When I say ‘weak evidence’, I mean it. If I met one person who had been practicing PUA for six months, and they hadn’t had significantly more sex in those six months than in the previous six months (as judged by some method other than self-reporting; probably by asking the person’s friends or roommates), that would be stronger evidence that PUA doesn’t work.
As I said above, I believe that you’re using the phrase “weak evidence” in a non-standard way, essentially violating Grice’s Maxim of Quantity. When something is as weak as in this case, people don’t call it weak, they call it trivial or negligible or exceedingly weak, or some other term to transmit the idea of just how very weak it is. When people say “weak evidence”, they mean that it’s definitely not strong or conclusive, but there’s nontrivial amount of it, it’s not vanishingly small.
You’re possibly right. I invite you to test your belief by consulting the search for “weak evidence” over LW. This probably sounds snarky, so let me quickly clarify that I don’t mean this in the sense “this search confirms I’m right”. I looked over the results briefly and saw that roughly half of them use further qualifiers like “astonishingly weak”, “very weak”, “extremely weak”and so on, which is consistent with what I’d said. But it’s also possible that many results use vanilla “weak evidence” to refer to what other results call “extremely/incredibly weak” etc., and then you’re right. I looked at a few “vanilla” uses in context and didn’t see that, but I didn’t look at enough.
I think I see different requirements for things working that for things not working.
People getting actively angry about something looks like a (weak) indicator that they are afraid there might be something to it. They do not get angry at 9/11 conspiracists, but laugh about them. They do get angry at the other party.
If you have someone trying to do something and be unsuccessful that tells you even less. I would not see it as indicator of not working. It is just no particular evidence.
But maybe I have the notion of what ‘weak’ means wrong.
If you have someone trying to do something and be unsuccessful that tells you even less. I would not see it as indicator of not working. It is just no particular evidence.
It almost has to be evidence; even if it’s just evidence that the person isn’t doing it right, then you push this case through your prior for how often they do things wrong. Unless your prior is very high, you’re still getting half the impact or more of the evidence. Although you could argue from the selection effect of “correct in your expectations of success” that you’re almost guaranteed to only notice cases where they are doing it wrong, not cases where it doesn’t work (because you don’t expect it to work where it won’t).
A person using the techniques and not having increased success—weak evidence
People hating PUA—trivial evidence
A bunch of people active on a PUA forum having increased success in aggregagate—moderate evidence (would be strong but selection on dependent variable)
But it works as more evidence than people denouncing it.
More evidence than a negative amount? How much is that worth?
Weak evidence! Weak! Weak.
For practical purposes, I regard weak evidence as equivalent to no evidence. It has the value of a penny lying on the road: I do not bother to pick it up. Weak evidence is not worth wasting time on.
More evidence than a negative amount? How much is that worth?
Further up in the thread, I claimed that (in cases like these) people denouncing it is weak evidence in favour.
It has the value of a penny lying on the road: I do not bother to pick it up.
Anecdote: I once refused to pick up a small denomination coin (we don’t have pennies in Australia anymore). A friend did and promptly discovered it was a penny, from 1938, and worth two dollars to collectors.
On practical matters, though, this is a case where training will greatly decrease effort and time to pick up (more so than for pennies). If you track how much time is spent on picking up pennies and put a value on that amount of time, and then how much money in pennies you have picked up, you will probably find it not worth it. I expect this weak-evidence-picking-up will prove worth it—if it even slightly addresses the common human problem of incorrectly assessing or entirely discounting large amounts of small evidence, it comes out positive for me.
Anecdote: I once refused to pick up a small denomination coin (we don’t have pennies in Australia anymore). A friend did and promptly discovered it was a penny, from 1938, and worth two dollars to collectors.
Still not worth it unless you have a personal interest in coins or expect to find or otherwise collect many such coins. (Of course many people would value the experience at more than $2 just for the novelty and the story.)
I expect this weak-evidence-picking-up will prove worth it—if it even slightly addresses the common human problem of incorrectly assessing or entirely discounting large amounts of small evidence, it comes out positive for me.
Can you give some examples of that, i.e. where large amounts of small evidence are being badly used?
This appears to be) an example of an accumulation of weak evidence badly used, but in the opposite direction. (I say “appears”, because I haven’t read the article it references, just that LW posting and its comments.)
Probabilities of multiple events are being multiplied without concern for whether they are independent. And that is the basic practical problem with accumulating weak evidence. Look at the vast amount of evidence for the existence of Santa Claus! Even if each story only offers a microbit of evidence....well, no. The stories are not independent of each other. Collectively they prove no more than a handful of them do.
