Vaccination nation

[secundem_artem]

...

Counties that voted for Trump have disproportionately lower vax rates & higher death rates.

Darwin at work.

If I understand correctly, the county averages are multiplied by county population. This biases standard errors, which are not reported, downwards.

Proper weighted regression would use population size to correct the variance.

Little need to get technical: Look at the scatter plots. The high Trump voting counties have death rates all over the map. That's a clue that the statistical analysis is pure genius, lucky, or wrong. :-)

Not as much noise looking at the lower vax rates in Trumpish counties.  You can argue fine points of methodology (that said, they had some number crunchers from Johns Hopkins as advisers) but the trends seem clear to me.  It's clear that vaccines do reduce the risk of hospitalization and death.  So lower vax rates can be reasonably correlated with higher death rates.

Beyond a reasonable doubt? Maybe not.  The preponderance of the evidence?  Yes.

[dismalist]

...

Counties that voted for Trump have disproportionately lower vax rates & higher death rates.

Darwin at work.

If I understand correctly, the county averages are multiplied by county population. This biases standard errors, which are not reported, downwards.

Proper weighted regression would use population size to correct the variance.

Little need to get technical: Look at the scatter plots. The high Trump voting counties have death rates all over the map. That's a clue that the statistical analysis is pure genius, lucky, or wrong. :-)

Not as much noise looking at the lower vax rates in Trumpish counties.  You can argue fine points of methodology (that said, they had some number crunchers from Johns Hopkins as advisers) but the trends seem clear to me.  It's clear that vaccines do reduce the risk of hospitalization and death.  So lower vax rates can be reasonably correlated with higher death rates.

Beyond a reasonable doubt? Maybe not.  The preponderance of the evidence?  Yes.

I'm not questioning a relationship between vax rates and death rates, but rather between Trump rates and death rates. [Just off the cuff, I was thinking that the Trump rates are a confounder! :-)] The newspaper article is to be appreciated for showing the scatter plots in an informative way. But it ain't good statistical analysis. Don't have to publish their detailed results in a newspaper article, but could refer to a working paper or something.

One mustn't think that everybody at any university knows everything. What's needed is arguments, not testimonials.

It's not preponderance of the evidence -- it's misleading at best, wrong at worst. My guess is it's meaningless.

[mamselle]

My guess is it's meaningless.

Eponymously...

;--}

M.

[nebo113]

...

Counties that voted for Trump have disproportionately lower vax rates & higher death rates.

Darwin at work.

If I understand correctly, the county averages are multiplied by county population. This biases standard errors, which are not reported, downwards.

Proper weighted regression would use population size to correct the variance.

Little need to get technical: Look at the scatter plots. The high Trump voting counties have death rates all over the map. That's a clue that the statistical analysis is pure genius, lucky, or wrong. :-)

Not as much noise looking at the lower vax rates in Trumpish counties.  You can argue fine points of methodology (that said, they had some number crunchers from Johns Hopkins as advisers) but the trends seem clear to me.  It's clear that vaccines do reduce the risk of hospitalization and death.  So lower vax rates can be reasonably correlated with higher death rates.

Beyond a reasonable doubt? Maybe not.  The preponderance of the evidence?  Yes.

I'm not questioning a relationship between vax rates and death rates, but rather between Trump rates and death rates. [Just off the cuff, I was thinking that the Trump rates are a confounder! :-)] The newspaper article is to be appreciated for showing the scatter plots in an informative way. But it ain't good statistical analysis. Don't have to publish their detailed results in a newspaper article, but could refer to a working paper or something.

One mustn't think that everybody at any university knows everything. What's needed is arguments, not testimonials.

It's not preponderance of the evidence -- it's misleading at best, wrong at worst. My guess is it's meaningless.

I may not always agree with you, but I appreciate and respect your clear, rational posts, which are neither snippy nor snarky.

[Sun_Worshiper]

...

Counties that voted for Trump have disproportionately lower vax rates & higher death rates.

Darwin at work.

If I understand correctly, the county averages are multiplied by county population. This biases standard errors, which are not reported, downwards.

Proper weighted regression would use population size to correct the variance.

