3 ms·

Which isn't weird at all - if you trust that people avoiding crowds also lead to fewer having the accidents which would otherwise have gotten them dead.

ajross 6y ago
Or prevents them from contracting any other disease that would have otherwise killed them. In fact even these numbers are undercounts of the "true" death total due to the pandemic, as we'd obviously expect baseline deaths to have dropped significantly over the short term. How much is for the next generation of epidemiologists to argue about. But FWIW: Sweden is absolutely an outlier here, and given the choice between "Swedish reporting is fundamentally better" and "Swedish death counts are wrong in some banal way", I'll choose the latter. Say it with me everyone: we don't make risk analysis decisions on the basis of outlier data.
- henrikschroder 6y ago
  Mortality data and excess mortality data goes back decades, it's not like we suddenly started counting these things yesterday because of the coronavirus. For Sweden's death count to be wrong now, they would also have to have been wrong going back decades, without anyone noticing. Or, the total all-cause death count would have to start being wrong about now because of unspecified reasons. That doesn't make sense.
  - ajross 6y ago
    Doesn't the logic go the other way around though? For Sweden's numbers to be unimpeachible now, they would have had to have never been wrong going back decades. But in fact public numbers get reported incorrectly all the time, they just get corrected. This is happening everywhere right now, watch all the various numbers on Worldometer jump around like crazy. What you're doing is pointing to this one, single number, as collated by the New York Times in one story, which doesn't match any other numbers in the same data set. And you'redemanding that we treat it as proof of your particular policy views on this one country. And I'm saying: well, OK, that could be true. But it's probably just a mistake.
    - henrikschroder 6y ago
      There's a huge difference between all-cause mortality numbers being wrong, and covid-19-attributed deaths being wrong. The only way the all-cause mortality numbers can be significantly wrong is if you suddenly find piles of corpses and skeletons around, and that they belong to people who magically weren't being missed or noticed by friends, family, neighbours, etc. And the only way for governments to hide these numbers would be by digging mass graves, shoveling them full of dead bodies, and hoping no-one notices. This simply does not happen, modern countries do not get this number wrong. If there's any public statistic you can count on going back centuries, it would be the number of births and the number of deaths. Excess mortality is simply graphing the number of deaths backwards through time, to find trend curves and spikes in the trend curves. Those numbers are as correct as the underlying numbers. In comparison, correctly attributing a death to covid-19 is very difficult right now. Tests are in short supply, so a lot of countries aren't testing already dead people, because those tests can be better used elsewhere, to save lives. There might be political pressure to downplay the severity of covid-19, so cause of death gets fudged here and there. Cause of death is also not always straight-forward, a lot of people are dying with covid-19, not from covid-19, how should they be counted? Different countries have different rules here. The NYTimes article is saying that a bunch of countries have unexplained excess mortality, quite a lot of it actually. And since the most likely cause of those deaths are in fact covid-19, it means that those countries are under-reporting, i.e. their numbers are garbage. In comparison to those countries, Sweden's covid-19 deaths align pretty well with Sweden's excess deaths. The official number is still probably wrong, but there's absolutely no way that Sweden's covid-19 numbers are more garbage than those other countries, and there's absolutely no way that Sweden is under-reporting to the same extent.
      - ajross 6y ago
        > The only way the all-cause mortality numbers can be significantly wrong is if you suddenly find piles of corpses and skeletons around Or if someone types a number incorrectly; or the reporting deadline changes for a month causing a few offices miss it and get counted in the wrong month; or if the Times writers didn't understand the specific categories being reported and missed some data... Again, I'm not saying you're wrong. I'm saying you're hanging your entire argument on one outlier number. And, statistically (see what I did there?) that doesn't tend to work out for you.

ajross 6y ago

Or prevents them from contracting any other disease that would have otherwise killed them. In fact even these numbers are undercounts of the "true" death total due to the pandemic, as we'd obviously expect baseline deaths to have dropped significantly over the short term. How much is for the next generation of epidemiologists to argue about. But FWIW: Sweden is absolutely an outlier here, and given the choice between "Swedish reporting is fundamentally better" and "Swedish death counts are wrong in some banal way", I'll choose the latter. Say it with me everyone: we don't make risk analysis decisions on the basis of outlier data.

henrikschroder 6y ago
Mortality data and excess mortality data goes back decades, it's not like we suddenly started counting these things yesterday because of the coronavirus. For Sweden's death count to be wrong now, they would also have to have been wrong going back decades, without anyone noticing. Or, the total all-cause death count would have to start being wrong about now because of unspecified reasons. That doesn't make sense.
- ajross 6y ago
  Doesn't the logic go the other way around though? For Sweden's numbers to be unimpeachible now, they would have had to have never been wrong going back decades. But in fact public numbers get reported incorrectly all the time, they just get corrected. This is happening everywhere right now, watch all the various numbers on Worldometer jump around like crazy. What you're doing is pointing to this one, single number, as collated by the New York Times in one story, which doesn't match any other numbers in the same data set. And you'redemanding that we treat it as proof of your particular policy views on this one country. And I'm saying: well, OK, that could be true. But it's probably just a mistake.
  - henrikschroder 6y ago
    There's a huge difference between all-cause mortality numbers being wrong, and covid-19-attributed deaths being wrong. The only way the all-cause mortality numbers can be significantly wrong is if you suddenly find piles of corpses and skeletons around, and that they belong to people who magically weren't being missed or noticed by friends, family, neighbours, etc. And the only way for governments to hide these numbers would be by digging mass graves, shoveling them full of dead bodies, and hoping no-one notices. This simply does not happen, modern countries do not get this number wrong. If there's any public statistic you can count on going back centuries, it would be the number of births and the number of deaths. Excess mortality is simply graphing the number of deaths backwards through time, to find trend curves and spikes in the trend curves. Those numbers are as correct as the underlying numbers. In comparison, correctly attributing a death to covid-19 is very difficult right now. Tests are in short supply, so a lot of countries aren't testing already dead people, because those tests can be better used elsewhere, to save lives. There might be political pressure to downplay the severity of covid-19, so cause of death gets fudged here and there. Cause of death is also not always straight-forward, a lot of people are dying with covid-19, not from covid-19, how should they be counted? Different countries have different rules here. The NYTimes article is saying that a bunch of countries have unexplained excess mortality, quite a lot of it actually. And since the most likely cause of those deaths are in fact covid-19, it means that those countries are under-reporting, i.e. their numbers are garbage. In comparison to those countries, Sweden's covid-19 deaths align pretty well with Sweden's excess deaths. The official number is still probably wrong, but there's absolutely no way that Sweden's covid-19 numbers are more garbage than those other countries, and there's absolutely no way that Sweden is under-reporting to the same extent.
    - ajross 6y ago
      > The only way the all-cause mortality numbers can be significantly wrong is if you suddenly find piles of corpses and skeletons around Or if someone types a number incorrectly; or the reporting deadline changes for a month causing a few offices miss it and get counted in the wrong month; or if the Times writers didn't understand the specific categories being reported and missed some data... Again, I'm not saying you're wrong. I'm saying you're hanging your entire argument on one outlier number. And, statistically (see what I did there?) that doesn't tend to work out for you.