Journal of Family Psychology 2007, Vol. 21, No.

4, 655 664

Copyright 2007 by the American Psychological Association 0893-3200/07/$12.00 DOI: 10.1037/0893-3200.21.4.655

Social Experiences and Daily Routines of African American Infants in Different Socioeconomic Contexts
Hillary N. Fouts
University of Tennessee

Jaipaul L. Roopnarine
Syracuse University

Michael E. Lamb
University of Cambridge
Sixty-two 3- to 4-month-old African American infants from lower, middle, and upper socioeconomic status (SES) backgrounds were each observed in naturalistic contexts for 12 hr. The social experiences of infants in the 3 groups were similar in many ways: Infants from all backgrounds slept and were vocalized to for similar amounts of time. However, infants in the upper SES families engaged in more self-play, vocalized less, fussed less, had fewer but longer naps, and fewer but longer bouts of social interaction than did infants in the middleand lower SES families. Infants in the upper SES families also received more verbal affection and soothing responses to their fussing and crying than did the other infants, whereas infants in the lower SES families interacted more with extended kin than did infants in the upper SES families. These results underscore the need to study African American families in a variety of socioeconomic contexts because families in more advantaged circumstances may greatly differ from those who are more disadvantaged, especially in terms of reliance on extended kin as caregivers. Keywords: infant behavior, social interactions, extended kin, SES, African American families

Little is known about socioeconomic differences in early infant experiences. The purpose of this study was to examine how the everyday experiences and social behavior of 3to 4-month-old African American infants varied, depending on their families socioeconomic status (SES), by naturalistically observing infants and caregivers in three different SES contexts (low, middle, and upper). Most previous studies of African American children have tended to focus disproportionately on disadvantaged families (Graham, 1992), even though substantial numbers of African American families live in middle- and upper SES contexts. One earlier study revealed more heterogeneity among African American parents than had previously been recognized (Roopnarine, Fouts, Lamb, & Lewis-Elligan, 2005), but few

Hillary N. Fouts, Department of Child and Family Studies, University of Tennessee; Jaipaul L. Roopnarine, Department of Child and Family Studies, Syracuse University; Michael E. Lamb, Faculty of Social and Political Sciences, University of Cambridge, Cambridge, England. This study was supported by funds from the National Institute of Child Health and Human Development awarded to Michael E. Lamb. We are grateful to the families who participated in the present study. Correspondence concerning this article should be addressed to Hillary N. Fouts, Department of Child and Family Studies, University of Tennessee, 1215 West Cumberland Avenue, Knoxville, TN 37996-1912. E-mail: hfouts@utk.edu

consistent SES differences were evident when the parents gender was controlled in the analyses. Contrary to the widespread belief that parents in low-SES families are not adequately involved with their children (Magnusson & Duncan, 2002), parents spent similar amounts of time with their infants in all SES groups (Roopnarine et al., 2005). Building on Roopnarine et al.s (2005) study of parenting behavior, we focused in this study on infant behaviors and interactions with all social partners, not only parents. Specically, we examined variations in the structure of infants awake and asleep periods, responses to the infants fussing and crying, infants social engagement and the structure of their dyadic social interactions, and the identity of the infants social partners. Like other researchers advocating cultural-contextual and adaptive-resilient perspectives, we assumed that childrearing and family patterns evolve over time to meet the demands of specic sociocultural and economic ecologies (Low, 2005; Ogbu, 1981; Super & Harkness, 1997). Instead of assuming that families with particular characteristics (e.g., single-parent, low-income) are decient, therefore, we presume that they have adapted to particular sociocultural and economic ecologies. In this study, we examined not only how infants were treated but also how infants behaved in these differing ecologies. Ogbu (1981) proposed that researchers interested in cross-cultural comparisons should rst study within-ethnic group variations in behavior across a variety of contexts to ensure that subsequent cross-cultural comparisons avoid
655