Little need to get technical: Look at the scatter plots. The high Trump voting counties have death rates all over the map. That's a clue that the statistical analysis is pure genius, lucky, or wrong. :-)

Not as much noise looking at the lower vax rates in Trumpish counties.  You can argue fine points of methodology (that said, they had some number crunchers from Johns Hopkins as advisers) but the trends seem clear to me.  It's clear that vaccines do reduce the risk of hospitalization and death.  So lower vax rates can be reasonably correlated with higher death rates.

Beyond a reasonable doubt? Maybe not.  The preponderance of the evidence?  Yes.

I'm not questioning a relationship between vax rates and death rates, but rather between Trump rates and death rates. [Just off the cuff, I was thinking that the Trump rates are a confounder! :-)] The newspaper article is to be appreciated for showing the scatter plots in an informative way. But it ain't good statistical analysis. Don't have to publish their detailed results in a newspaper article, but could refer to a working paper or something.

One mustn't think that everybody at any university knows everything. What's needed is arguments, not testimonials.

It's not preponderance of the evidence -- it's misleading at best, wrong at worst. My guess is it's meaningless.

I wouldn't say it is meaningless or misleading. It is a piece of evidence that would be a starting point for doing a more sophisticated statistical analysis - one that would include control variables and thoughtful model specifications. That said, it is certainly the case that this is not particularly convincing. I actually do an exercise in my stats class where students find and present plots like this one from the news media and we break down what they can and cannot tell us as a class. I'd bet this is the thing that stays with students most after the semester is over.

[mahagonny]

Counties that voted for Trump have disproportionately lower vax rates & higher death rates.

Darwin at work.

Republicans are older than democrats.  https://www.pewresearch.org/fact-tank/2020/10/26/what-the-2020-electorate-looks-like-by-party-race-and-ethnicity-age-education-and-religion/
"The median age among all registered voters increased from 44 in 1996 to 50 in 2019. It rose from 43 to 52 among Republican registered voters and from 45 to 49 among Democratic registered voters."
Older people are much more likely to die when they get COVID than are younger ones.
Also, more men than women voted for Trump. Men have lower life expectancy.

ETA: Does living longer mean you're smarter?

...

Counties that voted for Trump have disproportionately lower vax rates & higher death rates.

Darwin at work.

If I understand correctly, the county averages are multiplied by county population. This biases standard errors, which are not reported, downwards.

Proper weighted regression would use population size to correct the variance.

Little need to get technical: Look at the scatter plots. The high Trump voting counties have death rates all over the map. That's a clue that the statistical analysis is pure genius, lucky, or wrong. :-)

Not as much noise looking at the lower vax rates in Trumpish counties.  You can argue fine points of methodology (that said, they had some number crunchers from Johns Hopkins as advisers) but the trends seem clear to me.  It's clear that vaccines do reduce the risk of hospitalization and death.  So lower vax rates can be reasonably correlated with higher death rates.

Beyond a reasonable doubt? Maybe not.  The preponderance of the evidence?  Yes.

I'm not questioning a relationship between vax rates and death rates, but rather between Trump rates and death rates. [Just off the cuff, I was thinking that the Trump rates are a confounder! :-)] The newspaper article is to be appreciated for showing the scatter plots in an informative way. But it ain't good statistical analysis. Don't have to publish their detailed results in a newspaper article, but could refer to a working paper or something.

One mustn't think that everybody at any university knows everything. What's needed is arguments, not testimonials.

It's not preponderance of the evidence -- it's misleading at best, wrong at worst. My guess is it's meaningless.

I wouldn't say it is meaningless or misleading. It is a piece of evidence that would be a starting point for doing a more sophisticated statistical analysis - one that would include control variables and thoughtful model specifications. That said, it is certainly the case that this is not particularly convincing. I actually do an exercise in my stats class where students find and present plots like this one from the news media and we break down what they can and cannot tell us as a class. I'd bet this is the thing that stays with students most after the semester is over.

Are you sure?

[Sun_Worshiper]

Counties that voted for Trump have disproportionately lower vax rates & higher death rates.

Darwin at work.

Republicans are older than democrats.  https://www.pewresearch.org/fact-tank/2020/10/26/what-the-2020-electorate-looks-like-by-party-race-and-ethnicity-age-education-and-religion/
"The median age among all registered voters increased from 44 in 1996 to 50 in 2019. It rose from 43 to 52 among Republican registered voters and from 45 to 49 among Democratic registered voters."
Older people are much more likely to die when they get COVID than are younger ones.
Also, more men than women voted for Trump. Men have lower life expectancy.

