HVGIQ: Thailand

Thailand has a population of over 64 million people, which is comparable to the United Kingdom or France. It is the world’s 21st most populous country.

IQ and the Wealth of Nations (2002) and IQ and Global Inequality (2006) both assign Thailand an IQ of 91 based on one clinical trial from 1989. Lynn and Vanhanen (2012, p. 417) add two more studies that push their estimate down to 88. Some of their numbers are reported inaccurately and these books overlook larger, higher quality, and more recent intelligence test studies for this nation. In particular—as I noted on Gene Expression almost a decade ago—the government of Thailand has implemented broad national IQ surveys, which have been reported in relation to U.S. and U.K. test norms.

In this post I review almost 50 studies of intelligence and scholastic achievement for the nation of Thailand. Many representative studies since the 1990s give conflicting results, making a solid estimate of Thailand’s IQ a difficult task.

⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻ HᏤ ⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻⎻

Some personal details are relevant to this analysis: I worked on a commercial art project in Thailand for the better part of a year in 2004. I didn’t expect Thailand to offer the same kind of recreational academic reading that I’d always enjoyed in American college towns, but the Thai University libraries that I explored were unexpectedly well stocked with books about psychometrics. There were other clues that Thailand had a penchant for cognitive measurement: it was one of the only developing countries to consistently participate in the early cross-national student assessments of the 1970s and 1980s. This inspired me to poke around for intelligence test studies in the Thai-language psychology journals. As expected, Thai psychologists had eagerly adopted Western intelligence tests. The journals supplied me with several early test standardizations heretofore unknown to Western academics. A number of the studies and government reports reviewed in this post were published in Thai, and any mistaken translations of the research are entirely my fault.

                 SECTION I: INTELLIGENCE TEST STUDIES IN THAILAND                 

■ Ia. 1950s-1970s

I looked through the Journal of the Psychiatric Association of Thailand, which is not held in any Western library. The second volume contained a standardization of the Raven’s Coloured Progressive Matrices from the 1950s. Supa Malakul (1957) normed the CPM on 1,385 schoolchildren from four different provinces. This sample had an IQ of 82.1, which is the earliest IQ score for Thailand.

The CPM was standardized again in the mid-1960s with 1,438 children from Bangkok (Talapat & Suwannalert, 1966a). This later sample had an IQ of 82.1.

Talapat & Suwannalert (1966b) also collected Thai norms for the Draw-a-Man test. They evaluated drawings from 892 Bangkok schoolchildren using the recently developed Goodenough-Harris scoring system (Harris, 1963). These children had an IQ of 97.5.

It’s interesting how much higher scores were on this test than on the Raven, given the Draw-a-Man performance of other mainland Southeast Asians during the same time period. Burmese children tested with the Draw-a-Man in 1962 had an IQ of 106.8, and Vietnamese children tested with the Draw-a-Man in 1965 had an IQ of 98.9. Mainland Southeast Asians appear to have some sort of advantage on this test. Practice is an unlikely explanation, given that the Vietnamese sample had had no prior experience drawing pictures. (And it’s not obviously a test that generates small cross-cultural differences; cf. Cayman Islands, Jamaica, and Puerto Rico.)

The Raven test results are probably overestimating Thai IQ during this period as well. No rural children were included in the CPM norms by Malakul (1957) or Talapat & Suwannalert (1966a). In the early 1960s a Danish psychologist administered the CPM to 47 1st graders in a village in Northeast Thailand (Poulsen, 1982). Mean scores are not reported, but 62% had IQs below 75. Poulsen believed that scores would improve as children were exposed to more schooling, but when he tested 51 4th graders from the same village in 1978, 80% had IQs below 75.

An English-language study from a Thai medical journal provides one more data-point for this time period: Rajatasilpin et al (1970) tested 70 primary school children from suburban Bangkok with the Wechsler Intelligence Scale for Children. IQ on the WISC was 88.3.

A much larger study provides some data for older, rural Thai adults, giving us a lower-bound IQ score for Thailand. The World Bank organized a survey of 1,462 farmers from the Chiang Mai valley in 1978 (Chou & Lau, 1987). Variables such as education, years of experience, agricultural knowledge, intelligence, and numeracy were measured to find the best predictors of farm productivity. The farmers were 17-60 years old, with an average age of 50.

The Coloured Progressive Matrices was used as a measure of native intelligence. The CPM was originally designed for younger children, and Wicherts et al (2010, p. 37) have criticized both Richard Lynn and myself for calculating IQ scores for adult samples tested with the CPM. They argue that these IQ conversions are inappropriate because A) there are no adult norms for this test, and B) the test exhibits a ceiling effect—it is too easy for adults, and there are not enough items to compensate for errors, so a small number of incorrect responses can radically deflate IQ scores.

These objections have not persuaded me. For one, a ceiling effect on this test is not inherently more of a problem for adults than it is for children (temporal and ethnic differences on this test rival age differences). To the extent it is an issue, it would be a greater problem for newer (norm inflated) studies of higher IQ populations than it would be for older studies of lower IQ populations (since few to no people in the latter group are able to answer all of the items correctly). Second, the CPM is routinely used as a test for adults, and the Raven’s manuals do, in fact, provide norm tables for older age groups.

I’ve selected an adult U.S. sample as the best comparison group for the Thai farmers (Panek & Stoner, 1980). The Americans were rural, almost the same age (53), and tested with the CPM at almost the same time as the Thai subjects (~1979). Compared with the Americans, the Chiang Mai farmers had an IQ of 71.8.

Chou & Lau find that schooling—not intelligence—is overwhelmingly the variable most associated with profit and productivity in their regression analysis. Another study of 39 Chiang Mai valley farmers (presumably a subset of the same sample) finds that CPM scores were also unrelated to the ability to predict crop yields and prices (Grisley & Kellogg, 1983).

Low test validity wasn’t restricted to farmers. Harinasut & Suwanlerd (1958) found a negligible association between CPM scores and grades for Thai primary school children. And Somnapan (1972) found a low correlation between Thai SPM scores and college entrance exams in the 1960s.

The Progressive Matrices was apparently not a very useful test in Thailand during these decades.

__________________________________________

■ Ib. 1980s-1990s

The only reference for Thailand in Lynn and Vanhanen’s first two books (2002, p. 222; 2006, p. 309) is the iron supplementation study by Pollitt et al. (1989). Lynn does not report a sample size for this study in the first book, and misreports the sample size as 2,268 in his other two books. While Pollitt et al screened 2,268 children from the Chon Buri province for inclusion in this study, they excluded the children who did not meet certain health or age requirements. So only 1,358 children were tested with the Coloured Progressive Matrices. Lynn also misreports the sample age as 8-10 in all three books, when the actual age range is 9-11. The correct sample size and age range are reported in the abstract of the study. This sample had an IQ of 91.1, which Lynn reports correctly.

There is a better study for Thai IQ from this time period, although the technical papers were not released in English (I first learned about the study in a Thai newspaper, and discussed this data on GNXP back in 2005). The Thai Institute of Public Health included the Test of Nonverbal Intelligence (TONI) in their Second National Survey of Health: 1996-1997. The TONI was administered to 3,846 children, ages 6-13, from across the country. Thai children exhibited an IQ of 89.9 on this test (TIPH, 1998).

Another study from the mid-1990s looked at a population from a southern region with abnormally high levels of environmental arsenic. Siripitayakunkit et al (1999) administered the Wechsler Intelligence Scale for Children to 529 primary school children. Their IQ was 73.7.

Finally, Thai norms for the Raven’s Standard Progressive Matrices were collected for the first time in the late 1990s. 400 adults of all ages exhibited an SPM IQ of 89.5 (Phattharayuttawat et al, 2000).

__________________________________________

■ Ic. 2000s-2010s

A majority of the studies of Thai intelligence were published after the 1990s.

A study by Sungthong et al (2002) looked at 427 10 year-olds from Songkhla, a province in the far south that borders Malaysia. These children had an IQ of 75 on the Test of Nonverbal Intelligence.

It has been suggested that Islam could be a cultural impediment to intellectual development (e.g. Templer 2010, p. 204), and that this might partially explain low IQ scores in Muslim countries. It is therefore important to compile data for ethnically matched people with different religions. IQ was reported for 133 Thai Buddhists and 294 Thai Muslims attending the same schools in this region. The IQ of the Muslims was 76, and the IQ of the Buddhists was 75. To the extent that this can be taken as a meaningful difference, any number of factors might explain this result (e.g. minority disadvantage), but it does not lend support to the hypothesis that Islam depresses intelligence.

This was also one of the two references for Thailand’s IQ that Lynn and Vanhanen (2012, p. 417) added to their dataset, but it is not clear that Lynn read this study at all. He doesn’t tabulate any details from the study except the test instrument and the IQ, and he misreports the IQ as 88! [1]

Updated norms for the Coloured Progressive Matrices were collected by Phattharayuttawat et al (2003). The CPM IQ for 900 Thai primary schoolchildren was 106.1! This is the first study to suggest a radical increase in Thai IQ. But this rise may or may not be limited to the Progressive Matrices test.