656

FOUTS, ROOPNARINE, AND LAMB

representing one group as decient relative to another group. Studies that examine a cultural or ethnic group in a range of SES contexts facilitate cross-cultural or majorityminority comparisons, and we thus focused on African American families in different socioeconomic contexts rather than on different ethnic groups. In line with the cultural-contextual perspective, Lareau (2003) conducted an ethnographic study of childhood and social class among U.S. families from diverse ethnic backgrounds. Lareau concluded that upper-middle-class parents deliberately tried to stimulate their childrens development and foster their cognitive and social skills (p. 5), whereas lower- and working-class parents placed more emphasis on providing comfort, food, and shelter. In upper-middle-class homes, for example, children had very structured days and were often directed to engage in activities that fostered new skills even during play time, whereas children in lower- and working-class families had more unstructured and unsupervised periods during which they could choose what to do and also spent more time interacting with extended kin. Lareau (2003) suggested that these differences are best viewed as transmitting differential advantages to children (p. 5) rather than as being better or worse than one another. The upper-middle-class contexts may foster success in schools, but children from lower- and working-class families may be more comfortable making independent decisions and have a better understanding of social and kin networks. Although SES may have substantial effects on childrearing strategies, other scholars have suggested that African American families have child-rearing practices that are distinct from other minorities and European Americans (Bartz & Levine, 1978; Heath, 1989; Thornton, Chatters, Taylor, & Allen, 1990). For example, Bartz and Levine (1978) exemplied how lower- and working-class African American parents had different child-rearing expectations compared with Latino and European American parents in lower- and working-class contexts. Specically, Bartz and Levine found that the African American parents in their study expected early autonomy with respect to childrens bodily functions and emotions, valued both high support and control over childrens behavior, and emphasized egalitarian family roles. Although interest in the development of children from minority groups has grown in the last decade (Quintana et al., 2006), few studies have focused on infants. This is especially problematic because early experience has a considerable impact on later socioemotional development (Fraley, 2002). Three- to 4-month-olds are especially interesting because their behavior is increasingly characterized by discriminating sociabilitythe emergence of preferences for different social partners. This age is also marked by more coordinated behavior, less variable levels of arousal, and longer periods of time awake when they are alert and engaged in social interactions (Lamb, Bornstein, & Teti, 2002). The review of relevant literature that follows reects the paucity of information about African American infants by focusing on studies involving children, rather than infants.

Most researchers have focused on parents (especially mothers), and so far less is known about infant interactions with nonparental kin. Nonparental kin provide a great deal of infant care in many cultures throughout the world (Harkness & Super, 1992; Rogoff, Mistry, Goncu, & Mosier, 1991), and such kin commonly help care for older African American children (Early & Burchinal, 2001; Jackson, 1993; Uttal, 1999). We sought to determine whether this was also true with respect to infants, not only to see how early in life kin became involved in child care but also to identify possible socioeconomic differences in the reliance on kin care and the effects of varying kin involvement on early social experiences. The effects of SES on levels of social interaction and verbal stimulation in infancy are unclear, although Brody and Flor (1997) proposed that nancial resources directly affect maternal state (such as depressive symptoms and self-esteem), which, in turn, affects mother child harmony. Financial resources should also affect the behavior of caregivers other than mothers as well as the everyday routines of families. Several researchers have suggested that high levels of verbal stimulation are a distinctive characteristic of parent child interaction in upper socioeconomic contexts regardless of ethnicity (Gottfried, 1984; Hart & Risley, 1995; Hoff, 2003), but SES does not predict levels of verbal stimulation in all ethnic groups (Leyendecker, Lamb, Scholmerich, & Fricke, 1997). For example, levels of ma ternal vocalization to infants in low-, middle-, and upper SES African American families did not differ signicantly (Roopnarine et al., 2005). Likewise, Leyendecker et al. (1997) did not nd SES differences in overall levels of infant and maternal vocalization in Costa Rican families, although there were SES differences in specic contexts (e.g., feeding, caretaking, and social interactions). Overall, few conclusions can be drawn about the effects of SES and ethnicity on levels of verbal stimulation and interaction because researchers have tended to focus narrowly on mother child dyads rather than on overall social experiences and have tended to confound SES and ethnicity (Graham, 1992). Feiring and Lewis (1981) reported that European American 3-month-old infants in upper-middleclass families vocalized more frequently and were more often engaged in dyadic social interactions with their mothers than did European American infants in middle-class families. These ndings suggest that socioeconomic circumstances may inuence infants as young as 3 months of age, but further research carefully distinguishing between the effects of ethnicity and SES is clearly needed. There has been more research on responses to fussing and crying than on responses to other infant behaviors. In their recent analyses of the National Institute of Child Health and Human Development Early Childcare Research Network data set, Bakermans-Kranenburg, van IJzendoorn, and Kroonenberg (2004) concluded that the predictive associations between maternal sensitivity and attachment security were evident in both African American and European American contexts. Poverty appeared to impede maternal sensitivity to fussing and crying, however. Perhaps as a result, African American children had substantially lower

