Are stereotypes often true?

And why are some stereotypes false?

A 2009 paper by Lee Jussim, Thomas R. Cain, Jarret T. Crawford, Kent Harber, and Florette Cohen argued that most stereotypes are true, at least in the sense of representing valid estimates of statistical differences between various groups. In the introduction, the authors suggested that when conducting research on this topic they had faced some hostility from their colleagues:

Perhaps equally understandable, but scientifically untenable, is the corresponding belief that because stereotypes contribute to these many malignant outcomes, that they must also be—in the main—inaccurate. The tacit equation is, if stereotypes are associated with social wrongs, they must be factually wrong. However, the accuracy of stereotypes is an empirical question, not an ideological one. For those of us who care deeply about stereotypes, prejudice, and social harmony, getting to the truth of these collective cognitions should guide inquiry about them.

Unfortunately, this has not always been our experience. Because of his inquiries into stereotype accuracy, the first author has been accused by prominent social psychologists of purveying “nonsense,” of living “in a world where stereotypes are all accurate and no one ever relies on them anyway,” of calling for research with titles like “Are Jews really cheap?” and “Are Blacks really lazy?,” of disagreeing with civil rights laws, and of providing intellectual cover for bigots.

Clearly there are some differences between groups. Indeed, if groups did not differ in any way, then it’s hard to imagine how groups could even exist. And it seems plausible that human beings are smart enough that the tendencies they observe would be at somewhat accurate, at least on average. So, I’m not going to criticize the claim that stereotypes are more often than not consistent with statistical differences between various groups. Whether the implications of those stereotypes are interpreted correctly is quite another matter.

In this post, however, I’m much more interested in the cases where stereotypes do not seem to be accurate. How do inaccurate stereotypes develop? And why do stereotypes change over time? Is it because they were always inaccurate, or because they have become inaccurate?

Let’s start with a stereotype that has clearly changed over time. If you read books from the nineteenth century, you may run across the stereotype that Jewish people did not make good soldiers. Today, I almost never encounter that stereotype. One possible explanation is that in the late 1800s, Jews didn’t have their own country and were thus viewed as “rootless cosmopolitans”. Today, they have a homeland in Israel, and the Israeli military is widely viewed as a highly formidable army. Have Jewish people changed, or did society misjudge their character back in the 1800s?

Another common stereotype in the 1800s was the claim that Jews are greedy. Today, Jewish people are especially heavily involved in various philanthropic causes. In addition to donating more money than other groups, they contributed a disproportionate amount of time and effort to causes like the 1960s civil rights movement. Where did the previous stereotype come from? One possibility is that a combination of Christian usury prohibitions and the need to be able to flee a country during pogroms led to Jewish people being disproportionately involved professions such as banking. Perhaps people assumed that bankers were greedy, and then transferred that stereotype to Jewish people.

I would argue that many false stereotypes come from people transferring true stereotypes to seemingly related claims that are not in fact valid. Thus consider the stereotype of the “dumb jock”. Is it true that good athletes are less intelligent, on average? Some studies suggest that the reverse is true:

Elite athletes are generally smarter than us – cognitive sciences can explain why

How might the dumb jock stereotype have developed? Suppose that Ivy League colleges are willing to accept less talented students on athletic scholarships. If the average Harvard student has an IQ of 135, and the average Harvard athlete has an IQ of 120, then the student athletes might be viewed as ”dumb” by their peers. Notice how a correct stereotype (student athletes are dumber than their peers) leads to an inaccurate stereotype (athletic people are dumber than the average person.)

I wonder if something similar occurs with beautiful women. Isn’t there a sort of “dumb bimbo” stereotype? Now suppose that beauty, especially female beauty, allows one to be more successful than otherwise. In that case, beautiful people might rise to higher than average positions in society, where they are surrounded by people with above average intelligence. Have you ever watched a news show on election night with its handsome male newscaster and beautiful female newscaster, and then seen the camera switch over to an unattractive guy interpreting some poll numbers? Don’t we implicitly assume that the nerdy statistician must be very smart—how else would he have gotten that job?

Most people reading this blog are intellectuals with above average intelligence. Many of our friends and colleagues are also smarter than average. So even if we regard a pretty female talking head at Fox News as “dumb”, she’s very likely well above average in intelligence. And that doesn’t even account for the fact that it’s hard to appear smart when you are under pressure and there’s a TV camera in your face—indeed I’ve struggled to provide coherent answers when I’m interviewed on TV.

To summarize, athletes and models may be dumber than the average person in their income level, and still be smarter than the average person in society as a whole.

Now let’s consider another possible stereotype. I cannot be certain, but I suspect that most people believe African-Americans are poorer than Japanese people. In fact, it seems likely that African-Americans are richer than the Japanese (albeit poorer than Japanese-Americans.) This Statista graph suggests than in 2023, median family income of blacks was 30% below the US average, and even further below the white average:

I wasn’t able to find exactly comparable data for Japan, but the IMF says that in PPP terms (ie. adjusted for difference in the cost of living), per capita GDP in Japan is 39% below the US average. Even though median income and per capita GDP are different concepts, the difference between 30% and 39% is large enough that I suspect Japanese incomes do trail the average income of African-Americans. AI overview reports this data point for 2021, which further supports my claim:

In 2021, Japan's median household income, adjusted for purchasing power parity (PPP), was $45,601, according to GlobalData.

