Has it been scientifically proven racists and homophobes are stupid? Er, no.

A study from last year suggests something that most of us decent hummus-munchers will laugh at our hands behind: racists and homophobes are stupid, and the stupidity of their being racist and homophobic is mediated by them also being right-wing.

The paper, Bright Minds and Dark Attitudes: Lower Cognitive Ability Predicts Greater Prejudice Through Right-Wing Ideology and Low Intergroup Contact (Hodson & Busseri, 2012; paywalled, alas) suggests it has found a predictive between low general intelligence in childhood and greater levels of prejudice later life. The link was not direct, though. It was mediated by right-wing ideology. For homophobia, the link was also mediated by a low level of contact with gay people. The sample was large, with data from almost 16,000 people, and a longitudinal design was used, which is more robust than simply testing for a correlation. Sounds compelling?

Well no. In fact, it’s one of those studies that makes me want to set things on fire, it is so poorly conducted.

Participant problems

Let’s return for a moment to our participants. It is very important to note that the researchers did not directly collect any of the data they studied, and therefore are relying on existing datasets. This gives them little control over the measures used, and limited information about the participants which could prove to be pertinent.

They are mostly, shall we say, of a certain generation: one dataset of participants had their intelligence measured at the age of 10 or 11 in 1958, while the others had their intelligence measured at the age of 10 or 11 in 1970. In these groups, racism and right-wing beliefs were measured when the participants were in their early thirties, meaning that all of this was measured, at best, more than two decades ago. That long ago, Britain was a very different place, and I’m glad things don’t look so much that way any more, with white people fucking everywhere being all white supremacist. There was no information provided about the ethnicity of the participants, but given the time, it is safe to assume that the overwhelming majority of them were white.

Meanwhile, the sample for measuring homophobia was also pretty lacking, consisting of less than 300 US university students, who are hardly known for being representative of the general population. Unlike the participants included in the racism study, this project was, as far as I can discern, not longitudinal, but a cross-section, making it much harder to draw conclusions about anything being predictive.

Mangled measurements

Oh dear, where to begin? Every single measure used here is a whole can of worms, difficult to measure at the best of times.

In general, measuring intelligence is a fucking nightmare. Nothing is particularly satisfactory, and everything is likely to lead to raised eyebrows and sighs. This is, at least in part because it’s difficult to even agree on what intelligence is, let alone how best to measure it. It is also pertinent to note here that the vast majority of the participants had their intelligence measured decades ago, and that these measures may not necessarily be favoured any more, following a very long period of refinement and academic critique. Also, there’s loads more that, frankly, I’m already too tired of discussing the clusterfuck that is measuring intelligence to discuss, but please do pop into the comments with your thoughts on the matter because there’s lots  to talk about.

And do you know what is just as contentious as measuring intelligence? Measuring prejudice. It was noted quite a while ago that directly asking people about unpopular beliefs (and, of course, overt prejudice is hardly fashionable) will tend to lead to denial due to social desirability–people say what they think others want to hear. This is further complicated by the fact that over the years, what prejudice actually looks like has changed considerably. It is no longer “blacks need not apply” and “there goes the neighbourhood”, but, rather, dog whistles and unconscious biases and benevolent sexism and so forth. Obviously, this was not measured since most of the measuring was done so long ago. Instead, to measure racism participants were asked to indicate agreement or disagreement with statements such as “I wouldn’t mind working with people from other races”. Measures of homophobia were similarly direct.

Ultimately, if we pretend that the measure of intelligence was all right, the best we can conclude from this research is that low intelligence is indirectly associated with actually admitting to the fact you’re a fucking racist.

“Oh my god, they used Baron & Kenny” “You bastards”

Surely, at least the statistics are fairly robust?

Nope. Oh gods, no they aren’t.

Regular followers of this blog may have noticed I have certain nemeses, from certain feminists who reject intersectional analyses because it’s too hard, to Brendan O’Neill. Here’s another of my nemeses: the Baron and Kenny method for mediation analyses.

It is perhaps the most popular method for testing mediation, and it is popular because it is simple enough to do with a fairly basic statistical package without having to delve into writing syntax or running stats for hours. Basically, you need to run a few tests. You need to see if there is a significant correlation between the independent variable (in this case, intelligence) and the mediator (in this case right-wing beliefs). Then you check if there is a significant correlation between the mediator and the dependent variable (racism or homophobia in these studies). Finally, for a mediation to exist, there needs to be no significant relationship between the independent and dependent variables when controlling for the other two tests.

This diagram from Hodson & Busseri might make it easier to visualise:

Apologies for the lack of description. I am really slow at being able to interpret graphical representations of things. Does anyone want to volunteer to describe this image?


Anyway, there’s a lot of problems with this overly simplistic approach which the very-interested can read all about here. In short, multiple mediators can cancel each other out and basing things on significance might be a very trivial change indeed.

tl;dr: Take everything you see using Baron & Kenny with a vat of salt.


It might be nice to believe that our enemies are stupid, but when you think about it, actually that’s rather sad. It reduces the possibility for education if it prejudice is largely driven by a relatively fixed factor. It stops us from trying to understand where prejudice comes from and why people are prejudiced. It is really rather bleak.

The good news is, that study is largely a nonsense, so let’s get back on with smashing prejudice.

3 thoughts on “Has it been scientifically proven racists and homophobes are stupid? Er, no.”

  1. Another problem I have with the whole “only stupid people are bigots” hypothesis is that it completely ignores the problem of institutionalised prejudice; rather pinning the problem solely on individuals. Also, it gives people who already take exception another excuse to hide behind – “oh, I’m not stupid so I can’t be a bigot.”

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

This site uses Akismet to reduce spam. Learn how your comment data is processed.