HVGIQ: Cambodia

Much like Burma, Cambodia is a populous Southeast Asian country with a toxic authoritarian government, and routinely shows up at the back of human development indices. Richard Lynn’s international IQ dataset likewise does not yet have a study for this country.

Lynn & Vanhanen (2002, p. 74) make an IQ estimate of 89 for Cambodia by averaging together scores from its regional neighbors Thailand (91) and the Philippines (86). IQ and Global Inequality (2006, p. 56) revises this estimate to 91 by averaging together scores from three neighbors: Laos (89), Thailand (91), and Vietnam (94). Lynn & Vanhanen’s most recent update (2012, p. 21) assigns Cambodia an IQ of 92, but it’s not entirely clear why, as this book no longer lists the nations used to generate regional estimates. Presumably it’s the rounded average of Laos (89) and Vietnam (94).

In this post I present several different intelligence studies from Cambodia, as well some data for U.S. immigrants. These studies suggest that Cambodia has one of the lowest IQs in the world, but their achievement test scores in the U.S. exceed blacks and Hispanics.

⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻ HᏤ ⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻

The oldest study I’ve located for Cambodia is Fergusson et al (1995 ); Cattell’s Culture Fair Intelligence Test was administered to 23 students at the Institute of Economic Science in Phnom Penh and to 47 students at the Maharishi Vedic University in Prey Veng province. Remarkably, the IQ of these college students was 70.9.

More recently, Janssen & Geiser (2012 ) administered the Standard Progressive Matrices to 306 students at the University of Cambodia in Phnom Penh. In comparison with British norms, their IQ was 86.4.

Taking the average of these two tests gives us an IQ of 78.7 for Cambodian University students. Assuming the typical difference of 1 standard deviation between college students and the general population (Herrnstein & Murray, 1994, pp. 151-152), this would imply that the average IQ in Cambodia is 63.7!

There is only one existing study capable of verifying this estimate; Naudeau et al (2011 ) report intelligence test data for 4,015 preschool-age children from 7 Cambodian provinces. The researchers translated the Spanish version of the Peabody Picture Vocabulary Test into Khmer (You might remember that the TVIP was standardized in Mexico and Puerto Rico in the early 1980s). The reported IQ of the Cambodian children on this test was 82.7 (p. 20), but this is the sample’s IQ in relation to Mexican and Puerto Rican test norms. On top of that we have nearly 30 years of norm inflation. Adjusting for the Flynn Effect and U.S. norms brings the IQ down to 64.5, which is within one point of the estimate made from University student performance.

This number must be an underestimate of the population average, but probably not by much. While the children included in this study were socioeconomically mixed, the communities they live in were pre-selected for certain disadvantages, including a poverty rate of 30%. Since the national poverty rate is 30%, this means that no children from above average regions of Cambodia were tested. Only a study with good sampling could tell us how this omission really affected the average. But we can get a rough idea of what’s missing by looking at the socioeconomic differences in IQ within this sample (pp. 25-30): children from the upper quartile of wealth have IQs that are 4 or 5 points above the study average. Taking this as an upper bound, it is unlikely that Cambodia’s average IQ would be more than a few points higher than 64.5—at least on this particular test.

One additional study for Cambodia is O’Donnell et al (2012 ), which tested 237 young orphans on the second edition of the Kaufman Assessment Battery for Children. I would like to know how these children performed, but only the raw scores are reported, and I don’t have the conversion norms for this test. Regardless, orphans are a highly disadvantaged population and would not be an appropriate sample for a national IQ score.

Table I: IQ test scores in Cambodia

Admin Sample Age N Test IQ Reference
1993 MA 19 70 CCF 71 Fergusson et al, 1996
2011 MA 19 306 SPM 86 Janssen & Geiser, 2012
2008 SD 3-5 4015 TVIP 65 Naudeau et al, 2011


Did Pol Pot Lower Cambodia’s National IQ?

In 1975 the Khmer Rouge took power in Cambodia and committed large-scale atrocities on the model of Mao’s Great Leap Forward. The Communist dictator Pol Pot evacuated the cities with forced marches into the countryside and orchestrated a mostly intra-ethnic genocide of educated professionals in an attempt to transform Cambodia into an agrarian socialist utopia. This resulted in the mass-murder of ¼ of the Cambodian population!