ETA: Does living longer mean you're smarter?

Your point about confounding variables is valid and (as I noted above) these scatterplots and descriptive statistics should be scrutinized and critiqued. But there have been plenty of studies (which include control variables) showing that vaccines save lives. If Trump voters are getting vaccinated at lower rates, then they are consequently dying at higher rates - even while holding things like age and gender constant.

...

Counties that voted for Trump have disproportionately lower vax rates & higher death rates.

Darwin at work.

If I understand correctly, the county averages are multiplied by county population. This biases standard errors, which are not reported, downwards.

Proper weighted regression would use population size to correct the variance.

Little need to get technical: Look at the scatter plots. The high Trump voting counties have death rates all over the map. That's a clue that the statistical analysis is pure genius, lucky, or wrong. :-)

Not as much noise looking at the lower vax rates in Trumpish counties.  You can argue fine points of methodology (that said, they had some number crunchers from Johns Hopkins as advisers) but the trends seem clear to me.  It's clear that vaccines do reduce the risk of hospitalization and death.  So lower vax rates can be reasonably correlated with higher death rates.

Beyond a reasonable doubt? Maybe not.  The preponderance of the evidence?  Yes.

I'm not questioning a relationship between vax rates and death rates, but rather between Trump rates and death rates. [Just off the cuff, I was thinking that the Trump rates are a confounder! :-)] The newspaper article is to be appreciated for showing the scatter plots in an informative way. But it ain't good statistical analysis. Don't have to publish their detailed results in a newspaper article, but could refer to a working paper or something.

One mustn't think that everybody at any university knows everything. What's needed is arguments, not testimonials.

It's not preponderance of the evidence -- it's misleading at best, wrong at worst. My guess is it's meaningless.

I wouldn't say it is meaningless or misleading. It is a piece of evidence that would be a starting point for doing a more sophisticated statistical analysis - one that would include control variables and thoughtful model specifications. That said, it is certainly the case that this is not particularly convincing. I actually do an exercise in my stats class where students find and present plots like this one from the news media and we break down what they can and cannot tell us as a class. I'd bet this is the thing that stays with students most after the semester is over.

Are you sure?

Yes, I'm sure I would bet that this is the case.

[mahagonny]

Counties that voted for Trump have disproportionately lower vax rates & higher death rates.

Darwin at work.

Republicans are older than democrats.  https://www.pewresearch.org/fact-tank/2020/10/26/what-the-2020-electorate-looks-like-by-party-race-and-ethnicity-age-education-and-religion/
"The median age among all registered voters increased from 44 in 1996 to 50 in 2019. It rose from 43 to 52 among Republican registered voters and from 45 to 49 among Democratic registered voters."
Older people are much more likely to die when they get COVID than are younger ones.
Also, more men than women voted for Trump. Men have lower life expectancy.

ETA: Does living longer mean you're smarter?

Your point about confounding variables is valid and (as I noted above) these scatterplots and descriptive statistics should be scrutinized and critiqued. But there have been plenty of studies (which include control variables) showing that vaccines save lives. If Trump voters are getting vaccinated at lower rates, then they are consequently dying at higher rates - even while holding things like age and gender constant.

But if that would mean (1) Darwin's principle is at work, then one should note most of the  people dying of COVID are late in childbearing years or past those years.
(2) people who choose vaccination are doing so because they are smarter, as opposed to simply having a different outlook, then I can show you people who are 90, 95, 100 years old or more who wouldn't impress you that they done the intelligent thing by living so long.

As for the bet -- you're on. I don't think you can hide your far left politics, but maybe I'm wrong.*

*e.g. you might be looking for ways to prove or reinforce the perception that Trump voters are stupid.

[Sun_Worshiper]

Counties that voted for Trump have disproportionately lower vax rates & higher death rates.

Darwin at work.

Republicans are older than democrats.  https://www.pewresearch.org/fact-tank/2020/10/26/what-the-2020-electorate-looks-like-by-party-race-and-ethnicity-age-education-and-religion/
"The median age among all registered voters increased from 44 in 1996 to 50 in 2019. It rose from 43 to 52 among Republican registered voters and from 45 to 49 among Democratic registered voters."
Older people are much more likely to die when they get COVID than are younger ones.
Also, more men than women voted for Trump. Men have lower life expectancy.