SOCIAL EXPERIENCES OF AFRICAN AMERICAN INFANTS

657

Attachment Q-sort scores than European American children because African American children were overrepresented in lower SES groups. In the present study, we focused on responses by all caregivers rather than only on maternal sensitivity, but emphasized responses to fussing and crying because parental sensitivity to fussing and crying has been shown to affect later socioemotional development (Ainsworth, Blehar, Waters, & Wall, 1978; de Wolff & van IJzendoorn, 1997). Although Bakersmans-Kranenburg and colleagues (2004) suggested that motherinfant attachments might develop differently in African and European American families because multiple caregiving networks are more common in the former, socioeconomic variations (and their possible impact on socioemotional development) remain unexplored. There is emerging evidence that family functioning and relationships with other caregivers may inuence the quality of parentinfant attachment (Caldera & Lindsay, 2006; Cowan, 1997), and although we did not explore the effects on attachment, we did examine variations in responsiveness when there were multiple caregivers. As is the case with verbal stimulation and social interactions, less attention has been paid to the sleeping routines of infants with different SES backgrounds. The daily routines of infants revolve predominantly around their sleepwake cycles so that napping patterns are central to their daily routines. There is some evidence that sleep-wake cycles differ across SES contexts. Leyendecker, Lamb, Scholmer ich, and Fracasso (1995) reported that the circadian distribution of sleep among middle-class European American infants reected an emphasis on schedule, whereas infants in small-scale societies characterized by low levels of education and nancial resources are often allowed to sleep and breastfeed whenever they choose (e.g., Hewlett & Lamb, 2005; Super & Harkness, 1982). Research on middle-class European American infants daily routines may not accurately reect the experiences of infants in other ethnic groups in the United States. In the present study, we attempted to ll substantial gaps in the literature on child-rearing and infant social development by examining African American families in a broad range of socioeconomic contexts. We also attempted to represent infant social interactions with all caregivers, rather than focusing only on motherinfant interactions, and because so little is known about the associations between SES and infants daily routines, we documented infants sleepwake cycles. Pursuant to these concerns, we asked three questions: (a) Do 3- to 4-month-old African American infants behave differently and have different daily sleep routines depending on their socioeconomic context? (b) Do the infants social experiences and the extent to which they are engaged in social interactions vary depending on their SES? and (c) Do the number and types of individuals interacting with infants vary systematically depending on the socioeconomic context? On the basis of the few studies of infant behavior and daily routines, we expected infants in the higher SES contexts to have more sustained periods of social interaction and wakefulness (Leyendecker et al., 1995). Even though

many researchers have reported that European Americans in higher SES contexts tend to talk to their infants and children more than in lower SES contexts (Brody & Flor, 1997; Gottfried, 1984; Hart & Risley, 1995; Hoff, 2003), we did not predict signicant differences because some studies have shown that SES does not always predict verbal interaction with infants and children, especially among African Americans (Roopnarine et al., 2005). On the basis of the numerous reports that families in lower SES contexts often have extended kin both living in their households and playing prominent roles in the care of older children (Early & Burchinal, 2001; Jackson, 1993; Uttal, 1999), we expected that infants in the lower SES contexts would have more interactions with nonparental kin than infants in the other groups.

Method Participants
Sixty-two families from three different socioeconomic backgrounds participated in this study. Each parent selfidentied as African American. None of the parents were recent immigrants to the United States, and English was the primary language spoken in each household. All of the families had infants between 3 and 4 months of age. The families were recruited through advertisements in neighborhood newspapers, churches, health clinics, day care centers, and by word-of-mouth. Participating families came from different neighborhoods in a midsize northeastern city and were paid $60 for their participation. SES was assessed using indices of educational attainment, income, and occupation (Entwisle & Astone, 1994), and 20 of the participating families were categorized as lower SES, 21 as middle SES, and 21 as upper SES. Table 1 describes features of the families, including details about employment, yearly income, and household composition. The lower SES families met the federal criteria for living in poverty, and the parents had obtained less education than had the middle- and upper SES parents. In fact, the majority of parents in the lower SES group had either attended high school but not nished or had nished high school and taken some college courses (25% and 60%, respectively); very few parents had college degrees (7.5%). In the lower SES group, parents who were employed typically worked in semiskilled jobs such as factory work, fast food restaurants, claims agents, delivery, hair salons, and waiters at restaurants. Seventy-ve percent of these families made less than $10,000 annually. The average salary range for the middle SES families was higher ($35,000 $55,000), and the majority of parents had taken some college classes or had graduated from college (42.9% and 35.7%, respectively). Parents in the middle SES group were typically employed as managers, mechanics, drivers, engineers, sales and customer service representatives, tness trainers, day care providers, and nurses. Upper SES families had the highest salaries of the three groups (average of above $80,000), and the majority of parents had college degrees or postgraduate degrees (33.3% and 57.1%, respectively). The majority of

