Hacker News Reader

Comment by peteretep | Hacker News Reader

Fundamentally, though, the problem with articles like this is that even if he's right, there are very few situations where knowledge of a difference in distribution are that useful, and very very many where it is misused to justify existing (and unjustifiable) bias.

If you knew that men were on average 5 IQ points lower than women (which is not what he says, and probably not true, but whatever), it wouldn't be very useful information for actually making choices. Male student underperforming in class? That's 'cuz men are dumber. Emerging gender-skew in medical school students towards women? That's 'cuz men are dumber. Etc etc.

View on HN

If someone told me that a certain truth is not useful, I would be inclined to disbelieve them. The basis of your argument seems to be "I can't think of any uses" but that can hardly be considered "the problem with articles like this". I can think of quite a few applications, but in many cases I have experienced, whenever relevant truths have been ignored, delusions have thrived.

OP didn't say "not useful" - it was couched more moderately, which I imagine still allows for the possibility of unforeseen greater uses.

But the specific point made seems sound to me. The hypothetical 5 point difference would not be a sufficient explanatory factor for most scenarios.

OP literally typed the words "wouldn't be very useful". He then goes on to imply that average people would be confused by facts and should therefore not seek/be given them. This is a bad idea.

    > If someone told me that a certain
    > truth is not useful, I would be
    > inclined to disbelieve them

Sure. But I didn't say that. There are some situations in which it's useful to know the Thai alphabet has 44 consonants, but not only is that generally not very useful knowledge for the vast majority of people, it might also lead you to conclude learning to read Thai will be significantly harder than reading English, which is probably not true.

A truth is useful. A given person's knowledge of that truth, not always, especially given that not all truths are simple enough to be understood by the layperson -- how else did we get to a place where evolution is "humans are descended from monkeys"?

Stereotypes are literally heuristics for "difference in distribution". Curiously you used just such a heuristic when you said "articles like this".

If knowledge of distributions isn't useful, then why did you rely on knowledge of distributions to make a blanket judgment about articles like this?

A stereotype is fixed and oversimplified idea of a particular thing.

An example is much, much different than that.

Distribution alone doesn't help, but if you knew the distribution of IQ and its effect on academic success (a fairly big if), you could statistically use it to determine whether groups underperform relative to their potential and hence, decide whether corrective action is needed.

Gender-skew of over 10%? That's more than expected given the IQ distributions of men and women. Let's look for what causes that.

I would argue that understanding the underlying distribution is useless when looking at a single datum (a man or a woman) but critical for understanding things en masse. The author's point is that we can't be so quick to judge seemingly unfair trends or groupings which might actually be the product of statistics. He admits readily that there are exceptions to be found on all fronts.

The actual problem with this article is that experimental evidence for the variability hypothesis is not consistent across cultures. There's no particular reason to believe it is intrinsic... other than to support articles like this, anyway.

It doesn't have to generalize across cultures, it just needs to be true in our own culture.

I disagree that it wouldn't be useful. In the example you gave, knowing that there's an underlying difference would prevent ultimately futile attempts to affirmative action the numbers into being equal.

Yeah, people don't need to know the truth. There are others who can make decisions based on that truth for them.

Or isn't that the sort of point you were going for?

Huh, billions of dollars are spent on misguided feminist support programs. These are choices that matter.

Reply on news.ycombinator.com