Another look at exclusions and race

November 17, 2019

About a year ago I wrote this post looking at the data on permanent exclusions and ethnicity.

I had often seen it claimed that the differing rates of permanent exclusions in different ethnic groups were evidence of racism and grounds to reduce, restrict or stop exclusions.

In that post I observed the following:

  • The numbers involved are actually tiny. This means that a) no ethnic group is affected massively by permanent exclusions, and b) if somebody claims there is racism driving these figures, they are talking about the actions of far less than 1% of heads, which could be individually scrutinised, not “institutional racism” or “unconscious bias” throughout the system.
  • While individuals should always be judged as individuals, there is absolutely no reason to assume that the extreme behaviours that result in exclusion will be constant across different demographics. Different ethnic groups do not have the same history, social class, religion, geographical location, type of school attended and so on. To assume the same behaviours from all demographics would be wishful thinking.
  • The differences between ethnic groups above, conform to no obvious pattern of widely held prejudices. Does anyone believe that black African children experience far less racial prejudice than black Caribbean children? Or that Pakistani Muslim children are subjected to less racial prejudice than white children?

Since then I’ve had a few more things to think about.

The first is that anecdotally, at least one teacher has told me of their experience of exclusions being affected by prejudice. This does not mean that it is common, but it is something that should not be dismissed even if it is unlikely to be affecting the averages. However, if it is the case that some school leaders let prejudice affect their decisions on exclusions, then this is not necessarily a case to make all heads less able to exclude. It would be a reason to make grounds for exclusion clearer and more objective. It should be possible to say, “Well, you didn’t exclude a white child for the same thing.”. However, if schools did set clearer grounds for exclusions, it is likely to drive exclusions up, not down. It is easy to make exceptions to keep kids in school, far harder to push them out with no good reason. Greater consistency in the use of exclusions would be a “no excuses” approach, not an “a lot more excuses” approach. There is also no reason to assume a reduction in exclusions would change relative rates of exclusions.

A further development is that the Timpson Review tried controlling for all other factors when looking at the different rates of exclusion between different ethnic groups.

The discrepancies between ethnic groups were massively reduced, with only 3 groups being more likely to be excluded than white children. This suggests my argument from the previous post was correct, it also reduces still further the number of heads whose decisions might have contributed to the disparities from a few hundred (out of tens of thousands) to a few dozen, making talk of “institutional racism” even less plausible.

Does the remaining disparity provide good evidence of racism? Not really. We still wouldn’t expect total uniformity of exclusion rates after controlling for all those other factors, even in the complete absence of prejudice. This research controlled for the data schools keep, but not for all possible factors. It won’t control for peer group or for crime in the local area, or any other factor schools don’t keep records on. Any imperfections in the data will also have an effect as controlling for other factors requires accurate data.  We still cannot assume racism is the explanation for the remaining disparities. What we do know from this, is that prejudice can, at worst, only be a marginal factor in permanent exclusions and cannot be a reason to restrict or reduce exclusions.

And finally, there is the issue of just how subjective attempts to prove prejudice from data can be. In a completely different context (football), research by Silberzahn and Uhlmann looked at how the same data can lead to drastically different conclusions.

Twenty-nine teams involving 61 analysts used the same data set to address the same research question: whether soccer referees are more likely to give red cards to dark-skin-toned players than to light-skin-toned players. Analytic approaches varied widely across the teams, and the estimated effect sizes ranged from 0.89 to 2.93 (Mdn = 1.31) in odds-ratio units. Twenty teams (69%) found a statistically significant positive effect, and 9 teams (31%) did not observe a significant relationship. Overall, the 29 different analyses used 21 unique combinations of covariates. Neither analysts’ prior beliefs about the effect of interest nor their level of expertise readily explained the variation in the outcomes of the analyses. Peer ratings of the quality of the analyses also did not account for the variability. These findings suggest that significant variation in the results of analyses of complex data may be difficult to avoid, even by experts with honest intentions.

How could teams of researchers differ in their conclusions, when looking at the same set of data? A lot came down to what controls were considered relevant and what statistical tests seemed appropriate. I don’t suggest that this makes it impossible to ever demonstrate prejudice. But if professional statisticians can differ as to what data shows, we should perhaps be more sceptical at the certainty with which the politically motivated draw simple conclusions about racism from raw figures. And that applies whether the conclusions are being reached about school exclusions, the level of diversity of speakers at a conference or in arguments about whether people’s lists of favourite tweeters are representative or not.

If people wish to believe that exclusions are evil, or that schools are institutionally racist, they will. But the evidence is simply not there in the figures for permanent exclusions, and there is absolutely no reason to think reducing, restricting or banning permanent exclusions will address any injustice.


  1. White people and their “burden of proof” is just more evidence of institutionalised racism.

    I’m surprised you haven’t worked this out yet.

  2. You’re a racist. You shouldn’t be let near children.

  3. […] discuss why those who claim exclusions are racist are wrong (as I’ve already done that here, here and here). However, I do suspect that the national debate on school exclusions has been massively […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: