Child Development, July/August 1999, Volume 70, Number 4, Pages 1005–1016

Cultural Differences in Maternal Beliefs and Behaviors:

A Study of Middle-Class Anglo and Puerto Rican Mother-Infant Pairs
in Four Everyday Situations
Robin L. Harwood, Axel Schoelmerich, Pamela A. Schulze, and Zenaida Gonzalez

This study examines cultural patterning in situational variability in mother-infant interactions among middle-
class Anglo and Puerto Rican mothers and their 12 to 15-month-old firstborn children. Forty mothers were in-
terviewed regarding their long-term socialization goals and childrearing strategies, and videotaped interacting
with their infants in four everyday settings: feeding, social play, teaching, and free play. Results suggest that:
(1) Anglo mothers place greater emphasis on socialization goals and childrearing strategies consonant with a
more individualistic orientation, whereas Puerto Rican mothers place greater focus on goals and strategies
consistent with a more sociocentric orientation; (2) coherence was found between mothers’ childrearing beliefs
and practices, with Puerto Rican mothers more likely to directly structure their infants’ behaviors; and (3) situ-
ational variability arose in mother-infant interactions, but this variability showed a cultural patterning consis-
tent with mothers’ long-term socialization goals and childrearing beliefs.

INTRODUCTION tion of internal attributes . . . and (b) behaves prima-

rily as a consequence of those internal attributes”
The past decade has brought increased interest
(Markus & Kitayama, 1991, p. 224; see also Kessen,
among researchers in understanding the cultural con-
1979; Sampson, 1989; Shweder & Bourne, 1984;
text of childhood. Along with this concern has come a
Spence, 1985). This construal of the self is described as
heightened consideration of appropriate theoretical
a key component of the beliefs and practices which
frameworks for the study of culture and child devel-
organize perceptions of and interactions with chil-
opment. In particular, interpretive approaches have
dren in America, thus constituting a primary aspect of
been articulated among several researchers (Cole,
the cultural context of childhood in this country (Hark-
1996; Harkness & Super, 1996; Shweder, 1996). Despite
ness & Super, 1996; Harwood, Miller, & Lucca Irizarry,
their diversity, interpretive approaches generally share
1995; Rogoff et al., 1993; Schieffelin & Ochs, 1986).
the assumption that human beings construct meaning
In contrast, many other cultures are described as
through their cultural symbol systems, with language
“sociocentric” or “interdependent” in that they em-
being one of culture’s most powerful symbol systems.
phasize the fundamental connectedness of human
Many of these approaches go on to assert that this
beings to one another: “Experiencing interdepen-
construction occurs within a matrix of social interac-
dence entails seeing oneself as part of an encompass-
tion, in which the child as participant actively pro-
ing social relationship and recognizing that one’s be-
duces and reproduces culturally meaningful patterns
havior is determined, contingent on, and, to a large
of beliefs and behaviors (e.g., Corsaro & Miller, 1992;
extent organized by what the actor perceives to be the
Goodnow, Miller, & Kessel, 1995; Rogoff, Mistry, Goncu,
thoughts, feelings, and actions of others in the relation-
& Mosier, 1993; Schieffelin & Ochs, 1986).
ship” (Markus & Kitayama, 1991, p. 227). Again, this
As interpretive approaches become more widely
emphasis on interdependence is depicted by many re-
used and recognized, certain themes appear to recur
searchers as a key component of the beliefs and prac-
and to demand continued refinement. One salient
tices that organize perceptions of and interactions
question involves the use of terms like “individualistic/
with children in a variety of non-Western cultures,
independent” or “collectivistic/interdependent” as
thus constituting a primary aspect of the cultural con-
heuristic devices to characterize broad-level cultural
text of childhood in these countries (cf. Greenfield &
belief systems and practices (cf. Greenfield & Cock-
Cocking, 1994; Kurtz, 1992; Tobin, Wu, & Davidson,
ing, 1994; Markus & Kitayama,1991; Shweder &
1989; Triandis, Marin, Lisansky, & Betancourt, 1984).
Bourne, 1984; Triandis, Bontempo, Villareal, Asai,
Critics of this approach have maintained that such
& Lucca, 1988). Briefly, American culture is often de-
global characterizations are unidimensional and do
scribed as “individualistic” in that it conceives of the
individual as an “independent, self-contained, auton- © 1999 by the Society for Research in Child Development, Inc.
omous entity who (a) comprises a unique configura- All rights reserved. 0009-3920/99/7004-0016
1006 Child Development

not tap “the multitude of social orientations that con- 2. Although previous studies have suggested that
stitute the social life of individuals in most cultures” long-term socialization goals represent salient
(Wainryb, 1995, p. 390). The use of phrases like “the cultural constructs (Harwood, 1992; Harwood
individualism of American culture” or “the interdepen- et al., 1995), research has not directly examined
dence of Japanese culture” has drawn critical rejoin- these goals in relation to mothers’ interactions
ders highlighting instances where American mothers with their infants. Consistent with the idea that
emphasize cooperation, and Japanese or Indian par- culture consists of meaningful patterns of be-
ents provide their children with choice (Derne, 1995; liefs and behaviors (Corsaro & Miller, 1992;
Mines, 1994; Nucci, 1994). Rogoff et al., 1993; Schieffelin & Ochs, 1986),
The reality, of course, is that no substantive re- we thus hypothesized that coherence would
searcher (to our knowledge) claims that cultures are exist between mothers’ beliefs regarding so-
monolithic, homogeneous entities devoid of internal cialization goals and the ways in which they
variation, or that “individualistic” cultures lack a con- organize interactions with their infants. In par-
cept of relatedness, and “sociocentric” cultures lack a ticular, it was predicted that Anglo mothers
concept of personal choice. As Miller, Fung, and would structure interactions with their infants
Mintz (1996) note, however, one of the major chal- in such a way as to encourage independence
lenges in cultural psychology today consists of “find- and self-confidence, whereas Puerto Rican
ing effective ways to represent the intricate pattern- mothers would structure their interactions in
ing of similarities and differences among cultures— such a way as to encourage respectfulness and
without subduing the complexity of particular mean- attentiveness to others.
ing in each” (p. 239). More specifically, if childhood’s 3. It was anticipated that situational variability
cultural context is indeed patterned at a broad level, would arise in mother-infant interactions, but
then how do we understand and represent internal that this variability would show a cultural pat-
variations in parental beliefs and practices? terning consistent with mothers’ long-term so-
cialization goals.

Culture and Situational Variability

Selection of Study Participants
Situational effects have long been recognized as a
The reasons for choosing middle-class Anglo1 and
source of internal variation in the beliefs and behav-
Puerto Rican mothers include: (1) Middle-class Anglo
iors of individuals, and indeed it could be argued that
mothers were expected to provide an index of the
parental beliefs and practices show considerable vari-
mainstream culture that has shaped the articulation
ability across situations, thus belying any attempt at
of theories of child development within the United
broad-level group characterizations such as “individ-
States (Kessen, 1979), thus providing an appropriate
ualistic” or “interdependent.” Alternatively, it is pos-
comparison point for the study of other groups; and
sible that situational variability in mother-infant in-
(2) Middle-class Puerto Rican mothers in San Juan
teractions will be culturally patterned, thus suggesting
represent an educated, professional group in a society
that within-group variation exists, but needs to be ex-
that, due to its commonwealth status, shares the tech-
amined within the larger context of broad-level group
nology, industry, and urbanization of the United
differences.
States. Nonetheless, it was anticipated that Puerto
In this study, we sought to investigate the cultural
Rican mothers would demonstrate beliefs concordant
patterning of situational variability in mother-infant
with a cultural meaning system different from that of
interactions by examining middle-class mother-infant
the middle-class Anglo mothers. In particular, Puerto
pairs in northeastern Connecticut and in San Juan,
Puerto Rico. Briefly, there were three hypotheses:
1 We have chosen the term “Anglo” to describe the white
1. Consistent with previous research (cf. Harwood
American women of nonHispanic European ancestry who par-
et al., 1995) indicating that Anglo-American cul- ticipated in these studies. Although “white Americans of non-
ture can generally be described as more “indi- Hispanic European ancestry” is the most accurate name for this
vidualistic,” whereas Puerto Rican culture can group, it is far too cumbersome for repeated use. In addition, al-
generally be described as more “sociocentric,” it ternative labels like “Euro-American,” or “U.S.” ignore the fact
that many Puerto Ricans are not only also of (Hispanic) Euro-
was expected that the two groups would differ
pean ancestry, but are also all United States citizens. In addition,
in the extent to which they emphasize socializa- “Anglo” has a long history of use as a cultural term contrasting
tion goals and childrearing strategies consonant the English-speaking Americas with the Spanish-speaking
with these broad constructs. Americas.
 Harwood et al. 1007

Rican culture is generally considered to emphasize a Table 1 Demographic Characteristics of Participants

more sociocentric, and Anglo culture a more individ- Anglo Puerto Rican
ualistic, view of the person and of interpersonal rela-
tionships (Diaz Royo, 1974; Harwood et al., 1995; Characteristic M SD M SD
Lauria, 1982). Significantly, Puerto Rico’s status as a
U.S. commonwealth has made socioeconomic status Child’s age (months) 13.7 1.1 13.3 1.3
Mother’s age 31.5 5.0 27.9 4.6*
more comparable across the two cultures than is nor-
Mother’s education (years) 16.5 1.9 15.9 1.5
mally the case, thus facilitating a comparison of these Percent mothers employed 78.0 78.0
two middle-class groups. No. hours worked/week 32.1 12.2 37.3 6.3
No. people in household 3.0 0.0 3.2 0.4
Father’s education (years) 16.2 2.4 15.6 2.2
Father’s age 32.8 6.0 30.3 5.8
METHOD
Hollingshead score 53.3 9.1 50.9 8.9
Participants Percent mothers married 95.0 78.0

Forty middle-class mothers (Anglo 5 22, Puerto * p , .05.

Rican 5 18) participated as paid volunteers. All moth-
ers were at least 20 years of age, and had one firstborn
Analyses were performed on 10 demographic vari-
infant between the ages of 12 and 15 months. There
ables (child’s age, mother’s age, mother’s education
were no group differences in the age of each mother’s
in years, maternal employment status, hours worked
child (M 5 13.5 months), and gender distribution
outside the home, total number people in household,
(Anglo 5 11 female, 11 male; Puerto Rican 5 10 fe-
father’s education in years, father’s age, household
male, 8 male) was similar across the two groups. The
Hollingshead score, and mother’s marital status). As
Anglo mothers were drawn from urban areas within
can be seen in Table 1, results indicated that the
eastern Connecticut, were White American mothers
groups were demographically similar, differing sig-
of nonHispanic European ancestry who had been
nificantly on just one variable: Anglo mothers were
born, reared, and educated in the United States, spoke
older than were Puerto Rican mothers, mean 5 31.5
English as their first language, and had male partners
and 27.9 years, respectively, p , .05. The two groups
who met these same ethnic criteria. The Puerto Rican
were thus comparable, both comprised of highly ed-
mothers were drawn from the San Juan metropolitan
ucated, financially stable mothers.
area (Puerto Rico’s largest city, situated in the north-
eastern quadrant of the island), had been born,
Procedure
reared, and educated in Puerto Rico, spoke Spanish as
their first language, and had male partners who met Interviews were conducted in two sessions in
these same ethnic criteria. In both settings, mothers mothers’ homes by ethnically matched, trained inter-
were recruited primarily by an invitational letter from viewers. The first session included videotaped obser-
the principal investigator that was distributed through vations of each mother interacting with her infant in
the offices of pediatricians who agreed to identify for four everyday settings. The second session included
us mothers who met the sampling criteria; in both set- an interview with each mother regarding her long-
tings, a few mothers also were solicited by this same term socialization goals and childrearing beliefs. The
method through day care centers and in mother-in- videotaped observations were done prior to the inter-
fant activity groups. views in order to avoid biasing mothers’ interactions
with their infants during the observations. The in-
terview sessions were tape-recorded and tran-
Socioceconomic Status and Demographic Characteristics
scribed verbatim for further analysis. Each of the
Hollingshead’s (1975) Four Factor Index was used two sessions took about 1½ hours to complete, and
to determine socioeconomic status (SES). This scale were conducted approximately 1 week apart. The
computes a weighted average score based on both par- Spanish version of the interview protocol was ob-
ents’ occupations and educational levels when both tained through a committee of three bilingual, bi-
are employed outside the home, and on the income- cultural master’s level research assistants at the Be-
receiving parent only in other cases. In this study, havioral Sciences Research Institute, University of
“middle-class” mothers were defined as women living Puerto Rico Medical School. Spanish interviews
in a home where the household occupational prestige were transcribed and translated into English in the
score fell within Levels I and II (major and minor pro- first author’s laboratory by bilingual, bicultural
fessional) of Hollingshead’s scale. (Puerto Rican) graduate students.
1008 Child Development

Maternal Beliefs Overall agreement reached a level of .78 (range 5 .72–

.84, Cohen’s k).
Long-Term Socialization Goals
Childrearing strategies. Mothers’ responses to the
To provide an index of long-term socialization question regarding what parents could do to encour-
goals, mothers were administered an abbreviated age or discourage the development of specific quali-
version of the Socialization Goals Interview (SGI), a ties in their children were coded into seven categories
semistructured, individually administered interview derived from the most common types of responses
consisting of four open-ended questions in which generated by mothers in both groups. These in-
parents are asked to describe the qualities they (1) cluded: (1) Model the desirable behavior, (2) Provide
would and (2) would not like their children to possess Opportunities for child to develop the quality on his
as adults, and to describe toddlers they know who or her own, (3) Provide an Emotional Environment
possess at least the beginnings of those (3) positive that will allow the quality to flourish, (4) Praise the
and (4) negative qualities (Harwood, 1992). Because child for desirable behavior, (5) the child learns
previous research (Harwood et al., 1995; Harwood, through the parent’s Direct Instruction, (6) the charac-
Schoelmerich, Ventura-Cook, Schulze, & Wilson, teristic is innate or Inborn, and thus the parents’ influ-
1996) has indicated that the first two questions, which ence is negligible, (7) Other: the parent gives an am-
probe for adult socialization goals, may provide a bet- biguous or otherwise uncodeable response, such as,
ter index of broad cultural belief systems than the two “I’ll worry about that later.” Fewer than 5% of moth-
questions regarding desirable and undesirable child ers’ responses across both groups were coded as Other.
behavior, mothers in this study were asked only the Reliability in coding mothers’ responses to the
first two questions. question regarding childrearing strategies was calcu-
lated between the third author and one independent
judge blind to the study hypotheses as well as to each
Childrearing Strategies
mother’s sociocultural group on 50% of the sample.
To investigate the specific childrearing strategies Overall agreement reached a level of .92 (range 5 .90–
that mothers espouse for accomplishing their ex- .95, Cohen’s k).
pressed socialization goals, each mother also was
asked to describe what parents could do to help a
Mother-Infant Interactions
child come to possess the desirable qualities, or to not
come to possess the undesirable qualities, that she In order to investigate coherence between mater-
mentioned. An open-ended format rather than a stan- nal beliefs and practices, as well as to examine cross-
dardized questionnaire was used for this task in order situational variability in maternal practices, mothers
to obtain, as much as possible, mothers’ indigenous were videotaped interacting with their infants in four
beliefs on this topic. everyday situations:
1. Feeding. Mothers were instructed to, “Feed
Coding of Interviews your child as you normally would,” and video-
taped for the first 10 min of the feeding session.
Socialization goals. As in previous research using the
Few feeding sessions lasted beyond the allotted
SGI (Harwood et al., 1995, 1996), mothers’ responses to
10-min time period.
the questions regarding long-term socialization goals
2. Social Play. Mothers were instructed to, “Play
were coded at the level of individual word and phrase
with your child without using toys,” and video-
descriptors into one of six mutually exclusive catego-
taped for 5 min.
ries identified through previous work as culturally rel-
3. Teaching. Mothers were supplied with materi-
evant to Anglo and Puerto Rican mothers (Harwood,
als and instructed to spend 3 min on each of the
1992; Harwood et al., 1995): (1) Self-Maximization, (2)
following three tasks:
Self-Control, (3) Lovingness, (4) Decency, (5) Proper
Demeanor, and (6) Miscellaneous. Fewer than 5% of “Teach your child to draw a line on the pad of
mothers’ responses across both groups were coded as paper with a crayon”;
Miscellaneous, attesting to the broadly encompassing “Teach your child to stack one block on top of
character of the five content categories. another”;
Reliability in coding mothers’ responses to the SGI “Teach your child to push one toy with another
was calculated between two independent judges along the floor.”
blind to the study hypotheses as well as to each Total time spent videotaping mothers in the
mother’s sociocultural group on 50% of the sample. teaching situation was 9 min.
 Harwood et al. 1009

4. Free Play. Mothers were supplied with a variety haviors and settings specific to each situation also
of developmentally appropriate toys and in- were coded:
structed to, “Play with your child as you normally
1. Feeding. Mother feeds the infant by directly
would using toys.” Videotaping lasted for 8 min.
placing spoon in infant’s mouth; mother en-
courages autonomy in feeding (e.g., places
Coding of Mother-Infant Interactions spoon in infant’s hand); and infant feeds self.
2. Social Play. Duration of time each mother spent
The videotaped mother-infant interactions were
in one of four types of dyadic social games:
coded in real time using Interact (Dumas, 1993), a
clapping/singing games; bouncing games (e.g.,
software program that computes frequencies of dif-
swinging child up and down in parents’ arms);
ferent types of behaviors, and durations of different
hiding-chasing games; and touching/turn-taking
types of settings. In our mother-infant interactions,
games (e.g., tickling, or games involving some
we chose to code two types of data: (1) frequency of
kind of back-and-forth interaction between the
behaviors (maternal nonverbal, maternal verbal, and
mother and infant, such as tugging a diaper
infant) common across the situations; and (2) fre-
back and forth).
quency of behaviors and duration of settings specific
3. Teaching. Duration of total time infants spent
to each situation.
on-task for each of the three tasks: crayon, block,
Maternal nonverbal behaviors. In order to investigate
and pushtoy; and duration of total time mothers
situational variability in mothers’ nonverbal strate-
spent attempting to teach the tasks.
gies for organizing interactions with their infants
4. Free Play. Duration of time each dyad spent in
across four everyday situations, the frequency of five
the following states of coordinated activity:
specific types of maternal nonverbal behaviors was
mother and infant play together; infant plays
examined in each situation as follows. The mother: (1)
alone while mother watches; mother attempts to
attempts to signal infant’s attention by tapping or
shift infant’s attention to a new toy while infant
pointing, (2) praises infant by clapping and cheering,
is still playing with another toy; mother at-
(3) positions infant, (4) restrains infant when infant at-
tempts to gain infant’s attention with a new toy
tempts to pull away or wander off, and (5) offers af-
while infant is not involved in any play activity;
fection by hugging/kissing infant.
and no play activity.
Maternal verbal behaviors. In order to investigate
situational variability in mothers’ verbal strategies for The videotapes were coded in the first author’s
organizing interactions with their infants across four laboratory in Connecticut by three trained graduate
everyday situations, the frequency of five specific students. In particular, a native English-speaking stu-
types of maternal verbal behaviors was examined in dent was the primary coder for the Anglo videotapes,
each situation as follows. The mother: (1) signals in- and a bilingual Puerto Rican student was the primary
fant’s attention by calling infant’s name, (2) praises coder for the Puerto Rican videotapes. Reliability was
infant, (3) offers infant affection by using term of en- established: (1) between the two primary coders by
dearment, (4) directly structures infant’s behavior by having the bilingual student code 9% of the Anglo
commanding infant to perform or not perform a videotapes in addition to all the Puerto Rican tapes,
given action, and (5) indirectly structures infant’s be- and (2) between the two primary coders and a third
havior by suggesting that infant perform or not per- independent judge, by having an additional bilingual
form a given action. Puerto Rican student code 27% of the Anglo video-
Infant behaviors. Although we were primarily in- tapes, and 33% of the Puerto Rican videotapes. Be-
terested in mothers’ behaviors in this study, in order cause the coders were viewing videotapes of mother-
to control for the possibility that group differences in infant interactions among two linguistically different
maternal behavior reflect differences in infant behav- groups, it was not possible to blind them to each
ior, we also examined three behaviors thought to in- mother’s sociocultural group membership; however,
dex the infant’s overall activity level and cooperation all three students remained blind to the hypotheses of
with the mother: (1) resists maternal intervention, the study. Overall agreement reached a level of .70
displaying negative affect; (2) disengages from direct (range 5 .54–.85, Cohen’s k).
physical contact with the mother without signs of
negative affect; and (3) wanders away by retreating
RESULTS
out of arm’s length of mother.
Situation-specific behaviors and settings. In addition In order to control for possible gender effects, prelim-
to behaviors examined across the situations, a few be- inary analyses were performed with child’s sex en-
1010 Child Development

