Riddle me this Vaxman

Look I not responsible for you apparently failing high school mathematics (and it must have been an epic fail). Your knowledge of statistics is so woeful you don't realize what rubbish it is you've just spouted.

The area of statistics your talking about is called inferential statistics. One the primary uses of inferential statistical modelling is to gain as estimate of a value for an unknown population characteristic based on another sample. Furthermore it can be used to analyze any type of unknown variable in any time period.

In this case the 'unknown population' are asymptomatic COIVD cases in the untested population (Y). Put as simply as possible (after all it is you I'm dealing with) a projection for that population size can be made using data obtained from tested patients. From the sample group X (tested patients) you can extract a figure for for asymptomatic cases. After adjusting relevant variables such as age, gender or the degree to which testing increases or decreased the likelihood of catching COVID (if it does either) you get your answer. And as always the larger the sample of tested patients you have the more reliable your inference for the untested population is going to be (and we have a very large sample size when it comes to people tested for COVID) . There will be margin of error in that estimate but then its statistics - there always is. The point is once you have the answer for the tested group you can mathematically infer an answer for the untested group (Y).

And yes, you still don't know what your talking about.

 

I think we both are. At some point, on other specific issues the facts will no doubt support your arguments. Just not this particular one.

 

I love listening to experts!!Dont always understand but appreciate the education

 

You say you're done and begin arguing immediately after.

Disingenuous.

 

Your meme is a bit misleading
https://www.visualcapitalist.com/history-of-pandemics-deadliest/
Some of those plaques lasted decades, happened pre hospital in a time when they would just nail up the door of the hovel and if you were alive at the end of the month - lucky you



Bottom line with COVID - it ain’t over yet

 

Not arguing. If one party in a debate states that the other is 'done' by default then they must be 'done' to since the debate can't continue with only one side. And you were the person who stated I was 'done here' lest you forget, not me.

 

Still arguing.

 

the most you can obtain is a statistical model, and I can make that model reflect (pretend) any result I want to convey LOL

BTW where is your statistical model?

The fact remains crucial data was never retrieved, at best all you have to offer without hard evidence is advertising for your unscientific 'opinion'.

Thats why we have 'double blind' testing when we want 'real' answers.

Did I miss something?

 

All you have in here is your google search engine, name calling and ad hominem attacks. Not sure why the Mods allow you to continue with this.

Anyhoo, my post left a mark. There are NO statistics on how many people have had Covid- whether symptomatic or asymptomatic.

 

Indeed. No telling how many people have had Covid- even those who have had it multiple times, whether vaxxed or unvaxxed.

 

The problem with this explanation is that to be accurate, the tested population must be representative of the larger population you want to apply your results to. In the case of asymptomatic SARS-CoV-2 infection there is no way to create a representative sample. Asymptomatics are not likely to pursue testing. Random voluntary testing is not representative either because a large share of the general population doesn’t have any interest in being voluntarily tested. I have seen attempts at creating a somewhat representative sample by testing for antibodies from blood collected for other purposes such as outpatient healthcare. But this sample is biased towards those needing/wanting healthcare, not the general population. At the end of the day, mandatory mass testing of a (or many to actually be really representative) representative population is the only way to come up with a representative sample you can apply your “math” to.

Im sorry, but your opinion here is not based on any expertise. It’s contradictory to all accepted methodology of representative sampling. Your error is not the math, it’s in understanding sampling, human behavior, and epidemiology.

 

There are (and can never be) no direct statistics on anything that hasn't been counted. Your literally arguing that you can't count the number of words in a book that hasn't been written! To which the proper reply probably something like 'no **** Sherlock!'

That said you can use inferential statistics to draw some conclusions about the untested asymptomatic COVID sufferers. That is just a incontrovertible fact. FYI you could also in theory potentially do the same thing with the number of words in an unwritten book if you know more about the author, the subject matter & type of book etc as well other variables etc. At a guess? The margin of error might be significant but its still doable. It's not my problem you can't deal with facts when they're laid out for you. You could of course always try and prove me wrong about the use of statistics in such situations.

As for the search engine? Yes, I use search engines. That's how I learn stuff ? How do you get information on a computer?

 

Again, we will never know the amount of people who have been infected with Covid.

So any statistical models based on this are estimates and/or flawed.

‘That is the fact that you cannot overcome……..no matter how much you burn electricity running those search engines.

 

That is a good point. Asymptomatic patients won't ordinarily go for testing etc. But many will if they were close contacts of someone who is symptomatic and tests positive. Say because they are family members of that person or its mandated by work etc. That's how some of the first studies on this issue came about. Hospitals for example were regularly conducting mandatory testing of staff and asymptomatic patients showed up in the results. Other large organizations did as well. In some countries with stringent testing and case tracking mandates during the early stages of the pandemic people would be notified of the fact they were close to a positive case at a particular time and location and should go and get tested. Of course this broke down when case numbers finally exploded with Omnicrom etc.but again asymptomatic carriers showed up. Case in point kids, especially kids.