658

FOUTS, ROOPNARINE, AND LAMB

Table 1 Characteristics of Families

Characteristic Female infant Male infant % of rstborn infant % of infants with at least one sibling Mothers age (mean) Fathers age (mean) % father resident % both parents employed % father-only employed % mother-only employed No employment Average yearly income rangea
a

Lower SES (n 20) 10 10 35 65 23.05 25.58 20 30 30 15 25 below $10,000

Middle SES (n 21) 12 9 19.05 80.95 30.29 34.25 76 42.86 42.86 14.29 0 $35,00$55,000

Upper SES (n 21) 12 9 42.86 57.14 34.86 37.26 86 42.86 47.62 14.29 0 Above $80,000

Note. SES socioeconomic status. Participants reported their income range rather than their exact yearly income.

parents in the upper SES group were employed in professional positions such as physicians, lawyers, teachers, school principals, and heads of companies. In the lower SES group, none of the mothers and fathers were married to each other at the time of observation, whereas 61.9% of parents in the middle SES group and 85.7% of parents in the upper group were married to each other. Very few fathers in the lower SES group lived in the household with their infants. Mothers and fathers also tended to be older in the middle- and high-SES groups. In each group, there were few rstborn infants.

Procedure
Each focal infant was observed in unstructured circumstances for 3 hr on four different days (8 11 a.m., 11 a.m.2 p.m., 2 6 p.m., 6 8 p.m.) so that each infant was observed over a total of 12 hr. The observations were conducted by ve adult graduate and undergraduate students who each self-identied as African American. Prior to the rst observation session, the observer visited the family to explain and answer questions about the procedures. During this visit, demographic data were collected. Families were told that during observations, they were free to go about their normal activities and schedules. Mothers were not required to take leave from work or to be present during the observations. However, most mothers were present and not working outside of the home during observations. This was most likely an artifact of the young age of the infants. The behavioral coding protocol was based on that developed by Belsky, Gilstrap, and Rovine (1984) and has been used to observe 3- to 4-month-old infants in many different cultural and ethnic contexts in Central Africa, Central America, Germany, Canada, and the United States (e.g., Fracasso, Lamb, Scholmerich, & Leyendecker, 1997; Hewlett, Lamb, Shannon, Leyendecker, & Scholmerich, 1998; Leyendecker et al., 1997; Roopnarine et al., 2005). Using a 20-s observe and 10-s record time sampling technique, observers used a checklist to record the occurrence of

infant behaviors; behaviors directed at the infants (all individuals were coded); dyadic behaviors; the location, room, or space in which the infant was located; and the identity of all those who were present. Each 20-s observation is referred to throughout the remainder of the article as an interval of observation. Each 3-hr observation period was broken up into three segments containing a 45-min observation period followed by a 15-min break. Each observe (20 s) and record (10 s) segment was indicated through an earphone connected to a tape recorder worn by the observer. Before conducting live observations, observers were trained in the observational technique for 3 weeks by observing 45-min video recordings of 3- to 4- month-old African American infants in their homes. Observers were only permitted to begin eld observations once they agreed with one another regarding at least 90% of the occasions on which each behavioral code was used. Interobserver reliability was reassessed intermittently during 10% of eld observations, and Cohens kappa was used to calculate interobserver reliability. For identifying caregivers and social partners, the observers agreed 100% of the time during reliability checks. Cohens kappa coefcients were as follows: infant smiles .63, infant vocalizes .72, infant plays with self .91, infant fuss .83, infant cry .84, infant sleep .82, vocalization to infant .85, stimulate/ arousal of infant .72, verbal affect .81, physical affect .63, verbal soothe .88, and physical soothe .84. Infant smile and physical affect have relatively low reliability scores. This is perhaps due to the co-occurrence of these behaviors with other behaviors that may have obscured demarcation of smiling and physical affect. For example, physical affect often co-occurred with verbal affect, and infant smile often occurred with social engagement behaviors. In this study, six infant behaviors, eight behaviors directed toward infants, and instances of dyadic social interactions between infants and others were examined (see Table 2 for denitions).