tered as a factor. There was only one significant gen- and 34.6, respectively, p , .05). Therefore, data were
der effect among all analyses: mothers of girls were transformed to represent percentages of the total
more likely to give verbal signals during free play number of answers for each subject, and the resulting
than mothers of boys. Due to the general lack of sig- scores were treated as compositional data by log-
nificant gender effects, this variable will not be dis- transforming the scores to remedy the problems of non-
cussed further. independence of variation associated with data of this
type (Aitchison, 1986; see Bersoff & Miller, 1993). Anal-
yses were then performed on the transformed data.
Maternal Beliefs
Two of the seven categories (Inborn and Other) to-
Socialization goals. An analysis of variance (ANOVA) gether represented less than 6% of mothers’ responses,
indicated no group differences in total number of de- and thus were dropped from the final analysis.
scriptors generated by Anglo and Puerto Rican mothers Specifically, to compare mothers’ open-ended re-
in response to the open-ended questions regarding sponses regarding parenting strategies, a 2 3 5
long-term socialization goals (M 5 18.59 and 17.84, re- (Group 3 Category-Type) MANOVA was performed.
spectively). Because of the relatively small number of This analysis yielded a significant Group 3 Category-
responses generated, however, categories were com- Type interaction, F(5, 34) 5 9.05, p , .01. Follow-up
bined for the purposes of analysis, according to previ- one-way ANOVAs indicated that the Anglo mothers,
ous research (Harwood et al., 1995, 1996), which has in- compared to the Puerto Rican mothers, were (1) more
dicated that the categories of Self-Maximization and likely to generate childrearing strategies falling into
Self-Control tend to be more characteristic of the re- the categories of Modelling, p , .05, and Provide Op-
sponses of Anglo mothers, whereas the categories of portunities, p , .01; and (2) less likely to generate
Decency and Proper Demeanor tend to be more childrearing strategies falling into the category of Di-
characteristic of the responses of Puerto Rican moth- rect Instruction, p , .001. The categories of Provide
ers. Thus, the categories used in this analysis were: Optimal Emotional Environment and Praise did not
Self-Maximization/Self-Control; Decency/Proper De- yield significant group differences.
meanor; and Lovingness.
In order to investigate cultural differences in moth-
Mother-Infant Interactions Common across the Settings
ers’ long-term socialization goals, a 2 3 3 (Group 3
Category-Type) multivariate analysis of variance Infant behaviors. Preliminary analyses revealed that
(MANOVA) was performed on the number of mothers’ infant behaviors between the two groups differed on
descriptors falling into the three experimenter-derived just two variables across the four situations. In partic-
categories. This analysis yielded a significant Group ular, compared to Anglo infants, Puerto Rican infants
3 Category-Type interaction, F(3, 36) 5 6.5, p , .01. were less likely to retreat out of arm’s length during
Follow-up one-way ANOVAs revealed significant teaching (M 5 2.86 and 0.33, respectively, p , .01),
cultural differences in the number of descriptors and more likely to engage in resistive behaviors dur-
generated for two of the three categories. Consistent ing feeding (M 5 .50 and 5.28, respectively, p , .01).
with previous research, the Anglo mothers, com- This low incidence suggests that group differences in
pared to the Puerto Rican mothers, were more likely mothers’ behavior were not a result of group differ-
to generate descriptors falling into the combined ences in infant behavior. Because of this low inci-
Self-Maximization/Self-Control category (M 5 9.95 dence, and because our primary interest was in ma-
and 5.89, respectively, p , .05), and less likely to gen- ternal behavior, the analyses described below will
erate descriptors falling into the combined Decency/ focus on maternal variables.
Proper Demeanor category (M 5 3.82 and 9.17, re- Maternal nonverbal behaviors. Preliminary analyses
spectively, p , .01). There were no group differences indicated that the use of highchairs rendered the ma-
in the use of the category Lovingness (M 5 4.82 and ternal nonverbal behaviors of interest uncommon
2.78, respectively). among both Anglo and Puerto Rican mothers; feed-
Childrearing strategies. An ANOVA was performed ing therefore was omitted as a situation from this
on the total number of answers generated by mothers analysis. To investigate cultural differences in mothers’
in response to the question regarding what parents nonverbal strategies for organizing interactions with
can do to encourage or discourage the development their infants, a MANOVA with group (2) as a between
of desirable and undesirable qualities in their chil- factor, and situation (3) and behavior (5) as repeated
dren. This analysis indicated that Anglo mothers factors was performed on the frequency of different
gave a larger number of answers to this portion of the types of maternal nonverbal behaviors occurring in
interview than did Puerto Rican mothers (M 5 59.2 each of the situations except feeding. This MANOVA
 Harwood et al. 1011

