A Longitudinal Twin Study of Stability of Components from Bayley's Infant Behavior

Record
Author(s): Adam P. Matheny, Jr.
Source: Child Development , Apr., 1983, Vol. 54, No. 2 (Apr., 1983), pp. 356-360
Published by: Wiley on behalf of the Society for Research in Child Development

Stable URL: https://www.jstor.org/stable/1129696

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Society for Research in Child Development and Wiley are collaborating with JSTOR to digitize,
preserve and extend access to Child Development

This content downloaded from

51.6.62.24 on Mon, 12 Jun 2023 10:39:54 +00:00
All use subject to https://about.jstor.org/terms
A Longitudinal Twin Study of Stability of
Components from Bayley's Infant Behavior
Record

Adam P. Matheny, Jr.

University of Louisville School of Medicine

MATHENY, ADAM P., JR. A Longitudinal Twin Study of Stability of Components from Bayley's
Infant Behavior Record. CHILD DEVELOPMENT, 1983, 54, 356-360. Factor scores from Bayley's
Infant Behavior Record, obtained from about 300-400 infant twins at 6, 12, 18, and 24 months
were selected to represent 3 aspects of infant behavior: task orientation, test affect-extraversion,
and activity. Age-to-age correlations for the scores yielded a simplex pattern of low to mod-
erate order. Further age-to-age analyses by twin pairs generally showed that the identical twins
as compared with same-sex fraternal twins, generated more within-pair similarities for profile
of the factor scores across 2 overlapping age ranges: 6-18 months, and 12-24 months. Overall,
the findings indicated that there is reordering of individual differences for aspects of infant
behavior from one age to the next, but the sequence of reordering is somewhat structured in
accord with genotypic similarity. The longitudinal findings are also discussed within the con-
text of change and continuity of the individuality of infant temperament.

Interest in the Infant Behavior Record Cowan, Johnson, & Maxwell, 1981) of 125
(IBR), the third component of the Bayley Mexican-American infants at 12 months did
Scales of Infant Development (Bayley, 1969), not demonstrate the same factor structure;
has led to a series of studies in which the in-however, there was evidence that the first fac-
strument's psychometric properties and its tor represented both task orientation and test
clinical efficacy have been examined. The IBR affect-extraversion. A most recent large study
consists of 30 items representing broad de-of about 1,200 Dutch infants (van der Meulen
scriptions of infant behaviors observed during & Smrkovsky, Note 1) has replicated the re-
infant mental testing. Twenty-five of the items sults reported by Matheny (1980). Factors
are rating scales pertaining to social behaviors, denoting task orientation, test affect-extraver-
activity, attention and task-oriented behaviors, sion, and activity were isolated from a pooled
motor skills, specific sensory interest, andcorrelation matrix representing the age range
mouthing. The items are descriptively invari- from 2 to 30 months.
ant for the entire range of ages covered by
In addition to the reports of the isolation
the Bayley scales; however, distributions of
of underlying dimensions of the IBR, several
the ratings for each scale have been examined
reports on twins (Freedman, 1965; Goldsmith
at several ages (Bayley, 1969; Dolan, Math-
& Gottesman, 1981; Matheny, 1980; Matheny,
eny, & Wilson, 1974), and there are develop- Dolan, & Wilson, 1976) have shown that fac-
mental changes in the distributions of ratings.
tors, as well as individual scales, from the IBR
In order to detect the primary features of or its earlier versions have genetic influences.
infant behaviors observed during testing, sev- The salient findings from those studies sug-
eral investigators have factor analyzed the rat- gest that aspects of focal attention, goal di-
ing scales from the IBR. Matheny (1980) ob- rectedness, socially defined behaviors, emo-
tained IBR measures from between 300 and
tionality, and level of activity demonstrably
400 infant twins tested at 3, 6, 9, 12, 18, and
24 months and found three factors to be re- vary according to the similarity of genotype.
current at all of the six ages. The factors were These findings buttress a general conclu-
labeled task orientation, test affect-extraver-sion that there are genetically induced indi-
sion, and activity. A subsequent study (Mc- vidual differences in infant behaviors; how-
This research was supported in part by research grants 90-C-922 from the Office of Child
Development, BNS76-17315 from the National Science Foundation, and a grant from the
John D. and Catherine T. MacArthur Foundation. I thank R. Arbegust, S. Bateman, H. Dallum,
A. Dolan, P. Gefert, E. Harpring, M. Moseson, and S. Nuss for their contributions. Requests
for reprints should be sent to Adam P. Matheny, Jr., Child Development Unit, Health Sciences
Center, University of Louisville, Louisville, Kentucky 40292.
[Child Development, 1983, 54, 356-360. @ 1983 by the Society for Research in Child Development, Inc.
All rights reserved. 0009-3920/83/5402-0023$01.00]

