« HAPPY NEW YEAR! | Gene Expression Front Page | 30000 BC »
January 01, 2004

Post-festive musings

Well, I go away for a week or so, and when I get back I find everyone else has been blogging away like beavers. Why weren’t you all celebrating Kwanzaa?

While I was away I saw The Return of the King, which was much-discussed in earlier posts. I won’t comment on the merits of the film or the book, except to say:

(a) Andy Serkis deserves some kind of Oscar for his role as Gollum. I had assumed that the voice was electronically distorted, but I saw a TV interview where he did it live - rather creepy. And although what you see of Gollum is CGI, it is modelled on Serkis’s movements and expressions.

(b) Liv Tyler: - where was she??? All the publicity for the film features Liv prominently, but she is on screen for about 5 minutes max, and only has a handful of lines. As Liv is one of my favourite babes - sorry, I mean actresses - this was a severe disappointment.

BTW, there is an article in a recent issue of New Yorker magazine arguing that Tolkien owed more to Wagner’s Ring than he was willing to admit (he said ‘Both rings are round, but the similarity ends there’.)

Turning to less serious subjects, at least two recent newspaper articles have referred to ‘research’ into national IQ levels by a team at the University of Vienna Medical School. Chris Brand, in his provocative but enjoyable blog, describes it as replicating the work of Richard Lynn and Tatu Vanhanen.

Well, until proven otherwise, I will take replicate as meaning copy. (Oxford English Minidictionary definitions: replicate = make a replica of; and replica = an exact copy.)

The newspaper articles quote the ‘Vienna’ IQ levels (to the nearest point) for 8 countries (UK, USA, Hong Kong, Australia, New Zealand, China, Belgium and Ireland), and give rank orders for several others. These are all identical to those of Lynn and Vanhanen. Bearing in mind the numerous judgement calls and technical adjustments involved in these figures (see the continuation) the probability that two sets of researchers would independently come up with exactly the same figures for as many as 8 countries is remote. So I would guess (being a charitable kind of guy) that the Vienna team have simply quoted L & V’s data for some purpose of their own, and the journalists have misunderstood what they actually claim to have done.

L & V’s data are selected from a variety of sources, using different tests carried out at different dates. Judgement is required as to the best tests to use for this purpose and the validity of the data. The raw data are then adjusted to allow for the Flynn Effect and expressed in terms of a notional UK mean of 100 (for more detail see my post 'In like Flynn'). In some cases further adjustments are made to allow for unrepresentative samples (e.g. one of the raw data entries for China is ‘arbitrarily reduced’ by 6 points). In these circumstances, it would be generous to suppose that the probability of two researchers independently getting the same mean IQ (to the nearest point) for any given country is as high as .5. (This is equivalent to having only two equiprobable alternatives to choose from. If there were as many as three viable alternatives, with probabilities of, say, .25, .5, and .25, the odds against two researchers independently getting the same results would be closer to 2:1.) The odds against two researchers independently getting the same results for as many as 8 countries (without any discrepancies) are therefore longer than 250:1.

Posted by David B at 09:13 AM