yielded significant main effects for group, F(1, 37) 5 Table 3 Mean Frequency of Maternal Verbal Behaviors by
11.58, p , .01, situation, F(2, 74) 5 3.27, p , .05, and Situation
behavior, F(4, 148) 5 43.79, p , .001, as well as a sig- Puerto
nificant Situation 3 Behavior interaction, F(8, 296) 5 Anglo Rican
15.45, p , .001.
Follow-up one-way ANOVAs (see Table 2) indi- Behavior M SD M SD p
cated that, compared to Anglo mothers, Puerto Rican
Feeding
mothers: (1) exhibited a higher overall frequency of Verbal signal attention 1.7 2.2 9.7 13.4 **
nonverbal behaviors during social play, p , .01, and Verbal praise 3.6 5.7 .6 1.5 *
teaching, p , .05, but not during free play; and (2) Verbal affection .6 .8 .9 1.6
were more likely to restrain their infants during so- Command 4.0 3.9 18.1 16.2 ***
cial play, p , .001, and to position their infants dur- Suggest 7.2 4.4 10.7 8.9
Overall verbal 17.1 9.4 40.0 24.3 ***
ing teaching, p , .05. No group differences arose in
Social Play
mothers’ use of nonverbal behaviors toward their in-
Verbal signal attention 1.2 1.5 6.3 6.2 ***
fants during free play. Verbal praise 1.6 2.1 .6 .9
Maternal verbal behaviors. To investigate cultural dif- Verbal affection 1.2 2.2 1.5 2.2
ferences in mothers’ verbal strategies for organizing Command 6.8 5.3 10.3 8.8
interactions with their infants, a MANOVA with group Suggest 5.7 4.5 2.8 2.9 *
Overall verbal 16.3 8.5 21.4 12.3
(2) as a between factor, and situation (4) and behavior
(5) as repeated factors was performed on the fre- Teaching
Verbal signal attention 14.8 8.0 37.1 15.9 ***
quency of different types of maternal verbal behav- Verbal praise 9.3 5.6 4.8 4.8 *
iors occurring in each of the four situations. This Verbal affection .9 1.9 3.2 4.3 *
MANOVA yielded significant main effects for group, Command 21.7 15.4 34.3 17.9 *
F(1, 37) 5 7.64, p , .01, situation, F(3, 111) 5 95.95, p , Suggest 28.4 9.6 6.9 6.9 ***
.001, and behavior, F(4, 148) 5 58.34, p , .001, as well Overall verbal 75.1 25.0 86.4 30.9
as significant Group 3 Behavior, F(4, 148) 5 32.61, p , Free Play
Verbal signal attention 7.5 4.2 16.0 9.5 ***
Verbal praise 4.8 3.1 1.3 2.0 ***
Table 2 Mean Frequency of Maternal Nonverbal Behaviors by Verbal affection 1.3 2.1 3.0 3.4
Situation Command 9.8 9.4 26.2 16.2 ***
Suggest 19.6 8.8 3.6 2.9 ***
Puerto Overall verbal 43.0 19.4 50.1 23.0
Anglo Rican
* p , .05; ** p , .01; *** p , .001.
Behavior M SD M SD p

Social Play .001, Situation 3 Behavior, F(12, 444) 5 17.32, p ,

Nonverbal signal attention 1.3 3.3 2.7 4.4
.001, and Group 3 Situation 3 Behavior, F(12, 444) 5
Position infant 2.4 1.8 3.7 2.6
Restrain infant’s movements 1.6 1.4 5.1 4.6 *** 13.36, p , .001, interactions.
Nonverbal praise .1 .5 .2 .6 Follow-up one-way ANOVAs (see Table 3) indi-
Nonverbal affection 1.8 2.0 2.6 3.8 cated that, compared to the Anglo mothers, Puerto
Overall nonverbal 7.2 4.7 14.3 9.4 ** Rican mothers were more likely to: (1) signal their in-
Teaching fants’ attention in all four situations, (2) offer affection
Nonverbal signal attention 6.4 4.9 9.1 6.7 during teaching, (3) issue directives in the form of com-
Position infant .9 1.2 2.4 2.3 *
Restrain infant’s movements 1.8 2.2 2.6 2.5
mands in all situations except social play, and (4) emit
Nonverbal praise .4 .9 .7 1.4 a greater overall number of verbal behaviors towards
Nonverbal affection .3 .8 .6 1.0 their infants during free play. Conversely, compared to
Overall nonverbal 9.8 6.6 15.4 6.7 * the Puerto Rican mothers, Anglo mothers were more
Free Play likely to: (1) praise their infants’ efforts in all situations
Nonverbal signal attention 5.7 5.7 5.5 4.8 except social play, and (2) issue directives in the form of
Position infant .9 1.2 1.0 1.5
suggestions in all situations except free play.
Restrain infant’s movements .6 .7 1.3 2.4
Nonverbal praise .2 .4 .7 1.5
Nonverbal affection .6 1.2 .7 1.2 Situation-Specific Maternal Behaviors
Overall nonverbal 8.0 7.0 9.1 6.6
Behaviors specific to feeding. To investigate group dif-
* p , .05; ** p , .01; *** p , .001. ferences in the organization of feeding, a 2 3 3 (Group 3
1012 Child Development