For instance, the Wechsler Scale of Intelligence for Children-III was also adapted and normed in Thailand (Wanitrommani et al., 2004). The test was administered to a representative sample of 3,300 children, ages 6-16. Their WISC IQ was 94.5. This is certainly higher than the IQ scores found in the previous two studies using the WISC in Thailand (Rajatasilpin et al, 1970: 88.3; Siripitayakunkit et al, 1999: 73.7), but these were hardly comparable norm samples.

One thing is for sure, the Test of Nonverbal Intelligence, which has been used multiple times on large representative samples, shows a large gap with the U.S. norms, and there is no sign of improvement among Thai children since the 1997 government survey. For example, Nichara Ruangdaraganon collected norms for the Test of Nonverbal Intelligence-III. The standardization included 3,135 children, ages 6-12, and 3,150 adolescents, ages 13-18. The children had a TONI IQ of 86, and the adolescents had an IQ of 84.6 (Ruangdaraganon, 2004). This is 1/3 of a standard deviation lower than the TONI IQ of the 1997 survey children.

Two more studies show relatively high IQs on the Coloured Progressive Matrices. A doctoral dissertation by Niyot Sangtongluan (2003) looked at 396 primary school children in Bangkok. Their CPM IQ was 99.

A second and much larger CPM standardization gives a more modest and plausible IQ score than the 2003 norms, although the results still suggest that Thais are now close to the UK norms on the Progressive Matrices. Seven graduate students from Mahidol University collected local norms for the Coloured Progressive Matrices and the Advanced Progressive Matrices in seven different regions of Thailand, and reported each of their results in separate dissertations. The results from all seven CPM standardizations are combined and summarized by Sukhatunga et al (2006b). The CPM IQ of 3,848 children, ages 6-11, was 95.6. This is closer to the results of the WISC standardization than the TONI norms. However, the APM norms suggest something else entirely: The APM IQ of 5,702 adolescents, ages 12-18, was 104.6 (Sukhatunga et al, 2006a). If we average together the two age groups, we get a Thai IQ of 101, which is higher than the U.K. norms.

There have been four subsequent studies with the CPM, but all of them contain scores below these norms. An unpublished master’s thesis by Wuthisak Nimmalangkun (2006) has data for 100 Chiang Mai elementary school children. Their CPM IQ was 87.4.

Another thesis by Woraya Sroythong (2008) includes data for 380 children from Bangkok. Their IQ on the CPM was 94.

A dissertation by Amornpun Thavornsuwanchai (2008) reports data for 748 primary school students from Chonburi (an Eastern province on the Bay of Bangkok). Their IQ on the CPM was also 94.

Finally, there is data from a zinc and iron supplementation study in the Khon Kaen province in Northeast Thailand (Pongcharoen et al, 2011). The 560 children were not selected for disadvantages but had a CPM IQ of 75. Their IQ on the WISC was 87.

Additional studies with the Test of Nonverbal Intelligence also fail to corroborate the scores from the Progressive Matrices standardizations. A study of 319 teenagers reported a TONI IQ of 88 (Isaranurug et al, 2006).

A large, representative sample of youths, ages 6-14 years-old, were given the Test of Nonverbal Intelligence for the government’s 4th National Survey of Health: 2008-2009 (Aekplakorn, 2009). Average IQ on the TONI was 87.8, which is 2.1 points lower than the 1997 survey.

Finally, a special government survey of population intelligence from 2011 contains perhaps the largest sample size of any global IQ study. Nearly 1000 students were surveyed from each of Thailand’s 76 provinces: 72,780 children, ages 6-15, were tested with the Standard Progressive Matrices—parallel version (TDMH, 2011). The average IQ of this massive sample was 96.5.

__________________________________________

■ Id. Estimated IQ for Thailand

Table I lists 23 studies with intelligence test data for Thailand. You can hover your mouse over the variables at the top for definitions and procedures.

Table I. IQ test scores in Thailand

Admin

Age

N

Test

FC

IQ

SD

Reference
~1956
7-12
1385
CPM
+6.9
82.1
Malakul, 1957
~1965
7-12
892
DAM
0.6
97.5
Talapat & Suwannalert, 1966a
~1965
7-14
1438
CPM
+4.2
82.1
Talapat & Suwannalert, 1966b
~1969
7-11
70
WISC
6
88.3
Rajatasilpin et al, 1970
1978
17-60
1462
CPM
+0.3
71.8
Chou & Lau, 1987
~1988
9-11
1358
CPM
2.7
91.1
Pollitt et al, 1989
1995
6-9
529
WISC
13.8
76.6
11.5
Siripitayakunkit et al, 1999
1997
6-13
3846
TONI
2.1
89.9
14.9
Thai Institute of Public Health, 1998
~1999
18-50+
400
SPM
2.1
89.5
Phattharayuttawat et al, 2000
2000
10
427
TONI
3
75
12
Sungthong et al, 2002
~2002
5-11
900
CPM
6.9
106.1
Phattharayuttawat et al, 2003
2003
7-12
396
CPM
7.2
99
11
Sangtongluan, 2004
2004
6-12
3135
TONI
2.1
86
Ruangdaraganon, 2004
13-18
3150
84.6
 
2004
6-11
3848
CPM
7.5
95.6
Sukhatunga et al, 2006b
2004
12-18
5702
APM
7.5
104.6
Sukhatunga et al, 2006a
2004
6-16
3300
WISC
3.9
94.5
Wanitrommani et al, 2004
~2005
13-15
319
TONI
2.4
88
14.9
Isaranurug et al, 2006
2005
7-9
100
CPM
7.8
87.4
Nimmalangkun, 2006
2006
7-11
390
CPM
8.1
93.9
9
Sroythong, 2008
2006
7-12
748
CPM
8.1
94
Thavornsuwanchi, 2008
2007
9
560
CPM
8.4
75.1
Pongcharoen et al, 2011
WISC
4.8
87
9.8
 
2009
6-14
5993
TONI
3.6
87.8
14.3
Aekplakorn, 2009
2011
6-15
72780
SPM
2.1
96.5
16.4
Thai Dept. of Mental Health, 2011

__________________________________________

The inconsistencies in this data make a national IQ estimate problematic. Several recent standardizations of the Raven’s Progressive Matrices have produced IQ scores that meet or rival Western norms: 106.1 on the CPM (Phattharayuttawat et al,2003), 104.6 on the APM (Sukhatunga et al, 2006a), and 96.5 on the SPM (TDMH, 2011).

On the other hand, three large standardizations of the Test of Nonverbal Intelligence have produced IQs of 89.9 (TIPH, 1998), 85.3 (Ruangdaraganon, 2004), and 87.8 (Aekplakorn, 2009).

So which test should we trust? Which test should we distrust? Table I has data for two other tests. A standardization of the Draw-a-Man suggested Thai IQ was already close to Western norms in the 1960s. A standardization of the WISC in the mid-2000s suggests something intermediate to the RPM and TONI scores. Meanwhile, the achievement test data reviewed in Section II suggests that the TONI results are more accurate.

The environmental arsenic study by Siripitayakunkit et al (1999) is the only sample that I consider inappropriate for generating an IQ estimate. This leaves us with 22 studies and 23 normal samples. The median IQ from these samples is 89.5. The weighted average from the same samples is 94.6. If we remove the massive 2011 survey—simply to gauge its effect on the data—the weighted average is 91.2.

How do possible differences over time affect the data? I examine cohort differences in Section IIId. For now, the median IQ from 8 studies before the year 2000 is 88.9, and the median IQ from 15 studies between 2000 and 2011 is 93.9.

The median score from the newer studies is close to the Wechsler norms, and provides a fair compromise between the RPM and the TONI: so 15 recent and reasonably representative samples give us an IQ of 93.9 for Thailand.

__________________________________________
__________________________________________

                 SECTION II: ACHIEVEMENT TEST STUDIES IN THAILAND                 

Since the 1990s there have been accelerating efforts by global educators to compare school children and adults from different nations on their proficiency in academic skillsets like math, reading, and science. Psychometricians have labored to design school achievement tests that are unbiased across many different languages and cultures. International rankings on these tests are highly correlated with national IQ scores; the two kinds of tests appear to be measuring much the same thing (Rindermann, 2007).

The first cross-national achievement study in 1964 looked at math performance in 11 different nations. All were developed economies. Few poor nations participated in these early comparisons, but Thailand is an exception. There were 15 international assessment events between 1964 and 1995: Thailand participated in 7, which is more than any other developing nation (Iran participated in three. India participated in two early assessments, and has avoided them ever since).

So we have 40 years of achievement data for Thai schoolchildren, beginning with their participation in the First International Science Study (FISS): 1970-1971.