SOCIAL EXPERIENCES OF AFRICAN AMERICAN INFANTS

659

Table 2 Denitions of Behavioral Codes

Behaviors Infant smiles Infant vocalizes Infant plays with self (self-play) Denitions The infant smiled while awake or in response to external stimulation. Smiles during sleep were not coded. By denition, infant smiles could not co-occur with sleep, fuss, or cry. The infant was awake and made sounds or vocalized, but vocalizations reecting irritable states, nonlinguistic vocalizations (hiccups, coughs, sneezes), or sounds clearly due to physical effort or noises produced while child was sucking were not coded. The infant paid attention or entertained her- or himself or distracted her- or himself with objects she or he could look at or touch with parts of her or his own body, or the infant played with her or his body without objects. Infant play could not co-occur with sleep, fuss, or cry. The infant was awake and showed signs of agitation or upset; she or he was bothered and emitted moans, whines, and whimpers but did not cry. The infant was awake and made vocalizations, gestures, and motor behaviors that left no doubt that the infant was crying. This variable was not coded if the infant was only whimpering, complaining, or restless. The infant was sleeping with eyelids shut. A person vocalized to the infant using short, simple dialogue, vocalizing slowly and baby talking some words so that the infant would pay attention or talk normally. The person could make animal sounds for the child, sing to her or him, or change the intonation as a way of calling or drawing her or his attention. Action by which a person attempted to draw or maintain the infants attention on an event in order to maintain alertness or the level of positive affect. The person might touch the infant with a nger to call attention, play with her or him, or make other efforts to stimulate the infant: For example, making him jump on her lap; moving him, pulling limbs, etc. Stimulation was not coded if the individual was soothing the infant. A person expressed positive affect to the infant, either verbally (praise, terms of endearment) or by smiling in response to a behavior. This variable was reserved for clear demonstrations of love and affection, which went beyond simple states of pleasure. A person expressed affect physically (touching, hugging, kissing, makes loving gestures, or patting infant affectionately) in clear demonstrations of love and affection. Simultaneous verbal and/or physical affect (verbal plus physical affect), as dened above. Through verbal and facial expressions and gestures to distract the infant, a person tried to quiet or calm an irritable or crying infant. If the infant calmed for at least 20 s, then verbal soothe was no longer coded. A person tried to physically quiet or calm an irritable or crying infant, for example, by rocking, patting, picking up, pushing the infant in a carriage, or showing the infant a toy. This behavior was only coded while the infant was irritable or crying. If the infant calmed for a period of 20 s, then physical soothe was no longer coded. The combined score of verbal and/or physical soothing. Social interaction was coded whenever an infant expressed a social behavior (smiling or vocalizing) while receiving a social bid from someone (vocalizing, stimulate/arouse, affection). Social interactions were measured in terms of the number of intervals (each coding every 30 s) observed as well as the number of bouts (2 or more contiguous intervals of social interaction). Social partners were coded for each social interaction. Social partners were classied as the infants mother, father, sibling, relative, and nonrelative (e.g., family friends). Analysis focused on the proportion of social interactions in which each partner was involved.

Infant fuss Infant cry Infant sleep Vocalizing to infant

Stimulate/arouse infant

Verbal positive affect Physical positive affect Combined positive affect Verbal soothe Physical soothe

Combined soothe Social interaction

Social partners

Statistics
Scores for infant behaviors, social engagement, and social interactions were prorated to represent the proportion of intervals that the infants were awake. Sleep patterns, however, were quantied as proportions of the total number of intervals observed. Multivariate analyses of variance (MANOVAs) were conducted to determine whether infant behaviors, behaviors directed toward infants, proportion of soothing to fussing and crying, soothing styles, affection styles, social interactions, social partners, and sleep patterns differed as a function of SES. A MANOVA, in which SES was entered as the independent variable, was conducted for each set of dependent variables, as shown in separate sections of Tables 3 and 4. All

descriptive information for the dependent variables is shown in these tables as well. Following each MANOVA yielding signicant effects, univariate analyses (ANOVAs) were conducted. T tests or nonparametric t tests were used as appropriate to make pairwise comparisons between SES groups. Cohens f was used to quantify effect sizes. Cohen (1988) has categorized small effect sizes as ranging from .10 to .24, medium from .25 to .39, and large as .40 or greater.

