2 votes

The NYT's partisan tale about COVID and the unvaccinated is rife with sloppy data analysis

15 comments

  1. [12]
    moriarty
    (edited )
    Link
    The guy is exaggerating the confounders beyond belief. Yes, age is correlated with more Republican leaning, but if you check the studies he points to, the difference between Republican and...

    The guy is exaggerating the confounders beyond belief.

    Indeed, age appears to be the “strongest predictor of mortality” from COVID-19, with one’s risk of death increasing exponentially with age.

    Yes, age is correlated with more Republican leaning, but if you check the studies he points to, the difference between Republican and Democrat voters is 2% for 65+ and 4% for 50-64. Not really enough to explain away the differences the NYT article was pointing to. You know what is an even stronger predictor of covid mortality than age? Not getting the vaccine. And not getting the vaccine is very clearly political and correlated with partisan leaning. But he's right that correlation doesn't prove causation. You know how we can prove causation in this case? GOP representatives telling their constituents that covid is a hoax, that masks are a hoax and that the vaccine is a hoax.
    He then makes a point of cherry-picking Utah as an example to show that a red state with young population isn't impacted by covid. He of course fails to mention that Utah has one of the highest vaccine rates among red states, and has a staggering 92% fully vaccinated rate among the older population.

    Come on. This isn't a serious analysis of the NYT article. He spent more time self-righteously bashing the "liberal agenda" than he did seriously analyzing the data. Unsurprising that we find this on Glenn Greenwald's platform.

    27 votes
    1. [2]
      meff
      Link Parent
      If you're talking about the Pew Research page, Pew is aggregating over the entire US population. Looking at county level patterns would probably paint a much more complex picture. I'm also not...

      Yes, age is correlated with more Republican leaning, but if you check the studies he points to, the difference between Republican and Democrat voters is 2% for 65+ and 4% for 50-64.

      If you're talking about the Pew Research page, Pew is aggregating over the entire US population. Looking at county level patterns would probably paint a much more complex picture. I'm also not convinced with the example of Utah. That said, at an initial glance, the case of Texas looks pretty indicative to me of political party correlating with vaccine hesitancy, as Northern Texas has a young population yet a pretty low rate of vaccination. This is just an initial glance though, it's really just a scientific stab in the dark.

      It's not lost on me that this Outside Voices (is that what it's called? I don't really Substack so idk) piece is pretty bad. The line

      Gaba runs a Patreon page where he disseminates his writings to subscribers, which are not submitted to reputable peer-reviewed publications.

      really gave me a laugh when it's obvious that Substack follows the same model and this criticism piece isn't peer-reviewed either. But let's not use some skin-deep findings to try to fight this critique. If this piece is gaining notoriety, we can let other pieces bolster the analysis of the NYT or properly debunk this Outside Voices article, not throw pot shots at the NYT article's critics.

      9 votes
      1. moriarty
        (edited )
        Link Parent
        Yeah, I totally agree with you that there is a more complex picture for age-dependency. However, since I'm unwilling to dedicate more time to consider yet another cooky unsubstantiated claim...

        Yeah, I totally agree with you that there is a more complex picture for age-dependency. However, since I'm unwilling to dedicate more time to consider yet another cooky unsubstantiated claim (there's simply not enough hours in the day) by tracking the county-level dependencies, we can ballpark his claims with aggregated state-level dependencies. And the picture there is nowhere near enough to explain the extreme covid death rate differences between states. The difference in the fraction of elderly people between states is simply not that extreme. But even if it was, we can cherrypick another set - Vermont, Maine and New Hampshire are the top 3 states with the oldest population in the US and yet the lowest death rate in the country. And you're right, Utah, Alaska and Texas are the states with the youngest population and yet (with the exception of Utah) some of the highest death rates.

        7 votes
    2. [9]
      userexec
      Link Parent
      Tildes is probably the only place I see substack articles posted anywhere, and I feel like more often than not it's some edgy premise piece that ends up being apologism. Am I missing something...

      Tildes is probably the only place I see substack articles posted anywhere, and I feel like more often than not it's some edgy premise piece that ends up being apologism. Am I missing something about people's love of substack? The whole concept of subscription newsletters just seems outlandishly niche to be showing up so much, but then I get that there are plenty of ways to use the internet.

      9 votes
      1. spit-evil-olive-tips
        Link Parent
        yup. Substack is a Y Combinator funded company. their "editorial direction" is more or less in line with what you'd expect if Medium were run by Hacker News commenters. obviously, hashtag not all...

        more often than not it's some edgy premise piece that ends up being apologism.

        yup.

        Substack is a Y Combinator funded company. their "editorial direction" is more or less in line with what you'd expect if Medium were run by Hacker News commenters.

        In recent months, a slew of accomplished writers have migrated to Substack, including Matt Taibbi, Glenn Greenwald, Matthew Yglesias, Heather Havrilesky, Roxane Gay, Andrew Sullivan, Bari Weiss and others.

        obviously, hashtag not all substack writers...but it has become a pretty common home for the "professional contrarian" pundit class.

        12 votes
      2. [6]
        nothis
        Link Parent
        Substack is tied to the Slate Star Codex drama, which was mostly about solving the world through cold reason and logic (Slate Star wasn't that bad (if self-indulgent) and now found a home on...