The FISS evaluated three different age groups in 16 countries. Thailand participated in the 4th and 8th year assessment events. In relation to the UK scores, Thai 10 year-olds had an Achievement Quotient of 91.8, and Thai 14 year-olds had an AQ of 95.8 (Keeves, 1992).

Thai 13 year-olds participated in the Second International Mathematics Study (SIMS): 1980-1982. I calculate an AQ in relation to US scores here because that is the country available in my data source (Engelhard, 1990). Thailand had an AQ of 95.3.

Thai 14 year-olds participated in the Second International Science Study (SISS): 1983-1984. Thailand had an AQ of 94.9. (Keeves, 1992) [2]

Thailand participated in the Second International Reading Study (SIRS): 1990-1991. I calculate an AQ in relation to US scores, because the UK did not participate. Thai 15 year-olds had an AQ of 90.5. (Elley, 1992)

Thailand participated in the Third International Mathematics and Science Study (TIMSS): 1994-1995. Thai 10 year-olds had an AQ of 93.3. Thai 14 year-olds had an AQ of 99.6. (Martin & Kelly, 1997)

A continuation of this study, called TIMSS-Repeat, occurred in 1999. Thai 14 year-olds had an AQ of 93.7. (Martin et al, 2000)

Thai 15 year-olds participated in the Programme for International Student Assessment (PISA) in 2000, 2003, and 2006. Their AQ scores were 85.8, 86.7, and 87.4, respectively. (OECD, 2003; OECD, 2004; OECD, 2007)

Thai 14 year-olds participated in the Trends in International Mathematics and Science Study (TIMSS): 2007. Their AQ was 89.8. (Olson et al, 2008)

Thailand participated in TIMSS 2011. Thai 10 year-olds had an AQ of 89.4. Thai 14 year-olds had an AQ of 86.8. (Mullis et al, 2012)

Most recently, Thailand participated in PISA 2012. Thai 15 year-olds had an AQ of 90.3 (OECD, 2013a).

Comparative Achievement Quotients for Thailand are listed in Table II. The median Achievement Quotient from 12 studies and 15 samples is 90.5.

Table II. Achievement test scores in Thailand

Admin

Age

N

Test

AQ

Reference
1971
10
1810
FISS
91.8
Keeves, 1992
1971
14
1932
FISS
95.8
Keeves, 1992
1982
13
3740
SIMS
95.3
Engelhard, 1990
1984
14
3780
SISS
94.9
Keeves, 1992
1991
15
2753
SIRS
90.5
Elley, 1992
1995
10
5862
TIMSS
93.3
Martin & Kelly, 1997
1995
14
11643
TIMSS
99.6
Martin & Kelly, 1997
1999
14
5732
TIMSS
93.7
Martin et al, 2000
2000
15
5340
PISA
85.8
OECD, 2003
2003
15
5236
PISA
86.7
OECD, 2004
2006
15
6192
PISA
87.4
OECD, 2007
2007
14
5579
TIMSS
89.8
Olson et al, 2008
2011
10
4448
TIMSS
89.4
Mullis et al, 2012
2011
14
6124
TIMSS
86.8
Mullis et al, 2012

2012
15
6606
PISA
90.3
OECD, 2013a

__________________________________________

There is evidence that achievement scores have increased over time in developing nations (Meisenberg & Woodley, 2013). This is not the case for Thailand, where a straightforward reading of the data is that scores have been declining since the 1970s (Figure 1). Part of this can be explained by secular expansions in school enrollment. Lower ability students are also kept in the school system by changes in compulsory education laws. Thailand raised the schooling requirement from 6 to 9 years in 2002 (Meisenberg & Woodley, 2013, p. 814), but it is difficult to see any effect of this specific regulation on TIMSS or PISA performance. Scores have risen (from a low initial level) across all four of the PISA events from 2000-2012. Meanwhile, scores decreased on TIMSS from 1995-1999, and again from 1999-2007, and again from 2007-2011.

Figure 1. Achievement test scores in Thailand: 1971-2012

Rindermann et al (2009) use the collective scores from PISA and TIMSS assessments to study the tails of the achievement distribution. The mean AQ for Thailand from these studies was 90.1. The 5th percentile of Thai achievement was 71.1, while the 95th percentile was 110.

A reader at Steve Sailer’s website used this data to generate standard deviations for the list of nations. Thailand had one of the smallest standard deviations (ranking 81 of out of 90 nations). This is reflected in the 5th percentile, which compares favorably to other nations with a similar mean AQ (e.g. the mean in Israel is 92.6, but the 5th percentile is 64.6).

The average standard deviation of Thai IQ from three large population surveys was 14.9, which is no different from the Western norms (TIPH, 1998; Aekplakorn, 2009; TDMH, 2011).

.

__________________________________________
__________________________________________

                 SECTION III: THE DEMOGRAPHY OF THAI ABILITY                 

This section investigates various subgroup differences among Thai people. There is considerable curiosity about the Thai Chinese minority—they are 14% of the population but hold more than 80% of the nation’s capital. Unfortunately, none of the studies cover ethnicity.

__________________________________________

■ IIIa. University Students

There are a number of intelligence studies for Thai University students. The earliest report by Ravipan Somnapan (1972) looked at the correlation between Standard Progressive Matrices and GPA among 58 graduate students in the 1960s. The correlation was .12 and the sample had an IQ of 96.3.

Tipawadee Amawattana (1985) adapted and standardized the Otis-Lennon School Ability Test for 3056 freshman at 7 different Universities. A try-out with 183 Thammasat University and American Field Service students showed an IQ of 115 on a translated but unmodified version of the test. The adapted version of the test was highly correlated with Thai College Entrance Exam scores: .64.

Paiboon Tevarak (1997) administered the Advanced Progressive Matrices to 1000 undergraduates at Chulalongkorn University in 1985. This sample had an IQ of 116.7.

Sontirat et al (1988) developed norms for Cattell’s Culture-Fair Intelligence Test using 2453 freshman at Chulalongkorn and Kasetsart Universities in Bangkok. This sample had a CCF IQ of 95.9.

Manus Jintanadilokkul (1990) investigated the relationship between APM scores and Spearman’s g among 401 students at the University of the Thai Chamber of Commerce. This sample had an IQ of 112.

Phattharayuttawat et al (1994) created norms for the APM using 491 University students in Bangkok. This sample had an IQ of 114.3

Chalitta Kyiyanan (1995) created APM norms using 960 undergraduates attending six different Universities in Bangkok. This sample had an IQ of 115.5.

The final study is actually a pre-University sample of elite students (Prohmpetch, 2004): the APM was given to 832 adolescents who were attending selective admissions high schools in Bangkok. This sample had an IQ of 114.8.

IQ scores for Thai University students are listed in Table III.

Table III. IQ test scores of University students in Thailand

Admin

Age

N

Test

FC

IQ

Reference
1963
Adult
58
SPM
+4.8
96.3
Somnapan, 1972
1982
Adult
183
Otis
0.9
115
Amawattana, 1985
1985
Adult
1000
APM
1.8
116.7
Tevarak, 1997
~1987
16-24
2354
CCF
11.1
95.9
Sontirat et al, 1988
~1989
Adult
401
APM
3
112
Jintanadilokkul, 1990
~1993
Adult
491
APM
4.2
114.3
Phatthrayuttawat et al, 1994
~1994
16-25
960
APM
4.5
115.5
Kiyiyanan, 1995
2003
15-20
832
APM
7.2
114.8
Prohmpetch, 2004

__________________________________________

Based on 8 samples the median IQ of Thai University students is 114.6.

This is 1/3-1/2 of a standard deviation higher than would be predicted from the national IQ (Section Id). Research traditionally showed that college students had an IQ of 115 (Herrnstein & Murray, 1994, pp. 151-152), but the reason for this is simply because only the top 15-20% of the population used to attended college. Obviously, as tertiary enrollment expands, the average college person becomes more average. Since the 1960s, college enrollment in the U.S. has grown to ~40% and consumed talent down to the 60th percentile, so the average college graduate in the U.S. now has an IQ of 105 or lower.

If Thailand’s IQ is 94, then these Thai college students must represent the top 8% of the population. But tertiary enrollment was ~20% in Thailand during the 1980s (The median year for these studies is 1988).

An explanation for this is that these studies are from Thailand’s most selective Universities. Lesser colleges outside the Bangkok area would have IQs of 107-110.

In support of this, Amawattana (1985, p. 88) showed that Otis-Lennon scores of the five Universities in the Bangkok Metropolitan Region were about ½ standard deviation higher than the scores of the two Universities located in other parts of the country. This is relevant because all eight samples from Table III are students in Bangkok.

Of course, if the average Thai adolescent really has an IQ of 104.6 on the Advanced Progressive Matrices (Sukhatunga et al, 2006a), then it’s actually a puzzle that these scores aren’t much higher.

