung cancer without smoking, and people can still die of heart attacks without being stressed.

I think the difference may be because we've been hearing about the harms of smoking for decades now. So when we hear "smoking kills X people" we understand they don't mean that smoking is the cause of death written on the death certificate, we understand that smoking has a demonstrable causal link to a variety of reduced lung health.

While most people understand that stress probably isn't good for you, we haven't had decades of ad campaigns and public health initiatives dedicated towards it. So the general public just needs time to catch up before we put stress in the same bucket as smoking, lead paint, and asbestos.

Actually "smoking kills X people a year" is also wrong. Yes, there is directly link between smoking and physical damage to the body. But two issues involved:

1. people usually do not instantly die after smoking. There is a long process from smoking to death.

2. people who smoke are different from people who do not smoke. There could be other contributing factors associated with people who smokes.

"smoking kills X people a year" was never a scientifically accurate claim. The accurate claim should be smoking increase * chance of ##, or on average people who smoke live ## years shorter.

However, there are cases where you can say smoking kills someone. For example, a woman who frequently smokes during the pregnancy gives birth to a baby born with severe life-threatening defects and the baby dies immediately after birth. In this case it could be OK to say the mother's smoking killed the baby, because:

1. There is a relatively short period of time involved.

2. There may be no other significant contributing factor to the birth defects, if we can rule out genetic factors.

Otherwise, in most cases, it is not scientifically accurate to claim smoking kills X people a year. In some cases, smoking actually could save people, who would otherwise use more detrimental means to release stress or anxiety.

What about "virus killed x amount of people last year", when the vast majority of these deaths were patients older than the median life expectancy with significant existing health problems prior to the infection?

I do agree people need to be more critical of science reporting. And often science needs to do a much better job at both sciencing and explaining the results to people who aren't experienced in interpreting them.

I think my two tests are still valid on the claim:

1. is the length of the period of those people from getting infected to death relatively short?

2. are there any other significant factors contribute to those people's death in that period?

Even if deaths were patients older than the median life expectancy with significant existing health problems prior to the infection, they can still pass those two tests, so most of them were still killed by the virus.

Just to be clear my question was rhetorical. You have to do an autopsy to have a reasonable guess about the cause. A 90 year old has a 13% chance of dying for any reason within a year (in the US, see here: https://www.ssa.gov/oact/STATS/table4c6.html). If it's not the virus it may be their liver, or their heart, or, and this is interesting now, chronic respiratory diseases. This has been a leading cause of death for a long time (https://www.cdc.gov/nchs/fastats/leading-causes-of-death.htm).

>are there any other significant factors contribute to those people's death in that period?

Multiple things can contribute. Our body is an extremely complex organism. Putting a death down to 100% this or that when in reality it's a lot more complicated isn't just imprecise, it's often flat out wrong.

And just because events coincide doesn't mean there's causality. If 1000 people died on a day with severe thunderstorms around the country that doesn't mean the thunderstorms killed them all. When old people with a virus infection die it's reasonable to assume there could be a connection, but you can't just say everyone who tested positive in a given time died of the virus.

For 90 year olds, the COVID caused death probably happens in a few weeks, instead of within a year, right?

I agree "events coincide doesn't mean there's causality", but I think the causation here is still pretty strong, even for 90 year olds.

> 1. is the length of the period of those people from getting infected to death relatively short?

What about HIV/AIDS? Does it kill anyone?

The methodology accounts for this. If nobody was overworked worldwide, there would be 745k less dead people.

Yeah this is the scary part about this. These claims embolden people who don't take the time to read the work or don't understand it to make outrageous unscientific claims like we know for a fact that there would be 745k less dead people if x y or z.

This then further emboldens the misinformed people on the other side of the debate to be categorically anti science because they have easy go to evidence of "scientific claims" that are easily debunkable.

Whereas this absolutely is not even what they are claiming and it would never be in the realm of science under any circumstances to make this claim the way you are stating it.

> Fig. 1. Conceptual model of the possible causal relationship between exposure to long working hours and ischemic heart disease and stroke.

All they are doing is tying a statistical extrapolation of some data regarding the correlation of prevalence of ischemic heart disease and stroke with length of work week and then extrapolating, proposing a POSSIBLE CAUSAL MODEL for this and then tying this first layer of data to

> ESTIMATES of population exposed to long working hours

They absolutely did not systematically prove causality of anything and certainly not by a long shot did they prove the causality of overwork in specific numbers of deaths.

