5 ms·

Also consider age distribution/population density in areas you compare. Death rate is exponential in age, so small differences in age distribution gives big di

• asabjorn 5y ago
  Death rate to confirmed cases is TX 1.7% (51k deaths over 2.9m cases), Florida 1.56% (36k deaths over 2296 cases), CA 1.64% (62k deaths over 3.74m cases) [source: google coronovirus stats]. So data does not confirm your conclusion.
  • dragonwriter 5y ago
    Without stats on the age distribution of confirmed cases (even before considering the absence of analysis of confounding factors, liking case timing within the pandemic, etc.), I’m not sure how you can say the data doesn’t confirm the conclusion since you don’t have data on the proposed independent variable at all.
    • asabjorn 5y ago
      Yes, age distribution matters as older people are at higher risk. As some signal for this Florida is USAs old people home, so you would expect that if masks had a large positive mitigatory effect then Florida with its low mask use would do worse than CA and NY that has low mask use. However, we do not see this.
      • dragonwriter 5y ago
        > As some signal for this Florida is USAs old people home, so you would expect that if masks had a large positive mitigatory effect then Florida with its low mask use would do worse than CA and NY that has low mask use. Again, this has a problem of failing to address case timing and distribution, with CA and NY being hard hit and having COVID into the community before there was much awareness or any policy response (e.g., targeted control measures directed at elder facilities) or personal behavior response or evolution in treatment protocols, which radically changes the dynamic vs. states that weren't hit until after public awareness and health care competence was more advannced. And presumbaly, you mean “CA and NY with high mask use”, otherwise this argument is nonsense. OTOH, no one has cited any data supporting the “CA and NY are representative of high mask use states while FL (and TX, as claimed in other posts) are representative of low mask use” argument, which seems to be based on stereotyping based on dominant political parties in each state and leanings of political leaders of each party on mask mandates. What data there is on this doesn’t seem to support that this stereotyping accurately reflects mask use, OTOH; i.e., a study on mask use with self-reported data for May through August of last year had mask use as: AUG: CA 74.6% > FL 71.6% > NY 69.6% > TX 69.1% JUL: NY 80.6% > CA 74.3% > FL 66.9% > TX 66% JUN: NY 56.4% > CA 48.7% > FL 45.1% > TX 42.4% MAY: NY 46.1% > CA 45.2% > TX 43.8% > FL 43.5% APR: CA 38.1% > NY 36.1% > FL 35.0% > TX 32.4% And, notably, all four states were in the top half of the country in mask use in all 5 months with data. https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0249891 https://journals.plos.org/plosone/article?id=10.1371/journal...
        • asabjorn 5y ago
          I tend to trust state policy and political leanings more as an indicator then a self reported study. Can you please dig into why you trust this self reported study on the politically charged mask use question? Especially when the study is performed by academics that is perceived to be too often left-wing activists by republicans and many libertarians? (And according to Haidt have a provably largely left leaning bias) Self reported results are prone to preference falsification and hidden bias in whom answers. (“Yes, mom, I did my homework” ;) ) Self reported results of politically charged questions are effectively polls. To indicate the problem you can observed how inaccurate polls around politically charged choices has become when compared to actual choices. The normal problems with self reported results are probably aggravated by how academics, major authorities, media, and tech companies have openly without cover used force to promote the “right” answer and suppressed the “wrong”. The use of force to enforce specific conclusions had been an extraordinary and very noticeable break from western traditions.
          • dragonwriter 5y ago
            > I tend to trust state policy and political leanings more as an indicator then a self reported study. I don’t, in the absence of strong enforcement of policy (which did not exist in, e g., California COVID policy), especially on issues of strong partisan posturing; even solid red or blue states tend to be pretty closely balanced in the population; majoritarianism, reinforced by gerrymandering and related mechanisms, produces strong and durable political dominance from slight population imbalances, so “state dominant party” stereotypes and state policy vastly exaggerate differences in distribution of individual belief, preference, and, in the absence of effective compulsion, behavior. > To indicate the problem you can observed how inaccurate polls around politically charged choices has become when compared to actual choices. They...haven’t, really, become particularly inaccurate. Poll-based predictions of binary outcomes of things that poll very close to even have become somewhat less reliable than they very briefly were before (they were a fairly new practice compared to just reporting polls without predictions, anyway, so its not like there was a well-established baseline) because when polls are near even, very slight changes in accuracy have an outsize effect on binary predictions.