Herein lays a second pathway to that conclusion. I’ll provide a slightly less emotive, but equally valid example.

Back in the mid-90’s Professor Philip Rushton was cited within The Bell Curve. Part of the book investigated relationships between race and IQ. Rushton cited studies which seemed to conclude that black people had lower IQ’s than whites. In essence, what Rushton said was

“Stupidity rates are significantly higher in black communities — almost everywhere in the world.”

This book has since been cited by white surpremacists as academic proof of a definitive, biological link between low IQ and racial genetics. The results spread like wildfire in that community, igniting and justifying hatred through science, which Rushton et al continued to research and publish paper on. Persecuting action was all justified now. Communities can become intolerant to Black people in the name of eradicating stupidity. Black youth were demonised, communities segregated, and job options closed for them, because “they wouldn’t be able to reach management or drive the company forward”. But there was one small problem.

When it came to replicating the results, other researchers found zero evidence for the claims made by Rushton, when accounting for environmental factors, especially education. Hence, Rushton’s results on a genetic basis, simply were not true!

Image for post
A slide from one of Rushton’s talks on IQ in which he fails to account for the covariate

The Fatal Flaws

The work was examined by several peer-reviewing researchers. What they found was Rushton had totally failed to account for the role of education in his conclusions. Indeed, he took into account none of (none exhaustive list):

  • Poverty
  • Systemic underinvestment
  • Education
  • Systemic & temporal inequality

In one seminal talk, Philip Rushton presented a graph which showed the distribution of black university student IQ results versus whites. The graph appeared definitive and Rushton’s results appeared valid, at least superficially. Black IQ test results were significantly worse than white results.

However, when researchers examined the methodology and crucially, both the geography and the timing of the research, they found the claims of Rushton et al, astonishing. Why?

Cohort Covariates

Every piece of research is a journey between a hypothesis and a conclusion. The conclusion is empirically tested through the scientific method (double, blind, randomised controlled trials) and central to that is the ability to test against another [control] group, accounting for equality in every other variable but the variable(s) under test.

A covariate is a variable or path that can lead someone on a different route from the hypothesis to the conclusion. Giving scientists a “second path” to the conclusion. For example:

If I go outside and I walk back in wet, many people would come to the conclusion “it is raining outside”. Yet, it could be a number of other factors:

  • we were splashed by a car
  • upstairs neighbours threw a bucket of water over us
  • we walked into the sprinkler system whilst it was watering the lawn

[Keep it clean people!]

Each of those individual points provides a covariate, making the conclusion that “it is raining outside” scientifically unsound. Because it might not be raising outside. You have to do more research to truly determine that. For example, testing the “outside” by looking further out of the door, phoning a neighbour or something else.

During the rule of apartheid, black segregation, separate education and systemic inequality were the norm and indeed, mandated by law. The entire generation of black people in Rushton’s research with next to no exceptions grew up through a segregated upbringing and the results were stark for them, given the points above.

This entirely confounded Rushton’s work, since Rushton only used students from one single university (covariate 1 — not randomised), in South Africa (covariate 2 — systemic inequality), during the time of apartheid, which ended in 1991 (covariate 3 — failed to account for temporal inequality). Education was known to have a bearing on IQ results, so should have been accounted for.

Yet, Rushton’s research could only locate black people in South Africa affected by Apartheid and a lower quality education system. ONLY! Yet they used them in the research as if they had been brought up on an equal footing with South African whites all their lives. At best, 18 year old university students would have spent 15 of those years living under black segregated regimen. Thus coming to the conclusion that “controlling for other variables” (as established, they didn’t) black people have a lower IQ than white people, is blatantly false and scientifically unsound!

In true scientific circles, Rushton’s work was ridiculed. It was amateur, irresponsible, incredible and some thought, downright malicious. In wider society it was deemed dangerous, and quite rightly too. Sure, the science was ropy but the impact of that “scientific racism” was lethal! It was directly used as a reason for abuse against Black Americans and Africans for years.

Scientific Wording

In research circles, when testing a hypothesis, there are usually two alternative, mutually exclusive hypotheses you want to examine.

Null Hypothesis — A statement which is basically that “when I have an X characteristic , it makes no difference to the Y characteristics”. Hence:

“When I have a black geneology, it makes no difference to my IQ”

Alternat[iv]e Hypothesis — The statement you’re actually trying to prove.

“When I have a black geneology, my IQ is lower.”

The problem is that Rushton’s work was a “failure to reject the null hypothesis”. WTF does that mean? It means that you cannot accept the alternative hypothesis.

Let’s return to our “getting wet” example. We have two hypotheses to test there.

The Null Hypothesis — “You went outside and you didn’t get wet because it was raining”

The Alternative Hypothesis — “You went outside and got wet because it was raining”

The Null Hypothesis in the case of a sprinkler system, say, has not been disproven. We know that we cannot accept the alternative hypothesis because we cannot disprove that you walked into a sprinkler system, were splashed by a car etc.

The same is true of Rushton’s work, which is shocking! Absolutely shocking for a professor. The problem is he wasn’t a true scientist and this is perhaps what makes his conclusions even more dangerous. He had no credibility nor authority to make those statements nor come to those conclusions, but did anyway and in a scientifically invalid way. A bad scientist interprets the results Rushton found and comes the conclusion Rushton found. A good scientist says “we don’t know until we do more research” (which for the record, was done, it is conclusively education related and it shut down Rushton’s work completely).

What does this have to do with your statement about crimes and race? Your statement fails to account for several covariates and indeed, much of the research around at the moment also acknowledges one crucial factor:

Nobody studied enough disadvantaged white communities in enough detail to make any conclusion that it is exclusively a black problem. Collectively, all such research has failed to reject a null hypothesis stating “black people, for the simple act of being black, make no difference to crime” and the body of work definitely does not entitle them to claim that black people and crime are related. They haven’t rejected under-investment nor any other covariate.

Anecdotally, there are areas of Salford in Greater Manchester, and Glasgow in Scotland with extremely high deprivation rates. These “schemes” are like projects in the US, but unlike the USA, they are almost exclusively white areas. Yet, the same behaviours, the same high crime rates and the same rates of violence are seen in these communities, as there are in black communities “around the world”. Providing some evidence that there is absolutely no relationship between violent crime and colour, even if we don’t take into account any police prejudice. Black people are generally concentrated in areas of the world with high absolute poverty rates which we know have an effect on crime levels. Providing yet another covariate. Hence, your inference, is categorically unsound.

Written by

EA, Stats, Math & Code into a fizz of a biz or two. Founder: Automedi & Axelisys. Proud Manc. Citizen of the World. I’ve been busy

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store