iq test performance of black children adopted by white ... · pdf fileiq test performance of...

IQ Test Performance of Black ChildrenAdopted by White Families

SANDRA SCARR University of MinnesotaRICHARD A. WEINBERG University of Minnesota

ABSTRACT: The poor performance of black childrenon IQ tests and in school has been hypothesized toarise from (a) genetic racial differences or (6) cultural/environmental disadvantages. To separate genetic fac-tors from rearing conditions, 130 black/interracial chil-dren adopted by advantaged white families were studied.The socially classified black adoptees, whose naturalparents were educationally average, scored above theIQ and the school achievement mean of the white popu-lation. Biological children of the adoptive parentsscored even higher. Genetic and environmental de-terminants of differences among the black/interracialadoptees were largely confounded. The high IQ scoresof the socially classified black adoptees indicate malle-ability for IQ under rearing conditions that are relevantto the tests and the schools.

It is well known that black children reared by theirown families achieve IQ scores that average abouta standard deviation ( IS points) below whites(Jensen, 1973; Loehlin, Lindzey, & Spuhler, 197S).This finding is at the heart of a continuing contro-versy in the educational arena. Recent studies(Cleary, Humphreys, Kendrick, & Wesman, 1975)confirm the hypothesis that low IQ scores predictpoor school performance, regardless of race. Thus,more black children than white children fail toachieve academically and to earn the credentialsrequired by higher occupational status, with its

The present study was supported by the Grant Founda-tion and the National Institute of Child Health and Hu-man Development (HD-08016).

This study was conducted with the full collaboration ofthe Minnesota State Department of Public Welfare, Adop-tion Unit, directed by Ruth Weidell and assisted byMarjorie Flowers. Their help was invaluable. The addi-tional support of the Open Door Society, Lutheran SocialService, and The Children's Home Society, all of Minne-sota, facilitated the study.

We are very grateful for the assistance of Louise Carter-Saltzman, Harold Grotevant, Margaret Getman, MarshaSargrad, Patricia Webber, Joanne Bergman, WilliamThompson, and CaroJ Nelson.

Requests for reprints should be sent to Sandra Scarr,Institute of Child Development, University of Minnesota,Minneapolis, Minnesota 55455.

concomitant social prestige and economic security(Husen, 1974; Jencks, 1972).

In an attempt to remedy the alarming rate ofschool failure, compensatory educational programs,which were directed particularly at black children,were introduced in the 1960s. At the same time,but for different reasons, a more intensive 'inter-vention began: the adoption of black children bywhite families. Whereas compensatory educationalprograms involve the child for a few hours per day,transracial adoption alters the entire social ecologyof the child. Parents, siblings, home, peers, school,neighborhood, and community—the child's rearingenvironment—are transformed by adoption.

The existence of transracial families offers muchto the scientific study of social milieus and intel-lectual performance (Grow & Shapiro, 1974; Loeh-lin et al., 197S). Transracial adoption is the humananalog of the cross-fostering design, commonly usedin animal behavior genetics research (e.g., Mano-sevitz, Lindzey, & Thiessen, 1969). The study oftransracial adoption can yield estimates of bio-logical and sociocultural effects on the IQ test per-formance of cross-fostered children.

The results of a transracial or cross-fosteringstudy require careful interpretation. Black chil-dren reared in white homes are socially labeled asblack and therefore may suffer racial discrimination.Because of the unmeasured effects of racism, poorIQ test performance by black children in whitehomes cannot be uncritically interpreted as a resultof genetic limitations. In addition, equal perform-ance by black and other adoptees cannot be in-terpreted as an indication of the same range ofreaction for all groups. Again, the unknown effectsof racism may inhibit the intellectual developmentof the black adoptees. However, equally high IQsfor black and other adoptees would imply that IQperformance is considerably malleable.

Upper-middle-class white families have an ex-cellent reputation for rearing children who perform

726 • OCTOBER 1976 • AMERICAN PSYCHOLOGIST

well on IQ tests and in school. When such familiesadopt white children, the adoptees have been foundto score above average on IQ tests, but not ashighly as .the biological offspring of the same andsimilar families (Burks, 1.928; Freeman, Holzinger,& Mitchell, 1928; Leahy, 1935; Munsinger, 197Sb;Skodak & Skeels, 1949). How do the IQ testscores of black children adopted by white familiescompare to the scores of both white adoptees andthe biological children"of these parents?

If black children have genetically limited in-tellectual potential, as some have claimed (Jensen,1973; Shockley, 1971, 1972), their IQ performancewill fall below that of other children reared inwhite upper-middle-class homes. On the otherhand, if black children have a range of reactionsimilar to other adoptees, their IQ scores shouldhave a similar distribution. The concept, range ofreaction, refers to the fact that genotypes do notusually specify a single phenotype. Rather, geno-types specify a range of phenotypic responses thatthe organism can make to a variety of environ-mental conditions.

This is an investigation of the IQ test perform-ance of black and interracial black children adoptedby white families in Minnesota. The present studyis part of a larger investigation of the psychosocialfunctioning of transracial adoptive families. In-tellectual, personality, and attitudinal tests wereadministered to the parents and all children overthe age of 4 years. Extensive interviews were con-ducted with the parents, and ratings of the homeenvironment were made.

Minnesota has been in the forefront of inter-racial adoption. Although the black population ofthe state is small (.9% in 1970), there were toomany black and interracial children available foradoption and too few black families to absorb them.Minority group children—black, American Indian,Korean, and Vietnamese—have consequently beenadopted by white families in large numbers. Fur-thermore, in recent years, many nonwhite childrenhave been adopted from other states.

The climate for interracial adoption changeddramatically in the late 1950s and early 1960s be-cause of the efforts of public and private agenciesand the pioneering white adoptive parents. Sev-eral agency and parent organizations were formedto promote the adoption of black and interracialblack children. The most influential, continuingorganization is the Open Door Society of Minne-sota, formed in 1966 by adoptive parents of so-cially classified black children. The founding presi-

dent of the Open Door Society is a leadingcolumnist on one of the Minneapolis daily news-papers who frequently writes about his multiracialfamily. The intellectual and social climate ofMinnesota is generally conducive to liberal andhumanitarian movements such as interracial adop-tion.

Goals of the Study

We posed five major questions in the study:

1. What is the estimated reaction range for IQscores of black/interracial children reared in typicalblack environments or in white adoptive homes?