Several studies also provide intelligence by field of study. For instance, IQ data from Amawattana (1985, p. 96) suggests the following hierarchy of majors:

  MEDICINE 132.7      
  SCIENCE 119.5      
  HUMANITIES 108      
  EDUCATION 107.1      

Sontirat et al (1988, p. 56) also supplied IQ data for different majors:

  APPLIED SCIENCES 101.2      
  HUMANITIES 94.6      
  EDUCATION 92.6      

Prohmpetch (2004, p. 57) reports the same difference at selective admissions high schools:

  SCIENCE PROGRAM 118.8      
  ARTS PROGRAM 113.3      

This order of Science > Humanities > Education is found in G.R.E. scores in the United States. It would seem the average IQ of different majors is not peculiar to individual countries, but is functionally related to the inherent difficulty of different disciplines.

__________________________________________

■ IIIb. Sex

Some reports have suggested that Thailand has atypical sex differences in academic ability. For example, an analysis of PISA 2003 math scores found that Thailand and Iceland were the only two countries where girls were more likely than boys to perform above the 99th percentile (Hyde & Mertz, 2009). Further, one review of Thai intelligence studies concluded “Girls… have higher IQ score than boys in all studies.” (Visanuyothin & Arunruang, 2012)

Since the Human Varieties dataset contains more studies, it is worth revisiting this question. Table IV lists 10 studies of the general population that report scores for males and females. Standard scores are shown as they were reported, and raw scores have been converted from available norm tables without other adjustments. A negative difference indicates lower female scores.

Table IV. Male and female IQ differences in Thailand: general population

Male IQ
Female IQ
Difference
Test
Male N
Female N
Reference
96.1
91.2
-4.9
WISC
44
22
Rajatasilpin et al, 1970
91.9
90.8
-1.1
TONI
1971
1875
Thai Institute of Public Health, 1998
79
78
-1
TONI
199
228
Sungthong et al, 2002
114
113
-1
CPM
Phattharayuttawat et al, 2003
99.3
97.4
-1.9
CPM
1970
1878
Sukhatunga et al, 2006b
109.5
110.5
1
APM
2582
3120
Sukhatunga et al, 2006a
110
107
-3
CPM
Sroythong, 2008
93.4
92.8
-0.6
WISC
284
276
Pongcharoen et al, 2012
86
87
1
CPM
 
90.8
92.1
1.3
TONI
2968
3025
Aekplakorn, 2009
97.7
99.9
2.2
SPM
34420
38242
Thai Dept. of Mental Health, 2011

__________________________________________

Most of these references find that males have higher IQs than females. “All studies” included in the review by Visanuyothin & Arunruang (2012) do not show higher female IQ: The large 1997 health survey found higher scores for males (TIPH, 1998).

The Raven standardizations reported by Phattharayuttawat et al (2003), Sukhatunga et al (2006a), and Sukhatunga et al (2006b), were not included in their review, and two of the three show higher IQ for males.

Two large and recent surveys did find higher IQs for females, including the massive 2011 population survey, which indicated that IQ was over two points higher for females. A weighted average of these studies gives this reference full authority. The median difference from 10 different studies, on the other hand, gives males an advantage of 1 IQ point.

Additional data comes from Thai University students. Table V has 5 different studies of University students that report scores for males and females.

Table V. Male and female IQ differences in Thailand: University students

Male IQ
Female IQ
Difference
Test
Male N
Female N
Reference
104
97.9
-6.1
Otis
1072
1984
Amawattana, 1985
107.8
106.3
-1.5
CCF
1260
1091
Sontirat et al, 1988
117
115
-2
APM
148
253
Jintanadilokkul, 1990
119
118
-1
APM
220
271
Phattharayuttawat et al, 1994
120
119.4
-0.6
APM
480
480
Kiyiyanan, 1995

__________________________________________

If females have a higher average IQ this gap should grow larger at higher IQ levels. But all of the University studies show higher IQ for males, and the male advantage is larger in the University studies than it was for the general population. The median difference from 5 studies is 1.5 points in favor of males. To the extent that more females attend college, part of this might reflect the greater selectivity of the male sample; but this difference is also reflected in the study with higher male enrollment (Sontirat et al, 1988).

Although the data is open to different interpretations, I do not think this evidence supports a higher IQ for females in Thailand.

Counter-intuitively, a higher female IQ in a population can suggest a poor general environment, while a higher male IQ can indicate a salutary environment. This is because male development is more responsive to environmental conditions (good and bad) than female development (Jensen, 1971).

__________________________________________

■ IIIc. Age

Is the ~6 point IQ gap between Anglophone and Thai children smaller at early ages and does it widen over time?

Some social scientists have suggested that African-Americans suffer from a cumulative deficit, and that this is reflected in smaller intelligence differences at younger ages. They hypothesize that A) intelligence gaps are created by poor environments, and that B) poor environments incrementally damage cognitive development throughout childhood and adolescence.

In earlier posts I have provided evidence that IQ gaps do not widen over time for two U.S. ethnic groups. The IQ gap between white and black Americans is about the same size prior to school entry as it is at the end of high school. I’ve found a similar pattern for Puerto Ricans living in the mainland United States.

However, I also showed that children in Puerto Rico lose ground on Americans over time. Differences between the populations are smaller at younger ages. This suggests that part of the IQ gap between Puerto Rico and the United States can be explained by an inferior environment.

             Supplement 1. Age, birth cohort, & IQ in Thailand             

To investigate this same question for Thailand, I’ve taken the 22 studies from Section Id, and tabulated the data for all individually reported age groups (Supplement 1). This creates 72 samples. If a paper gave an age range, but did not report data for yearly age groups, I assigned a middle value (e.g. a study with 7-9 year olds is used to represent 8 year-olds); these estimates are labeled ‘Averaged’ in the Method column.

Unfortunately this dataset does not currently contain IQ scores for Thai preschool children, and most of the age groups above 14 are represented by only one study (Sukhatunga et al, 2006a), so my analysis is restricted to ages 6-14. This leaves 65 samples.

Figure 2. Thai IQ, Ages 6-14 years

There is no indication of developmental differences between Anglophone and Thai children (Figure 2). Thai IQ is comparatively stable from childhood to mid-Adolescence.

__________________________________________

■ IIId. Birth Cohort

How has Thailand’s IQ changed over time? To make an estimate of the Flynn Effect we need two different standardizations of the same test across different decades. We have four different standardizations of the Coloured Progressive Matrices (Malakul, 1957; Talapat & Suwannalert, 1966a; Phattharayuttawat et al, 2003; Sukhatunga et al, 2006b). The norms by Talapat & Suwannalert (1966a) were limited to Bangkok, and Phattharayuttawat et al (2003) do not provide details about their sampling. Malakul (1957) and Sukhatunga et al (2006b) provide the longest time series and are the most comparable samples. The norms by Malakul (1957) were collected in ~1956 in four provinces, while the norms described by Sukhatunga et al (2006b) were collected from seven provinces in 2004. These two norms are separated by nearly 50 years.

IQ on the 1956 norms was 82.1, while IQ on the 2004 norms was 95.6. This is an increase of 13.5 points (.90σ).

If we remove the Flynn adjustments, which mask the contemporaneous rise in U.K scores, IQ in Thailand increased by 27.9 points on the Coloured Progressive Matrices over 48 years.

This is a rate of 0.58 points per year (or 5.8 points per decade), which is about twice as fast as the Flynn Effect in the United States and England (0.3 per year/3 per decade).

But how generalizable has this gain been across different tests and studies? Did the comparative IQ of Thailand really increase by almost one standard deviation since the 1960s?

Another way to evaluate this is to look at the IQ scores from different yearly birth cohorts. Supplement 1 includes data for 72 different Thai samples born between 1928 and 2003 (75 years). Tests were administered to these subjects between 1956 and 2011 (55 years).

Figure 3. Thai birth cohort & IQ: 1944-2003

The trendline in Figure 3 suggests a similar story to the CPM norms, with IQ increasing by about 10 points over 50 years.

One problem is that there is little data from the 1960s-1980s, with the two major comparison groups being people born in the 40s-50s and in the 90s-00s. There was some sort of increase in those 50 years, but it’s not obvious that it was continuous. If we look at just the data since the 1970s, there is no evidence for increasing IQ.

Figure 4. Thai birth cohort & IQ: 1986-2003

Figure 4 looks at 44 samples born between 1986 and 2003. This evidence suggests a decline in Thai IQ since the 1980s. This is consistent with the trend in achievement data, which exhibits declining scores since the 1970s (see Figure 1). You’ll notice that many of the 1980s scores are above 100—that’s because they are all teens from the APM standardization by Sukhatunga et al (2006a). Removing these samples does not erase the downward trend.

Indeed, we can also calculate a Flynn Effect from two Thai standardizations of the Test of Nonverbal Intelligence: the 2nd National Survey of Health in 1997 (TIPH, 1998), and the 4th National Survey of Health in 2009 (Aekplakorn, 2009). IQ on this test dropped by 2.1 points over 12 years. This suggests a Flynn Effect of 0.1 point per year/1 point per decade. This is three times slower than the Flynn Effect on Western norms, which suggests that IQ in Thailand is decreasing by 2 points per decade.

