Hacker News Reader

Comment by dmreedy | Hacker News Reader

If there are, they would be as difficult to recognize as germ theory to a 16th century doctor. We simply don't know enough yet about the mechanism.

View on HN

Yeah, it's interesting. We have decades of research into standardized tests in the academic setting, and yet most people generally feel that these metrics have very limited utility.

The notion that a startup has A: deep insights into the nature of cognitive function that others do not; and B: devised and tested a paradigm that sustainably improves it, was always a little shocking to see claimed.

You don't need germ theory to notice that washing your hands before performing surgery improves outcomes.

It helps. It's much easier to notice something and take it seriously when it fits a theory you already have. Kind of like if your house is haunted, you might not notice that the sounds you hear at night fit that theory unless you already have it in mind. If you try to understand the noises in terms of your house settling, a fugitive living in your attic, or an animal breaking in at night to steal food, you won't notice, or won't take seriously, the patterns that would be obvious to someone who believes in ghosts.

Yes, but you do need to start moving towards germ theory to understand why washing your hands with charcoal instead of soap results in dead patients again. Of course scientific theory and scientific observation are bootstrapped off of each other in the course of a dialectic, but raw statistics only get you so far until deductive work is required for the next steps, and the cycle repeats. It's easy to spot and distill in hindsight what is less obvious in the act; and when you've got as many variables as the human experience does, and as many ethical questions when you experiment with them, the process becomes all the more vague.

can you explain a bit about the charcoal/soap thing?

This is, if you will, a very germ-theoretic perspective. Why would you notice, by eye, this correlation and not that one? More importantly, why would you collect these data rather than those? Scientific paradigms are "nefarious" because they control experimental design, so unparadigmatic data only come about by accident. And accidental results are often as not a flaw in the design or instrumentation, or else "noise."

Of course it's easier if you have a viable theory to deduce efficient treatments. But the correlation between hand washing and medical outcome was noticed before germ theory, this is why I mentioned it. It didn't become popular until later though.

https://en.wikipedia.org/wiki/Ignaz_Semmelweis

Fair enough, and that guy's work was pretty close (in time) to the "germ revolution." The accounting of scientific revolutions is that anomalous results start getting common enough that people get uncomfortable calling them noise or faulty methodology. I'm not an expert in the history of paradigms of disease, but these results coming out were almost certainly part of what chipped away faith in miasma theory and, if you will, created the unease required for revolution.

Importantly, Semmelweis couldn't offer an explanation for why washing hands helped, which seems to have greatly hindered his credibility. I said you were offering a germ-theoretic perspective, because we do (and must do) the same thing: How else should we separate the spurious correlations from the important ones? How do we separate superstition from good practice? How do we tell the crackpots from the creative? If we suspend the germ theory, what if there was something else going on at the hospital, and hand-washing was just something this guy noticed and latched on to? Then you'd be giving people the wrong advice---which is, after all, one of many reasons that correlation is not causation.

Reply on news.ycombinator.com