Also check out their own ratings of the confidence in the data they used:

> Ischemic heart disease (Li et al., 2020a) c 41–48 hours/week 20 studies (312,209 participants) 0.99 (0.88–1.12) Low quality Inadequate evidence of harmfulness No

49–54 hours/week 18 studies (308,405 participants) 1.01 (0.82–1.25) Low quality Inadequate evidence of harmfulness No

≥55 hours/week 22 studies (339,680 participants) 1.17 (1.05–1.31) Moderate quality Sufficient evidence of harmfulness Yes

Stroke (Descatha et al., 2020) d 41–48 hours/week 12 studies (265,937 participants) 1.04 (0.94–1.14) Low quality Inadequate evidence of harmfulness No

49–54 hours/week 17 studies (275,181 participants) 1.13 (1.00–1.28)e Moderate quality Limited evidence of harmfulness No

≥55 hours/week 7 studies (162,644 participants) 1.35 (1.13–1.61) Moderate quality Sufficient evidence of harmfulness Yes

Most of their data is by their own admission judged to be of low quality for proving this point and in two cases over a high threshhold they established that moderate quality data was sufficient to establish "evidence of harmfulness"

Basically it is difficult to communicate how many different levels you are incorrect on.

Good grief.

You're referring to a headline from an online magazine health section, not "science," nor even the study you are criticizing (which you clearly haven't even glanced at).

> *CONTRIBUTES*

What's the insight here? Do you really think that people can possibly die from "overwork," full stop, or are you being deliberately obtuse for the sake of wedging your politics into the discussion?

> In fact, what are the odds that overwork doesn't correlate with living in a higher air pollution environment?

Somehow this still works on HN.

It comes down to semantics of “kill”. I wonder if there's anyone can believe long work hours are the sole cause of mortality, regardless of whether they read the title.

The usual interpretation, long work hours are the major factor in the mortality of 745k people, is a valid interpretation of the research results.

To say it only contributes to mortality along with many other factors, without pointing out the portion of correlation, is to say nothing at all, which isn't what the report tries to communicate.

Can you find anyone who reads that headline and thinks overwork popped out behind them with a gun and shot them, or anything remotely like it?

> "It's just obviously a nonsense claim"

Yes, obviously, and even more obviously after reading the article and the linked WHO press release which says "The study concludes that working 55 or more hours per week is associated with an estimated 35% higher risk of a stroke".

Attacking straw people based on a title of a newpaper headline isn't pro-science.

I am clearly not proposing that they are claiming an abstract concept held a physical weapon to kill anyone. Actually that's the point though. You would have to believe something nonsensical like this for it to be possible for the headline to be true as it is written.

The headline is "Overwork Killed x" and that implies in no uncertain terms that overwork directly and solely killed people which is again, obviously nonsense.

The attribution effect is implied, is the point the parent was making, and while it's good to be aware of your point, I personally feel like the title is the least important part to critique given the obvious implication. If your point carries into the analysis or synthesis, that is a problem.

> Conclusions — WHO and ILO estimate exposure to long working hours (≥55 hours/week) is common and causes large attributable burdens of ischemic heart disease and stroke.

The study itself uses the word "causes", so I I guess some people misread it to say it causes heart disease and death, rather than contributes to.

Factor analysis is a thing often used in health science and is hardly controversial. Sure, you have to control and be careful with causality - but contribution certainly is possible to establish.

> *CONTRIBUTES*

Thank you for pointing that out. It's an easy mistake to make given how very often one can see that terrible disease, "Overworkitis," on death certificates... /s

Not to be too cynical, but I wonder what the result would look like if they were able to do the math all the way. It seems like they can tell us how many people die early from heart disease due in part to overwork, but we don't know how the value provided by their work extended the lives of others. For example, if overwork created a higher GDP, which resulted in better nutrition, medical outcomes, etc. maybe the net effect of overwork is positive. (Now that I think about it, that could get very dark if we embraced the idea, though.)

You'd probably also find some who's early demise would be considered socially beneficial. However, I think it would be hard to know good or bad impacts without extreme and obvious examples. We've also had heroes who looked to be doing a lot of good throughout their life, only to turn out to be monsters behind closed doors.

Also something like GDP is very macro, and not all GDP activity is socially beneficial, so GDP contribution would be a flawed way to make such an assessment. For example, someone scamming seniors out of their retirement money through shady business tactics wouldn't be a net positive to society, but could contribute millions to GDP.

That would be interesting to see. I for one have no real idea what the result would be.

My $0.02 based on nothing more than personal experience is that it seems to be a net negative in practice. I'm sure that some degree of overwork in certain industries has raised GDP, but there are also notorious examples of systemic overwork reducing productivity, like medical providers (mistakes) and Japan (Parkinson's Law), which seem to resonate more with my experience.

Who knows, though? It could easily be the opposite, for all I know. It would be interesting if someone was able to do this analysis.

Anti-science is a meme. Everything 'science' the general public reads is political, and if you disagree with their politics you're anti science. In STEM there's a blind spot where its easy to forget that the general public aren't equipped to think analytically 24-7 and they are preyed on constantly.

No. It is not all political.

How many people did overwrought headlines kill?