        Substack is tied to the Slate Star Codex drama, which was mostly about solving the world through cold reason and logic (Slate Star wasn't that bad (if self-indulgent) and now found a home on Substack under the name Astral Codex Ten). That type of writing appeals to people disillusioned with the more irrational sides of human behavior but it's also a path towards stubborn wrongness if all that thinking is based on subtly flawed axioms. A lot of conservative authors have claimed that brand of writing for themselves by now, mostly starting from "the libs are wrong, I can feel it in my gut!" and then working their way backwards towards a "rational" sounding article. Substack is their safe space.

        9 votes
        1. [5]
          meff
          Link Parent
          Hm idk. I've certainly seen my share of Left stuff on it (I'm not a believer of Marxism myself but have read a lot of Marxist thought over the years) and will occasionally read some Substacks on...

          Hm idk. I've certainly seen my share of Left stuff on it (I'm not a believer of Marxism myself but have read a lot of Marxist thought over the years) and will occasionally read some Substacks on Marxist analyses of current events. I've also been linked some pretty cool posts on Substack in the past. I don't really use Substack frequently though so I don't know if these are outliers or the norm.

          3 votes
          1. [4]
            nothis
            Link Parent
            Well, of course they have their Marxists. It's coming from the "both sides" school of thought where there are basically no wrong ideas as long as a large (or powerful) enough group of people...

            Well, of course they have their Marxists. It's coming from the "both sides" school of thought where there are basically no wrong ideas as long as a large (or powerful) enough group of people believes them so you better give a platform to all. You know, like inviting a climate scientist and a coal mine operator to discuss not politics but climate science.

            I was one of those early internet people, optimistic about the internet as a platform of free thought and ideas. But it's become pretty clear that, without curation, you'll have it taken over by propaganda. People realize that platforms that can reach millions are powerful tools that can be abused. It ruined reddit, facebook and twitter and will ruin any place that claims to have "neutral" values over the long term. It's a naive interpretation of free speech that doesn't work.

            Substack is pretty young. I think there's some good writing on there, occasionally. But its appeal to "crowd sourced rationality" is a trap. True rationality is hard-fought and under constant attack. Universities have built up a strong culture of rational thought over centuries. Letting truth_seeker_69 run his own blog without opinionated oversight won't lead to anything good. That's why I'm skeptical of places like Substack praising its "lightweight" moderation. It won't end well and it's IMO not the future of quality online discussion. You need moderation. Heavy, opinionated moderation. Badly.

            6 votes
            1. [3]
              skybrian
              Link Parent
              Each substack is a separate newsletter with a separate audience. Generalizing across all writers who use Substack seems similar to generalizing across all bloggers? Taking the average is going to...

              Each substack is a separate newsletter with a separate audience. Generalizing across all writers who use Substack seems similar to generalizing across all bloggers? Taking the average is going to miss most of what's interesting about the writers.

              We had similar discussions when blogs became popular. Sure, anyone can start a blog and some blogs are bad. It's an old story.

              Not crossing the streams seems wise; you can support your favorite writers without feeling like you're supporting other, worse writers or whatever nasty audiences they attract. Like with other blogging, how good the moderation is will depend on the person running the blog. So far it seems decent for the ones I read.

              Inevitably, at some point there will be pressure to remove the worst substacks, but it seems less likely to taint unrelated writers if they're kept separate.

              3 votes
              1. [2]
                nothis
                Link Parent
                I'm not criticizing each individual writer, I'm criticizing substack as a platform that is designed to make reaching a large audience easier. Similarly, facebook would probably be harmless if it...

                I'm not criticizing each individual writer, I'm criticizing substack as a platform that is designed to make reaching a large audience easier. Similarly, facebook would probably be harmless if it made it harder to form groups over a certain size. As soon as you make it your goal to give a person the stage to preach to 1 million+ people with a the press of a button, you have some responsibility. Substack is denying that responsibility in its rules.

                It's ok to provide such a platform. But then you have to moderate.

                2 votes
                1. skybrian
                  Link Parent
                  What about the rules don’t you like? Looking at the Publisher Agreement and Content Guidelines, they’re clearly set up legally to do content moderation. They have the right to remove newsletters...

                  What about the rules don’t you like? Looking at the Publisher Agreement and Content Guidelines, they’re clearly set up legally to do content moderation. They have the right to remove newsletters at any time.

                  All content hosting platforms need to worry about abuse. How good a job will they actually do? I guess we will have to wait and see.

      3. mycketforvirrad
        Link Parent
        We tag all substack articles for this very reason, after some people requested an ability to be able to filter them from their feeds.

        We tag all substack articles for this very reason, after some people requested an ability to be able to filter them from their feeds.

        6 votes
  2. [3]
    skybrian
    Link
    These are good points, putting into words some of my own skepticism about graphs I’ve seen. But I’m a bit disappointed that they didn’t redo the data analysis and find out what the corrected...

    These are good points, putting into words some of my own skepticism about graphs I’ve seen. But I’m a bit disappointed that they didn’t redo the data analysis and find out what the corrected numbers are.

    Hopefully someone will actually do the homework.

    3 votes
    1. [2]
      babypuncher
      Link Parent
      I think it's because they know that the corrected numbers won't make red states look any better. Their own methodology is flawed, as this user pointed out

      I think it's because they know that the corrected numbers won't make red states look any better. Their own methodology is flawed, as this user pointed out

      3 votes
      1. skybrian
        Link Parent
        I don't think we know what the corrected numbers would look like. People can speculate, sure, and they might even guess right, but it's not a substitute for actually doing the homework.

        I don't think we know what the corrected numbers would look like. People can speculate, sure, and they might even guess right, but it's not a substitute for actually doing the homework.

        2 votes