This content downloaded from

51.6.62.24 on Mon, 12 Jun 2023 10:39:54 +00:00
All use subject to https://about.jstor.org/terms
 Adam P. Matheny, Jr. 357
ever, there remains the question of how much, were available for 300-400 infants, depending
if any, of the change or stability of behavioral on the age of the visit. The most complete
individuality during development can be at- longitudinal data were found for the test ages
tributed to genetic influence. This question of 6, 12, 18, and 24 months; therefore, those
is posed in the context of an ongoing discus- ages were selected to represent the 2-year in-
sion (e.g., Kagan, 1971) regarding the degree fant period.
to which idiosyncratic features of change in
behavior can be attributed not to innate char- Measures.-In order to create equivalent
measures at every age, the salient factors ob-
acteristics (traits) but to exogenous features
tained from the previous analyses (Matheny,
that foster developmental change.
1980) were examined to see which IBR rat-
In the present study, evidence for the ing scales had recurrently high loadings at
continuity of behavioral dimensions of the IBR every test age. The salient factors were those
was investigated first by examining the age-to- isolated by the factor analyses of Matheny
age correlations of infant twins tested at 6, (1980) and van der Meulen and Smrkovsky
12, 18, and 24 months, and then by examin- (Note 1): task orientation, test affect-extra-
ing the degree to which identical and fra- version, and activity. The rating scales se-
ternal twin pairs were concordant for patterns lected to represent those factors were those
of change over several ages. which had factor loadings of 0.60 or higher
as determined from the factor analyses by
Method Matheny (1980). Those scales were then
used to estimate the factor scores by giving
Subjects.-The infants were participantsunit weights of 1 to each rating scale, and
in a longitudinal study of twins (described the
by estimated factor score was obtained by
Wilson, 1978), and a complete description summing
of the ratings.' For example, the factor
the characteristics of the present sample test
of affect-extraversion had the highest load-
infant twins can be found in Matheny (1980).ings from the following rating scales: social
In brief, the infant sample consisted of 436
orientation-examiner, cooperativeness, fear-
twins who were recruited from families repre-
fulness, and emotional tone. The ratings on
senting the entire socioeconomic range found
those scales, as completed for each infant,
in the Louisville metropolitan area. The dis-
were summed (ratings of fearfulness were re-
tribution of occupations of head of household
versed), and the sum represented the esti-
(Reiss 1961) was generally comparable mated
to factor for each infant. The same pro-
that in the metropolitan area. cedure was carried out for each of the factors
Zygosity was determined for the same- at each of the four ages. The factors and the
sex twins by establishing concordance or dis- IBR rating scales representing those factors
cordance on 22 or more antigens. Discordance are shown in Table 1.
on any of the comparisons classified a twin Interobserver reliabilities for the esti-
pair as fraternal. For practical reasons, mated the factor scores were calculated from two
twins were not blood-typed until 3 years of
age; therefore, all of the infant data had beensets of IBR ratings completed by independent
collected prior to an objective determination observations of 57 infants tested at any one
of zygosity.
of the four ages. Those ratings were combined
according to the method described above, and
Testing.-The twins were administered the estimated factor scores for the two sets of
the Bayley scales individually by separate ex- ratings for each infant were correlated. The
aminers during each visit to the study center interobserver reliabilities for task orientation,
at 3, 6, 9, 12, 18, and 24 months of age (? 1 test affect-extraversion, and activity were 0.82,
week). The testing schedule was arranged so 0.87, and 0.79, respectively.
that examiners did not test the same twin on
successive visits. Thus, during concurrent and Results
subsequent tests, there was independence
from both twins being rated by the same ex- Age-to-age correlations.-The correla-
aminer. After each test, the examiner com- tions showing the stability of the factor scores
pleted the IBR according to the instructions from age to age are shown in Table 2. The
provided in the Bayley manual. From the total correlation matrix has two features of interest:
sample of infants, complete IBR protocols the pattern is generally simplex in form, and
1 The approach used to estimate factor scores is discussed by Nunnally (1967) and Tryon
and Bailey (1970). Tryon's method-summing the scores of variables and then standardizin
the simple means-was used in order to preserve the equivalence of measures across ages.