__________________________________________

■ IIIe. Diaspora

There is not a large amount of cognitive data for Thais outside of Thailand. There are over 200,000 Thai Americans—by far the largest Diaspora population. Many of these either came as foreign brides or are the mixed children from these relationships. (Tiger Woods being the most famous example. The infamous white supremacist Charles Murray also has two half-Thai children). Many others came as high skill immigrants. I have no cognitive data for Thai Americans.

There is a book about the adjustment of 116 Thai children that were transracially adopted and raised in the Netherlands in the 1970s-80s (Hoksbergen et al, 1987). This book does not include IQ data, but mostly provides soft evidence from sources like teacher and parent questionnaires. The authors conclude that the children were well adapted to local academic demands: “… school records indicate that their performance in the cognitive field ranges from average to good. They are perfectly able to keep up with their classmates at school … The fact that 6% of the children are thrown back on special education can hardly be called remarkable, because a similar percentage of Dutch children is in special education, as well” (Hoksbergen et al, 1987, pp. 62-63).

Oddly enough the only IQ data for Thais outside of their home country appears to be a small sample from Pakistan.

Afzal Imam (1986) compared volunteers from 5 ethnic groups on the Rod-and-Frame Test (a measure of field dependence) and the Standard Progressive Matrices. There are few details about these subjects, but it is implied that they are students at the University of Karachi (… “none had taken any courses in psychology”). Their SPM scores are very low for University students, but there is no indication that the test was administered incorrectly.

__________________________________________

    Nationality IQ N    
             
  THAI 74.2 15    
             
  IRANIAN 73.2 27 Table VI. IQ of 5 ethnic groups in Pakistan  
             
  PAKISTANI 73 28    
             
  PALESTINIAN 67.2 43    
             
  SOMALIAN 63.2 14    

__________________________________________
__________________________________________

The Thai subjects had the highest IQ score, while the Somalians had the lowest IQ score, which suggests a Rushtonian pattern can even be observed for ethnic groups living in South Asia.

__________________________________________

■ IIIf. Region

Regional differences in Thailand are estimated from 5 sources: 1) A giant 2011 survey of every province, 2) 6 representative surveys conducted between 1998-2011, 3) a study of University students, 4) 11 research studies from various locations, and 5) PISA 2012.

__________________________________________

1. A VERY BIG STUDY

By far the best study for determining regional differences is the government’s 2011 survey of intelligence, which examined nearly 1000 children in each of the 76 provinces (TDMH, 2011).

I’ve taken the scores from every province, adjusted them around my own average of 93.9, and created an IQ map for Thailand’s 76 provinces (Figure 5), and an IQ map for Thailand’s 7 major regions (Figure 6).

Most of the studies divide the country into seven different regions, with the populous capital Bangkok classified as its own region.

Bangkok is clearly the most intelligent region, but there is a hot spot of intelligence in the whole area surrounding the Bay of Bangkok, which includes many of the provinces in the lower Central and Eastern region. The Bangkok Metropolitan Region, which contains 22% of Thailand’s population, has an IQ of 101.4.

Scores are low in the South and lowest in the Northeast. There is a very modest relationship between latitude and intelligence in Thailand. The correlation between IQ and latitude for 76 provinces is .12. For comparison, the correlation between achievement scores and latitude for 61 provinces in Vietnam was .33.
__________________________________________

2. SIX SURVEYS OF INTELLIGENCE

Another method for estimating regional differences is to look at scores from all the major surveys since 1998. This data is provided in Table VII.

__________________________________________

Table VII. IQ for regions of Thailand

   
TIPH 1998
Ruangdaraganon 2004
Mahidol 2004 (CPM)
Mahidol 2004 (APM)
Aekplakorn 2009
TDMH 2011
 
                 
   Bangkok
94.4
92.5
100.8
104.3
93.6
102.4
(98)
   Central
89.1
86.7
93.8
102.5
86.3
99.3
(93)
   East
87.5
105
98.8
   North
83
82.1
89.7
102.5
85.3
98
(90.1)
   Northeast
84.3
83.8
91.8
98
90
94
(90.9)
   West
87.5
102.5
98.6
   South
90.6
86
94.1
108.8
84.8
94.8
(93.2)
   
 
   Country
89.9
85.3
95.6
104.6
87.8
96.5
(93.3)

__________________________________________

__________________________________________

Regional data from the Mahidol University study are not reported in Sukhatunga et al (2006a; 2006b), I had to extract this data from the 7 source dissertations (Moleechart, 2005; Mungkhetkland, 2005; Runseawa, 2005; Sirisakpanit, 2005; Palakas, 2005; Udomphol, 2005; Intuptim, 2005).

The mean IQ of the six studies in Table VII is 93.3. On average scores are 4.7 points higher in Bangkok. This is consistent with the 2011 study, but there is not a lot of agreement between the studies on other regions. For example, half of the studies find above average scores in the South (The increasingly dubious APM norms find a higher IQ in the South than in Bangkok). The region with the lowest scores is the North. The Northeast has the second lowest scores.

__________________________________________

3. A UNIVERSITY STUDY

Another source of data is the study of 3056 freshman at 7 Universities by Amawattana (1985, p. 101). Students were given a locally adapted IQ test (set mean 100) and scores were reported by region of origin:

  EAST 102.8      
  BANGKOK 102.3      
  NORTH 99.8      
  CENTRAL 99.2      
  NORTHEAST 98.6      
  WEST 98.1      
  SOUTH 95.5      

__________________________________________

4. ELEVEN RESEARCH STUDIES

Some of the studies I reviewed in Section 1 reported IQ scores in specific locations of Thailand:
__________________________________________

  BANGKOK 93.9

Talapat & Suwannalert, 1966a; Talapat & Suwannalert, 1966b; Rajatasilpin et al, 1970; Sangtongluan, 2004; Sroythong, 2008

   
         
             
  EAST 92.6 Pollitt et al, 1989; Thavornsuwanchi, 2008    
             
  NORTHEAST 81.1 Pongcharoen et al, 2011    
             
  NORTH 79.6

Chou & Lau, 1987; Nimmalangkun, 2006

   
             
  SOUTH 75 Sungthong et al, 2002    

__________________________________________

5. PISA 2012

Finally, there is regional data for the PISA 2012 assessment (OECD, 2013b). The Acievement Quotients are shown below.

  BANGKOK 92.7      
  CENTRAL 85.9      
  UPPER NORTH 91.8      
  LOWER NORTH 86.8      
  UPPER NORTHEAST 87.9      
  LOWER NORTHEAST 86.4      
  SOUTH 85.2      
           
  COUNTRY 90.3      

__________________________________________

Unlike the other studies, PISA makes a distinction between the “Upper North” and the “Lower North” and finds that scores in the Upper North rival Bangkok. The 2011 intelligence survey found higher scores in the province of Lampang than in Bangkok (see the big bright spot in Figure 5), but otherwise does not suggest high scores in the ‘Upper North’, or much of a difference between the upper and lower parts of the region. PISA also shows somewhat lower scores in the South than in the Northeast.

Giving some consideration to all five sources of data, I conclude that IQ is highest in Bangkok followed by the Eastern region. IQ is at the Thai average in the Western and Central regions, and on the low end of average in the North. IQ is below average in the South, and lowest in the Northeast.

Rural/urban differences in Thailand were reported in one recent study (Pongcharoen et al, 2012). The IQ of urban schoolchildren was 89.1, and the IQ of rural schoolchildren was 82.5. This is a difference of 6.6 points.

The gap between rural and urban children virtually disappeared in developed nations in the decades after WWII. The United States had a similar rural/urban gap of 6.5 points in the 1940s, but it fell to 2 points by the 1970s (Loehlin, 2000, p. 181). If and when this happens in Thailand it should positively affect national IQ.

__________________________________________
__________________________________________

                 SECTION V: CONCLUSIONS                  

There are a number of high quality IQ studies for Thailand, but this has paradoxically made it harder to estimate a national IQ, because the tests do not agree with each other. Several standardizations of the Raven’s Progressive Matrices have suggested an IQ close to 100, while several standardizations of the Test of Nonverbal Intelligence have suggested an IQ close to 90. Smaller research studies with the Progressive Matrices also tend to show lower IQs. A standardization of the WISC gives an intermediate score.

The median IQ from 15 samples tested since the year 2000 is 93.9, which I believe is a safe estimate for Thailand’s current national IQ (Section Id).

The average for the 8 samples tested before the year 2000 is 88.9. I examined these secular differences more closely in Section IIId. IQ increased by nearly 2 standard deviations on the Coloured Progressive Matrices between 1956 and 2004—this is about 6 points per decade, while the Flynn Effect in the US and the UK is only 3 points per decade. However, time series data from 72 samples does not suggest relative increases in Thai IQ since at least the 1980s. In fact, the data suggest that Thai IQ may actually be declining relative to Western norms over the last several decades.