How many people didn't get vaccines because they live in a culture that is skeptical of the validity of science. And if this is the science they are getting exposed to, I don't blame them for being skeptical.

Plus in a world population of about 8 billion, the WHO has nothing better to do than worry about factors "contributing to" the deaths of 745k people? That should be considered a embarrassing waste of effort.

Maybe they can focus their smarts by announcing a pandemic a month earlier next time.

It wasn't a pandemic back then because they no longer had a definition for pandemics. We're lucky they found the dictionary when they did.

> Overwork CONTRIBUTES to the death of more than 745k per year. It did not solely and directly cause those deaths.

Try telling HN to remember this when talking about covid death figures.

> Overwork CONTRIBUTES to the death of more than 745k per year. It did not solely and directly cause those deaths

Slightly off-topic, but the truth is the same can be said of Covid 19. That is, presence of or correlation is not necessarily causation. The best example is, people who were morbidly obese; with high blood pressure, Type 2 Diabetes, other complications, etc. Yet Covid gets sole credit for the deaths?

All that aside, a lack of time (e.g., from working a lot) leads to "compromises." Fast food, not exercise, no mental downtime, etc.

This is an unreasonable standard to apply to any cause of death.

Take cocaine overdose, causing a heart attack. Might have had a heart murmur, asthma, obesity, and it's always possible that without those conditions the patient would have survived. Cause of death was still a cocaine overdose.

There's a separate question, which shouldn't be conflated, whether hospitals were over-reporting COVID deaths because of financial incentives. It's possible there was some of that, but while you can move numbers around a bit, you can't hide death itself, and the US had substantially more deaths last year than would have been expected had the pandemic not come along.

Worth noting that by that standard, overwork doesn't cause death in any case. People die of strokes, heart attack, disease, but they don't die of overwork: or smoking, for that matter.

The study of what contributes to mortality has little to do with the study of what causes it, is what I'm saying. COVID is assuredly a case of the latter.

What is the financial incentive for reporting a COVID death vs something else? Why does such incentive exist? (I am assuming this is a US thing)

"Two weeks to flatten the curve" was due to fears that hospitals were going to be swamped with thousands of covid patients. Not only would there be the issue of not enough beds, ventilators, and trained medical personnel, the hospitals would quickly run out of funds to operate. The government wanted to keep this from happening so they created a fund that would pay hospitals money quickly for each covid patient they treated.

This was to avoid the time consuming process of dealing with insurance claims and the uninsured. It unintentionally created a perverse incentive for hospitals to claim as many deaths as possible as covid deaths. That doesn't mean that all or even most hospitals are doing this, at least not intentionally but everyone working in the hospital is aware that it needs financial stability. Even if they're not intentionally sticking their thumbs on the scale to attribute deaths to covid that shouldn't be, there's going to be a bias for it.

As for the amounts, you can probably find those with a web search but be aware that there's lots of bad information out there so be careful about the source you check.

Unreasonable? Why? The Covid hyperbole has beaten up the data (read: facts) beyond recognition. That is far from reasonable.

- Why was the fact that Covid disproportionately affected those 65 & over brushed aside?

"Adults 65 and older account for 16% of the US population but 80% of COVID-19 deaths in the US, somewhat higher than their share of deaths from all causes (75%) over the same period."

https://www.kff.org/coronavirus-covid-19/issue-brief/what-sh...

Note: Covid rates and deaths are falling off because people 65 & over have gotten the jab. Not necessarily because other less vulnerable groups have; including children.

- Why was the fact that comorbidies - a medical science standard - suddenly abandoned?

https://www.msn.com/en-us/health/medical/cdc-94percent-of-co...

- Why was Covid's disproportionate affect on communities of color brushed aside?

https://www.psychologytoday.com/us/blog/talking-about-health...

- Why was the response to Covid and its contribution to an increase in poverty brushed aside?

https://www.nbcnews.com/news/us-news/8-million-americans-sli...

https://www.brookings.edu/blog/future-development/2020/10/21...

----

Hiding behind aggregation and hyperbol in order to scare Karen is a dangerous tactic; a spin far too many people accepted with out question. The problem with fear is its limits the field of vision - literally, as well as to facts beyond the narrative.

Yes. Covid exists. That said, in terms of the response and data transparency there has been a significant amount of unreasonable manipulation.

Please don't confuse how "popular" a fact is with its legitimacy. Just the same, the popularity of propaganda doesn't increase its legitimacy.

The burden of causally connecting a testable covid case to death in the short term from known symptoms of that viral infection and the burden of causally tying a subjective concept with debatable criteria like "overwork" to specific deaths is night and day different.

In either case, in all cases, correlation is not causation. We can debate the degrees of difference between overwork and Covid but speaking of correlation as if it's causation is unacceptable. Simply because one agrees with the correlation doesn't magically make it causation. Yes, this attitude has been normalized, but that doesn't make it correct.

Reply on news.ycombinator.com