As I said you 'd need a large sample size (which sadly COVID gives us) before you can have some confidence in the results. If you look at one of the links I posted you'll see that its a meta study of lots of smaller studies conducted in various countries. But again my point is its doable if you have the data and we've been collecting it now for what 3 years almost?

 

You don't need to know every case, that's the power of statistics. What is true is that the larger sample size you have the better off you are. But as I keep saying the the end of the day you'll get a number. That number will however have a plus and minus error range associated with it. But you will never get certainty, never that. Only an estimate.

 

Yes, you do need to know the total amount of infected in order to compute these statistical models. Because without that number, it is all estimates, that is subject to being erroneous.

We also do not know how many people who have contracted Covid multiple times, whether vaxxed or unvaxxed.

 

As far as the first bit goes you don't need exact numbers but you do need large ones i.e the most recent confirmed numbers you can get. And of course sadly with COVID we do have numbers.

I believe there's some (a little) data on that second bit but I'd have to search for it. And its certainly true you can catch COVID multiple times (albeit the first infections should grant you some degree of immunity so I would expect there to be months separating infections in all but rare cases. Not an expert but I would also expect your chances of being asymptomatic would go up with subsequent infections as long as they happened while you still had high levels of immunity from your last case. But all these issues are why medical statisticians and epidemiologists work in a complex field and (should) get paid well.

 

There is not any data on this. It is all conjecture and guesstimates. It is not complex to see that.

 

Mass testing of hospital staff can not be applied to the general population, but yes it can help predict asymptomatic rates. But that data can not be used to predict how many asymptomatic infections in the general population have been missed which I believe was the original question.

On testing of asymptomatics in the US it would vary a lot by region. On the coasts I suspect people were more likely to be tested just for the fun of it. In the Midwest and west a lot of people just don’t care. I probably know 35-40 people personally who were exposed to known cases (often repeatedly in the household etc.) who never tested. Those 35-40 likely had contacts who had no interest in testing either. I can’t project those numbers onto mandatory healthcare testing any more than healthcare testing can be applied to my scenario. At the end of the day, the claim we have no idea how many asymptomatic cases existed and recovered without being documented or predicted by non-representative samples. We can make a guess but it’s not going to be very accurate.

Your meta analysis was very clear.

The study makes no claim about the total number of asymptomatics in the US who recovered without being vaccinated. It can’t. It can only predict what it claims in the title—the percentage of asymptomatic infections in tested and confirmed cases.

The asymptomatic rate of any population can’t predict the number of undiagnosed and untested asymptomatics in a population.

 

There is is a hard data on positive COVID cases and including asymptomatic ones from across the world That is fact, like it or not. Just like there's hard data on virtually every other disease known to mankind and a host of other medical issues. There's an entire field of statistics devoted purely to medical research for Gods sake.

And the fact the data sets for COVID are not 'perfect' or complete or whatever term you want to use is irrelevant. They don't have to be produce useful/meaningful results - that's the beauty of statistics. You don't need perfect information to get meaningful results with real world applications. And again statistics is not conjecture or 'guesstimates'. It's a one of the most (if not the most ) widely used and important branches of mathematics on the planet! Literally millions of people rely on it for their livelihoods every day because it works. That is the thing that 'is not complex'. Or rather it would be if you could be bothered to do your own research in to the subject.

 

No, there is not. There is only reported cases, which you linked earlier. There is no way of knowing how many people have been infected with Covid.

Statistics is about understanding data. Without this crucial figure, total amount of people infected, any statistical models are flawed, i.e. incomplete.

I see your back with the snarkiness. Man, you must be a hoot at parties.

 

This is nub of the issue. And I can only respond that it can IF the sample size is large enough. For any given population you will know who have been tested both as a round number and as a %. That same figure will also tell who isn't tested. If (and only if) the % of the population that has been tested is high enough (statistically valid) do you get a good basis for assessing the asymptomatic rate in the tested population. But the thing is if you do get that number then you have a basis for calculating the asymptomatic rate in any subset of the population you want to study, including untested people. Provided you also allow for all the relevant variables and use the right statistical tools. AND accept that your figures will have a margin of error.

Have to go now BTW we can continue later. I might also get our resident statistician to way in.

 

'There is no way of knowing how many people have been infected with Covid . Look up the word epidemiology (hint there is). And I generally do great at parties thank you. In part because I don't make sweeping claims about subjects I obviously have no knowledge of.

 

Ahh, there it is- another ad hominem attack.

I do know what I am talking about, and I am chuffed to bits about it!

If there was data on how many people have been infected with Covid, then your search engine would have found it. But, there is not.

We both know, you cannot overcome that crucial fact.

 

https://ourworldindata.org/covid-cases