This also appears to be the case with cross-national achievement test data, which I reviewed in Section II. Thai scores have fallen since the 1970s, even while scores have been rising in other parts of the developing world. Data from 12 studies and 15 samples between 1971-2012 give Thailand an Achievement Quotient of 90.5, which is suspiciously similar to the median IQ from all 23 normal samples (89.5). Either Thailand has lower IQ than my estimate from current samples, or it underperforms on international assessment tests relative to its ability level.

Despite apparent declines, I find no evidence for cognitive deprivation in two theoretical bellwethers: 1) A comparison of male and female scores suggests that male IQ is higher (Section IIIb)—deprivation is thought to advantage females vis-à-vis males; 2) A cross-sectional analysis of age differences suggests Thai intellectual development across childhood is stable relative to Western norms (Section IIIc)—deprivation is thought to progressively lower IQ as children age.

I looked at regional differences in IQ in Section IIIf. IQ is highest in Bangkok and in the East, and lowest in the South and in the Northeast. Despite a brainy central coastline, a lackluster North, and a lagging Northeast, there is still a modest correlation between latitude and IQ for Thailand’s 76 provinces: .12. This pattern is repeatedly found within and between countries.

There is apparently a sizable urban/rural gap in Thailand. There is data to suggest similar gaps used to exist in developed countries, and that they disappeared rapidly in the middle of the 20th century. This seems like a potential avenue for future rises in national IQ. There is not much in this post that speaks to the innate potential of Thai people; one transracial adoption study suggested that Thai children raised in the Netherlands had no special difficulties with Western academic demands.

On the other hand, the regional and latitudinal slope in IQ scores is consistent with certain genetic theories (e.g. Gregory Cochran’s mutational load hypothesis).

Leaders and scientists in Thailand are far more overtly concerned and curious about their national intelligence than their counterparts in any other country, so at least the future promises us more data to help unlock the mysteries of Thai intelligence.

__________________________________________
__________________________________________

__________________________________________
__________________________________________

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


Aekplakorn, W. (2009). Report on the Fourth National Health Examination Survey: 2008-2009. Bangkok, Thailand: National Health Security Office.

Amawattana, T. (1985). Adaptation and restandardization of the Otis-Lennon school ability test for Thai university students. Unpublished doctoral dissertation, University of Missouri–Columbia, USA.

Chou, E.C., & Lau, L.J. (1987). Farmer ability and farm productivity: a study of farm households in the Chiang Mai valley, Thailand 1972-1978. Report No. EDT62. Education & Training Department, Washington, DC, USA: World Bank.

Elley, W.B. (1992). How in the World Do Students Read? IEA Study of Reading Literacy. Hamburg, Germany: IEA.

Engelhard, G., Jr. (1990). Gender differences in performance on mathematics items: Evidence from the United States and Thailand. Contemporary Educational Psychology, 15, 13-26.

Grisley, W., & Kellogg, E.D. (1983). Farmers’ subjective probabilities in Northern Thailand: an elicitation analysis. American Journal of Agricultural Economics, 65, 74-82.

Harinasut, P., & Suwanlerd, S. (1958). The further investigation of Progressive Matrices Test in Thai children. Journal of the Psychiatric Association of Thailand, 3, 72-83.

Harris, D.B. (1963). Children’s drawings as measures of intellectual maturity: A revision and extension of the Goodenough Draw-a-Man Test. New York, USA: Harcourt, Brace & World.

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

Hoksbergen, R.A.C., Juffer, F., & Waardenburg, B.C. (1987). Adopted children at home and at school: The integration after eight years of 116 children from Thailand in the Dutch society. Lisse, Netherlands: Swets & Zeitlinger.

Hyde, J.S., & Mertz, J.E. (2009). Gender, culture, and mathematics performance. Proceedings of the National Academy of Sciences, 106, 8801-8807.

Imam, A. (1986). Field dependence: a cross-cultural study. Pakistan Journal of Psychology, 17, 3-20.

Intuptim, W. (2005). The characteristics of the Colored Progressive Matrices (CPM) and the Advanced Progressive Matrices (APM) in Thai students age 6-18 years old: the eastern region. Unpublished master’s thesis, Mahidol University, Thailand.

Isaranurug, S., Klinman, S., Nantamongkolchai, S., & Plubrukan, R. (2006). Implications of family protective-risk index for screening cognitive development of children aged 13-15 years. Journal of the Medical Association of Thailand, 89, 1427-1433.

Jensen, A.R. (1971). The race X sex X ability interaction. In R. Cancro (Ed.), Intelligence: Genetic and environmental influences, (pp. 107-161). New York: Grune & Stratton.

Jintanadilokkul, M. (1990). The relationship between APM scores and Spearman’s g in University students. Journal of the University of the Thai Chamber of Commerce, 11, 56-69.

Keeves, J.P. (1992). Learning Science in a Changing World. Cross-National Studies of Science Achievement: 1970 to 1984. The Hague, Netherlands: IEA.

Kyiyanan, C. (1995). The normative scores of the Advanced Progressive Matrices for the Thai Undergraduate students in Bangkok. Unpublished master’s thesis, Mahidol University, Thailand.

Loehlin, J.C. (2000). Group differences in intelligence. In R.J. Sternberg (Ed.), Handbook of Intelligence (pp. 176-193). Cambridge, UK: Cambridge University Press.

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

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

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

Malakul, S. (1957). Standardization of Progressive Matrices for Thai children. Journal of the Psychiatric Association of Thailand, 2, 7-25.

Martin, M.O., Gregory, K.D., & Stemler, S.E. (2000). TIMSS Technical Report: IEA’s Third International Mathematics and Science Study at the Eighth Grade. Boston, USA: International Study Center, Boston College.

Martin, M.O., & Kelly, D.L. (1997). Third International Mathematics and Science Study Technical Report, Volume II: Implementation and Analysis. Boston, USA: Center for the Study of Testing, Evaluation, and Educational Policy, Boston College.

Medrich, E.A., & Griffith, J.E. (1992). International mathematics and science assessments: What have we learned? US Department of Education, Office of Educational Research & Improvement, National Center for Education Statistics.

Meisenberg, G., & Woodley, M.A. (2013). Are cognitive differences between countries diminishing? Evidence from TIMSS and PISA. Intelligence, 41, 808-816.

Moleechart, W. (2005). The characteristics of the Colored Progressive Matrices (CPM) and the Advanced Progressive Matrices (APM) in Thai students age 6-18 years old: northeast region. Unpublished master’s thesis, Mahidol University, Thailand.

Mullis, I.V.S., Martin, M.O., Foy, P., & Arora, A. (2012). TIMSS 2011 International Results in Mathematics. Chestnut Hill, USA: TIMSS & PIRLS International Study Center, Boston College.

Mungkhetklang, C. (2005). The characteristics of the Colored Progressive Matrices (CPM) and the Advanced Progressive Matrices (APM) in Thai students age 6-18 years old: Bangkok (The Metropolitan). Unpublished master’s thesis, Mahidol University, Thailand.

Nimmalangkun, W. (2006). A validity study of Draw a Person: A quantitative scoring system. Unpublished master’s thesis, Mahidol University, Thailand.

OECD (2003). Literacy skills for the world of tomorrow. Further results from PISA 2000. Paris, France: OECD.

OECD (2004). Learning for tomorrow’s world. First results from PISA 2003. Paris, France: OECD.

OECD (2007). PISA 2006: Science Competencies for Tomorrow’s World, Vol. 1. Paris, France: OECD.

OECD (2013). PISA 2012 Results: Excellence through Equity. Paris, France: OECD.

OECD (2013). Southeast Asian Economic Outlook 2013: With Perspectives on China and India. Paris, France: OECD.

Olson, J.F., Martin, M.O., & Mullis, I.V.S. (Eds.). (2008). TIMSS 2007 Technical Report. Chestnut Hill, USA: TIMSS & PIRLS International Study Center, Boston College.

Opper, S. (1977). Concept development in Thai urban and rural children. In P.R. Dasen (Ed.), Piagetian psychology: Cross-cultural contributions (pp. 89-122). New York, USA: Gardner Press.

Palakas, S. (2005). The characteristics of the Colored Progressive Matrices (CPM) and the Advanced Progressive Matrices (APM) in Thai students age 6-18 years old: the western region. Unpublished master’s thesis, Mahidol University, Thailand.

Panek, P.E., & Stoner, S.B. (1980). Age differences on Raven’s Coloured Progressive Matrices. Perceptual & Motor Skills, 50, 977-978.

Phattharayuttawat, S., Sukhatunga, S., Channarong, P., & Luckome, M. (1994) The normative scores of intelligence by Advanced Progressive Matrices (APM) in Thai college students. Journal of Thai Clinical Psychology, 25, 66-80.