When I say ‘weak evidence’, I mean it. If I met one person who had been practicing PUA for six months, and they hadn’t had significantly more sex in those six months than in the previous six months (as judged by some method other than self-reporting; probably by asking the person’s friends or roommates), that would be stronger evidence that PUA doesn’t work.
As I said above, I believe that you’re using the phrase “weak evidence” in a non-standard way, essentially violating Grice’s Maxim of Quantity. When something is as weak as in this case, people don’t call it weak, they call it trivial or negligible or exceedingly weak, or some other term to transmit the idea of just how very weak it is. When people say “weak evidence”, they mean that it’s definitely not strong or conclusive, but there’s nontrivial amount of it, it’s not vanishingly small.
I do not believe it is non-standard for LessWrong. I admit I’m guilty of tailoring my posts to fit the LessWrong-specific audience.
You’re possibly right. I invite you to test your belief by consulting the search for “weak evidence” over LW. This probably sounds snarky, so let me quickly clarify that I don’t mean this in the sense “this search confirms I’m right”. I looked over the results briefly and saw that roughly half of them use further qualifiers like “astonishingly weak”, “very weak”, “extremely weak”and so on, which is consistent with what I’d said. But it’s also possible that many results use vanilla “weak evidence” to refer to what other results call “extremely/incredibly weak” etc., and then you’re right. I looked at a few “vanilla” uses in context and didn’t see that, but I didn’t look at enough.
I think you find more people for who it does not work, than for whom it works. But I doubt this works as evidence that it does not work.
There are many factors that influence success. One would be doing it wrong.
Not much, no. But it works as more evidence than people denouncing it. Weak evidence! Weak! Weak.
I think I see different requirements for things working that for things not working.
People getting actively angry about something looks like a (weak) indicator that they are afraid there might be something to it. They do not get angry at 9/11 conspiracists, but laugh about them. They do get angry at the other party.
If you have someone trying to do something and be unsuccessful that tells you even less. I would not see it as indicator of not working. It is just no particular evidence.
But maybe I have the notion of what ‘weak’ means wrong.
It almost has to be evidence; even if it’s just evidence that the person isn’t doing it right, then you push this case through your prior for how often they do things wrong. Unless your prior is very high, you’re still getting half the impact or more of the evidence. Although you could argue from the selection effect of “correct in your expectations of success” that you’re almost guaranteed to only notice cases where they are doing it wrong, not cases where it doesn’t work (because you don’t expect it to work where it won’t).
My take -
A person using the techniques and not having increased success—weak evidence
People hating PUA—trivial evidence
A bunch of people active on a PUA forum having increased success in aggregagate—moderate evidence (would be strong but selection on dependent variable)
More evidence than a negative amount? How much is that worth?
For practical purposes, I regard weak evidence as equivalent to no evidence. It has the value of a penny lying on the road: I do not bother to pick it up. Weak evidence is not worth wasting time on.
Further up in the thread, I claimed that (in cases like these) people denouncing it is weak evidence in favour.
Anecdote: I once refused to pick up a small denomination coin (we don’t have pennies in Australia anymore). A friend did and promptly discovered it was a penny, from 1938, and worth two dollars to collectors.
On practical matters, though, this is a case where training will greatly decrease effort and time to pick up (more so than for pennies). If you track how much time is spent on picking up pennies and put a value on that amount of time, and then how much money in pennies you have picked up, you will probably find it not worth it. I expect this weak-evidence-picking-up will prove worth it—if it even slightly addresses the common human problem of incorrectly assessing or entirely discounting large amounts of small evidence, it comes out positive for me.
Still not worth it unless you have a personal interest in coins or expect to find or otherwise collect many such coins. (Of course many people would value the experience at more than $2 just for the novelty and the story.)
Can you give some examples of that, i.e. where large amounts of small evidence are being badly used?
This appears to be) an example of an accumulation of weak evidence badly used, but in the opposite direction. (I say “appears”, because I haven’t read the article it references, just that LW posting and its comments.)
Probabilities of multiple events are being multiplied without concern for whether they are independent. And that is the basic practical problem with accumulating weak evidence. Look at the vast amount of evidence for the existence of Santa Claus! Even if each story only offers a microbit of evidence....well, no. The stories are not independent of each other. Collectively they prove no more than a handful of them do.