2. Do interracial children (with one black andone white parent) perform at higher levels on IQtests than do children with two black parents; thatis, does the degree of white ancestry affect IQscores?

3. How do the IQ scores of socially classifiedblack children reared in white homes compare tothose of other adopted children and biological whitechildren within the same families; that is, do dif-ferent racial groups, when exposed to similar en-vironments, have similar distributions of IQ scores?

4. How well do socially classified black childrenreared in white families perform in school?

5. How accurately can we predict the IQ testperformance of adopted children from the educa-tional characteristics of their natural parents, fromthe educational, intellectual, and other character-istics of their adoptive homes, and from their place-ment histories?

The Families

The 101 participating families were recruitedthrough the Newsletter of the Open Door Societyand by letters from the State Department of PublicWelfare Adoption Unit to families with blackadopted children, 4 years of age and older, whowere adopted throughout the state of Minnesotathrough Lutheran Social Service and Children'sHome Society. These agencies have placed themajority of black and interracial children in thestate. We were unable to ascertain how manytransracial adoptive families learned about thestudy from the Newsletter, because the mailing listof about 300 includes agencies, social workers, andinterested citizens. In addition, we do not knowhow many of these families were also contacted bythe State Department of Public Welfare. The SUP-

AMERICAN PSYCHOLOGIST • OCTOBER 1976 • 727

TABLE 1

Recruitment of Families

Method

Department of Public Welfare letters

Not eligible to participate 46°Unknown

Letter undelivered 43No response 41

EligibleNot participating

In another study 3Don't approve of study 2Child appears white 3Personal reasons 3No reason given 3Live too far away 10Yes, but changed their minds 6

Participating 68Total letters sent 228

Open Door Society

Not eligible to participate 19s

EligibleNot participating

Live too far away 4Yes, but changed their minds 1

Participating 33Total responses 57

ft Most because their black children were under 4 years of age.

port of the Open Door Society was important, how-ever, in affirming the legitimacy of the study.

The State Department of Public Welfare mailed228 letters to transracial adoptive families. Insome cases a family received more than one letterif they had adopted more than one child. Table 1describes the results of the mailing. Of the 136families known to be eligible for participation in thestudy, 74% did participate.

The 101 participating families included 321 chil-dren 4 years of age and older: 145 biological chil-dren (81 males, 64 females) and 176 adopted chil-dren (101 males, 75 females), of whom 130 aresocially classified as black and 25 as white. Theremaining 21 included Asian, North AmericanIndian, and Latin American Indian children.

All of the adopted children were unrelated to theadoptive parents. Adopted children reared in thesame home were unrelated, with the exception offour sibling pairs and one triad adopted by thesame families.

The sample of families live within a 150-mileradius of the Twin Cities (Minneapolis-St. Paul)metropolitan area. Although nearly all of the chil-

dren were adopted in Minnesota, 68 were born out-side of the state. Through interstate cooperation,the child placement agencies arranged for the adop-tion of many nonwhite children from other states.Table 2 gives the out-of-state origins of the sample.

Procedures

Most of the information was obtained directly from mem-bers of the adoptive families. Some additional data on thenatural parents and the children's preadoption history wereobtained by State Department of Public Welfare personnelfrom the adoption records. Achievement and aptitude testscores were supplied by school districts for all of the school-aged children to whom such tests had been administered.

THE IQ ASSESSMENT

Both parents and all children in the family over 4 yearsof age were administered an age-appropriate IQ test aspart of an extensive battery of intellectual, personality, at-titudinal, and demographic measures. Children under 4years of age were excluded because IQ tests are less predic-tive of later IQ at younger ages. By 4, the correlation ofIQ with adolescent scores is about .7. The tests were ad-ministered in the family home during two visits by a teamof trained testers. The examiners were all graduate stu-dents who had completed at least a year-long course inpsychoeducational assessment and who had participated ina training session on assessment for this study. Among the21 examiners were 6 males and 15 females, including 2blacks. Testers were assigned randomly to members ofthe family.

TABLE 2

Out-of-State Origins of the Adopted Children

Origin n

Other adopted

KoreaVietnamCanada (Indian)Ecuador (Indian)

Black and interracial adopted

IllinoisIowaKentuckyMassachusettsNew YorkNorth DakotaOhioTexasUtahWashingtonWisconsin

White adopted

Massachusetts

Total

7152

419

11312212

16

1

68


Both parents and all children 16 years of age and olderwere administered the Wechsler Adult Intelligence Scale(WAIS; Wechsler, 1955). Children between 8 and 15 weregiven the Wechsler Intelligence Scale for Children (WISC;Wechsler, 1949), and children between 4 and 7 were ad-ministered the Stanford-Binet Intelligence Scale, Form L-M(Terman & Merrill, 1972).

All scoring of protocols and computations of IQ scoreswere done by a graduate student with extensive experiencein administering and scoring IQ measures. This studenthad no contact with the families and with the examinersexcept to clarify questionable responses. In no case wasthe scorer aware of the child's race or adoptive status.

THE ADOPTION RECORDS

The Director of the Adoption Unit, State Department ofPublic Welfare, abstracted the following information fromthe records of the adopted children and their families:

1. The child: (a) birthdate; (b) number and dates ofpreadoption placements, unless the child was in the adop-tive home at 2 months of age; (c) evaluation of the qualityof preadoption placements, rated by the authors on a scaleof 1 = poor to 3 = good; 4 = placement only in the adop-tive home; (d) date of placement in adoptive home.

2. The natural parents: (a) age at birth of child; (b)educational level at birth of child as an estimate of intel-lectual functioning, since IQ scores were not available; (c)occupation of mother; (d) race.

The race of the two natural parents was used to classifytheir child's race. If a child had one or two black parents,he was considered socially black.

FAMILY DEMOGRAPHICS

As part of the interview portion of the testing session, eachparent was asked his or her birthdate, last school gradecompleted, occupation and whether it was full time or parttime, range of income, and date of marriage. Occupationswere coded for prestige using the scale developed from theNational Opinion Research Center (NORC) survey (Reiss,1961).

THE SCHOOL DATA

With parental consent, forms requesting recent aptitudeand achievement test scores were mailed to the schools ofall school-aged children participating in the study; 100%of the forms were returned. Because school districts usea variety of tests,1 comparable scores were combined acrosstests. For aptitude tests, a total score was generated. Forachievement tests, a vocabulary, a composite reading, andan arithmetic score were used.

