wedrifid comments on Review: Selfish Reasons to Have More Kids - Less Wrong

17 Post author: jsalvatier 29 May 2012 06:00PM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (257)

You are viewing a single comment's thread. Show more comments above.

Comment author: RichardKennaway 31 May 2012 05:12:04PM *  1 point [-]

As for heritability studies, you are certainly right that there is a lot of shoddy work, and by necessity they make a whole lot of wildly simplifying assumptions. If there existed only a handful of such studies, one would be well advised not to take them very seriously. However, the amount of data that has been gathered in recent decades is just too overwhelming to dismiss,

Piling up shoddy evidence does not make good evidence. (And it still doesn't if you -- that's the impersonal "you", not you in particular -- call it "Bayesian evidence".)

especially taking into account that often there have been considerable ideological incentives to support the opposite conclusions.

There are considerable ideological incentives on both sides.

The Sesardic book you recommended is in my university library, but when I went to look at it, I found at least a shelf-foot of books on the subject, some (I could tell just from the authors' names) on one side, some on the other. So I didn't bother looking any further and left all the books there. I could read Sesardic saying what you say he says, but then I could read Kamin arguing the opposite, and in that situation, to form a view of my own with any real basis I'd have to research the subject enough to write a book of my own. I have other things to do. Such is the nature of controversies: they cannot be settled by saying "read this book".

One observation though, that I haven't seen made on either side. Failing to find strong genetic causes for something does not imply that it's the environment; failing to find strong environmental causes does not imply that it's the genes; failing to find either does not imply that it's the interaction of genes and environment. I believe I've seen (but no cites) all three wrong arguments being made from time to time. All that failing to find the causes implies is that we have failed to find the causes.

Comment author: wedrifid 31 May 2012 07:59:13PM 5 points [-]

Piling up shoddy evidence does not make good evidence.

Um... yes, often it does.

Comment author: RichardKennaway 31 May 2012 08:03:14PM 0 points [-]

For example, 100 people saying "Um... yes, often it does" would not sway me a jot.

Do you have an example?

Comment author: wedrifid 31 May 2012 08:08:26PM *  4 points [-]

Do you have an example?

1 study with too small a sample size to even make the findings significantly significant is shoddy evidence. 100 such studies is (probably) good evidence.

Second example:

For example, 100 people saying "Um... yes, often it does" would not sway me a jot.

It should (unless all 100 people are known to be fools.)

Comment author: RichardKennaway 01 June 2012 07:46:55AM *  -2 points [-]

Um... yes, often it does.

Do you have an example?

1 study with too small a sample size to even make the findings significantly significant is shoddy evidence. 100 such studies is (probably) good evidence.

That isn't an actual example, and there are many more ways of being shoddy than merely a limited sample size.

It should (unless all 100 people are known to be fools.)

This presumes explicitly that they are not, and tacitly that they are all independent, have not all come to the same conclusion for the same bad reasons. These are very substantial presumptions and need substantial justification to make such a meta-analysis worth anything at all. The presumption that one would have to be a "fool" to be wrong about something is just rhetorical spin.

100 pieces of crap is still just a pile of crap, however you spin it.

Comment author: wedrifid 01 June 2012 07:58:06AM *  5 points [-]

and there are many more ways of being shoddy than merely a limited sample size.

"Give me an example! ... Ha! Your example isn't a fully general enumeration of the entire class... Fooled you!!"

Comment author: RichardKennaway 01 June 2012 08:10:29AM -2 points [-]

Your example wasn't an example at all, just a made-up scenario of many small pieces of evidence adding up to large evidence. You originally said

Um... yes, often it does

(emphasis mine.) When asked to back that up, you just made stuff up.