Behavior) MANOVA was performed. The three feed- MANOVA was performed. This analysis yielded a
ing behaviors examined were: mother feeds infant, significant group 3 game-type interaction, F(4, 32) 5
mother encourages autonomy in feeding, and infant 3.3, p , .05. Follow-up one-way ANOVAs (see Table
feeds self. This analysis yielded a significant Group 3 4) indicated that, compared to the Anglo mothers,
Behavior interaction, F(3, 34) 5 38.7, p , .001. As can Puerto Rican mothers were significantly more likely
be seen in Table 4, follow-up one-way ANOVAs indi- to engage in touching/turntaking games with their
cated that Anglo mothers, compared to Puerto Rican infants, p , .05. Group differences for clapping/
mothers, were more likely to encourage autonomy in singing, hiding/chasing, and bouncing games did not
feeding (M 5 7.64 and 1.39, respectively, p , .01), reach significance.
and less likely to directly feed their infants (M 5 8.36 Behaviors specific to teaching. A t test of total time
and 41.94, respectively, p , .001). Concomitantly, spent on-task in the teaching situation yielded a sig-
Anglo infants were more likely than Puerto Rican in- nificant group difference, with Anglo mother-infant
fants to self-feed (M 5 23.82 and 2.78, respectively, pairs spending on average 330.2 s on-task, compared
p , .001). with an average of 194.1 s on-task among the Puerto
Behaviors specific to social play. An ANOVA indicated Rican mother-infant pairs. A by-task breakdown of
no significant group difference in total amount of on-task time indicated that in both groups, infants
time that the Anglo and Puerto Rican mothers spent spent approximately 87% of their time on-task during
playing dyadic social games (clapping/singing; bounc- the crayon portion of the teaching situation, but that
ing; hiding/chasing; touching/turntaking) with their on-task time dropped to an average of 67% among the
infants (M 5 93.1 s and 112.1 s, respectively). Types of Anglo infants, and 35% among the Puerto Rican in-
social games played were thus analyzed according to fants for the block and pushtoy portions. Moreover, a
number of seconds mothers spent engaged or at- t test of total time spent on the teaching situation (see
tempting to engage their infants in each type of game. Table 4) indicated that Anglo mothers persisted with
To investigate group differences in types of social this situation longer than did Puerto Rican mothers
games played, a 2 3 4 (Group 3 Game-Type) (M 5 459.4 s and 371.2 s, respectively, p , .05).
Behaviors specific to free play. To investigate group
differences in the coordination of mother-infant activ-
Table 4 Mean Frequency of Behaviors, and Duration of Set- ity during free play, a 2 3 5 (Group 3 Activity)
tings Specific to Each Situation MANOVA was performed. The five activities exam-
ined were: (1) mother and infant play together with
Puerto
object, (2) infant plays alone while mother watches,
Anglo Rican
(3) mother attempts to shift infant’s attention to a new
Behavior/Setting M SD M SD p toy while infant is still playing with another toy, (4)
mother attempts to gain infant attention while infant
Feeding is unoccupied, and (5) no play activity. This analysis
Mother spoon-feed infant 8.4 8.7 41.9 11.5 *** yielded a significant Group 3 Activity interaction,
Mother encourage autonomy 7.6 8.3 1.4 2.9 **
Infant feed self 23.8 12.6 2.8 4.3 ***
F(4, 152) 5 5.25, p , .01. Follow-up one-way ANOVAs
(see Table 4) indicated that, compared to the Anglo
Social Play
Hiding/chasing games 35.1 40.4 15.4 31.8 mothers, Puerto Rican mothers spent (1) more time
Clapping/singing games 15.6 28.6 31.3 32.5 attempting to shift or gain their infants’ attention, p ,
Bouncing games 22.2 25.6 15.6 21.8 .05, and (2) less time watching while their infants
Touching/turntaking games 20.2 21.5 49.8 46.7 * played alone, p , .01.
Teaching
Time on-task crayon 135.8 38.5 107.8 38.1 *
Time on-task blocks 98.4 49.8 47.2 37.8 ** DISCUSSION
Time on-task pushtoy 96.1 51.8 39.1 33.0 ***
Total time on-task 335.5 77.9 204.7 80.9 *** Consistent with a broad characterization of Anglo
Total time in situation 459.4 128.9 371.2 82.3 * cultural belief systems as more individualistic, and
Free Play Puerto Rican cultural belief systems as more socio-
Play together 190.6 85.8 181.9 73.7 centric, the Anglo mothers in this study were more
Infant play, mother watch 207.7 92.9 125.3 57.9 ** likely to generate long-term socialization goals falling
Attempt to shift attention 7.6 15.9 25.4 29.7 *
into the combined categories of Self-Maximization
Attempt to gain attention 2.3 10.7 12.0 18.4 *
No play activity 37.5 37.6 61.9 50.7 and Self-Control, whereas the Puerto Rican mothers
were more likely to generate long-term socialization
* p , .05; ** p , .01; *** p , .001. goals falling into the combined categories of Proper
 Harwood et al. 1013

Demeanor and Decency. Moreover, when asked what Puerto Rican mothers were more likely than Anglo
parents could do to encourage or discourage these mothers to structure their infants’ behavior directly.
qualities in their children, mothers were likely to gen- This was evident in all four situations and expressed
erate childrearing strategies that corresponded to in a variety of verbal and nonverbal maternal behav-
these broad cultural constructs. In particular, Anglo iors, including greater tendencies to physically posi-
mothers were more likely to say that they would tion and restrain the infant, to signal the infant’s atten-
model positive behaviors and provide their children tion and issue direct commands, and to spoon-feed
with opportunities that allow them to learn for them- the infant. Conversely, consistent with a greater em-
selves, whereas Puerto Rican mothers were more phasis on individualism, Anglo mothers were more
likely to say that they would teach their children likely than Puerto Rican mothers to praise the infant
through the direct exercise of parental authority. verbally, to attempt to structure the infant’s behavior
These results suggest a greater emphasis among the indirectly through the use of suggestions, and to en-
Anglo mothers on childrearing strategies that indi- courage autonomy in feeding.
rectly structure the child’s learning experiences, thus It could be argued that Puerto Rican and Anglo
optimizing the child’s own sense of autonomy and mothers’ behavioral differences reflect temperamen-
personal choice; and a greater focus among the Puerto tal differences in the infants. In particular, Puerto
Rican mothers on childrearing strategies that directly Rican mothers might be more directive because their
guide the child’s learning experiences, thus highlight- infants are either more or less active than Anglo in-
ing the child’s sense of interpersonal obligation. fants. There were few group differences in infant be-
Of course, it should be noted that statements fall- havior, however, and those that arose did not form a
ing into the categories of Proper Demeanor and Direct coherent pattern. These findings do not suggest a pat-
Instruction were not absent among the Anglo mothers, tern whereby either group of infants appears to be
any more than statements falling into the categories more or less active than the other group. It thus seems
of Self-Maximization and Provide Opportunities unlikely that differences in maternal behavior are ex-
were absent among the Puerto Rican mothers. What plicable as responses to preexisting group differences
is significant is not the presence or absence of each of in the temperament or activity level of Puerto Rican
these response categories, but their relative patterning versus Anglo infants.
across the two groups. Long-term socialization goals It also should be noted that group commonalities
and childrearing strategies that emphasize the individ- in mother-infant interactions also occurred. Consistent
ual’s own agency were foregrounded for the Anglo with the finding of no group differences in emphasis
mothers, whereas goals and strategies that focus on the on Lovingness as a socialization goal, or in Providing
individual’s obligations and connectedness to others an Optimal Emotional Environment as a childrearing
were foregrounded for the Puerto Rican mothers. strategy, there were generally no group differences in
Group commonalities in socialization goals and mothers’ physical or verbal affection to their infants.
strategies also emerged. In particular, there were no There was one exception: Puerto Rican mothers were
group differences in the use of adjectives falling into more likely than Anglo mothers to show verbal affec-
the category of Lovingness when describing long- tion to their infants during teaching.
term goals, and no significant differences in the use of The finding of Puerto Rican mothers’ greater ver-
Provide Optimal Emotional Environment and Praise bal affection during teaching is interesting in light of
as childrearing strategies. These findings suggest the lesser tendency of Puerto Rican infants to be on-
that the capacity for emotional intimacy, and the desire task during this situation, and to spend less time
to provide the child with an emotionally supportive overall on the teaching tasks. Group differences in on-
environment, is equally valued among mothers in task time during teaching can be interpreted in at
both groups. Where the two groups differ is in the least two possible ways: (1) The nondirective meth-
relative emphasis placed on dimensions consonant ods used by the Anglo mothers are more effective at
with the broad cultural constructs of individualism sustaining infant attention in these tasks than the
and interdependence. more directive methods used by the Puerto Rican
Greater differential emphases on individualism mothers, or (2) the teaching tasks were less salient
and interdependence also were evident when exam- and meaningful for the Puerto Rican dyads than they
ining mother-infant interactions across the four every- were for the Anglo dyads. The latter interpretation is
day situations, providing evidence of coherence be- supported by the fact that the proportion of infant
tween mothers’ childrearing beliefs, and the ways in on-task times for the two groups was roughly equiv-
which they chose to structure interactions with their alent at around 86% during the crayon task (which
infants in four everyday situations. In particular, most infants this age find inherently interesting), but
1014 Child Development