This content downloaded from

51.6.62.24 on Mon, 12 Jun 2023 10:39:54 +00:00
All use subject to https://about.jstor.org/terms
 358 Child Development
TABLE 1

RATING SCALES SELECTED TO ESTIMATE FACTORS EXTRACTED

FROM THE INFANT BEHAVIOR RECORD

FACTORS

Task Orientation Test Affect-Extraversion Activity

8. Object orientation (.75) 2. Social examiner (.73) 14. Activity (.85)

11. Goal directedness (.80) 4. Cooperativeness (.66) 21. Body motion (.82)
12. Attention span (.77) 5. Fearfulness (- .65) 25. Energy (.73)
7. Emotional tone (.78)

NoTE.-Numbers in parentheses represent the average of the factor loadings for the rating scales at 6, 12,
18, and 24 months.

TABLE 2
AGE-TO-AGE CORRELATIONS FOR IBR FACTORS
AT 6, 12, 18, AND 24 MONTHS
the analyses were divided into two overlap-
ping spans: 6, 12, and 18 months; 12, 18, and
24 months.

Factor The statistical analysis of concordance

and Age 12 Months 18 Months 24 Months for change was based on a repeated-measures
analysis of variance design (Haggard, 1958;
Task Orientation
Winer, 1962) adapted for twin data (Wilson,
Task orientation: 1979). Of interest is the systematic variance
6 months..... .32** .23** .27** attributed to the twin pairs' profile of changes
12 months.... .39** .35** in scores obtained at several ages. When age-
18 months.... .46**
to-age changes are large, as found in the pres-
Test Affect-Extraversion ent study, those changes may be systematic
for each pair, thereby generating high profile
Test affect- correlations. For a more complete explanation
extraversion:
6 months..... 27** .25** .10
of the method as applied to physical growth,
12 months..... 19"* .23**
see Wilson (1979).
18 months.... .35**
The infant twins were recombined into
Activity twin pairs. Opposite-sex pairs, pairs with in-
complete data, and pairs not blood-typed were
Activity:
6 months..... 29** .14* .18** excluded from the analysis. For the 6-18 age
12 months.... .34** .25** span, there were 60 identical and 35 fraternal
18 months.... .42** twin pairs; for the 12-24 age span, there were
66 identical and 40 fraternal twin pairs.
NOTE.--N' for Means for
correlations the estimated
range from factor291
scores to
were 339.
not
"* p <.05, one-tailed.
"** p < .01, one-tailed. significantly different when the identical and
fraternal pairs were compared for either of the
spans of ages.
the correlations become higher during t
ond year. During both years, the corre
The within-pair correlations for the pro-
are not particularly
file ofrobust, thereby
the factor scores obtained from the sug
that there are idiosyncratic features whic
identical and fraternal twin pairs are pre-
each infant, preclude
sented in Table strong predict
3. The correlations are con-
test-taking behaviors from one age to th
sistently higher for the identical pairs than for
Twin the fraternal
concordance pairs, andchange.-Hav
for it is apparent that the
established that there patterns of are
change within the two age spans
considerable c
for each of the factors are more coherently
across organized for the identi-we d
ages,
the next analysis toward cal twin pairs. Thus, it would appear that the
determining th
gruency within identical age-related reordering of individual
and differ-
fraternal
pairs for those changes. Thebyquestio
ences for the behaviors represented the fac-
whether the identical twins traced a more tors is more closely matched by both twins
similar pattern or profile of change over ages within identical pairs. By implication, the rela-
than did the fraternal twins. In order to maxi- tive stability (or instability) of the behaviors
mize the sample size available at several ages, seems to be regulated by genetic factors.