Even though Lynn’s books have always given a relatively high IQ estimate to Cambodia, there has long been speculation from the HBD corners of the Internet that Pol Pot’s extermination of the intelligentsia must have had an enormously deleterious effect on Cambodia’s genotypic IQ (See comments here, here, and here).

How much of Cambodia’s current IQ deficit might be explained by this genocide?

James Flynn recently addressed this very question in his book Intelligence and Human Progress (2013, pp. 42-44) and answered ‘not much’:

How much did Pol Pot do to lower the mean IQ of the Cambodian people? … This question can be settled by a few calculations. Pol Pot killed somewhere between 1.7 and 2.5 million people. I will put this at 2.1 million or 26% of Cambodia’s 8 million people (Kiernan, 2002). If he had done it using IQ tests, eliminating the top 26% would have lowered the IQ of the remaining parents by 6.4 IQ points and a good portion of this deficit would have been handed down to their children. However, as we have seen, he in fact used occupation as his criterion.
We do not know the correlation between the occupational status of the parent and the IQ of their (no longer to be born) children, but in a semirural society it would be below that of the United States. [Note: this claim is possibly false. See Malloy, 2008, p. 1085. –JM] At that time in the United States, it was 0.300 (Flynn, 2000b). If you eliminated the top 26% of the US population by occupation, the mean IQ of their children would drop by only 1.92 points. Moreover, Pol Pot did not really use a pure criterion of occupational status. For example, a lot of his henchmen doing the killing were intellectuals (Pol Pot attended the Sorbonne, although he did flunk all of his courses). When he tried to eliminate everyone who lived in the capital city of Phnom Penh, this included many in humble occupations. The genetic capital of the Cambodian people was lowered by not much more than an IQ point. The people were hardly stripped of intellectual talent.

Of course, this only wrestles with the hereditary cost of the Khmer Rouge. A few 1979 reports from one Cambodian doctor—used in the trial of Pol Pot—specifically discuss the environmental impact of the regime on cognitive ability: “The survivors of this generation (aged from one to four years) will necessarily have a highly reduced intelligence potential, with IQs no higher than 95.” (De Nike, 2000, pp. 335-338)


Intelligence and Achievement of Cambodian Americans

Cambodian immigration to the U.S. mostly occurred in a single wave in the late 1970-1980s as peasants fled the Vietnamese military occupation (few were so lucky in escaping the Khmer Rouge).

Clarke et al (1993 ) tested 39 adolescent refugees who had been living in the U.S., on average, for about 6 years. They were given the Peabody Picture Vocabulary Test-Revised and the vocabulary subtest of the Stanford-Binet Intelligence Scale. IQ on the PPVT was 49.7. Only a few had scores in the normal range, and 75% scored below 70. Their IQ on the Stanford-Binet subtest was 75.2. Taking an average of the two tests gives this sample an IQ of 62.5.

Large-scale annual achievement test data is available for Cambodian Americans. Standardized Testing and Reporting (STAR) is a state program for measuring student performance in California, and test scores from 1998-2013 are available online. Conveniently for Human Varieties, the results are also broken down by ethnicity, which includes a dozen different Asian nationalities.

For the years 2003-2006 I looked at the performance of over 65,000 Cambodian Americans on the reading and math sections of the California Achievement Test. In comparison with white students, Cambodians had an AQ of 91.7 (Reading = 89.2, Math = 94.1)

For the years 2009-2012 I looked at the performance of nearly 49,000 Cambodian Americans on the math and reading sections of the California Standards Tests. In comparison with white students, Cambodians had an AQ of 94.8 (Reading = 92.9, Math = 96.8)

A much smaller amount of data is also available from the Children of Immigrants Longitudinal Study, which gave the Abridged Stanford Achievement Test to 95 Cambodian school children living in San Diego, CA during 1992 and 1996 (Kim, 2002 ). Cambodians had an AQ of 89 (Reading = 84, Math = 94).

These three sources of data give Cambodian Americans a weighted Achievement Quotient of 93.

To the extent the results from Clarke et al (1993) represent the scores of new immigrants, and the results from the post-1990s Californian state exams are mostly second generation, we might take this as evidence that when Cambodians are nurtured in the high quality U.S. environment their standard scores rise to the low or mid 90s: similar to the (presumed) scores of closely-related populations like the Thais or Vietnamese.



⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻ REFERENCES ⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻

Clarke, G.N., Sack, W.H., Ben, R., Lanham, K., & Him, C. (1993). English language skills in a group of previously traumatized Khmer adolescent refugees. Journal of Nervous & Mental Disease, 181, 454-456.