dropped dramatically for the two remaining tasks to was unique among the four situations in that it did
around 67% for the Anglo infants, and to around 35% not yield any group differences in maternal nonverbal
for the Puerto Rican infants. In other words, both behavior. Group differences in mothers’ verbal behav-
groups of infants found the crayon task to be rela- iors, however, showed considerable cross-situational
tively more interesting than the other two tasks, but consistency. In particular, compared to Anglo mothers,
this was especially true for the Puerto Rican infants, Puerto Rican mothers were more likely to signal the in-
and this may be related to the relative lack of persis- fant’s attention and to use commands, and less likely to
tence on these tasks among the Puerto Rican dyads. It praise the infant or to use suggestions.
is conceivable that the teaching tasks were perceived These findings point to maternal responsiveness to
as less age-appropriate and less interesting by the differential situational demands among both groups
Puerto Rican mother-infant pairs than they were by of mothers. It is noteworthy, however, that the signif-
the Anglo dyads; the Puerto Rican mothers’ greater icant differences that occurred were in the expected
use of verbal affection during this situation may indi- direction. In other words, the differences were consis-
cate that they viewed themselves as trying to coax tent in all four situations with a greater degree of di-
their infants to do something that was difficult or un- rect structuring of the infant’s behavior among the
interesting. This possibility highlights the importance Puerto Rican mothers, and a greater offering of per-
of employing culturally meaningful tasks, as well as sonal choice to the infant among the Anglo mothers.
the importance of looking at mother-infant interac- What differed by situation was the specific strategy
tions across a variety of situations when attempting that mothers chose to communicate either greater di-
to understand cultural differences. rectiveness or greater choice.
As suggested by the example above, situational Together, these findings suggest that situational
variability did occur in mother-infant interaction, variability occurs in the instantiation of broad cul-
manifesting itself in at least three different ways: (1) tural constructs. This variability remained coherent in
Overall frequencies of maternal nonverbal and verbal this study with larger cultural goals, however, ap-
behaviors differed according to the situation, (2) some pearing to consist primarily of different manifesta-
behaviors showed significant group differences in tions of the same concern, or of situationally specific
one situation but not in another, and (3) certain be- interactional opportunities. Although mothers did
haviors and activities were unique to each situation not behave identically across all four situations, they
(such as types of social games played, or spoon-feeding did behave in ways that were consistent with cultural
versus self-feeding). goals and expectations. These findings add to our un-
In terms of overall activity levels, both groups of derstanding of culture’s role in child development by:
mothers were more verbally active during teaching (1) demonstrating coherence among long-term social-
and free play than they were during feeding and so- ization goals, maternal beliefs regarding strategies for
cial play. Although some of these differences reflect realizing those goals, and the structuring of mother-
differential lengths of observation, observation time infant interactions in four everyday situations; (2)
alone cannot account for these differences. For in- providing evidence for the usefulness of ‘individual-
stance, even if all social play maternal frequencies are ism’ and ‘sociocentrism’ as general heuristic devices
doubled, they still do not reach the levels exhibited by for understanding broad cultural differences in the
mothers during the teaching tasks. Mothers in both context of child development; and (3) suggesting that
groups thus appear to be differentially responsive to more localized, intragroup variations in childrearing
the demands of each situation. Significantly, however, beliefs and practices need to be understood within the
when group differences arose in maternal behavioral context of broader cultural belief systems, goals, and
frequencies, they were in the direction of the Puerto values. As Wiley, Rose, Burger, and Miller (1998) note,
Rican mothers being more active, particularly with an appreciation of the “subtle patterns of similarities
regard to nonverbal behaviors. and differences within and across cultural groups . . . is
Variability also arose in terms of mothers’ differen- essential if we are to move beyond dichotomous com-
tial selection of strategies for organizing each situa- parisons of sociocultural groups” (p. 844).
tion. This was particularly true for nonverbal behav- These findings, though suggestive, are limited by a
iors. For instance, Puerto Rican mothers were more small sample size of just 40 mothers, and concomi-
likely to restrain their infant’s free movements about tantly lowered reliability coefficients for maternal be-
the room during social play, but not during teaching haviors, two of which fell below .60. Because reliabil-
or free play; similarly, they were more likely to phys- ity coefficients, along with group differences, become
ically position their infants during teaching, but not more robust with sample size, the findings presented
during social play or free play. Moreover, free play here bear further study and replication with a larger
 Harwood et al. 1015

