IQ and Race: Part 2

I have a masters in organisational psychology, and have studied psychometrics at a reasonably high level, and use psychometric testing as part of my profession. So I am reasonably qualified to comment on this issue.

I thought it would be useful to give my comments on the current debate on IQ and race raised by Molyneux in his videos.

When you read part one of my article yesterday you probably thought that I came across as a bit of a lefty on this. But it is what I think is the correct way to look at the topic from a science perspective.

Problems with IQ tests.

A single IQ score is not very useful for a lot of purposes. So, when I test job applicants, the test I use measures: understanding of English, numerical ability, and logical ability. Understanding of English is important to me because most jobs I test for require that the candidate can communicate well in English. But here underlies a problem with tests:

Someone who is not good at English would probably do poorly on the rest of the test because that is written in English as well. So, changing the language will fix this? Not necessarily. For instance a question like: Unscramble the name of this city: LACNUKAD would probably be easy for a New Zealander because we are familiar with Auckland. But maybe not for a Russian (same question in Russian).

So we change these questions to suit different races/language groups right? Perhaps ask the Russian to unscramble the name of a Russian city. ?Well, that would make the test more valid for that specific group. The problem is that it makes it very difficult to compare between groups because the questions need to be shown to be equal in difficulty between different tests, which is not that easy to do. So, is the test given to a Russian person in Russian going to match reliably with a similar test given to a New Zealander in English? And how would we know this?

There have been a number of attempts to solve this issue by developing ?culture free? tests such as the Raven?s matrices. These are pattern solving tests. While these go some way to solving the problem, they do tend to narrow considerably the dimensions being tested. So, someone may be a brilliant playwright, but do hopelessly on one of these tests.

So there is still considerable debate about comparing people from very different cultural backgrounds. Hence, there isn?t by any means universal agreement that races can be ranked by IQ tests.

Can measured differences be due to genetics alone?

When examining an issue such as the relationship between IQ and race, it is important to weed out all the factors other than genetics that could be contributing to the result. For instance, poverty is negatively associated with IQ. Some groups are disproportionately represented in poverty statistics compared to other groups. As an extreme example, someone growing up in Somalia for instance. So, poverty could well skew the results. Factors associated with poverty such as poor nutrition can affect brain development, and these factors can be related to poverty. There are a wide range of factors such as this that can affect IQ results. To compare between races we need to weed these factors out.

What about a pigmy bushman who may score 70 on an IQ score??He must be thick right? Not necessarily. A pigmy has grown up in a completely different environment and may have learned to think in a completely different way to us. But, if the test were to include items that related to his environment, then there may be a completely different result. If we were tested on his test, we may perform very poorly as well, simply because the test is completely outside our way of thinking and solving problems.

Ideally, to get comparable results, we should be able to measure people of different races who have grown up in identical environments, identical levels of income, educational opportunities etc, in order to get directly comparable results that we could say with a degree of confidence is due to differences in racial genetics.

If we were able to do this on a big enough scale to produce valid results, then we could answer this question with a degree of confidence. However, I am not sure that in this PC world there is enough motivation or money available to answer this question in a definitive way. But, I suspect that if it were possible to study comparable IQ in this way, then the difference in results may well be considerably less than some of the results to which Molyneux refers.

So what are my concerns with Molyneux?s approach?

My concern is that his arguments are over-simplified and ignore a lot of relevant factors. This can lead people to make unjustified assumptions, for instance, that ?blacks are stupid? when there simply isn?t data available to make such an assumption.

I hope I have shown here that measuring IQ is not like measuring the weight of a bag of flour. IQ is a theoretical construct, and there is a lot of complexity and factors that contribute to results. So discussing this topic requires consideration of those factors.

That doesn?t mean there aren?t genetic differences between races in intelligence. But when it boils down, those differences might be quite small when everything is taken into account, and certainly no basis to make assumptions about individuals within those races.


by Tony Norriss