Results

Since the major focus of the study was to estimatethe level of IQ performance of the black adopteesand to account for that performance level, the

TABLE 3

Demographic Characteristics of the Adoptiveand Natural Parents

Characteristic

Income

Adoptive

n M

100 $15,000-17,500

SD

$5,000

Range

$5,000-> $35,000

Education

Adoptive fatherAdoptive motherNatural fatherNatural mother0

10110146

135

16.915.112.112.0

3.02.22.02.2

9-2212-218-176-18

1 Eight aptitude and 11 achievement tests were used bythe various school districts. A list is available from theauthors (see author note for address).

Age

Adoptive fathers'- 100 37.3 6.7 28-59Adoptive mothers'1 100 35.5 5.8 26-53Natural fathers0 55 26.3 6.6 16-44Natural mothers0 150 21.6 5.3 12-40

a If the 40 students are excluded, the mean is the same.h Current.• At birth of child.

nature and quality of the children's adoptive ex-perience were examined.

FAMILY CHARACTERISTICS

The adoptive families who participated in the studycan be characterized as highly educated and aboveaverage in occupational status and income. Table3 is a summary of selected demographic character-istics of the adoptive and natural parents.

The educational level of the adoptive parentsexceeded that of the adopted children's naturalparents by 4-5 years. The typical occupations ofthe adoptive fathers were clergyman, engineer, andteacher. Nearly half (46.5%) of the adoptivemothers were employed at least part time, typicallyas teachers, nurses, and secretaries. The mean edu-cational level of the natural parents was high schoolgraduation, which is close to the median for thatage cohort of the general population. Actually,the black mothers had one year less education thanthe black females in their age group (25-44).Fathers of the early-adopted black children hadslightly more. Table 4 shows the average educa-tional level of the white mothers of interracialblack children, the black mothers, and the blackfathers, compared to local and regional norms.(Because there were only two white fathers of in-terracial children, they have been omitted from thetable.) In contrast, the mean educational level ofthe adoptive parents was atypically high. Typical


TABLE 4

Educational Levels of the Natural Parents ofAdopted Children, Compared to Their Populations

Black mothersWhite mothers

of interracialchildren

Black fathers

Naturalparentsof the

adoptedchildren

10.8

12.412.3

Naturalparentsof theearly-

adoptedchildren

10.8

12.612.6

NorthCentralregion

11.9

12.512.0

Minneapolis-St. Paul"

12.0

12.512.0

Note. Levels given in years.ft Men or women, aged 25-44 years.

occupations of the natural mothers were officeworkers, nurse's aides, and students. Insufficientinformation was available on the occupations ofthe natural fathers.

PREADOPTIVE EXPERIENCE

Table 5 includes two measures of the children'spreadoptive placements: number and quality. Theinformation is presented for all adoptees and byrace.

TABLE 5

The Adopted Children

Preadoption

Number of placements"MSDRange

Quality of placements11'^MSDRange

Adoptive placement

Age of placement0

MSDRange

Time in adoptive home0

MSDRange

Current age°MSDRange

Alladopted

(» = 176)

1.061.040-6

3.17.63

1-4

22.4834.200-189

64.7033.508-199

87.1840,8048-257

White(» = 25)

.771.240-4

3.46.84

1-4

19.0432.800-94

104.2039.3022-187

123.2448.0069-257

Black/inter-racial

(« = 130)

1.02.93

0-6

3.18.50

2-4

17.9724.700-124

57.2525.508-199

74.2229.6048-201

Asian/Indian

(n = 21)

1.S71.120-4

2.50.73

1-4

60.7156.90

1-189

63.8138.209-137

124.5244.4052-218

• Information available for 156 children: 22 white; 120 black/inter-racial; 14 Asian/Indian.b Quality of placement was rated: 1 = poor to 3 = good; 4 = placedwhen less than 2 months old.0 In months.

Forty-four children were placed in their adoptivehomes by 2 months of age and were considered tohave had no previous placements. The remainingadopted children had from one to six previousplacements. Black children had a smaller numberof preadoption placements, and the quality of theirplacements was better than that of the Asian/Indian adoptees. Fewer black children were ininstitutions or were removed from homes for ne-glect or abuse, and more were in agency fosterhomes.

Only 18 of the 176 adopted children had everlived with their biological parents: 7 of the Asian/Indian adoptees, for an average of 85 months; 3 ofthe white children, for an average of 28 months;and 8 of the black children, for an average of 36months.

THE ADOPTIVE EXPERIENCE

As shown in Table 5, the average age of placementin the adoptive homes was 22 months, but themedian age of placement was 6 months. One hun-dred and eleven children, including 99 black andinterracial adoptees, were placed in their adoptivehomes during the first year of life. The Asian andIndian children were placed significantly later thaneither white or black children. The socially classi-fied black children, however, had lived with theiradoptive families for fewer years than the others,particularly than the white adoptees. Also shownin Table 5, black and interracial children werecurrently younger, on the average, than the others.

IQ SCORES OF ADOPTIVE PARENTS

As indicated in Table 6, the mean WAIS IQ scoresof the adoptive parents were in the high averageto superior range of intellectual functioning. Thedistribution of scores extends from the low averageto the very superior, with considerable restrictionof range. The scores were congruent with the veryhigh educational level of the group.

TABLE 6

WAIS IQ Scores of Adoptive Parents

Mother

WAIS M SD Range n M

Father

SD Range

VerbalPerformanceFull Scale

1009999

118.3115.9118.2

10.411.410.1

92-14486-14396-143

999999

120.7118.2120.8

10.610.910.0

92-14091-14993-140


TABLE 7

IQ Scores of the Natural Children of the Adoptive Parents

Total

Scale

Stanford-Binetwise


WAISVerbalPerformanceFull Scale

n

48

828282

141414

M

113.8

113.5119.5117.9

117.5117.7118.9

SD

16.7

13.114.912.7

11.010.811.2

Range

81-148

84-14768-14787-150

100-139103-137101-141

n

26

505050

555

Males

M

111.6

114.0120.5118.5

121.6121.8123.0

SD

16,5

12.812.510.8

13.914.614.9

Range

81-148

89-14782-14396-145

103-139104-137104-141

n

22

323232

999

Females

M

116.3

112.8117.8117.0

115.2115.4116.6

SD

16.9

13.618.115.3

9.28.18.6

Range

88-140

84-14468-14787-150

100-125103-125101-126

Note. WISC « Wechsler Intelligence Scale for Children. WAIS = Wechsler Adult Intelligence Scale.

