Last week, the conservative Heritage Foundation issued a new study purporting to show that letting illegal immigrants from south of the border become citizens would cost more than $6 trillion dollars in social benefits by 2050. Researchers from all segments of the political spectrum contested that finding. In the midst of the controversy, it turned out that one of the study's authors, Jason Richwine, had argued in his 2009 Harvard dissertation that immigration policy should focus on selecting and admitting individuals with higher IQs. Naturally, all hell broke loose, and the brave leadership at Heritage tossed the hapless Richwine overboard by the end of the week.
So what exactly did Richwine's dissertation say? And is there any truth to it?
For most of his analysis, he relies on the work of the University of Ulster psychologist Richard Lynn and the University of Tampere political scientist Tatu Vanhanen, which focuses on differential average IQs between nations. Consequently, Richwine's main argument is that Hispanic immigrants and their American descendants have low IQs because they have immigrated from countries whose citizens have low average IQs. As further evidence, Richwine cites a review of 39 studies by the Clemson psychologist Philip Roth and colleagues that reports that Hispanic-American IQs average 89.2 points. Richwine also reviews research suggesting that second-generation average Hispanic-American IQs and incomes do increase, but that there is not much of an increase in either over subsequent generations. Richwine agrees that these IQ deficits in poor countries are partially the result of bad nutrition, pervasive infections, and lack of adequate schooling, but he also suggests that there is a significant genetic component.
Early in the 20th century, various immigration restrictionists made a somewhat similar argument. While Richwine is aware of the debates over immigrant IQs that took place in that period, he concludes that "there was hardly any consensus at all about that topic." A brief look back at leading research at the time finds that Richwine is being overly dismissive: Now-discredited theories about ethnicity and IQ dominated the field.
For example: In 1919 Lewis Terman, one of the founders of intelligence testing, reported in his book The Intelligence of School Children that the American-born children of Spanish immigrants had an average IQ of 78. For the children of Portuguese immigrants, the average was 84; Italians, 84; northern Europeans, 105; and Old Stock Americans—descendants of British, German, and Dutch immigrants—106.
The Princeton psychologist Carl Campbell Brigham, using data from intelligence tests administered to army draftees during World War I, reported the results in his 1923 opus A Study of American Intelligence. The Yale psychologist Robert Yerkes, summarized Brigham's data in a 1923 article for The Eugenics Review, asserting that American army officers scored an average of 18 on Brigham's intelligence scale, whereas the average white draftee scored 13.77. Immigrants from England scored 14.87; from Germany, 13.88; from Norway, 12.98. Irish immigrants scored 12.32; Greeks, 11.90; Russians, 11.34; Italians, 11.01; Poles, 10.74; and "U.S. Colored," 10.41.
Similarly, the eugenicist Paul Popenoe claimed in a 1924 article that if a mental age of 20 years was taken as the point attained by a very intelligent adult, 16 years was the average normal adult of white American stock. Popenoe added, "The young men of Grecian birth measured below the average mentality of a 12-year-old American schoolboy; the larger contingents of Russia and Italy are not far above the level (10.37 years) of the American negro or of a white adult who is of 'dull mentality.'" And in 1926, after 5,000 Massachusetts schoolchildren were given intelligence tests, the Harvard researcher Nathaniel Hirsch published the results in his monograph, A Study of Natio-Racial Mental Differences. His list of ethnic groups, from smartest to dullest, was: "Polish Jews, Swedes, English, Russian Jews, Germans, Americans, Lithuanians, Irish, British Canadians, Russians, Poles, Greeks, Italians, French Canadians, Negroes, and Portuguese."
Richwine is right that not every IQ researcher in the early 20th century agreed that many immigrant groups were dim-witted. For example, the Hartford Public School researcher Gustav Feingold reported in 1924 that "American-reared children of foreign-born parents show insignificant mental differences among themselves." Similarly, K.T. Yeung reported in a 1921 article that there were "no striking differences in the intelligence of the Chinese and American children." Nevertheless, the restrictionist researchers were far more prominent in the public debate over immigration.
Richwine suggests that the research reporting low immigrant IQs played essentially no role in the passage of the highly restrictive 1924 Immigration Act. As evidence, he notes that there is practically no mention of such research in the congressional testimony related to that legislation. But it was clearly part of the larger public debate about immigration. In 1924, Feingold, for example, worried that as a result of the research into immigrant IQs, "Educators and congressmen have become alarmed at the possible inferiority of our future immigrants. The present immigration laws are proof that no uncertain fear was evoked following the publication of the data."