Results Infant Behavior

A MANOVA of the infant behaviors (shown in Table 3: smiling, vocalizing, self-play, fussing, and crying) revealed

660

FOUTS, ROOPNARINE, AND LAMB

Table 3 Mean Percentages of Infant and Social-Partner Behaviors Observed

Lower SES (n 20) Variable Infant behaviors Smiling Vocalizing Self-play Fuss Cry Infant sleeping Sleep during the day Number of sleep bouts per day Length of sleep bouts in minutes Social engagement behaviors of infantsa Vocalizing Stimulus-arousal Affection Soothing Relative rate of soothing to fussing and crying Types of soothinga Physical Verbal Types of affectiona Physical Verbal
a a

Middle SES (n 21) SD M 9.58 20.28 8.75 8.15 2.68 35.01 6.19 35.94 38.01 10.68 17.40 8.94 0.81 4.65 4.30 13.79 3.84 SD 6.39 7.76 7.46 3.98 2.43 10.04 2.14 18.75 12.96 7.05 12.98 6.60 0.34 4.59 2.60 9.80 7.43 M

Upper SES (n 21) SD 4.58 10.17 8.90 3.38 2.90 11.10 1.96 22.37 11.05 6.28 28.60 8.06 0.46 6.34 4.18 10.92 9.13

M 9.73 19.82 6.88 10.76 4.61 36.00 9.10 23.40 37.06 10.89 18.13 10.05 0.67 5.39 4.51 16.32 1.81

6.72 9.57 3.96 4.35 4.68 12.83 4.88 6.68 16.15 6.62 12.09 5.28 0.24 3.59 2.39 11.35 1.44

7.64 13.55 13.63 7.78 4.77 37.25 4.62 50.91 42.42 8.17 28.60 16.12 1.34 8.47 8.81 17.20 11.40

Note. SES socioeconomic status. Mean percentages of intervals that the infant was awake.

a signicant main effect for SES (Wilkss .69), F(10, 110) 2.27, p .05, f .46. Subsequent univariate analyses revealed SES differences for levels of infant vocalizing, F(2, 59) 3.47, p .05, f .35; self-play, F(2, 59) 4.94, p .01, f .41; and fussing, F(2, 59) 3.55, p .05, f .35, but not for smiling and crying. Pairwise comparisons showed that infants in upper SES families vocalized less than did infants in middle-, t(40) 2.41, p .05, or lower, t(39) 2.03, p .05, SES families but engaged in more self-play than infants in middle-, t(40) 1.93, p .06, or lower, t(27.93) 3.16, p .01, SES families. Levels of infant vocalizing and self-play were not signicantly different in the middle- and lower SES groups. Infants in lower SES families fussed signicantly more than infants in middle-, t(39) 2.03, p .05, and upper, t(39) 2.46, p .05, SES families, whereas fussing levels for infants in middle- and upper SES families were not significantly different. In summary, infants in upper SES families differed from infants in the other groups in several ways: They vocalized less, fussed less, and engaged in more self-play than infants in the lower SES contexts.

sleeping bouts, F(2, 59) 12.85, p .001, f .66. Infants in the three groups had three distinct sleep patterns: Infants in upper SES families had longer but less frequent bouts than infants in middle-SES, frequency t(40) 2.48, p .05; length t(40) 2.65, p .05, and lower SES families, frequency t(24.75) 3.83, p .001; length t(23.70) 5.39, p .001, and infants in middle-SES families had longer and less frequent bouts than those in the lower SES families, frequency t(25.76) 2.45, p .05; length t(25.22) 2.88, p .01. In summary, daily sleeping routines were distinct in each group, with fewer but longer naps enjoyed by infants in the higher SES contexts.

