20 July, 2004: Not like us?

[ Home page | Web log ]

No doubt each of my half-dozen readers sees the Guardian every day and there is little need to draw their attention to yesterday's front-page headline, titled, 90% of whites have few or no black friends. The story relates to a recent poll conducted by YouGov for the Commission for Racial Equality, for a report to be released later this week. Besides the headline number we learn,

(The poll asked respondents to consider `their closest 10 to 20 friends'.)

Data on the ethnicity of the UK population were collected by the 2001 census, and are presented in an appallingly non-accessible table on the National Statistics website. Reformatting this properly and crudely amalgamating categories, the proportions of the whole UK population in 2001 were,

Ethnicity %
White 92.1
mixed 1.2
Asian (various) 4.0
black 2.0
Chinese 0.4
all minority 7.9

(Note that I've followed the Guardian Style Guide in capitalising `Asian' and not `black'. I'm not sure I'm happy with that, but then I am quoting from their article....)

On the religion side, the fractions are,

Religion %
Muslim 2.7
Hindu 1.0
Sikh 0.6

but note that the numbers of practising members of those religions will be lower.

Suppose that the population were homogeneously mixed. How would we expect those headline numbers to turn out? We need one more assumption, which is the number of friends each respondent has. God know how you'd find that out, so let's plump for 15, in the middle of YouGov's suggested 10 -- 20 range. `Most' of this group is, let's say, three quarters: 12 or more. (More detail: the argument here is that if people make friends in a homogeneously mixed population, then the probability of each friend being from a particular ethnic group simply reflects the fractions of the population belonging to those groups; it is then simple to compute the fraction of people we would expect to report a particular number of their friends being of certain ethnicities.)

On this basis we would expect that:

Two obvious comments here:

  1. The population obviously isn't ethnically or religiously homogeneous, at all. This is especially obvious on the religious side: people who go to mosque or temple or church are likely to be friendly with other members of the same congregations.
  2. And yet, with one exception, the numbers reported in the Guardian's piece are pretty close to what you'd expect to get in a homogeneous population.

So, does that mean we should pat ourselves on the back and think happy thoughts about how integrated Britain is becoming? Sadly, no.

That exception in (2) is that members of the ethnic minorities tend to have more friends from the ethnic minorities than we would expect in a homogeneously mixed population. This is a measure of the extent to which Britain's population is not integrated: if we imagine a simplistic model with a white majority and a single non-white minority, which is split into an integrated population and an isolated non-white minority, where a `mixing fraction' of the minority's members have joined the integrated population, we find something like this:

Population mixing

Note that this model doesn't really work (it always predicts values too high for the fraction of white people having mostly white friends). But it does illustrate that the extent of population mixing has a huge effect on the number of non-white people reporting mostly non-white friends, and very little effect on the number of white people reporting mostly white friends.

The Guardian goes on to quote Trevor Phillips (chair of the CRE):

When it comes to race and religion this clearly demonstrates we are dealing with a difference of which most people in this country have no first-hand experience, and therefore it is not surprising that they can be misled about blacks, Gypsies and Muslims, and it's not surprising that for no apparent reason they can become hostile and racist.

This is superficially plausible. Certainly I remember when I first encountered that idiot Peter Cuthbertson I put his prejudices down to the fact that he grew up in 98.5% white Darlington. But from the above it's clear that, at least on the measures in YouGov's survey, the extent to which -- in aggregate, at least -- white people have ethnic minority friends is currently not that far off the extent to which they would do in an homogeneously mixed population. If there is a solution to white racism, I suspect that further integration alone isn't it (though obviously these results will differ from place to place).

Phillips proposes summer camps for 16-year-olds, to be used (basically) for social engineering. This idea makes me pretty uneasy, but clearly something has to be done to fix one horrifying problem mentioned in the Guardian's article:

In January, a MORI poll found that 41% of white people and 26% of ethnic minority people surveyed wanted the races to live separately.

(I can't find the MORI report but note that in a -- say -- 1,000-person survey, we'd expect about 79 ethnic minority respondents. One hopes that, like the YouGov survey here, the MORI one asked a separate ethnic minority sample; otherwise, the margin of error on that 26% would be huge.)

In other news

Congratulations to Francis and Julian on winning a New Statesman New Media Award with Public Whip. Since he's also published this photo of the Downing Street Says team from a separate event, I suppose I can too without looking too much of a self-publicist:

Look, mum! I went to Downing Street!

Elsewhere we discover that David Blunkett is organising a five-year plan (I wonder why nobody has thought of doing that before) and that fascists get just as lousy service from the banks as everyone else. The latter, I suppose, is heartening, in a way. And don't forget to send in your consultation responses to the Home Office by half-past-five today. It won't make a bit of difference, but it's the thought that counts.

Copyright (c) 2004 Chris Lightfoot; available under a Creative Commons License.