Wednesday
Aug172011

Affirmative Action and Asians

I promised in previously that my Website would contain a debate about one of the most controversial arguments in The Affirmative Action Hoax: that Whites are more victimized by affirmative action than Asians. I reported that I was engaged in an email debate over this subject with someone named Yan Shen, whom I have never met, and that I offered him 3005 words on this update to present his arguments. (That is the number of words in Section D of Chapter 19 of my book, which he would try to refute.) However, when I emailed him recently to send me his arguments, he did not reply. So, below I present further substantiation of my arguments. If at any time Mr. Shen sends his arguments, I will post them.

I must begin by qualifying the title of the section in my book in which I discuss this subject (Section D of Chapter 9): "Asians Are Beneficiaries of Affirmative Action." In that section, I document that Asians are the greatest beneficiaries of government programs that set-aside contacts for minority-owned businesses and that Asians are often beneficiaries of programs to increase the proportion of minorities in high-level positions. I also point out that 61 percent of Asians voted against the 1996 referendum in California that banned affirmative action. These facts are nearly always ignored in discussions of the effect of affirmative action on Asians. Nearly all discussions about Asians and affirmative action focus on university admissions, in which Asians are clearly not beneficiaries of affirmative action. In my analysis of university admissions, my argument is that Asians are not more victimized than Whites.

In order to substantiate that point, I will first discuss the ethnicity of the various groups that comprise American Asians, then the relative importance of verbal and mathematical tests, then the performance of Asians in universities.

Ethnicity of American Asians

The designation "Asian" is infuriatingly imprecise. I point out several times in my book that some of the national-origin groups that are labelled "Asian" are less successful academically and occupationally than White Americans, and some are more successful. I will here use the breakdown of Asian groups in the 2000 U.S. census, rather than the 2010 census because the facts in my book pertain to the period before 2010. The 2000 census recorded 11.9 million people who reported themselves as having either full or partial Asian ancestry, 4.2 percent of the U.S. population. The largest ethnic subgroups were Chinese (2.7 million), Filipinos (2.4M), Asian Indians (1.9M), Vietnamese (1.2M), Koreans (1.2M), and Japanese (1.1M). Other significant groups were Cambodians (206,000), Pakistanis (204,000), Laotians (198,000), Hmong (186,000), and Thais (150,000). I say on page 258 of my book that many (maybe most) Vietnamese in the United States are ethnically Chinese. I deduce that from the fact that emigrants from Communist countries are highly self-selected for intelligence. For example, until about 1990, nearly all emigrants from Communist Cuba were White. In fact, on the 1990 census, the average income and proportion of college graduates among Cuban Americans were higher than among the general White American population (Thernstrom 1997: 542). (I provide an example of the effects of selective immigration in my book, on page 249, footnote 203.) (When I cite a source by last name and date, full information is in the Bibliography of my book.)

An article on pages 255-76 of the Annual Review of Sociology 35, 2009, (Arthur Sakamoto, et al., "Socioeconomic Attainment of Asian Americans") supplies the average attainment of different Asian American groups on many measures of success. The differences are not as great as I had thought. For example, only four groups of Asian American adults have a lower proportion of college graduates than White American adults; and three of those four groups - Cambodians, Hmong, and Laotians - combined constitute only five percent of all Asian Americans (pages 258-9). The other group of Asian Americans who have a lower proportion of college graduates than Whites are Vietnamese. But their children do extremely well on academic tests (page 262).

The Relative Importance of Verbal and Mathematical Tests

In my book, I repeatedly point out that the contention that Asians suffer from greater discrimination than Whites in undergraduate admissions is usually based on comparisons between the combined Verbal+Math SAT scores of Asians and Whites at the same colleges. But that is a misleading measure since the Verbal section, on which Whites do better than Asians, is more important than the Math section, on which Asians do better than Whites. However, I do not provide any evidence for the greater importance of the Verbal than the Math SAT in my book. I will correct that defect now.

There are two reasons why the Verbal SAT is more important than the Math. One is obvious: the types of questions and problems on the Verbal section are relevant to many more university courses and many more occupations than the questions and problems on the Math. (In my book (pages 152, 303), I warn against confusing mathematical reasoning with accuracy of arithmetic computation.)

A more important reason for the greater importance of the Verbal section is that it is a more accurate measure of General Intelligence (i.e., it is more g-loaded) than the Math section. To state that non-technically: the Verbal section is a more accurate measure than the Math of ability at "analyzing, synthesizing, and manipulating information; distinguishing relevant from irrelevant information; and other types of abilities that are commonly called intelligent" (page 303 of my book). (The definitive discussion of General Intelligence (g) is Jensen 1998.)