Infant Experiences: Social, Affectionate, and Soothing Behavior Directed Toward Infants
A MANOVA of behaviors directed toward infants, including vocalizing, stimulate-arouse, affection, and soothing (see Table 3), revealed a main effect for SES (Wilkss .71), F(8, 112) 2.67, p .01, f .44. Subsequent univariate analyses showed SES differences for affection, F(2, 59) 4.20, p .05, f .37, and soothing, F(2, 59) 6.83, p .01, f .48, but not for vocalizing or stimulatearouse. Specically, infants in the upper SES families received more affection than did infants in middle-, t(40) 2.46, p .05, and lower SES families, t(36.76) 2.34, p .05, whereas infants in the middle- and lower SES families received similar amounts of affection. Likewise, infants in upper SES families received more soothing than did infants in middle-, t(38.50) 3.16, p .01, and

Infant Daily Routines

A MANOVA of the measures of infant sleeping patterns shown in Table 3 revealed a signicant effect for SES (Wilkss .63), F(6, 114) 4.95, p .001, f .51. Subsequent univariate analyses showed socioeconomic differences in the number of bouts of sleeping, F(2, 59) 10.04, p .001, f .58, and in the average length of

SOCIAL EXPERIENCES OF AFRICAN AMERICAN INFANTS

661

lower, t(34.67) 2.87, p .01, SES families even though infants in upper SES families fussed less than infants in middle- or lower SES families. Because of these SES differences in the infants experiences, we examined soothing and affection styles as well as the ratios of soothing to fussing and crying. An ANOVA of the ratios of soothing to fussing and crying shown in Table 3 revealed a main effect for SES, F(2, 59) 19.71, p .001, f .82. Infant fussing and crying were more likely to elicit responses in upper SES families than in middle-, t(36.76) 4.24, p .001, and lower, t(30.56) 5.83, p .001, SES families, whereas the middle- and lower SES groups did not differ signicantly with respect to the proportions of fussing and crying instances that received responses. A MANOVA of soothing styles (see Table 3: physical, verbal) revealed a signicant main effect for SES (Wilkss .69), F(4, 116) 6.57, p .001, f .46. Univariate analyses indicated that SES signicantly predicted levels of both physical, F(2, 59) 3.45, p .05, f .35, and verbal soothing, F(2, 59) 13.44, p .001, f .68. Infants in upper SES families received more physical and verbal soothing than infants in middle-, physical t(40) 2.24, p .05; verbal t(33.49) 4.20, p .001, and lower, physical t(31.95) 1.92, p .06; verbal t(32.12) 4.07, p .001, SES families, whereas levels of soothing in the middle- and lower SES groups did not differ signicantly. A MANOVA of affectionate behavior styles (see Table 3: physical, verbal) revealed signicant main effects for SES (Wilkss .71), F(4, 116) 5.34, p .001, f .42. Univariate analyses showed that SES was signicantly associated with levels of verbal affection, F(2, 59) 11.11, p .001, f .62. Infants in upper SES families received more verbal affection than did infants in middle-, t(38.41) 2.95, p .01, or lower, t(21.04) 4.75, p .001, SES families. In order to further examine the lack of differences in verbal affection between the lower and middle-SES contexts, a subsequent MANOVA of affectionate behaviors

was conducted, entering number of siblings as a covariate. The SES effect remained signicant, and number of siblings did not predict variation in levels of physical or verbal affection. In summary, infants in the upper SES context differed from those in the other contexts in that they received more soothing and affection than other infants. Differences were evident with respect to physical and verbal soothing, but only with respect to verbal (not physical) affection.

Social Interactions
A MANOVA of the measures of overall social interaction shown in Table 4 revealed a signicant effect for SES (Wilkss .72), F(6, 114) 3.45, p .01, f .42. Univariate analyses indicated that SES predicted the frequency of social interaction bouts, F(2, 59) 5.59, p .01, f .44, and the average length of bouts, F(2, 59) 6.51, p .01, f .47, but not the overall number of intervals spent in social interaction. In the upper SES families, the bouts of social interaction were longer than in the lower, t(34.21) 4.02, p .001, and middle-SES families, t(40) 2.31, p .05, but social interactions were less frequent in the upper SES than in the middle-, t(40) 2.91, p .01, and lower SES families, t(39) 2.83, p .01. Social interaction bout frequency and length were similar in middle- and lower SES contexts. In summary, social interaction bouts followed the same general pattern as sleeping bouts: Longer but fewer bouts occurred in the higher SES contexts.