Fergusson, L.C., Bonshek, A.J., & Masson, G.L. (1995). Vedic science based education and nonverbal intelligence: a preliminary longitudinal study in Cambodia. Higher Education Research & Development, 15, 73-82.

Flynn, J.R. (2013). Intelligence and Human Progress: The story of what was hidden in our genes. Oxford, UK: Elsevier.

De Nike, H.J., Quigley, J.B., & Robinson, K.J. (2000). Genocide in Cambodia: documents from the trial of Pol Pot and Ieng Sary. Philadelphia, USA: University of Pennsylvania Press.

Herrnstein, R.J., & Murray, C. (1994). The Bell Curve: Intelligence and class structure in American life. New York, USA: Simon and Schuster.

Janssen, A.B., & Geiser, C. (2012). Cross-cultural differences in spatial abilities and solution strategies—an investigation in Cambodia and Germany. Journal of Cross-Cultural Psychology, 43, 533-557.

Kim, R.Y. (2002). Ethnic differences in academic achievement between Vietnamese and Cambodian children: Cultural and structural explanations. Sociological Quarterly, 43, 213-235.

Lynn, R., & Vanhanen, T. (2002). IQ & the wealth of nations. Westport, CT: Praeger.

Lynn, R., & Vanhanen, T. (2006). IQ & global inequality. Augusta, GA: Washington Summit Publishers.

Lynn, R., & Vanhanen, T. (2012). Intelligence: A Unifying Construct for the Social Sciences. London, UK: Ulster Institute for Social Research.

Malloy, J. (2008). James Watson tells the inconvenient truth: Faces the consequences. Medical Hypotheses, 70, 1081-1091.

Naudeau, S., Martinez, S., Premand, P., & Filmer, D. (2011). Cognitive development among young children in low-income countries. In H. Alderman (Ed.), No Small Matter: The impact of poverty, shocks, and human capital investments in childhood development (pp. 9-50). Washington, DC, USA: The World Bank.

O’Donnell, K., Murphy, R., Ostermann, J., Masnick, M., Whetten, R. A., Madden, E., … & Whetten, K. (2012). A brief assessment of learning for orphaned and abandoned children in low and middle income countries. AIDS & Behavior, 16, 480-490.

0 thoughts on “HVGIQ: Cambodia

  1. Good post.

    To the extent the results from Clarke et al (1993) represent the scores of new immigrants, and the results from the post-1990s Californian state exams are mostly second generation, we might take this as evidence that when Cambodians are nurtured in the high quality U.S. environment their standard scores rise to the low or mid 90s: similar to the scores of closely-related populations like the Thais or Vietnamese.

    But why assume that these immigrants were representative of the Cambodian population? (Immigrants to the U.S. usually aren’t, save Mexicans.)

    • But why assume that these immigrants were representative of the Cambodian population?

      There was significant elite immigration from Vietnam, but the refugees from Cambodia were mostly rural peasants. Maybe they were below average, but when your GDP is $50 per capita, how can you even tell?

      • A reference on this point:

        Arthur Sakamoto and. Hyeyoung Woo, “The Socioeconomic Attainments of Second-Generation Cambodian, Hmong, Laotian, and Vietnamese Americans,* Sociological Inquiry, Volume 77, Issue 1, pages 44–75 (February 2007).

        To quote:

        In short, unlike most other Asian immigrants, Southeast Asians typically arrived in the United States as refugees under duress with essentially no economic assets or well-devised plans for the future. Many of the first arrivals had no prior family connections. Nor were they particularly selective in terms of possessing high levels of educational attainment or scarce occupational skills as stipulated by immigration law (Kao 1995). Refugees tend to have lower socioeconomic statuses than immigrants who are admitted for their labor market skills.

        The initial immigration of Southeast Asians is usually divided into two major waves. The first are those who entered the United States in 1975 or earlier (i.e., Wave 1), while the second (i.e., Wave 2) are those who entered after 1975. Wave 1 immigrants are mostly persons who entered the United States in 1975 (with the fall of Saigon and the withdrawal of U.S. military forces in Vietnam) because the immigration of these groups before 1975 was highly limited (Rumbaut 1995). Wave 2 immigrants are mostly from 1979 and later as the immigration of these groups between 1976 and 1978 was comparatively minor (Rumbaut 1995).