The higher g-loading of the Verbal than the Math SAT explains a fact that I mention on pages 281-3 of my book. I show there that coaching/preparation does little to raise SAT scores. But it raises Math scores more than Verbal scores. This probably surprised many readers, who assumed that the opposite should be true. However, although the form of the Verbal section is words, its substance is General Intelligence; and General Intelligence is the most completely genetically determined of all mental traits.

Unfortunately, I could find only one article on this subject, by Alexander Beaujean, et al., in Personality and Individual Differences 41, 2006 (pages 353-7), which is available at www.iapsych.com/articles/beaujean2006.pdf. In it, Beaujean, et al., reported the results of a study of the correlation of Verbal and Math SAT scores with scores on the Composite Intelligence Index, which consists of two verbal and two non-verbal subtests. Beaujean, et al., concluded, "SAT Total or SAT Verbal alone best predicted IQ;" "Using both SAT Math and SAT Verbal does not produce the optimal equation, as when both variables are included SAT Math loses significance as a predictor."

The participants in this study were 97 students at a private, Midwestern university who represented a wide cross-section of academic majors. But 94 percent were self-identified as Caucasian.

Fortunately, much more cogent and relevant data on this subject are available. On page 261 of my book, I discuss the predictive accuracy of the subtests of the Medical College Admission Test (MCAT). Because of its importance for the applicants involved and for the general society, the predictive accuracy of the MCAT has been more intensively studied than any other academic test. (The predictive accuracy of the SAT has been more extensively studied.)

I report on page 261 of my book that before 1991, the MCAT had five sections: Biology, Physics, Science Problems, Reading, and Quantitative. Extensive studies were done of how well each section predicted performance on the three parts of the National Board of Medical Examiners examinations (NBME). Here I will remind the reader that the courses in the first two years of medical school are devoted to the sciences relevant to medicine (anatomy, biochemistry, etc.). The courses in the third and fourth years involve specifically medical subjects (surgery, obstetrics/gynecology, pediatrics, etc.). Then, after a medical student has an MD degree, he works in a hospital as an intern. Part I of the NBME was taken at the end of the second year of medical school, Part II at the end of the fourth year, and Part III at the end of internship.

The Quantitative section of the MCAT, which resembled the Math SAT, was found to have no predictive value for any of the three parts of the NBME. The science sections had the highest predictive accuracy for Part I, as could be expected. The Reading section had slightly better predictive accuracy than the science sections for Part II. On Part III, which is clearly the most important, the Reading section had by far the highest predictive accuracy. (One point on the combined science sections added 9.70 points to the score on Part III; one point on Reading added 14.45 points on Part III.) Karen Glaser, et al., (1992: 404), who summarized these studies, concluded, "verbal ability reflected in the reading skills scores of an applicant to medical school are a more important indicator of later physician competence … than the applicant's ability to solve science problems." For this reason, the Association of American Medical Colleges radically changed the contents of the MCAT in 1991. It eliminated the Quantitative section, reduced the science sections from three to two, and added a Writing Sample, thus giving verbal ability equal weighting with science ability (Glaser, et al., 1992).

In the first sentence of the previous paragraph, I wrote that the Quantitative subtest was found to have no predictive value for any of the three parts of the NBME. In my book, I wrote that it had "little" predictive accuracy. But I overstated its value. In fact, "Scores on the quantitative skills subtest did not contribute to any prediction" (Glaser et al., 1992: 395). This total non-prediction includes Part I of the NBME, which examines what students learned in the science courses that comprise the first two years of medical school.

These facts make clear that even what I call above the "obvious" wider applicability of the Verbal than the Math SAT significantly understates the difference between them. We must assume that Math SAT has a great deal of predictive accuracy for mathematics-based subjects, like engineering; although even there, the Verbal+Math score must be a more accurate predictor than the Math score alone. But for the large majority of academic subjects, the Math SAT is completely irrelevant. That includes even non-mathematical scientific subjects.

University Performance

I point out on pages 261-2 of my book that in 1992, a year after the new MCAT was introduced, the three parts of the NBME was replaced by Steps I, II, and III of United States Medical Licensing Examination (USMLE). Of the students who entered medical schools in 1992, 93.4 of the Whites and 86.8 percent of Asians passed Step I of the USMLE on the first try; and 96.3 percent of Whites and 87.6 percent of Asians passed Step II on the first try (Case, et al., 1996: S91).

On pages 259-60 of my book, I discuss the relative performance of Asian and White law students on the bar examination. The most recent statistics I could find for the passing rates of Whites and Asians on a bar exam were from New York in 2006: 78.3 percent of the Whites and 71.6 percent of the Asians passed the New York bar exam on the first try (www.nybarexam.org).

