Discover more from Just Emil Kirkegaard Things
Transracial adoption study: Bruce et al (2009)
This is a post in the on-going series of comments on studies of international/transracial adoption. A global genetic/hereditarian model of cognitive differences and their socioeconomic effects implies that adoptees from different populations/countries/regions should show the usual group differences in the above mentioned traits and outcomes, all else equal. All else is of course not equal since adoptees from different regions can be adopted at different ages, experience different environment leading up to the adoption, possibly experience different environments after adoption thru no cause of their own (discrimination), and so on. It is not a strict test: finding the usual group differences can be explained by non-genetic factors, and finding no differences or unexpected ones could be consistent with a genetic model given strong non-genetic effects such as differences in adoption practices between origin countries/regions/populations. However, were such differences to be found relatively consistently in sufficiently powered studies, it would be an important prediction verified by the genetic model, broadly speaking. The question thus remains: what do the studies show?
Bruce et al (2009) - Disinhibited Social Behavior Among Internationally Adopted Children
Their abstract reads:
Postinstitutionalized children frequently demonstrate persistent socioemotional difficulties. For example, some postinstitutionalized children display an unusual lack of social reserve with unfamiliar adults. This behavior, which has been referred to as indiscriminate friendliness, disinhibited attachment behavior, and disinhibited social behavior, was examined by comparing children internationally adopted from institutional care to children internationally adopted from foster care and children raised by their biological families. Etiological factors and behavioral correlates were also investigated. Both groups of adopted children displayed more disinhibited social behavior than the nonadopted children. Of the etiological factors examined, only the length of time in institutional care was related to disinhibited social behavior. Disinhibited social behavior was not significantly correlated with general cognitive ability, attachment-related behaviors, or basic emotion abilities. However, this behavior was negatively associated with inhibitory control abilities even after controlling for the length of time in institutional care. These results suggest that disinhibited social behavior might reflect underlying deficits in inhibitory control.
While this does not seem immediately relevant, the authors do investigate IQ. The study design is a three-way comparison between adoptees in foster homes, institutional care and non-adoptees. The sample are small: 40 x 40 x 40. These are children who spent most of their lives first in foster homes and were then adopted, children who spent most of their lives in institutional care and then adopted, and biological children of the adoptive families for comparison. The groups were not equal in origin composition:
However, because countries have institutional or foster systems to care for wards of the state, the institutional care and foster care groups differed in terms of country of origin. The institutional care group was primarily from Eastern Europe (45%) and China (43%), whereas the foster care group was primarily from South Korea (80%).
Their cognitive measure is:
General cognitive ability.To provide an estimate of the children's general cognitive functioning, each child was administered the vocabulary and block design subtests of the Wechsler Intelligence Scale for Children, 3rd edition (Wechsler, 1991). These subtests are considered the best measures of verbal and nonverbal intelligence, respectively, and are highly correlated with the full scale intelligence quotient (Sattler, 1992). Raw scores on the subtests were converted into age-normed scaled scores. The scaled scores were then summed and transformed into full scale intelligence quotient equivalents.
They give the mean FSIQ by the above three groups, not origin groups. These are: Group Institutional care Foster care Control FSIQ 102.68 (16.25) 109.37 (12.93) 117.11 (15.88) Note: numbers in parentheses are SDs.
The high scores is presumably due to the FLynn effect. The IQ test is from 1991, but the study is from 2009, so there has been 18 years for raw score gains compared to the normative sample. An alternative idea is that the families were just above average SES/IQ which boosts the IQ scores of younger children. Nearly 100% of the adoptive families were Caucasian (presumably European) and were part of the Minnesota International Adoption Project Registry. According to Wiki, the non-Hispanic % of this state is 83%, so Caucasians are a bit overrepresented. In general, these elevation effects are not important when comparing groups within a study.
I contacted the first author to ask if she would give me some more data, and she obliged:
Please find the requested information below. For each region of origin, I provided the number of children, mean, and standard deviation for the Block Design Standard Score, Vocabulary Standard Score, and full scale IQ equivalent. I did not provide these statistics for regions with less than 5 children. Please let me know if you have any questions
Origin N FSIQ Vocabulary Block design Institutional care Eastern Europe (e.g., Russia, Romania, Ukraine) 18 94.84 (15.070) 8.39 (2.615) 8.39 (2.615) China 17 111.77 (13.774) 11.76 (2.969) 12.29 (3.274) Foster care South Korea 31/32 110.38 (12.457) 12.10 (2.970) 11.53 (2.940) Guatemala 5 104.64 (16.609) 12.40 (3.050) 9.20 (3.033)
Notes: Sample size for S. Korea was 31, 31, 32. No explanation given. Numbers in parentheses are SDs.
Again we see that East Asians do well, altho not better than the control children (mean=117). Eastern European do less well, but it is hard to say the exact expected mean since it is not stated how many come from which countries. The IQ of Lynn and Vanhanen (2012) gives 91 as the IQ of Romania, 96.6 for Russia and 94.3 for Ukraine. The Guatemalans certainly do better than expected (79 IQ), but N=5, and the demographics of Guatemala is very mixed making it possible that the adoptees actually had predominantly European ancestry.
The email reply
Hello Emil, Please find the requested information below. For each region of origin, I provided the number of children, mean, and standard deviation for the Block Design Standard Score, Vocabulary Standard Score, and full scale IQ equivalent. I did not provide these statistics for regions with less than 5 children. Please let me know if you have any questions Thank you for your interest in our publication, Jackie
Institutional care group: Eastern Europe (e.g., Russia, Romania, Ukraine) N Mean Std. Deviation Block Design Standard Score 18 9.83 3.884 Vocabulary Standard Score 18 8.39 2.615 IQ equivalent 18 94.84 15.070
China N Mean Std. Deviation Block Design Standard Score 17 12.29 3.274 Vocabulary Standard Score 17 11.76 2.969 IQ equivalent 17 111.77 13.774
Foster care group: South Korea N Mean Std. Deviation Block Design Standard Score 32 11.53 2.940 Vocabulary Standard Score 31 12.10 2.970 IQ equivalent 31 110.38 12.457
Guatemala N Mean Std. Deviation Block Design Standard Score 5 9.20 3.033 Vocabulary Standard Score 5 12.40 3.050 IQ equivalent 5 104.64 16.609
References Bruce, J., Tarullo, A. R., & Gunnar, M. R. (2009). Disinhibited social behavior among internationally adopted children. Development and psychopathology, 21(01), 157-171.