IQ SCORES OF THE NATURAL CHILDREN

OF THE ADOPTIVE PARENTS

The mean IQ scores of the natural children of theadoptive families were in the high average tosuperior range of intellectual functioning. As ex-pected from polygenic theory, when both parentshave high IQ scores, there is less regression towardthe population mean than under conditions of ran-dom mating. Table 7 gives the Stanford-Binet,WISC, and WAIS results for the natural children.Only the Wechsler scores had a restricted range.With tests combined, the total IQ score of the na-tural children averaged 11,6.7 with a standard devi-ation of 14.0.

THE IQ SCORES OF ADOPTED CHILDREN

The mean IQ scores of the adopted children werein the average range. As shown in Table 8, thescores on the three IQ tests, although for childrenat different age levels, were highly comparable.The adopted children did not perform as well as

TABLE 8

IQ Scores of Adopted Children

either the adoptive parents or their biologicalchildren.

For all of the groups of children, the Stanford-Binet (1972 norms) yielded a slightly lower meanscore than did the WISC or WAIS. Had the 1960Stanford-Binet norms been used, the average IQscores of the children would have been 7 pointshigher.

IQ SCORES OF ADOPTED CHILDREN BY RACE

Although adopted children of various ages wereadministered different tests, their performance wassufficiently comparable that we could combine theIQ scores across the three tests. Table 9 givesthe mean IQ scores by race.

Although all groups had comparable ranges andwere performing within the average range of in-tellectual functioning, the black and interracialchildren scored, on the average, between the whiteand Asian/Indian adopted groups. The scores ofthe socially classified black and white groups weresignificantly above the mean of the general popula-

Total

Scale

Stanford-BinetWISC


WAISVerbalPerformanceFull Scale

n

122

484848

666

M

106.5

101.2109.7105.8

98.3113.5105.2

SD

13.9

15.617.716.1

7.06.56.3

Range

68-144

66-14262-14364-140

86-107107-11994-113

n

69

303030

333

M

107.1

101.9111.0106.9

95.3113.0102.7

Males

SD

12.6

14.418.315.8

8.74.97.8

Females

Range

80-144

71-13962-14364-140

86-107107-11994-113

n

53

181818

333

M

105.6

100.2107.5104.1

101.3114.0107.7

50

15.5

17.917.016.8

2.19.52.9

Range

68-136

66-14280-14280-133

99-103108-125106-111

Note. WISC - Wechsler Intelligence Scale for Children. WAIS - Wechsler Adult Intelligence Scale.


TABLE 9

IQ Scores for Adopted Children by Race,with Tests Combined

IQ scores

Children SD Range

All adoptedBlack and interracialWhiteAsian/Indian

Early-adoptedBlack and interracialWhiteAsian/Indian"

130 106.3 13.9 68-14425 111.5 16.1 62-14321 99.9 13.3 66-129

999

110.4116.8

11.2 86-13613.4 99-138

« Only 3 cases.

tion. The Asian/Indian adopted children scoredexactly at the population mean. The means of thethree groups of adopted children differ significantly(p < .005). The children adopted during the firstyear of life scored higher than those adopted afterthe first year. The average score for the 111 early-adopted group was an IQ of 111; for the 65 lateradoptees, the mean IQ score was 97.5.

For those who hypothesize that blacks have lowerIQ scores than whites because of their African an-cestry, we compared socially classified black chil-dren with one versus two black natural parents.On the average, children with two black parentshave a higher degree of African ancestry than thosewith one black and one white parent. Table 10compares the IQ scores, placement histories, andnatural-parent education of children with one ortwo black parents. Socially classified black chil-dren with one parent of unknown, Asian, Indian,

or other racial background have been eliminatedfrom this analysis.

The 29 children with two black parents achieveda mean IQ score of 96.8. The 68 with only oneblack parent scored on the average 109.0. It isessential to note, however, that the groups alsodiffered significantly (p < .05) in their placementhistories and natural mother's education. Childrenwith two black parents were significantly older atadoption, had been in the adoptive home a shortertime, and had experienced a greater number ofpreadoption placements. The natural parents ofthe black/black group also averaged a year lessof education than those of the black/white group,which suggests an average difference between thegroups in intellectual ability. There were alsosignificant differences between the adoptive familiesof black/black and black/white children in father'seducation and mother's IQ. One can see in Table10 that the children with two black parents hadpoorer histories and had natural and adoptive par-ents with lower educational levels and abilities. Itwill be shown in the section on IQ variance thatthese characteristics largely account for the IQdifferences between black children with one or twoblack parents.

EXPECTANCY EFFECT

It is possible, though not likely, that the adoptiveparents' belief about the child's racial backgroundcould influence the child's intellectual development.If parents expected interracial children to scorehigher than children with two black parents, therecould be an expectancy effect. Twelve interracial

TABLE 10

Comparison of Adopted Children with One or Two Black Natural Parents

Black/black

Variable

IQAge at adoption11

Time in homeb

Quality of placementNumber of placementsNatural mother's educationNatural father's educationAdoptive father's educationAdoptive mother's educationAdoptive father's IQAdoptive mother's IQ

n

2929292727221529292928

M

96.832.342.22.91.2

10.912.116.514.9

119.5116.4

SD

12.833.114.3

.4

.71.91.42.72.3

10.37.5

Range

80-1301-1248-1202-40-36-14

10-1612-2112-20

106-137100-129

n

6868686464662068686668

Black/white"

Jl/

109.08.9

60.63.3

.812.412.517.215.3

121.4119.2

SD

11.511.217.4

.5

.91.82.22.82.0

10.110.5

Range

86-1360-52

33-1993-40-67-188-17

12-2111-2093-14096-143

" 66 black fathers, 2 black mothers.b In months.


children were believed by their adoptive parents tobe black/black. Only two black/black childrenwere believed to be interracial, and they have beenomitted from the analysis.