number of mothers. In addition, future research in Mansfield Road, Storrs, CT 06269; e-mail: harwood@
this domain would benefit from a closer examination uconnvm.uconn.edu. Axel Schoelmerich is currently
of patterns of within-group variability, and the ways at the Ruhr University in Bochum, Germany; Pamela
in which local variations can be understood as situ- A. Schulze is also at the University of Connecticut;
ated within broader cultural goals and values. Finally, and Zenaida Gonzalez is at the University of Puerto
as Super and Harkness (1986) have noted, parents’ Rico School of Medicine, San Juan.
cultural beliefs work in concert with other important
dimensions of everyday life to shape childhood’s
context. Further research in this domain needs to con-
tinue to examine the multiple paths through which REFERENCES
parents create settings in which children come to con- Aitchison, J. (1986). The statistical analysis of compositional
struct and reconstruct their own cultural understand- data. London: Chapman & Hall.
ings of self and others. These include not only paren- Bersoff, D. M., & Miller, J. G. (1993). Culture, context, and
tal beliefs, but also other aspects of the child’s setting, the development of moral accountability judgments. De-
such as number and type of social contacts, and velopmental Psychology, 29, 664–676.
amount of time spent engaged in a variety of activi- Cole, M. (1996). Cultural psychology: A once and future disci-
ties throughout the day (Leyendecker, Lamb, Schölm- pline. Cambridge, MA: Harvard University Press.
erich, & Fracasso, 1995). Corsaro, W. A., & Miller, P. J. (Eds.). (1992). Interpretive ap-
Rogoff et al. (1993, p. 1) stated that “development proaches to children’s socialization. New Directions for Child
occurs through active participation in cultural sys- Development, 58. San Francisco: Jossey-Bass.
Derne, S. (1995). Culture in action: Family life, emotion, and
tems of practice in which children, together with their
male dominance in Banaras, India. Albany: State University
caregivers and other companions, learn and extend
of New York Press.
the skills, values, and knowledge of their commu- Diaz Royo, A. (1974). The enculturation process of Puerto
nity.” Consistent with this, our findings suggest that Rican highland children. Dissertation Abstracts Interna-
children engage repeatedly in situationally specific in- tional, 35-12A, 7646. (University Microfilms No. 75-
stantiations of larger cultural values in a variety of 12506,324).
everyday settings. Situationally appropriate behavior Dumas, J. (1993). Interact Software System, v. 2.0. Lafayette,
may itself be culturally patterned, further illustrating IN: Purdue University.
the ways in which homogeneity and heterogeneity can Goodnow, J. J., Miller, P. J., & Kessel, F. (Eds.). (1995). Cul-
coexist coherently within a larger cultural belief system. tural practices as contexts for development. New Directions
for Child Development, 67. San Francisco: Jossey-Bass.
Greenfield, P. M., & Cocking, R. R. (Eds.). (1994). Cross-
ACKNOWLEDGMENTS cultural roots of minority child development. Hillsdale, NJ:
Erlbaum.
This research was made possible through a grant to
Harkness, S., & Super, C. M. (Eds.). (1996). Parents’ cultural
the first author from the National Institute of Child
beliefs sytems: Their origins, expressions, and consequences.
Health and Human Development (HD32800). We also New York: Guilford.
are grateful to Glorisa Canino of the University of Harwood, R. L. (1992). The influence of culturally derived
Puerto Rico School of Medicine for making available values on Anglo and Puerto Rican mothers’ perceptions
the resources of her laboratory for the collection of the of attachment behavior. Child Development, 63, 822–839.
Puerto Rican data. We would like to express appreci- Harwood, R. L., Miller, J. G., & Lucca Irizarry, N. (1995).
ation to Eugenio Ayala, Delia Collazo, Helena Mendez, Culture and attachment: Perceptions of the child in context.
and Stephanie Wilson, for their assistance with inter- New York: Guilford.
viewing, translation, and data coding. We also would Harwood, R. L., Schoelmerich, A., Ventura-Cook, E.,
like to thank Penn Handwerker for many lively discus- Schulze, P. A., & Wilson, S. P. (1996). Culture and class
sions concerning the conceptualization of culture. Por- influences on Anglo and Puerto Rican mothers’ beliefs
tions of this paper were presented at the biennial meet- regarding long-term socialization goals and child be-
havior. Child Development, 67, 2446–2461.
ings of the Society for Research in Child Development
Hollingshead, A. B. (1975). Four Factor Index of Social Status.
in Washington, DC, April, 1997, and of the International
Unpublished manuscript, Yale University, New Haven,
Society for Infant Studies in Providence, RI, April, 1996. CT.
Kessen, W. (1979). The American child and other cultural
ADDRESSES AND AFFILIATIONS inventions. American Psychologist, 34, 815–820.
Kurtz, S. N. (1992). All the mothers are one: Hindu India and the
Corresponding author: Robin L. Harwood, School of cultural reshaping of psychoanalysis. New York: Columbia
Family Studies, University of Connecticut, U-58, 348 University Press.
1016 Child Development

Lauria, A. (1982). Respeto, relajo, and interpersonal relations and the lure. In R. Jessor, A. Colby, & R. A. Shweder
in Puerto Rico. In F. Cordasco & E. Bucchioni (Eds.), The (Eds.), Ethnography and human development: Context and
Puerto Rican community and its children on the mainland meaning in social inquiry. Chicago: University of Chicago
(2nd ed., pp. 58–71). Metuchen, NJ: Scarecrow Press. Press.
Leyendecker, B., Lamb, M. E., Schölmerich, A., & Fracasso, Shweder, R. A., & Bourne, E. J. (1984). Does the concept of
M. P. (1995). The social worlds of 8- and 12-month-old the person vary cross-culturally? In R. A. Shweder, &
infants: Early experiences in two subcultural contexts. R. A. LeVine (Eds.), Culture theory: Essays in mind, self,
Social Development, 4, 194–208. and emotion (pp. 158–199). Cambridge, UK: Cambridge
Markus, H. R., & Kitayama, S. (1991). Culture and the self: University Press.
Implications for cognition, emotion, and motivation. Spence, J. T. (1985). Achievement American style: The re-
Psychological Review, 98, 224–253. wards and costs of individualism. American Psychologist,
Miller, P. J., Fung, H., & Mintz, J. (1996). Self-construction 40, 1285–1295.
through narrative practices: A Chinese and American Super, C. M., & Harkness, S. (1986). The developmental
comparison of early socialization. Ethos, 24, 237–280. niche: A conceptualization at the interface of child and
Mines, M. (1994). Public faces, private voices: Community and culture. International Journal of Behavioral Development, 9,
individuality in South India. Berkeley: University of Cali- 545–569.
fornia Press. Tobin, J. J., Wu, D. Y. H., & Davidson, D. H. (1989). Preschool
Nucci, L. (1994). Mothers’ beliefs regarding the personal do- in three cultures: Japan, China, and the United States. New
main of children. In J. G. Smetana (Ed.), Beliefs about Haven, CT: Yale University Press.
parenting: Origins and developmental implications, New Di- Triandis, H. C., Bontempo, R., Villareal, M. J., Asai, M., &
rections for Child Development, 66 (pp. 81–97). San Fran- Lucca, N. (1988). Individualism and collectivism: Cross-
cisco, CA: Jossey-Bass. cultural perspectives on self-ingroup relationships. Jour-
Rogoff, B., Mistry, J., Goncu, A., & Mosier, C. (1993). Guided nal of Personality and Social Psychology, 54, 323–338.
participation in cultural activity by toddlers and care- Triandis, H. C., Marin, G., Lisansky, J., & Betancourt, H.
givers. Monographs of the Society for Research in Child De- (1984). Simpatia as a cultural script of Hispanics. Journal
velopment, 58(Serial No. 236). of Personality and Social Psychology, 47, 1363–1375.
Sampson, E. E. (1989). The challenge of social change for Wainryb, C. (1995). Reasoning about social conflicts in dif-
psychology: Globalization and psychology’s theory of ferent cultures: Druze and Jewish children in Israel.
the person. American Psychologist, 44, 914–921. Child Development, 66, 390–401.
Schieffelin, B. B., & Ochs, E. (Eds.). (1986). Language social- Wiley, A. R., Rose, A. J., Burger, L. K., & Miller, P. J. (1998).
ization across cultures. Cambridge, UK: Cambridge Uni- Constructing autonomous selves through narrative
versity Press. practices: A comparative study of working-class and
Shweder, R. A. (1996). True ethnography: The lore, the law, middle-class families. Child Development, 69, 833–847.