Forum: Author Hangout

Thanks, Remus2. Interesting reading, but I didn't see anything in that link that asked epidemiologists what they'd do in certain circumstances.

but what would you personally do in certain circumstances. I missed the intro to this bit so I don't know who conducted the poll - but it sounded like they were asked to respond one of four ways to the series of circumstances. They'd do it (1) this summer (2) within 3 to 12 months (3) more than a year from now or (4) never.

You didn't specify the "certain circumstances," therefore the lack of answers.

Maybe THIS is what you were looking for?

Yes - Thank you, Reluctant Sir. That sure sounds like what I heard on the radio. Thanks for actually reading my question, as poorly as I may have worded it.

Another thing interesting is the survey respondents. The survey was sent out to 6,000 epidemiologists. That means less than 10% (8.51%) responded, and of the 511 that did, 385 of then were academics.

Which points out why the results of this survey are likely unreliable. What I remember of sampling theory is that it assumes a random sample represents an adequate cross sample of the population and is reasonable representative of relevant sub populations. If it doesn't, you can throw whatever results are obtained out the window. It's not that the answer produced is wrong; it's that you don't know if it is or it isn't. Sampling error and confidence limits can't be reliably estimated under such conditions. This is especially when the relevant sub populations may hold radically divergent views regarding the topic being investigated. Like between academics and non academics, for example.

But since no sample in modern times conforms to said principle, all of the "good" research firms just ignore the "inconvenient truth." Hence, one gets research reports of 97% of scientists agree that man is responsible for global warming; or that Hillary is going to win in a landslide (two of the more extreme examples). Or just about any political poll for either party in the U.S. But one must never let science stand in the way of grabbing a large fee dontchaknow.

Gotta love push polling.

As I read the story, it does not claim to be a scientific sample. Just that it was the results from 511 epidemiologists who responded.

As I read the story, it does not claim to be a scientific sample.

But if it's not a proper representative sample, you can't extrapolate results beyond the responders, so what's the point?

S'far as I can tell, the story does not extrapolate results beyond the responders. What it tells me is how 511 epidemiologists responded, and that's interesting in and of itself.
I remember way back yonder when news organizations did man-in-the-street polls without any pretense of being scientific. One of them (I assume tongue in cheek), went something like this: Two out of three people randomly selected on the streets of (city) said they planned to vote for (candidate) in next week's election. The third guy didn't understand the question.

S'far as I can tell, the story does not extrapolate results beyond the responders. What it tells me is how 511 epidemiologists responded, and that's interesting in and of itself.

Maybe. But whenever I see reference to a poll, I assume the report or article is going to make a point about the population mentioned in the article. That's the author's intent. Hence why else even mention that 6000 epidemiologists were surveyed if not to imply that, given the results, that's how they feel?

The intent of articles like that is to give the impression that a group surveyed believes a certain way. My post was an attempt to shed some light on why it doesn't show anything. I really dislike anyone, on either side of the topic, spewing what is nothing more than propaganda rather than fact. I really believe that sampling theory when followed can provide interesting insights into such topics. Proved that to myself with repeated computer simulation when I was doing this for a living. Sample design is hard to do correctly; it's real easy to do it incorrectly. Done correctly, they can provide insights into how populations really feel. But only when they're not based on a bullshit sample design that is done just to blow smoke up the public's ass. Like this one.

As I read the story, it does not claim to be a scientific sample. Just that it was the results from 511 epidemiologists who responded.

You're right, Bob. Add the the unscientific nature of the poll and I have to believe that the most who received it would throw it into the garbage. The top epidemiologists are so pursued by the media for their opinion on the pandemic that they run and hide when they see a microphone. A survey conducted by a non-scientific publication would be ignored by most. "Here, son, why don't you practice your writing skills by answering this stupid survey. You've heard the correct answers on the news for months now; you can easily do it!"

the top epidemiologists don't know what's going to happen

That's too true! I was reading yesterday that researchers at NIH have yet to verify that those who have had the virus now have immunity to it. How do you find an effective vaccine without knowing the cause of the virus, immunity factors, or even an effective test. I was hospitalized for several days last week. The first thing they did was test me for the virus. The comment when they gave me the results was "When you took the test an hour and a half ago, you didn't have the virus." That obviously implied that there was a chance that I had contracted the disease in the 90 minutes it took to process the test. This is truly scary stuff.

What they should have done is give you a test right before you left the hospital. If there's one place you can get a disease you didn't have before it's the hospital.

This is truly scary stuff.

If you are otherwise healthy it's not that scary.

At least 40% of the deaths in the US have happened in nursing homes. And several of the US states that were hard hit with deaths were sending COVID patients to nursing homes (and threatening to revoke the licenses of reluctant facilities that didn't want to take the COVID patients) to keep hospital beds open.

it's just that the odds of them doing so are incredibly low

It's not that the odds of them catching it are low. They are more and more documenting that the vast majority of healthy adult humans who contract covid19 are only developing relatively mild symptoms only marginally worse than a normal cold. These people didn't show up in the early figures because they don't seek medical care.