Interracial children believed to be the offspringof two black parents scored on the average at thesame level as interracial children correctly classifiedby their adoptive parents. The mean IQ score of43 correctly identified interracial children was 108.4(SD = 12.,6). The average IQ score of 12 inter-racial children believed to be black/black was108.6 (SD = 10.2), There was no evidence for anexpectancy effect.

THE CRITICISM OF SELF-SELECTION

Self-selection has been used to criticize the above-average IQ scores obtained in other adoptionstudies. Munsinger (197Sa) noted that obviouslyretarded and damaged infants are not likely to beadopted, a fact which raises the mean IQ of adop-tees above the population average. This bias isslight, however: If all infants with eventual IQscores of less than 60 (at most 3% of children)were eliminated from the adoption pool, the meanIQ of adoptees would be raised by only 1 IQ point.

Another bias could be the self-selection of fami-lies whose children appear normal in intelligenceand school work. The range of IQ scores in thisstudy contraindicates a strong bias in this regard,because IS of the 176 adopted children have IQscores of 85 and below. Furthermore, since 74%of those families known to be eligible did partici-pate and the average IQ of all adoptees was 106,the average IQ of children in the 26% of the fami-lies who did not participate would have to be un-reasonably low to explain mean results. If weconsider the sample to be composed entirely ofinterracial children, with white adoptees offsettingthose with two black parents, their average IQshould fall between those of black and white chil-dren in the region.

To lower the average adoptee's IQ to a hypo-thetical average of 95 for interracial children, thenonparticipants would have to have IQ scores thataverage 64, or in the retarded range. This is highlyunlikely for any sample of adopted children.

SCHOOL ACHIEVEMENT

The IQ assessments of the present study shouldbear a meaningful relationship to school achieve-ment. Slightly above average IQ test performance

TABLE 11

School Achievement Test Scores of Black/Interracial Adopted and Natural Children:Mean National Percentiles

Test -U 70ilc SD tt

Black adoptees

VocabularyReadingMathematicsAptitude (IQ)

57.2SS.O55.2

108.8

29. t28.629.95.9

2024195

Natural children

VocabularyReadingMathematicsAptitude (IQ)

73.174.571.3

119.6

11.725.822.611.7

48776939

should predict to slightly above average schoolachievement. The school data are also importantbecause they come from many different school dis-tricts and are uncontaminated by any biases thatmay have inadvertently influenced testing in ourstudy. Most importantly, they represent a "real-life" criterion of intellectual achievement.

Table 11 gives the mean national percentilescores for vocabulary, reading, and mathematicsachievement, and a total aptitude score expressedin IQ form, for the socially classified black adoptedand natural children of the adoptive families. Al-though the sample sizes were rather small, the blackchildren in school were performing slightly abovethe national norms on standard scholastic achieve-ment tests, just as their IQ scores would predict.The average IQ of the children with achievementtest scores was 104.9. The mean aptitude scoresof the 5 black adoptees who had been given school-administered group IQ tests were quite close totheir average scores on the WISC and Stanford-Binet. The correlation between aptitude and in-dividual IQ scores could not be calculated becauseof small sample size.

The natural children of the adoptive parentsscored higher than the adopted children on scho-lastic achievement tests,,as predicted by their indi-vidual IQ test scores. Furthermore, their group-tested IQ performance was also very close to theiraverage IQ as assessed in this study with individualtests. The correlation between the individual andgroup test scores of the 39 natural children was .78(p< .001).


TABLE 12

Correlations of Natural Parent Characteristics, Child's Adoptive Experience, Adoptive FamilyCharacteristics, and Child's IQ Scores for Black/Interracial Children

10 12 13

Natural parent characteristics1. Natural mother's race (117)2. Natural mother's education (107)"3. Natural father's education (37)

Adoptive experience4. Age at placement (130)5. Time in home (130)6. Number of placements (112)7. Quality of placements (112)

Adoptive family characteristics8. Adoptive mother's education (130)9. Adoptive father's education (130)

10. Adoptive father's occupation (129)11. Family income (129)12. Adoptive father's IQ (127)13. Adoptive mother's IQ (128)

Child's IQ14. Black adoptees (130)

.36 -.34 -.27-.45 .27 .37 -.31

.22 -.17 -.31 .50 -.21-.30 .26 .17 -.37 .15 -.65

-.10 .22 .12 -.10 .12 -.13 .02-.13 .26 .25 -.27 .26 -.14 .04 .56

.01 .07 .04 .00 .09 .01 -.05 .31 .29

.08 .16 -.06 .16 .12 -.04 -.06 .31 .04 .45-.01 .12 .33 -.19 .06 -.33 .08 .26 .47 .18 -.07-.18 .09 .29 -.01 .26 -.05 -.05 .53 .30 .21 .27 .21'

-.41 .31 .45 -.36 .30 -.36 .38 .22 .34 -.01 -.00 .18 .17

Note. Total N = 130. Numbers in parentheses are «s.• Students included.

Sources of Variance inBlack Adoptees' IQ Scores

The possible effects of the adoptive experience andof natural and adoptive family variables on IQscores were explored in correlational and regressionanalyses. To account both for the differences be-tween black/black and black/white children andfor the above-average performance of the blackadopted children on the IQ tests, we intercorrelatedtheir natural parents' education, natural mother'srace, their adoptive experience, adoptive familycharacteristics, and IQ scores. We were particu-larly concerned about the confounding of racialvariables with preadoptive and adoptive familyvariables that could affect the children's IQ per-formance. Selective placement of the children ofbetter educated (presumably brighter) naturalmothers with better educated adoptive families—asituation that creates genotype-environment cor-relations—also needed to be examined. The cor-relation matrix is presented in Table 12.2

2 The age of the child and the race and sex of the ex-aminer are omitted from the tables because they are uncor-related with the children's IQ scores (rs = .01, .06, and .01,respectively).

NATURAL PARENTS AND THECHILD'S ADOPTIVE EXPERIENCE

The educational and racial characteristics of thenatural mothers of the adopted children had agreat deal to do with when and by whom the chil-dren were adopted. Less well educated mothers,who were more often black, had children who wereplaced later for adoption, had spent less time inthe adoptive homes, and were adopted by familieswith lower educational and income levels. Thesame pattern held for natural fathers' education.(Since all but two of the known natural fatherswere black, father's race was omitted from theanalysis.)