That’s a lot lower than Statista’s $54,080 figure for black families in 2021. So why are blacks in America widely perceived as being poorer than the Japanese? Partly because blacks in America are far poorer than Japanese-Americans. Indeed, even white Americans are poorer than Japanese-Americans. But it is also partly because the black community experience relatively high rates of various social problems such as drug use, crime, homelessness, etc., compared to people in Japan. Again, that’s likely true of white Americans as well, who also experience more of these problems than the average Japanese citizen. Because these pathologies are correlated with poverty, people often draw the inaccurate inference that African-Americans are poorer than the Japanese.

To be clear, my point here is not that American blacks are “doing better” than Japanese. Rather that the stereotypes that people have are only loosely related to the actual situation. It’s perfectly reasonable to argue that the socioeconomic condition of African-Americans is worse than the socioeconomic condition of the Japanese, for all sorts of reasons. That is my view. But it is highly misleading to suggest that the differences are all about income.

BTW, a related stereotype is that most African-Americans are poor, whereas the actual poverty rate for blacks in 2022 was 17.1% (and trending lower), which is well above the rate for whites, but well below 50%. Roughly 83% of American blacks are not poor. Black Americans are disproportionately poor is a correct stereotype, but black Americans are poor is not a correct stereotype. If I’m right that most Americans would get the African American—Japanese income comparison wrong, it’s likely because back poverty occurs against a backdrop of unusual American affluence.

Now consider the stereotype that Germans are hard working. Later I’ll argue that in some respects this stereotype may be true. Nonetheless, in the most widely used sense of the term it is not only false; it is almost the exact opposite of reality. Germans have perhaps the lowest level of annual hours worked in the entire world:

So where did the hard-working German stereotype come from? Consider that Germany is a very affluent country with a highly productive workforce. In addition, Germans are often seen as being quite orderly and disciplined. Those presumably accurate stereotypes are often (correctly) associated with hard work. And perhaps during the 1360 hours that the average German is on the job, they do work very hard. Maybe they take fewer coffee breaks. Or perhaps they study harder when young, allowing them to become more productive and thus work fewer hours. But it is not true that Germans are hard working in the conventional sense of the term—working long hours.

Once again, we see possibly accurate stereotypes about a group leading people to infer closely related stereotypes that are not even close to being accurate. More broadly, there’s often an assumption that people in poor countries are not hard working. I suspect that this generalization is also false, that people in poor countries tend to work harder than people in rich countries. Here’s AI Overview:

In Latin America, the average annual hours worked varies by country, but generally falls between 1,900 and 2,200 hours per year. Mexico and Costa Rica tend to have some of the highest average annual hours worked in the region, while Chile is also known for its relatively high number of hours worked.

In addition, people in poor countries often work harder in the sense of doing difficult manual labor, rather than sitting at a desk in an air-conditioned office.

My concern is not so much with the accuracy of stereotypes as it is with the way that they are used. Thus in 1900, it was probably accurate for an American nativist to note that most Italian immigrants were coming from southern Italy, an area with a major organized crime problem (both then and now.) To some extent they brought this organized crime to America, and when I was a child (in the 1960s) the Mafia was still viewed as a very big deal, a far greater threat than “terrorism”. But in recent years, the Mafia has dramatically declined in importance, as Italian-Americans have assimilated, becoming much more like other Americans. (Even terrorism now seems to be declining in importance, so I’d compare attitudes toward the Mafia in the 1960s to attitudes toward terrorism 10 or 20 years ago.)

The initial stereotype may have been partly accurate but did not necessarily have the policy implications that nativists might have assumed back in 1900—especially those who saw criminality as some sort of “innate” characteristic of southern Italy.

Another problem is that stereotypes are often applied in an insulting fashion. White Americans often label black and Hispanic students as “lazy”, because they supposedly don’t study hard enough. Asian-American students are labeled “grinds”, because they supposedly study too hard. Perhaps unconsciously, whites seem to assume that the white level of effort is optimal. What a convenient assumption!

Similarly, the murder rate for whites is lower than for blacks or Hispanics, but higher than for Asian-Americans. Perhaps whites have also stumbled upon the optimal quantity of murder. Two hundred years ago, Thomas DeQuincey had this to say:

A golden mean is certainly what every man should aim at. But it is easier talking than doing; and, my infirmity being notoriously too much milkiness of heart, I find it difficult to maintain that steady equatorial line between the two poles of too much murder on the one hand and too little on the other.

Some educators might assume that the East Asian attitude toward studying is obviously best. Maybe so. But note that Korea and Taiwan have fertility rates of 0.8, whereas Sub-Saharan Africa is closer to 4.3. Is that difference related to the grind in academia?

PS. I wonder if the decline in the American mafia causes younger moviegoers to view a film like The Godfather differently from the way we viewed it when it was released back in 1972.