        Wave 1 Vietnamese immigrants tended to have somewhat higher socioeconomic origins than Wave 2 Vietnamese immigrants. As discussed by Rumbaut (1995), Wave 1 Vietnamese immigrants were substantially more likely than Wave 2 Vietnamese immigrants to have completed high school, to know some English, to be from an urban background, and to have had an upper white-collar occupation. Wave 1 Vietnamese immigrants were more likely to be associated with the South Vietnamese military or the U.S. military or to have been employed by the U.S. government. By contrast, Wave 2 Vietnamese immigrants were more likely to have never attended school, to have been farmers or fishermen, to know no English, and to have a rural background. Wave 2 Vietnamese immigrants were also sometimes known as the “boat people” in the American popular press during the 1980s because of their precarious and often tragic exodus via boats from Vietnam to nearby asylum countries.3

        Wave 1 Cambodian, Hmong, and Laotian immigrants also tended to have slightly higher socioeconomic origins than their Wave 2 counterparts. Both waves of these groups, however, tended to have lower levels of education than even Wave 2 Vietnamese. The school system in Laos at that time did not offer schooling beyond the 12th grade. Only tiny proportions of these populations were urban. The Wave 2 Cambodians have very low levels of education because many are the survivors of the “killing fields” controlled by the Khmer Rouge led by Pol Pot. The Hmong are particularly disadvantaged in part because the Laotian government at that time did not develop a school system in Hmong communities. The Hmong were mostly illiterate, and were engaged in hunting and in traditional slash-and-burn agriculture (Chan 1994; Kitano and Daniels 1995; Rumbaut 1995).

        As noted above, we focus our analysis on the second generation rather than on persons who immigrated to the United States as adults. Our investigation therefore does not include Wave 1 or Wave 2 adult immigrants per se but rather their children who were educated in the United States. It is a well-established finding, however, that children tend to have higher educational attainment to the extent that their parents have higher educational attainment and socioeconomic status (Featherman and Hauser 1978; Goyette and Xie 1999; Mare and Winship 1988). Furthermore, persons who complete more education tend to have higher incomes and wages (e.g., Farley 1996). We therefore expect that

        second-generation Southeast Asian Americans who have Wave 1 parents to tend to have higher socioeconomic attainments than those who have Wave 2 parents.

        The results:

        Among Cambodians, Wave 1 immigrants have a mean years of schooling of 11.91, while Wave 2 immigrants have a mean of 7.15. The difference between the two waves is thus 4.76 years which is the largest difference for any of the Southeast Asian groups. While Wave 1 Cambodians have mean years of

        schooling that is fairly similar to African Americans (i.e., 11.91 versus 12.14 for African Americans), the 7.15 mean years of schooling for Wave 2 Cambodians is quite low by U.S. standards. Wave 1 Cambodians of the “parental generation” have a poverty rate that is substantially lower than that for African Americans, while among Wave 2 Cambodians poverty is about 10 percentage points higher than among African Americans. Only one of five Wave 1 Cambodians speaks English poorly, while among Wave 2 Cambodians the corresponding figure is about two of three. It should be noted, however, that the sample sizes shown in Table 1 imply that slightly less than 10 percent of the “parental generation” of Cambodians is Wave 1 (over 90% are Wave 2).

        The rest of the paper has other data & insights germane to this discussion and worth reading in full.

    • I did miss that, but it contains no new information. Lynn’s paper simply summarizes the results from Janssen & Geiser (2012). He is attempting to use the IQ of University students for the average IQ of Cambodia. If he used this as a consistent inclusion criteria for his dataset he would get all kinds of wonky results. Many nations would have inflated national IQs.

  2. Thanks.

    How about doing Ukrainians, another region subjected in the 20th Century to massacres and exiling of people with above average intelligences: Jews, kulaks, aristocrats, and Communist Party officials (both by Stalin in the Great Terror and by Hitler in 1941)? And no doubt there has been some selective migration outward since the late 1980s: my son’s one Ukrainian friend in grade school, Andy, was the son of a genius aerospace engineer who owned the single nicest piece of land I’ve ever been on in the San Fernando Valley.

    The lone Ukrainian figure I’ve seen from Rindermann is 95, which sounds reasonable, but you are amazing at digging up more numbers.

  3. Pingback: More Links For June 2014 | Slate Star Codex

Leave a Reply

Your email address will not be published. Required fields are marked *