The black children's IQ scores were significantlycorrelated with the same placement and adoptivefamily variables. Children who were adopted

In the correlation and regression analyses (Tables 12-14), natural mothers who were students at the time of thechild's birth were included. Of the 107 mothers of blackchildren for whom we had educational data, 34 werestudents in high school or college. Since the mean educa-tional level of the natural mothers, with and without thestudents, was the same and since the correlation of naturalmother's education and child's IQ was higher when studentswere included, we decided to present the tables based onthe larger »s.


TABLE 13

Two-Step Multiple Regression of Biological and Adoptive Family Variables on the IQ Scores ofBlack/Interracial Children, Adoptive Variables First

Step

1. Social variablesAdoptive mother's educationQuality of placementsAdoptive father's IQAdoptive father's occupationFamily incomeAdoptive mother's IQAge at placementAdoptive father's educationNumber of placementsTime in home

Multiple R

.22

.44

.45

.46

.46

.46

.53

.56

.56

.56

R'

.05

.19

.20

.21

.21,21.28.31.31.31

JJ! change

.05

.14

.01

.00

.00

.01

.07

.03

.00

.00

Simple r

.22

.38

.18-.01-.00

.17-.36

.34-.36

.30

P<

.001

2. Biological variablesNatural mother's education1

Natural mother's race.57.59

.32

.35.01.03

.31-.41

.001

• Students included; natural mother's education entered first to leave residual variance for race.

earlier, who had spent more years in the adoptivehomes, who had fewer preadoptive placements, andwho had better quality placements had higher IQscores. In addition, adopted black and interracialchildren who had better educated and higher-IQadoptive parents had higher IQs. Thus, there wasan important confounding of the characteristics ofthe natural parents, the preadoption experience,and the adoptive family, all of which affected thelevel of the black/interracial children's intellectualfunctioning.

SELECTIVE PLACEMENT

Selective placement further confuses the sourcesof variance in the black children's intellectual func-tioning. As Table 12 indicates, the natural mother'seducational level is correlated with the adoptiveparents' educational level, between .22 and .26,suggesting that the adoption agencies practicedselective placement, based on the educational in-formation they had available. The correlations ofnatural mother's education and adoptive parents'IQ scores are not as high (.09 and .12), presumablybecause the agencies did not have the IQ dataavailable. Selective placement increases the simi-larity between natural parents and their (adopted)children and between the adoptive family and theiradopted children.

The biological and social factors, many of whichseparately and together can affect IQ scores, werelargely confounded in the sample of black and in-terracial adoptees. Therefore, we did not attempt

to estimate point values for the genetic and environ-mental contributions to IQ differences. Instead,we decided to present two regression analyses.

When the biological variables were put into theregression first, we could find out how much of theremaining variance would be accounted for by thesocial variables. When the social variables wereput into the regression equation first, we could de-termine how much of the remaining variance wouldbe determined by the biological variables. Tables13 and 14 present the two regression analyses (seeFootnote 2).

In Table 13, the social variables, including place-ment and adoptive family measures, were steppedin first. The natural family data, called biologicalvariables, were entered second into the regressionequation. In Table 14, the biological variableswere entered first, the social variables second. Bothsteps were statistically significant in both tables.

When the social variables were entered first, theyaccounted for 31% of the total variance in the IQscores of socially classified black adopted children.The biological variables added 4% of the variancewithout natural father's education and 11% withfather's education. (Because the sample of blackchildren with natural father information was small,n = 37, a separate regression including only thosechildren was done. The results for the other vari-ables were very similar, and father's education ac-counted for an additional 1% of the total IQvariance.)

, When the biological variables were entered intothe regression analysis first, natural mother's edu-


TABLE 14

Two-Step Multiple Regression of Biological and Adoptive Family Variables on the IQ Scores ofBlack/Interracial Children, Biological Variables First

Step Multiple R R1 change Simple r P<

1. Biological variablesNatural mother's education"Natural mother's race

2. Social variablesAdoptive father's occupationAdoptive father's IQAdoptive mother's IQQuality of placementsAdoptive father's educationFamily incomeAdoptive mother's educationNumber of placementsAge at placementTime in home

.31

.44

.44

.47

.48

.54

.58

.58

.58

.59

.59

.59

.09

.20

.20

.22

.23

.29

.34

.34

.34

.35

.35

.35

.09

.10

.00

.03

.01

.06

.05

.00

.00

.01

.00

.00

.31-.41

-.01.18.17.38.34

-.00.22

-.36-.36

.30

.001

.001

a Students included; natural mother's education entered first to leave residual variance for race.

cation and race accounted for 20% of the variancein the black children's IQ scores. (Natural father'seducation added 11%, but the sample size was toosmall to include in the full analysis.) The socialvariables, stepped in second, added 15% of theIQ variance.

It is impossible to distinguish the effects of theseparate social and biological variables, because24.S of the 35% of the variance accounted for wasshared by the so-called biological and social vari-ables. Using part correlations, we found thatnatural mother's race and adopted father's educa-tion each contributed 3% to the variance of thesocially classified black adoptee's IQ scores, and thequality of the children's preadoptive placementscontributed 2%. The remaining 1.4% of theunique variance was contributed almost equallyby the other "biological" and "social" variables.

In the case of natural mother's race, it is un-warranted to conclude that race stands solely forgenetic differences between the races. In this sam-ple, natural mother's race was correlated with manymeasured social variables; it is conceivably corre-lated with other unmeasured social variables. Racedoes make a small contribution to the sociallyclassified black children's IQ variance, independentof the other measures, but not necessarily inde-pendent of other environmental variables.

Another consideration in the interpretation ofthe regression analyses is the restricted range ofvariation in adoptive family characteristics. Paren-tal education, IQ scores, income, occupationalstatus, and other unmeasured family variables, such

as child-rearing practices, varied over half or lessof their normal range in the general population.Thus, the adoptive family variables accounted forless of the IQ variance among black and interracialadoptees than they would in a more varied adoptivepopulation. The importance of the social variablesis very likely to be underestimated.

Discussion

This study attempted to answer five questionsabout the impact of transracial adoption on the IQperformance of black and interracial childrenadopted into white homes. The first question fo-cused on the reaction range of IQ scores within theblack population. Would socially classified blackchildren reared in economically advantaged whitehomes score above those reared in black environ-ments?