After my book was published, I found a study of the relative undergraduate performance of Asians and Whites. It is a chapter in a book edited by Guofang Li and Lihsgung Wang entitled Model Minority Myth Revisited, published in 2008 by Information Age Publishers. The chapter that compares Asian and White undergraduates is on pages 233-51. Its title is "Whither the "Whiz Kids' Went," by Julia Dmitrieva, et al. They begin (pages 234-5) by reporting the results of the only two previous studies of this subject. One, published in 2001, concerned undergraduate performance at the University of California at Berkley. It included 291 Asian students, of whom 60 percent were Chinese, 14 percent Korean, 7 percent South Asian (i.e., Indian and Pakistani), and 19 percent other Asians. It found, "Asian American students had significantly lower self-reported GPA [Grade Point Average] than their European American counterparts." The other study, which was published in 2004, examined Asian-White differences in academic achievement of 3,500 undergraduates at New York University. The national origins of Asian American group were representative or Asians in the United States. It "found similar ethnic differences in grades [as the Berkley study], using official GPA records." (Dmitrieva et al., do not provide the specific results of these studies, only the overall results, which I quote.)

Dmitirieva et al., then (pages 238-42) report on the three studies they performed, all on undergraduates at the University of California at Irvine. The first study was of 785 students, 70 percent of whom were Asian Americans and 30 percent European Americans, which mirrored their representation at Irvine. Of the Asian students, 37 percent were Chinese, 16 percent Vietnamese, 15 percent Korean, 14 percent Filipino, 7 percent Thai, 6 percent South Asians, and 5 percent Japanese. The average high school GPA of the Asian students was significantly higher than that of the European American students; their average SAT scores were similar. In all four college years, the average GPA of the European American students was higher than that of the Asian American students. This difference was significant in the sophomore and senior years, but too small to be significant in the freshman and junior years.

The second study involved 400 students, of whom 74 percent were Asian Americans and 26 percent were European Americans. Of the Asians, 29 percent were Chinese, 18 percent Korean, 17 percent Vietnamese, 11 percent Filipino, 9 percent South Asian, 5 percent Japanese, and 13 percent other Asians. In this study also, the average Asian American student had lower average grades than European American students in all four college years; although the difference was significant only in the junior year.

The third study followed 452 students from the twelfth grade through their sophomore year in college. Forty-eight percent were Asian, 52 percent were European. The high school grades of the two groups were similar. The Asians did worse academically in their freshman college year, the two groups did not differ in their sophomore year.

In all three studies, the Asian-European difference remained after controlling for gender and parental educational attainment.

An obvious explanation for the poor college performance of Asian relative to White students might be that a higher proportion of Asians take courses in the natural sciences, which are more difficult and graded more stringently than Liberal Arts and social science courses. However, Dmitirieva et al., found (page 243) that among undergraduates at Irvine, Whites majoring in the natural sciences got higher grades than Asians majoring in the natural sciences in every year except senior year, in which their grades were similar.

A chapter in the same book as Dmitrieva, et al.,'s article (Model Minority Myth Revisited) proposes a possible explanation for the disjuncture between high school and university performance of Whites and Asians. On pages 184-6, David Dai cites several studies, conducted in both the United States and Asia, that compared creativity of Chinese and Whites. All except one found that Chinese are less innovative and creative - less able to generate new ideas and approaches - than Whites. For example,

Singaporean students are ... top performers ... in the world in TIMSS [Trends in International Math and Science Studies] ... Yet when they were asked to be creative and generate new ideas for a firm, many of them were simply lost, to the point that the leadership of the firm had to open a new sector in the United States to take care of more innovate endeavours for the firm.

Unfortunately, Dai did not know about the most comprehensive study of this subject: Richard Lynn's "Race Differences in Intelligence, Creativity and Creative Achievement," on pages 157-68 of the 2007 (48,2) issue of Mankind Quarterly. In it, Lynn came to the same conclusion as the studies that Dai cites.

Creativity - "thinking outside the box" - does not contribute to high school grades. In fact, it may lower them. Nor is it captured by the SAT or the tests for entrance to professional schools. But it is an important factor in university performance and on professional qualifying examinations.

However, the cause is irrelevant to the present discussion. What matters is that Asians do worse academically than Whites at the same colleges and do worse on qualifying examinations for professions. Consequently, the constant brouhaha about Asians being the primary or only victims of academic affirmative action is absurd.

Sunday
Feb132011

Recommendation: Who owns the future

I advise all readers to consult Pat Buchanan's article Who Owns the Future?