Phatthrayuttawat S., Sukhatunga, K., Chantra, J., Chaiyasit, W., Bunnagulrote, K., & Imaroonrak, S. (2000). The normative score of the Standard Progressive Matrices for the Thai population. Journal of the Psychiatric Association of Thailand, 45, 45-57.

Phatthrayuttawat S., Sukhatunga, K., Chantra, J., Chaiyasit, W., Bunnagulrote, K., & Imaroonrak, S. (2003). The normative scores of intelligence by Colored Progressive Matrices (CPM) in Thai population. Journal of Thai Clinical Psychology, 34, 58-70.

Pollitt, E., Hathiral, P., Kotchabhakdi, N.J., Missell, L., & Valyasevi, A. (1989). Iron deficiency and educational achievement in Thailand. American Journal of Clinical Nutrition, 50, 687-697.

Pongcharoen, T., DiGirolamo, A.M., Ramakrishnan, U., Winichagoon, P., Flores, R., & Martorell, R. (2011). Long-term effects of iron and zinc supplementation during infancy on cognitive function at 9 y of age in northeast Thai children: a follow-up study. American Journal of Clinical Nutrition, 93, 636-643.

Pongcharoen, T., Ramakrishnan, U., DiGirolamo, A. M., Winichagoon, P., Flores, R., Singkhornard, J., & Martorell, R. (2012). Influence of prenatal and postnatal growth on intellectual functioning in school-aged children. Archives of Pediatrics & Adolescent Medicine, 166, 411-416.

Poulsen, A. (1982). Raven Coloured Progressive Matrices Intelligence Test in Thailand and in Denmark. School Psychology International, 3, 191-194.

Rajatasilpin, A., Suepsaman, B., & Yamarat, V. (1970). Intellectual development and its relationship to the nutritional status among school children. Journal of the Medical Association of Thailand, 53, 788-793.

Rindermann, H. (2007). The g-factor of international cognitive ability comparisons: The homogeneity of results in PISA, TIMSS, PIRLS and IQ-tests across nations. European Journal of Personality, 21, 667-706.

Rindermann, H., Sailer, M., & Thompson, J. (2009). The impact of smart fractions, cognitive ability of politicians and average competence of peoples on social development. Talent Development & Excellence, 1, 3-25.

Ruangdaraganon, N. (2004). Growth development and intelligence of Thai children. Songkla, Thailand: Hat Yai Press.

Runseawa, B. (2005). The characteristics of the Colored Progressive Matrices (CPM) and the Advanced Progressive Matrices (APM) in Thai students age 6-18 years old: the southern region. Unpublished master’s thesis, Mahidol University, Thailand.

Sangtongluan, S. (2004). A study of the correlation of intellectual ability measures between draw a person test and the coloured progressive matrices test in primary school students in Bangkok. Unpublished master’s thesis, Mahidol University, Thailand.

Siripitayakunkit, U., Visudhiphan, P., Pradipasen, M., & Vorapongsathron, T. (1999). Association between chronic arsenic exposure and children’s intelligence in Thailand. In W.R. Chappell, C.O. Abernathy & R.L. Calderon (Eds.), Arsenic exposure and health effects (pp. 141-150). New York, USA: Elsevier.

Sirisakpanit, S. (2005). The characteristics of the colored progressive matrices (CPM) and the advanced progressive matrices (APM) in Thai students age 6-18 years old: the central region. Unpublished master’s thesis, Mahidol University, Thailand.

Somnapan, R. (1972). Progressive Matrices as a scholastic aptitude test for Thai graduate students (research bulletin, no. 11). Bangkok, Thailand: Bangkok Institute for Child Study.

Sontirat, S., Surinya, T., & Jamonman, S. (1988). The study for intelligence norms of Thai students. Kasetsart Journal: Social Sciences, 9, 47-67.

Sroythong, W. (2008). The validity study of the Test of Nonverbal Intelligence, Third Edition: TONI-3. Unpublished master’s thesis, Mahidol University, Thailand.

Sukhatunga, K., Kowasint, C., Phatthrayuttawat S., Chantra, J., Chaiyasit, W., Bunnagulrote, K., & Imaroonrak, S. (2006). Norms of the Advanced Progressive Matrices (APM) in Thai students age 12-18 years old. Journal of Thai Clinical Psychology, 37, 1-10.

Sukhatunga, K., Kowasint, C., Phatthrayuttawat S., Chantra, J., Chaiyasit, W., Bunnagulrote, K., & Imaroonrak, S. (2006). Norms of the Coloured Progressive Matrices (CPM) in Thai students age 6-11 years old. Journal of Thai Clinical Psychology, 37, 9-22.

Sungthong, R., Mo-suwan, L., & Chongsuvivatwong, V. (2002). Effects of haemoglobin and serum ferritin on cognitive function in school children. Asia Pacific Journal of Clinical Nutrition, 11, 117-122.

Talapat, N., & Suwannalert, S. (1966). Measuring the intellectual development of children from drawings. Journal of the Psychiatric Association of Thailand, 11, 137-156.

Talapat, N., & Suwannalert, S. (1966). Standardization of Progressive Matrices Test for Thai children. Journal of the Psychiatric Association of Thailand, 11, 123-136.

Templer, D.I. (2010). The comparison of mean IQ in Muslim and non-Muslim countries. Mankind Quarterly, 50, 188-209.

Tevarak, P. (1997). Measuring intelligence: psychological data of Chulalongkorn University students. Bangkok, Thailand: Chulalongkorn University.

Thai Department of Mental Health (2011). Survey of the intellectual level of students in Thailand: 2010. Bangkok, Thailand: Thai Department of Mental Health, Thai Ministry of Public Health.

Thai Institute of Public Health (1998). The Second National Health Examination Survey: 1996-1997. Nonthaburi, Thailand: Thai Institute of Public Health.

Thavornsuwanchai, A. (2008). The psychometric properties of the Development Test of Visual-Motor Integration 4th Edition, Revised (VMI 4R) in primary school students in Cholburi province. Unpublished master’s thesis, Mahidol University, Thailand.

Udomphol, S. (2005). The characteristics of the Colored Progressive Matrices (CPM) and the Advanced Progressive Matrices (APM) in Thai students age 6-18 years old : the northern region. Unpublished master’s thesis, Mahidol University, Thailand.

Visanuyothin, T., & Arunruang, P. (2012). Intelligence Quotient survey of Thai children in two decades. Journal of Mental Health of Thailand, 20, 67-78.

Wanitrommani, K., Channarong, P., & Wetwirun, C. (2004). The study of intellectual quotient (IQ) scores of Thai children ages 6-16 years using the WISC-III: Thai Version. Nonthaburi, Thailand: Department of Mental Health.

Wicherts, J.M., Dolan, C.V., Carlson, J.S., & van der Maas, H.L.J. (2010). Raven’s test performance of Africans: Average performance, psychometric properties, and the Flynn Effect. Learning & Individual Differences, 20, 135-151.

26 Comments

  1. Home to a large population of Thai Chinese, it is not surprising Bangkok has the highest iq.

  2. Wow, amazing work.

  3. I think you made an error- the national survey from 2011 gives a score of98.6, not 96.5. It says so right in the abstract: https://humanvarietiesfiles.files.wordpress.com/2014/07/2011-tdmh-spm-t.pdf

    “The study show the representative of estimated national IQs of Thai students aged 6-15 equals 98.6 (SD=16.4, 95%CI=98.5-98.7).”

    And from a report in Thai media: http://www.nationmultimedia.com/2011/07/08/national/Thai-students-found-below-global-average-30159779.html

    “Thai children’s average intelligence quotient (IQ) is just 98.59, a little lower than the world’s median of 100, according to new research.”

    • Jason Malloy

      July 19, 2014 at 12:50 am

      Jamison, thank you for keeping an eye out for errors.

      TDMH (2011, p. 22) created their IQ score using the Raven et al (2004) UK norms for the SPM-parallel. Thus I used a 2.1 point “Flynn Correction” to account for 7 years of norm inflation (0.3 points per year). These adjustments are reported in the FC column in Table I.

  4. Thanks for clarifying- in retrospect, it was a bit doubtful you’d make an error as big as that for the largest IQ test out of the whole country, but you made no mention of the meaning of “Flynn Correction” in this post, and I couldn’t see any other explanation for the discrepancy. You should clarify that in the main post.

    Also, while you say that none of the studies mention race, making it impossible to ascertain the differences between chinese and ethnic thais, maybe you can look up the demography of the provinces and see how well they correlate with IQ?

    • Jason Malloy

      July 20, 2014 at 5:39 pm

      As per your recommendation, I have added interactive footnotes to Table I which clarify my methods.

      I actually did look for Chinese representation by province while writing this post to perform the exact analysis you suggest. But the closest data I could find was a map from the 1930s (lowest representation in the North and Northeast), and it seemed like a weak method so I moved on.

