I am being lazy and not reading all the papers you referenced - do many of them discuss the viral load of the person who is infected? 

I worked on PCR covid testing during the pandemic and the viral load of samples that went through our labs would often be multiple orders of magnitude different between samples. Some people seemed to just have way more virus in them than others, it did not obviously correlate with symptoms. 

I was likewise very shocked at just how little was known about how virus actually spread on a practical level. How do people not know whether humidity affects how long virus remains active?, temperature?, dry vs wet surfaces?, rough vs smooth?, recently disinfected?, plastic/metal/material?, daylight (UV) vs artificial light, which of these factors is true regardless of particular virus?

Working with viruses is (I am told, I no expert) very hard - but I agree completely that this is an area that seems woefully underfunded vs related areas. I would be surprised if there were not cheap and effective 80/20 effects here that could be used to drastically cut infection rates in real world scenarios. 

I am not that close to any of it any more and maybe it has, but I have not seen any of the post-covid enquiries raise this knowledge gap either. 

I am not an HR professional. This is all experience as manager/director/multiple business owner.

I worked at one big corporate (Capital One Bank) who had both very rigorous recruitment processes and had very comprehensive annual appraisal processes for employees - plus of course hard data like whether they made it through appraisal, how long they worked there, promotions etc. 
They did analysis (they are renowned for analysing everything), looking for correlations between scores of various bits of the recruitment process and the above outcomes. 
There is literature on this sort of thing as well, some of it of dubious quality. 

But really if you did some or all the interviewing and were responsible for the hiring decision and then work with the person you hired then you often know within days or weeks whether the person is what you thought they would be. That's a generalisation and there are exceptions, there are people who are slow starters, a few who start well but have fatal flaws. 
But I would say with experience you can be confident 80% of the time, within two weeks of having either found someone really good, or bad enough you wish you hadn't hired them, or they are ok but not a superstar. 

Clearly none of this covers the people you didn't hire. But again with experience, and having made mistakes, you get to see the correlations. between good/bad hires and what you saw at interview.

I don’t have any rigorous principled argument for this, other than just the empirical personal observation that ignoring the feeling usually seems to be a mistake.

Likewise only empirical personal observation, but having interviewed and hired (or not) a lot of people. I would say I have got better over time - and separating first impression/prejudice from vibe is mostly achievable. And vibe definitely counts and is valid. 

There are also people who I trust to be much better than me at hiring good people, and they also consider vibe valid (but are also very rigorous about applying wide ranging interview&selection methods).... And there are people I trust to be awful at choosing people!

Normally though I would say you can find some way to put that vibe into a more specific concern/worry. And then you can normally create some form of question or test that might give a little more data on whether your concern is valid. 

Clearly not all situations give the luxury of devising such a thing, but often just a couple of minutes of conversation can be enough.