This content downloaded from

51.6.62.24 on Mon, 12 Jun 2023 10:39:54 +00:00
All use subject to https://about.jstor.org/terms
 Adam P. Matheny, Jr. 359
TABLE 3 the behavioral dimensions as reflecting char-
acteristics
TWIN CONCORDANCE FOR AGE-TO-AGE CHANGE of temperament, then it is apparent
IN IBR FACTOR SCORES from the stability correlations that there are
marked reorderings of individual differences
PROFILE CORRELATIONS of temperament from one age to another.
Taken at face value, these results, as well as
Test
those from a study by Rothbart and Derry-
AGE (Months) Task Affect-
AND ZYGOSITY Orienta- Extraver- berry (1981), would suggest that individual
(No. of Pairs) tion sion Activity differences in temperament, as measured at
one age or another, need not be markedly sta-
6, 12, and 18: ble. Apparently, there are individual differ-
Identical
(N = 60)..... .53* .35* .27 ences in the transitory features of tempera-
Fraternal ment-related behaviors as well.
(N = 35) ..... .18 .06 .19
12, 18, and 24: From this perspective, the profile correla-
Identical tions obtained from the identical and fraternal
(N = 66)......49* .37* .52* pairs are particularly instructive. The correla-
Fraternal
tions show, within the context of behavioral
(N = 40)..... ..21 .12 .18
transitions, that the sequences of individual
"* p < .05, one-tailed, identical R >differences
fraternal ofR.
change are partially regulated
by genetic influences. In effect, the data from
Discussion twins show that perturbations or instabilities
in temperament do not vitiate the concept of
The results from the present study temperament,
sup- but permit the concept to in-
port and extend previous data suggestingclude a constitutional basis for the regulation
that
there are genetic origins of individualof developmental change.
vari-
ations in behaviors rated on the IBR. Freed-
man (1965) pooled sequential ratings of 20
Reference Note
twin pairs assessed by a research version of
the IBR during the first year, and found that 1. van der Meulen, B. F., & Smrkovsky, M. Fac-
identical pairs were more concordant than fra- tor analysis of Bayley's Infant Behavior Rec-
ternal pairs for ratings of sociability, affect, ord. Paper presented at the International
and object-oriented behaviors. Matheny et al. Conference on Infant Studies, Austin, Texas,
(1976) replicated and extended those findings 1982.
by examining the behavioral ratings of about
80 pairs of twins assessed during the first 2
years. In later studies, factor analyses of theReferences
IBR (Matheny 1980) and an earlier version
Bayley, N. Bayley Scales of Infant Development.
of the IBR (Goldsmith & Gottesman, 1981)
New York: Psychological Corp., 1969.
demonstrated identical-fraternal twin pair dif-
Dolan, A. Brown, Matheny, A. P., Jr., & Wilson,
ferences for several dimensions representing,
among others, affective and social behaviors, R. S. Bayley's Infant Behavior Record: Age
trends, sex differences, and behavioral corre-
task orientation and persistence, and activity.
lations. ISAS Catalog of Selected Documents
Matheny (1980) further showed that the or-
in Psychology, 1974, 4, 9.
ganization of the profiles of factor scores was
Freedman, D. G. An ethological approach to the
systematically related to genotypic similarity.
genetic study of human behavior. In S. Van-
At every one of the six test ages during the
denberg (Ed.), Methods and goals in human
first 2 years, identical pairs had significantly
behavior genetics. New York: Academic
higher profile correlations than fraternal pairs.
Press, 1965.
The emergence of such prominent fea- Goldsmith, H. H., & Gottesman, I. I. Origins of
tures of infant behaviors and the demonstra- variation in behavioral style: A longitudinal
tion of genetic origins for individual variations study of temperament in young twins. Child
of those behaviors have been discussed within Development, 1981, 52, 91-103.
the context of research on infant temperament Haggard, E. A. Intraclass correlation and the anal-
(Goldsmith & Gottesman, 1981; Matheny, ysis of variance. New York: Dryden, 1958.
1980). Within that same context, the pres- Kagan, J. Change and continuity in infancy. New
ent findings help to trace more clearly the York: Wiley, 1971.
genetic contributions to the emergence and McGowan, R. J., Johnson, D. L., & Maxwell, S. E.
persistence of temperament. If one considers Relations between infant behavior ratings and

This content downloaded from

51.6.62.24 on Mon, 12 Jun 2023 10:39:54 +00:00
All use subject to https://about.jstor.org/terms
360 Child Development
concurrent and subsequent mental scores.
Developmental Psychology, 1981, 17, 542-
553.
Matheny, A. P., Jr. Bayley's Infant Behavior Rec-
ord: Behavioral components and twin analy-
ses. Child Development, 1980, 51, 1157-
1167.
Matheny, A. P., Jr., Dolan, A. B., & Wilson, R. S.
Twins: Within-pair similarity on Bayley's
Infant Behavior Record. Journal of Genetic
Psychology, 1976, 128, 263-270.
Nunnally, J. C. Psychometric theory. New York:
McGraw-Hill, 1967.
Reiss, A. J., Jr. Occupations and social status.
New York: Free Press, 1961.
Rothbart, M. K., & Derryberry, D. Development
of individual differences in temperament. In
M. E. Lamb & A. L. Brown (Eds.), Ad-
vances in developmental psychology (Vol.
1). Hillsdale, N.J.: Erlbaum, 1981.
Tryon, R. C., & Bailey, D. E. Cluster analysis.
New York: McGraw-Hill, 1970.
Wilson, R. S. Synchronies in mental develop-
ment: An epigenetic perspective. Science,
1978, 202, 939-948.
Wilson, R. S. Analysis of longitudinal twin data.
Acta Geneticae Medicae et Gemellologiae,
1979, 28, 93-105.
Winer, B. J. Statistical principles in experimental
design. New York: McGraw-Hill, 1962.

This content downloaded from

51.6.62.24 on Mon, 12 Jun 2023 10:39:54 +00:00
All use subject to https://about.jstor.org/terms