  5. “Leaders and scientists in Thailand are far more overtly concerned and curious about their national intelligence than their counterparts in any other country, so at least the future promises us more data to help unlock the mysteries of Thai intelligence.”

    To add to this, while it’s a virtual certainty the chinese minority has a higher IQ than ethnic thais, nearly all of the names on these studies are thai, so that might speak in favor of their real IQ as well.

    • From wiki: “Almost all Sino-Thais, especially those who came to Thailand before 1920s, possess a Thai surname, as was required by King Rama VI in order for them to become Thai citizens. The few who retain native Chinese surnames are either recent immigrants or resident aliens.”

      • Well, I stand corrected. But it’s not a certainty the authors are mostly chinese either. This could all be better resolved if the studies bothered to differentiate between ethnic thais and chinese.

        Is it really that hard to gauge the chinese makeup of the various regions? If the Bangkok metropolitan region contains a full 22% of the country’s population, then it’s likely a majority of the chinese population is contained there. And if that’s the case, then outside of that region, IQ isn’t very strongly correlated with the proportion of the population which is ethnic Thai. Wikipedia also claims, however, that atleast 40% of the Thai population is atleast partially chinese, and also claims this: http://en.wikipedia.org/wiki/Thai_chinese#History

        “However, early Chinese immigration consisted almost entirely of Chinese men who married Thai women. Children of such intermarriages were aptly called Sino-Thai[16] or known as Luk-jin (ลูกจีน) in Thai.[17] This tradition of Chinese-Thai intermarriage declined when large numbers of Chinese women began to emigrate into Thailand in the early 20th century.”

        So is this much of the chinese ancestry in ethnic thais that recent? That could easily further complicate things.

  6. You say Lynn and Vanhanen reported some numbers inaccurately…maybe they did, but keep in mind that they adjusted the numbers (adjusting scores if the norms were old..adjusting scores if the the test wasn’t normed on an exclusively European ancestry sample, etc) so the scores they report are often different from those listed in the references they cite, causing many people to falsely assume they’re inaccurate.

    • Jason Malloy

      July 22, 2014 at 11:13 am

      “adjusting scores if the norms were old..adjusting scores if the the test wasn’t normed on an exclusively European ancestry sample, etc”

      I know what corrections Lynn makes to his data. These are errors.

      Pollitt et al, 1989: sample size and sample age reported incorrectly (Lynn+Vanhanen 2006;2012 also report the test as the SPM. This is also incorrect; the study used the CPM).

      Sungthong et al, 2002: report an IQ of 78. Lynn reports the IQ as 88. A Flynn Correction pushes the estimate down to 75 (I don’t currently adjust norms for racial demographics, but this would also push the estimate down, not up). So Lynn is off by nearly 1 SD.

      • When you publish as much data as Lynn has, it’s inevitable that a few typos and miscalculations occur, but in my humble opinion, these errors are likely random & thus cancel out in his aggregate statistics.

  7. Jason Malloy

    July 23, 2014 at 12:50 am

    “it’s inevitable that a few typos and miscalculations occur”

    Not a few. I’ve found errors in most of the studies I’ve looked at. He’s incredibly careless, and I’m sorry but that matters. Future analyses will determine if the errors are random or systematic.

  8. You seem to have been much more positive towards Lynn in the past. Have you really changed your views on him? What do you honestly think of him at this point? What has made you notice all these errors you haven’t before?

    • Jason Malloy

      July 23, 2014 at 1:37 am

      “What has made you notice all these errors you haven’t before?”

      Analyzing the data for this website. My attitudes have developed and will continue to develop based on what I find.

      (As far as Lynn goes, I think I was first most appalled by that list of fake numbers from a hoax website that he submitted to P&ID in 2012. It took me 15 minutes to discover that that dataset was a fraud. Every number I looked at was fake. Lynn didn’t bother checking one study, and apparently the peer reviewers couldn’t be bothered either. I can’t believe that that paper still hasn’t been retracted.)

  9. Wow, that’s really bad. I’ve been used to shoddy scholarship from Lynn, but that’s really a whole different level. Have you brought this to the attention of P&ID?

    To me, as I said in a comment on your post about the Dominican Republic, one of the most contemptible (and very little remarked upon) cases of Lynn’s scholarship has been his praise of “Erectus Walks Amongst Us.” Admittedly, that book does raise some important questions about the evolution and variation in intelligence across human evolution and the lack of consideration for it, along with the politicization of OOA, and some of it’s ideas about archaic admixture have been confirmed in the past few years (whether it be neanderthals, denisovans or the fact archaic admixture in SS africans is one of the reasons for their remarkable levels of genetic diversity), but, for the most part, the book is an overwhelming load of garbage. It reads almost like a parody at times of certain themes in the study of racial differences, like the absurd lengths it goes to with correlating climate with intelligence differences or modern racial characteristics with far-flung, ancient hominid admixture (it is probably the first book within close to a century that argues khoisans look asian because of asian admixture, specifically ancient asian hominids), and the fact it variously cites “March of the Titans” is another point against it.

    If you ever have the stomach for it, give it a read, and try to bare in mind Lynn agrees with “99% of it” and what it says about him that he’d praise such an odious, illucid book, and when it’s also fairly clear the author, Fuerle, is a complete lunatic.

    • As an addendum, I meant to say “some of it’s ideas about archaic admixture have been *somewhat* confirmed”, but then again, pure OOA was never a universal belief, nor was especially likely to hold. Fuerle, for example, argued that africa’s remarkably high genetic diversity is entirely due to the fact africa has seen many immigrations of archaics into africa over the millenia, whereas the recent discovery of archaic admixture accounts for part of it, but hardly all.

  10. Charles Murray lived in Thailand for several years with the Peace Corps and he described his first wife as something like “half-Thai and all Chinese.” I bet he’d find this post interesting.

  11. Jason Malloy

    August 18, 2014 at 1:23 am

    No, I reported IQ data for the full sample from Table 1 (N=70). The underweight and matched subsamples in Table 2 (N=50) were selected out of this group. As were the riboflavin deficient and matched subsamples (N=20).

  12. Jason Malloy

    March 8, 2015 at 9:19 am

    1. Sangtongluan (2004) reports DAP IQ scores from local norms (p. 5). I would need access to the cited norms to make a proper conversion, but obtaining foreign test norms is difficult.

    2. The Pongcharoen et al (2011) test scores are reported separately in Table 1. If averaging over disparate tests is problematic, then national estimates from every post will require revision. (… and, of course, Lynn’s books and the research based on them will require these revisions as well.)

    3. I mention Lynn’s Opper citation in my first footnote. I have not included Piaget tests in my own reviews for several reasons, although I am not strictly opposed to their inclusion in this kind of dataset. There are a number of other Piaget studies for Thailand. The linked Surinya study, like many others that I excluded, does not report sufficient data to calculate an IQ score.

    • 1. Sangtongluan cites a study by one Titiwan Foo-trakoon, which gives Thai norms on the DAP. I assume you have found this study impossible to obtain?

      2. What’s problematic is averaging for a single group of testees, not averaging several distinct tests. For example, the Wechsler FSIQ isn’t equivalent to the average of the subtest indices.

      3. What are your reasons for rejecting Piagetian studies? I would suggest that, even if you do not want to include Piagetian studies in your review, you should upload the .pdfs of Piagetian studies in your archive so other researchers can utilize them.

  13. I agree that it’s difficult to make sense of Thailand’s statistics. What I don’t see however is the research of representativeness and attempts to exclude high scoring samples like it is often done in other parts of the world.

    In search of a representative adult sample, I’m only left with Phattharayuttawat et al, 2000’s mean 89.5 IQ.

  14. First of all thank the author immeasurably for the IQ literature in Thailand.

    I am studying children with Down syndrome now. I fear that beyond test writer bias (which you mentioned with reference to the demographics), there is clearly a problem in the narrow understanding of intelligence quotient int he Stanford – Binet (SB). It measures only 5 factors of cognitive ability: Fluid Reasoning, Knowledge, Quantitative Reasoning, Visual-Spatial Processing, and Working Memory. However, I have not seen any standardized multiple intelligence testing.

    Any comment?

    • There is no standardized test of multiple intelligences because Gardner’s theory is false (as are other similar conceptions). See here for a discussion of the largely unidimensional character of human intelligence.

  15. Good work. But your PISA data seems to be way off.

    Thailand scored 420 on PISA.
    That’s 100 points behind US whites (one standard deviation) who score 520 PISA.
    And will come as 85 IQ on US norms.

    Lastly, you are calculating PISA IQ on OECD average which is not the standard 100 IQ set by scientific community.

    Whites in USA and UK are two standard norms for IQ.

    Most OECD nations score much below US white norms like Mexico, Turkey, Israel, Chile, etc.

Leave a Reply

Your email address will not be published.

*

© 2017

Theme by Anders NorenUp ↑