Re: Hmm - I think someone may have a short career

"Peter H.M. Brooks" <peter@xxxxxxxxx> wrote in message
news:demadc$kc$1@xxxxxxxxxxxxxxxxxxxxx
> Lance wrote:
>> Thursday, 25 August 2005, 09:57 GMT 10:57 UK
>> 'Men cleverer than women' claim
>>
>>
>> Earlier this year, the president of Harvard University Lawrence
>> Summers sparked controversy when he suggested at a seminar that
>> one reason men outperformed women in maths and science was
>> genetics.
>>
>> Several guests walked out of the conference after hearing the
>> comments.
>>
>> Dr Summers, who has apologised repeatedly for his remarks, said
>> later that the shortage of senior female academics was partly
>> caused by child-minding duties, which restricted working hours.
>>
> These last three paragraphs seem to be tacked onto the article for no
> logical reason. They seem to connect recent research results to an apology
> from a Yank for not being PC - two quite different things.
>
> No doubt adding things that are utterly irrelevant is the BBC's way of
> showing 'balance'. On the one hand 2+2=4 on the other hand colourless
> green ideas sleep furiously.

The coverage on BBC Radio 4 was equally uninspired. When this report came
out, a number of alternative hypotheses come naturally to mind (well, my
mind at least), including:
1. The researcher has made a mistake in their analysis [sounds very
unlikely]
2. The meta-analysis has failed to correct fully for a bias in the conduct
of the original tests, i.e. a significant proportion of the tests were
biased against women in some way [possible but unlikely, as this would
surely be something the researchers would examine closely]
3. The meta-analysis suffers from a selection bias in the original tests.
For example, it is well known that on a wide range of performance measures
there is greater variation between males than between females. (As I
understand it, there are good evolutionary reasons why this should be so).
If it were the case that those towards the bottom of the scale were less
likely to take the IQ tests used in the meta-analysis then this would result
in the average male score being higher, even if the whole-population
averages were the same. [quite possible - this sort of thing is often missed
in meta-analyses]
4. Men are genuinely better than women on the sorts of abilities being
tested for in IQ tests [also quite plausible - there is increasing evidence
that men's and women's brains are (on average) better at different things]
5. Men are more intelligent than women in the broadest sense [I wouldn't
rule it out a priori (which I guess makes me non-PC!), but it would be
surprising to me and in any case it wasn't what the researcher was claiming]

The relative performance of boys and girls in exams appears to be partly due
to the exam system. It used to be the case that the entire mark depended on
final exams, which reward calculated risk taking, plus the ability to cram
and reproduce lots of facts in a short timescale. Now a substantial
proportion of the marks are gained through coursework, which rewards
application and diligence. There is also the factor that girls reach puberty
earlier than boys, meaning they tend to do better between the ages of about
10 and 16, although it is thought that boys catch up by 18.

Needless to say, instead of an intelligent discussion around, say,
hypotheses 3 and 4 and the changing nature of the exam system, most people
focus on hypothesis 5 and try to rubbish the research. Absolutely typical!


.