The average IQ score of black and interracialchildren, adopted by advantaged white families, wasfound to be 106. Early-adopted black and inter-racial children performed at an even higher level.This mean represents an increase of 1 standarddeviation above the average IQ of 90 usuallyachieved by black children reared in their ownhomes in the North Central region (Kaufman &Doppelt, in press). Furthermore, in the Minneapo-lis public school district, the average performanceof 4th-grade children on the Gates-MacGinitee vo-cabulary test at a school with 87% black andinterracial enrollment in 1973 was about the 21st


national percentile, which translates to an IQequivalent of about 90.

Since 68 of the 130 black children were knownto have one white parent and only 29 were knownto have had two black parents (the remainder wereof other mixed or unknown parentage), it may,seem misleading to compare the adoptees to blackchildren in the general population. Even if allof the black children were interracial offspring,however, a strong genetic hypothesis should notpredict that they would score well above the whitepopulation average. Nor should they score ashighly as white adoptees. In fact, the black andinterracial children of this sample scored as highlyon IQ tests as did white adoptees in previousstudies with large samples (Burks, 1928; Leahy,193S).

In other words, the range of reaction of sociallyclassified black children's IQ scores from average(black) to advantaged (white) environments is atleast 1 standard deviation. Conservatively, if weconsider only the adopted children with two blackparents (and late and less favorable adoptive ex-periences), the IQ reaction range is at least 10points between these environments. If we considerthe early-adopted group, the IQ range may be aslarge as 20 points. The level of school achieve-ments among the black and interracial adoptees isfurther evidence of their above-average perform-ance on standard intellectual measures.

The dramatic increase in the IQ mean and theadditional finding that placement and adoptivefamily characteristics account for a major portionof the IQ differences among the socially classifiedblack children strongly suggest that the IQ scoresof these children are environmentally malleable.

One reason for the substantial increase in testperformance of the black and interracial adopteesis that their rearing environments are culturallyrelevant to the tests and to the school. Amid theIQ controversy, some have argued that standardizedmeasures are inappropriate for children whose cul-tural background is different from that of the tests.While the rejection of IQ tests as predictors ofacademic success, on the basis of their culturalbias, is untenable (Jensen, 1974), we believe thatthe tests and the schools share a common culture towhich black children are not as fully acculturatedas are white children. However, the socially classi-fied black children in this study have been fullyexposed to the culture of the tests and the school,although they are still socially defined as black.

IQ COMPARISONS WITHIN THE BLACK GROUP

The second question concerned a comparison ofthe IQ scores of children whose parents were bothblack with black children of interracial parentage.The interracial children scored about 12 pointshigher than those with two black parents, but thisdifference was associated with large differences inmaternal education and preplacement history. Thepart correlations suggested that variation in therace of mothers accounted for 3% of the children'sIQ variance, but even this percentage of varianceprobably includes some additional and unmeasuredenvironmental differences between the groups.

For example, black mothers are known to be atgreater risk than white mothers for nutritionaldeficiencies, maternal death, infant mortality, andother reproductive casualties (Scarr-Salapatek &Williams, 1973). The prematurity rate amongblack mothers is more than double that of whjtes.These antenatal risks are often found to be asso-ciated with long-term developmental problemsamong the children. The interracial children, allbut two of whom have white mothers, were lesslikely to have suffered any of these problems.

COMPARISONS OF BLACK/INTERRACIAL,

ASIAN/INDIAN, AND NATURAL CHILDREN

OF THE ADOPTIVE FAMILIES

The third question asked for comparisons amongthe IQ scores of black/interracial, Asian/Indianadoptees, and the biological children of the adop-tive families. There were significant differences inIQ scores among the groups. The socially classi-fied black children scored on the average betweenthe white and Asian/Indian adoptees, but theseresults were confounded with placement variables.Among the early adoptees, there were too fewwhite and Asian/Indian children to make meaning-ful comparisons. The black/interracial early adop-tees, however, performed at IQ 110, on the average.

Compared to adopted children in previousstudies, the average IQ of 110 for the 99 early-adopted black/interracial children compares wellwith the 112.6 reported by Leahy (193S, p. 28S)for white adoptees in professional families.

The above-average IQ level of. adopted children,reported in all adoption studies, reflects both theirbetter-than-average environments and the elimina-tion of severely retarded children from the pool ofpotential adoptees. Although Munsinger's (197Sa)review concluded that adoptive family environ-


ments have little or no impact on the intellectualdevelopment of adoptees, past studies have notadequately tested this hypothesis. Because chil-dren who are selected for adoption are not grosslydefective, their predicted IQ level is slightly abovethat of the general population. In this study, how-ever, the adopted black /interracial children couldnot have been predicted to have average IQ scoresabove the mean of the white population unlessadoptive family environments have considerableimpact.

The biological children of the adoptive familiesscored above the average of the black/interracialearly adoptees. Not only have the biologicalchildren been in their families since birth, but theirnatural parents are considerably brighter than thoseof the adopted children, regardless of race.

SCHOOL ACHIEVEMENT

A fourth question focused on the school achieve-ment of the black/interracial adoptees and the bio-logical children in the adoptive families. Black/interracial adoptees were found to score slightlyabove average on school-administered achievementand aptitude tests, as predicted by their IQ scores.The natural children of the adoptive families scoredhigher than the socially classified black adopteeson school achievement measures, a finding whichis congruent with their higher IQ scores. Theschool achievement data provided validation forour IQ assessment.

GENETIC AND ENVIRONMENTAL SOURCES OF

IQ VARIANCE

The final question posed by the study dealt withthe relative contributions of biological and socialenvironmental measures to IQ differences amongthe socially classified black children. The place-ment variables, adoptive family characteristics, andgenetic background all contributed to the IQ dif-ferences among the black/interracial adoptees.Because the social and biological variables wereconfounded, it is very difficult to make a clearcomparison. Although this study has an unusualsample of children, we propose that genetic andsocial variables are usually confounded in families.Indeed, we suspect that genotype-environmentcorrelations are the rule and that they account fora sizable portion of the IQ variance in the generalpopulation.

In making any comparison between biological

and social variables, we must be concerned aboutthe quality of those measures. Although the adop-tive family variables are only indices of the quali-ties of the environment that have an impact onchildren, the natural parent data are even morelimited. It would have been advantageous to havecomparable IQ scores for the natural parents, ratherthan educational levels, although the latter correlateabout .7 with IQ in the general population (Jencks,1972).