In his review of Hirsch's book, the Tufts University psychologist George Van Ness Dearborn opined, "It is to be hoped that our present immigration law if not strengthened will be at least maintained indefinitely and stringently enforced, for the 'melting pot' is running over however slow the fusion of these alien stocks within it…. We have enough people but they have not enough education or intelligence."
Richwine acknowledges that this earlier work claiming that immigrant groups suffered intelligence deficits was premature and muddled. He notes, "There is no modern evidence of substantial IQ differences among American whites of different national backgrounds." In other words, the average IQs of the descendants of all those Poles, Jews, Italians, Russians, Greeks, and other low IQ immigrants have converged upward to the American average. But the lower average IQ of Hispanics—both American-born and new immigrants—is "persistent," he argues, so he doubts that they will so converge.
So what does Richwine want to do about immigration? Unlike the early-20th-century restrictionists, he rejects the notion of ethnic and national quotas. "Group differences in intelligence do exist," he writes, but "that does not mean that any individual should ever be judged on the basis of group membership." Richwine wants instead to give intelligence tests to individual prospective immigrants from wherever they may hail and to let the smartest ones in.
As noted above, Richwine relies a great deal on the research of Lynn and Vanhanen that finds persistent IQ deficits in many nations. Recently, American Conservative publisher Ron Unz analyzed their work and concluded that in fact, their data suggest that low average national and ethnic IQs are not persistent. Instead, average IQs rise as countries become wealthier. For example, Lynn and Vanhanen cite four studies from 1967 in which West Germans' average IQ scores ranged from 99 to 107 while East German scores were as low at 90. Now both eastern and western scores in Germany average around 102.
In 1961, according to Lynn and Vanhanen, the average Greek IQ was 88. According to other data, Croatians had an average of 90, Bulgarians were at 91, Romanians were at 94, Poles were at 92, and Sicilians were at 89. In 1972, the lowest IQ scores in Europe belonged to the Irish, who averaged just 87 points. These results are not far off from those that the early 20th century psychometricians found for immigrant groups coming to America. But as Unz points out, by the 21st century the incomes and educational attainments of American descendants of Greek, Slav, Italian, and Irish immigrants are now comfortably above those of the Old Stock white natives whose purity Brigham, Hirsch, and Yerkes were so anxious to defend.
And the Irish? Even Lynn and Vanhanen's data find they've largely caught up. In the 2009 results from the triennial Program for International Student Assessment, the Irish now outscore the British on reading and are very close on math and science. Interestingly, the current Hispanic-American IQ is slightly higher than the average Irish IQ in 1972. It isn't definitive, but using data from the Wordsum test in the General Social Survey, which correlates fairly well with IQ, Unz reports the IQ scores of second-generation Mexican Americans have likely risen a full 10 points in the past 20 years.
Hispanic educational achievements and incomes do lag behind those of white Americans. And certainly there is a substantial genetic component to intelligence; genes are, after all, the recipes that build bodies and brains in response to environmental cues. Yet as Unz's analysis of Lynn and Vanhanen's data shows, the average IQ of a population can change in a generation whereas its genetic makeup cannot. So what else might account for relatively lower Hispanic achievement so far in the U.S.?
The University of Texas economist Stephen Trejo suggests a number of possibilities. For example, Mexican immigration has lasted longer than immigration from any other country, promoting the growth and stability of culturally comfortable ethnic enclaves and slowing the process of assimilation. Trejo also proposes that earlier generations of unskilled immigrants faced a far less steep learning curve for moving up in a modern economy. In his 2005 book Italians Then, Mexicans Now, the Bard College sociologist Joel Perlman bolsters this point: "The crucial difference between the immigrant experience a hundred years ago and today is that relatively well-paid jobs were plentiful for workers with little education a hundred years ago, while today's immigrants arrive in an increasingly unequal America." Trejo also wonders if some fairly significant proportion of Mexican-Americans have simply already melded into the white population and so are not counted in the sorts of IQ, income, and education statistics cited by Richwine and other researchers.
Perlman concludes that "Mexican economic assimilation may take more time—four or five generations rather than three or four." Possibly so. But ultimately, modern Hispanic immigrants seem to be no stupider than the immigrant ancestors of other Americans.