Repeat Offender: Satoshi Kanazawa’s Other Greatest Misses

By Arturo R. García

Satoshi Kanazawa’s Monday blog post about black women and beauty standards, since taken down, was only the latest in a string of questionable contributions to both Psychology Today and his field.

In 2006, Kanazawa was accused of reviving eugenics-era theories after publishing a paper in England blaming low IQ levels for low life expectancy and high infant mortality rates in the continent of Africa – seemingly ignoring decades worth of political and social unrest. This led to him being called “the great idiot of social science” by renowned biologist PZ Myers in an article last year on Alternet.

Daniela Perdomo’s piece for Alternet focused on another study by Kanazawa, this one alleging that atheists are “more likely to acquire and espouse evolutionarily novel values and preferences (such as liberalism and atheism…) than less intelligent individuals.” Perdomo writes:

… Not only does Kanazawa wax over structural inequalities that may lead to varying IQ levels in American society, even the disparities he finds in this imperfect measure of intelligence are relatively miniscule. For the most part, he is not speaking of a difference of more than six IQ points between liberals and conservatives, atheists and believers — a negligible difference one would never notice in real person-to-person interactions.

Kanazawa isn’t the first to study the intelligence-religiosity nexus. Other studies have also found a three- to six-point IQ difference between atheists and religious believers, in the atheists’ favor. But those studies didn’t claim that atheists were more evolved, as Kanazawa presumes, and merely conclude that they are more skeptical due to a certain kind of schooling and cultural exposure (which might also account for why some people perform well on IQ tests), leaving room to account for why so many people — say, like William F. Buckley, Jr., the late conservative public intellectual — can be so religious and conservative and yet quite intelligent.

In February 2008, Kanazawa defined his position as “extremely purist” in a post in Psychology Today, saying findings can only be either true or false:

No other criteria besides the truth should matter or be applied in evaluating scientific theories or conclusions. They cannot be “racist” or “sexist” or “reactionary” or “offensive” or any other adjective. Even if they are labeled as such, it doesn’t matter. Calling scientific theories “offensive” is like calling them “obese”; it just doesn’t make sense. Many of my own scientific theories and conclusions are deeply offensive to me, but I suspect they are at least partially true.

Once scientists begin to worry about anything other than the truth and ask themselves “Might this conclusion or finding be potentially offensive to someone?”, then self-censorship sets in, and they become tempted to shade the truth. What if a scientific conclusion is both offensive and true? What is a scientist to do then? I believe that many scientific truths are highly offensive to most of us, but I also believe that scientists must pursue them at any cost.

It is not my job as a scientist to “use” scientific knowledge in any way to improve the human condition; that’s the job of politicians, policy makers, physicians, and other social engineers. Their goal of helping people and improving their lives is a noble and important (albeit nonscientific) one. Any successful intervention, however, must be based on the true understanding of nature. If these social engineers don’t know the true causes of what they are trying to create or eliminate, how can they possibly hope to succeed? By opposing and entirely disregarding certain scientific theories and conclusions a priori on ideological and political grounds, because they believe they could not and should not be true, they risk the chance they might not achieve their goal of helping people.

Less than a month later, however, he engaged in a rather unscientific – and genocidal – bit of speculation as to how the United States could have ended the “war on terror” more quickly, emphasis his:

Here’s a little thought experiment. Imagine that, on September 11, 2001, when the Twin Towers came down, the President of the United States was not George W. Bush, but Ann Coulter. What would have happened then? On September 12, President Coulter would have ordered the US military forces to drop 35 nuclear bombs throughout the Middle East, killing all of our actual and potential enemy combatants, and their wives and children. On September 13, the war would have been over and won, without a single American life lost.

That post is still active on PT’s website, while Monday’s has been pulled – justifiably, according to fellow PT blogger Mikhail Lyubansky. But it wasn’t because Kanazawa’s work arrived at an unpopular confusion, emphasis his:

The point is that there are also group differences, not in attractiveness (as Kanazawa claims), but in cultural messages about what is and is not attractive. Standards of beauty, like most other beliefs, are socialized and change not only from place to place but also over time. In both the United States and England, (where Kanazawa lives and works), standards of beauty are essentially “White” standards, because whites comprise the majority of the population and have disproportional control over both media and fashion. And while it is not just White respondents who are socialized this way (internalized racism has been well documented), it is certainly the case that White Americans and Europeans (who are less likely to have received more positive messages about Black beauty) would show the strongest anti-Black bias.

As long as this is understood and framed accordingly, there is no problem with the data Kanazawa reports. What they show is that because Black faces and bodies don’t fit mainstream White standards of physical attractiveness, both respondents and interviewers show an anti-Black bias. Unfortunately, Kanazawa fails to consider either sample bias or socializing effects. Even if he believes, as he apparently does, that human behavior is entirely “evolutionary”, good science requires a careful analysis of sample bias and an explicit discussion regarding the study’s generalizability. Without this kind of methodological analysis, Kanazawa’s entire premise — that there is such a thing as a single objective standard of attractiveness — is fatally (and tragically) flawed.

It is worth noting that Kanazawa repeats this same flaw of omission when he explains that the attractiveness results are not due to race group differences in intelligence, as though there are no scholarly critiques of IQ measures in general and their racial bias in particular.

These are not trivial omisisions. They are the necessary context that gives readers the information they need to draw their own conclusions.