ETA: Does living longer mean you're smarter?

Your point about confounding variables is valid and (as I noted above) these scatterplots and descriptive statistics should be scrutinized and critiqued. But there have been plenty of studies (which include control variables) showing that vaccines save lives. If Trump voters are getting vaccinated at lower rates, then they are consequently dying at higher rates - even while holding things like age and gender constant.

But if that would mean (1) Darwin's principle is at work, then one should note most of the  people dying of COVID are late in childbearing years or past those years.
(2) people who choose vaccination are doing so because they are smarter, as opposed to simply having a different outlook, then I can show you people who are 90, 95, 100 years old or more who wouldn't impress you that they done the intelligent thing by living so long.

As for the bet -- you're on. I don't think you can hide your far left politics, but maybe I'm wrong.*

e.g. you might be looking for ways to prove Trump voters stupid.

Your two points are so poorly worded that I can't entirely understand what you are trying to say, but I'm pretty sure neither are sensible responses to my post. This is probably because you don't understand statistics and, therefore, are not capable of carrying on a conversation about control variables.

I also never said anything about Trump voters being stupid. Just that they are dying at higher rates because many of them refuse to get vaccinated. Some people might think that refusing to take a life saving vaccine is stupid, but I did not say that.

I'm in a redish state and have taught plenty of conservative students. Nobody has ever mention my politics in teaching evaluations. I'm also amused that you think I'm "far left," since I support free trade/investment/migration.

[mahagonny]

I have never taken a course in statistics, it's true. It wasn't you who posted 'Darwin at work' but you think the comment is fine, apparently.

I'm in a redish state and have taught plenty of conservative students. Nobody has ever mention my politics in teaching evaluations. I'm also amused that you think I'm "far left," since I support free trade/investment/migration.

Enjoy your state of amusement, then.

ETA: I would be interested in what they might say in, say, five years.

[secundem_artem]

I have never taken a course in statistics, it's true. It wasn't you who posted 'Darwin at work' but you think the comment is fine, apparently.

I'm in a redish state and have taught plenty of conservative students. Nobody has ever mention my politics in teaching evaluations. I'm also amused that you think I'm "far left," since I support free trade/investment/migration.

Enjoy your state of amusement, then.

ETA: I would be interested in what they might say in, say, five years.

That was me.  If you want somebody to argue with, at least pick the right interlocutor.

[mahagonny]

I hesitated. Without remembering specifics, I have little reason to 'call out this' poster. I like them.

I find ridicule of Trump voters and the condoning of it a very sad comment on our times. I have a colleague at work, well an administrator whom I run into now and then, who voted for Trump. She's not stupid or bigoted. She just didn't like the alternative. She's pro-vaccination for everyone, but doesn't mean she's in favor of government or employer's overreach. We have had a total of three conversations, but for some reason she sensed that she could tell me. It's not something you ever hear from anyone at work. However, even in our very blue state, (and you could count the Trump bumper stickers on autos in our town on one hand) many actually did vote for Trump. Maybe one in five, one in four. I wonder if she feels isolated.

If I misread and it wasn't ridicule, then sorry, disregard.

Of course, some Trump voters sometimes ridicule Biden or Hillary voters. I wonder if it will always be like this.

[kaysixteen]

Look, plenty of folks in my orbit were 2-time Trump voters too, including more or less everyone at my church (one guy voted for Hillary but does not say whether he voted for DT last year, merely that he didn't vote for JB).   I get this, but I gotsta say that it is really really realllly hard to respect this obviously lack of discernment decision.   And, esp if the Trumper in question also regularly self-propagandizes himself with nonsense, and also often regales all and sundry with said nonsense, well....

[mleok]

While not every Trump voter is a covidiot, but it certainly seems that every covidiot was a Trump voter.

[lightning]

I got two shots last night--one for the flu and one for COVID-19 (booster shot). I asked for one shot in each upper arm.

I'm thinking that if I ever have to get a 3rd shot for something else at the same time as the other two, where would they stick the 3rd needle? My leg? Serious question.

(BTW, the COVID-19 shot left me really sore this morning, just like the 2nd dose, while the flu shot left negligible soreness in my arm.)