Because the social variables accounted for asubstantial portion of the IQ variance among black/interracial adoptees, it is likely that IQ performanceis malleable within the range of existing environ-ments. If all black chilren had environments suchas those provided by the adoptive families in thisstudy, we would predict that their IQ scores wouldbe 10-20 points higher than the scores are undercurrent rearing conditions.

SOCIAL IMPLICATIONS OF THE STUDY

Given the above-average IQ scores of black/inter-racial children adopted transracially, it may seemthat we are endorsing the adoption of black chil-dren by white families as a social policy. There isno question that adoption constitutes a massiveintervention, as noted earlier, and that it has afavorable impact on IQ scores. However, there isgood reason why transracial adoption is not apanacea for low IQ scores among black children.Only an infinitesimally small proportion of blackchildren will ever be available for adoption, and ofthose, many will and should be adopted by blackfamilies.

What we do endorse is that if higher IQ scoresare considered important for educational and occu-pational success, then there is need for social actionthat will provide black children with home en-vironments that facilitate the acquisition of in-tellectual skills tapped by IQ measures. Althoughthere has been some research describing the im-mediate environments of middle-, working-, andlower-class homes (Hess & Shipman, 196S; Kohn,19S9; White & Watts, 1973), there is still a needto investigate how families, such as these transracialadoptive families, constitute an ecological systemin which IQ skills are developed. The physicalenvironment, the amount and quality of parent-child interaction, the parents' attitudes and prac-tices in child rearing, the neighborhood and com-munity settings of the family, and the larger socialcontexts of employment, economic security, and


cultural values must all be considered in describingthe parameters of family effects.

Educational interventions alone are unlikely tohave the effects reported here for adoption. Schools,as presently constituted, cannot have the far-reach-ing, intensive impact of the family and home.

Our emphasis on IQ scores in this study is notan endorsement of IQ as the ultimate human value.Although important for functioning in middle-classeducational environments, IQ tests do not samplea huge spectrum of human characteristics that arerequisite for social adjustment. Empathy, socia-bility, and altruism, to name a few, are importanthuman attributes that are not guaranteed by a highIQ. Furthermore, successful adaptation withinethnic subgroups may be less dependent on theintellectual skills tapped by IQ measures than isadaptation in middle-class white settings.

This study was not designed to address the socialissues we have just highlighted. Rather, it wasintended to examine the effects of cross-fosteringon the IQ scores of black/interracial children. Themajor questions of the study concerned the relativeeffects of genetic background and social environ-ment on IQ levels and variations among sociallyclassified black children. The major findings of thestudy support the view that the social environmentplays a dominant role in determining the average IQlevel of black children and that both social andgenetic variables contribute to individual variationamong them.

REFERENCES

Burks, B. S. The relative influence of nature and nurtureupon mental development; a comparative study of fosterparent-foster child resemblance and true parent-truechild resemblance. Yearbook of the National Society forthe Study of Education, 1928, 27, 219-316.

Cleary, T. A., Humphreys, L. G., Kendrick, S. A., &Wesman, A. Educational uses of tests with disadvantagedstudents. American Psychologist, 1975, 30, 15-41.

Freeman, F. VS., Holzinger, K. J., & Mitchell, B. C. Theinfluence of environment on the intelligence, schoolachievement and conduct of foster children. Yearbook of

the National Society for the Study of Education, 1928,27, 101-217.

Grow, L. J., & Shapiro, D. Black children-white parents.New York: Child Welfare League of America, 1974.

Hess, R. D., & Shipman, V. C. Early experience and thesocialization of cognitive modes in children. Child De-velopment, 1965, 36, 869-886.

Husen, T. Talent, equality, and meritocracy. The Hague:Martinue Nijhoff, 1974.

Jencks, C. Inequality: A reassessment of the effects offamily and schooling in America. New York: BasicBooks, 1972.

Jensen, A. R. Educability and group differences. NewYork: Basic Books, 1973.

Jensen, A. R. How biased are culture-loaded tests? GeneticPsychology Monographs, 1974, 90, 185-244.

Kaufman, A. S., & Doppelt, J. E. Analysis of WISC-Rstandardization data in terms of the stratification vari-ables. Child Development, in press.

Kohn, M. L. Social class and the exercise of parentalauthority. American Sociological Review, 1959, 24, 352-366.

Leahy, A. M. Nature-nurture and intelligence. GeneticPsychology Monographs, 1935, 17, 237-307.

Loehlin, J., Lindzey, G., & Spuhler, J. N. Race differencesin intelligence. San Francisco, Calif.: Freeman, 1975.

Manosevitz, M., Lindzey, G., & Thiessen, D. Behavioralgenetics: Method and research. New York: Appleton-Century-Crofts, 1969.

Munsinger, H. The adopted child's IQ: A critical review.Psychological Bulletin, 1975, 82, 623-659. (a)

Munsinger, H. Children's resemblance to their biologicaland adopting parents in two ethnic groups. BehaviorGenetics, 1975, S, 239-254. (b)

Reiss, A. J., Jr. Occupations and social status. New York:Free Press, 1961.

Scarr-Salapatek, S., & Williams, M. L. The effects of earlystimulation on low-birth-weight infants. Child Develop-ment, 1973, 44, 94-101.

Shockley, W. Morals, mathematics, and the moral obliga-tion to diagnose the origin of Negro IQ deficits. Reviewof Educational Research, 1971, 41, 369-377.

Shockley, W. Dysgenics, geneticity, raciology: A challengeto the intellectual responsibility of educators. Phi DeltaKappan, 1972, S3, 297-307.

Skodak, M., & Skeels, H. M. A final follow-up study ofone hundred children. The Journal of Genetic Psychol-ogy, 1949, 75, 85-125.

Terman, L. M., & Merrill, M. Stanford-Binet IntelligenceScale. Boston, Mass.: Houghton Mifflin, 1972.

Wechsler, D. Wechsler Intelligence ^Scale for Children.New York: Psychological Corporation, 1949.

Wechsler, D. Wechsler Adult Intelligence Scale. NewYork: Psychological Corporation, 1955.

White, B. L., & Watts, J. C. Experience and environment.Englewood Cliffs, N.J.: Prentice-Hall, 1973.


iq test performance of black children adopted by white ... · pdf fileiq test performance of...

Documents