Social Interaction Partners

A MANOVA of the proportions of social interactions with various partners (see Table 4) revealed a signicant multivariate effect for SES (Roys largest root . 22), F(5, 56) 2.44, p .05, f .47. Univariate analyses showed SES differences in the numbers of social interactions with relatives, F(2, 59) 3.07, p .05, f .32, but not with

Table 4 Mean Percentages of Social Interactions Observed Between Infants and Various Partners
Lower SES (n 20) Variable Overall social interactions Social interactions during the day Number of interaction bouts per hour Length of interaction bouts in minutes Social interactions with different partnersb Mother Father Siblings Relatives Nonrelatives
a a

Middle SES (n 21) M 16.78 3.55 1.62 75.92 8.77 3.75 9.87 1.69 SD 8.85 1.50 0.81 27.09 18.61 6.72 17.08 6.02

Upper SES (n 21) M 12.80 3.08 2.16 75.12 19.15 0.65 3.94 1.20 SD 5.54 1.64 0.69 30.08 28.24 2.99 17.38 4.09

M 16.84 3.54 1.43 65.75 8.69 1.06 17.39 7.43

SD 8.99 1.52 0.44 24.99 16.61 2.77 17.71 19.55

Note. SES socioeconomic status. Mean percentage of intervals that an infant was observed awake. b Mean percentage of observed social interactions.

662

FOUTS, ROOPNARINE, AND LAMB

mothers, fathers, siblings, or nonrelatives. Infants in lower SES families had more social interactions with relatives than did infants in upper SES families, t(38.82) 2.45, p .05, with no differences between infants in upper and middle SES families or between infants in the lower and middle SES groups. In summary, infants in the lower SES families differed from those in the upper SES group: They interacted more with nonparental kin than infants in the upper group did.

Discussion
The results reported here make clear that the infants indeed behaved differently and had different social experiences, depending on their socioeconomic background. Even though infants in each SES group were equally engaged in social interactions overall (i.e., social interactions spanned similar numbers of intervals), the number and types of social partners varied systematically, depending on the socioeconomic context. Even though many SES differences were identied, our results do not support the notion that social deprivation characterizes lower SES contexts: Infants in all three groups received the same amount of social stimulation. In fact, the principal difference between SES groups was evident not in the amount of social stimulation but in the variety of people the infants interacted with, how long the interactions lasted, and the structure of the sleep-wake cycles. Bouts of social interaction with infants in upper SES families lasted longer but were less frequent than those in the middle- and lower SES contexts, and upper SES infants took fewer but longer naps. These ndings are consistent with those obtained in Lareaus (2003) ethnographic study of social class and race in the United States, even though Lareau studied older children. Lareau described distinct differences in the parenting strategies and daily routines of upper-middle-class (similar to our upper SES category) as opposed to lowerand working-class (similar to our middle SES category) parents regardless of ethnicity. Upper-middle-class parents tended to give their children very structured daily schedules, whereas lower- and working-class parents placed more emphasis on providing comfort, food, and shelter. Lareau observed that children in lower- and working-class families had much more free time to engage in activities of their choosing and spent time with extended kin daily. It is interesting that, even as early as 3- to 4-months of age, infants in different SES contexts experienced interactions and sleep-wake cycles structured much like the patterns that Lareau described. Unfortunately, we are unable to discern from our data whether the caregivers were intentionally structuring the infants days. Such parental intentions merit the attention of researchers in the future. Extended kin played strikingly different roles in the different SES groups as well. We are unable to determine whether the high involvement by extended kin in the lower SES context was due to the higher availability of extended kin, a higher sense of responsibility of extended kin, or both. These possible associations could be identied if caregiving ideologies of nonparental kin were examined in samples

varying more with respect to the presence of extended kin within each SES context. Nonetheless, the between-group differences in extended kin involvement with infants have notable implications. Namely, because researchers have tended to focus on motherinfant interactions, the extent and nature of social experiences may have been previously misrepresented, especially on the part of infants in lower SES families who have the broadest social networks. Among low-income African American single fathers, for example, stress associated with parenting may be mitigated by caregiving support received from kinship and nonkinship members (Hamer & Marchioro, 2002). From a culturalcontextual, adaptive-resilient perspective, our results would support the notion that having broad social and kin networks is valued and important for success in low-SES niches where economic resources are not reliable and families often depend on friends and kin to help them. Of course, these patterns of multiple caregiving during infancy also raise many questions beyond the scope of this study. For example, these results prompt us to ask what impact multiple caregiving systems might have on parent child relationships and later development? Furthermore, we cannot assume that a wide range of social partners either fosters or harms infant socioemotional development; more studies are needed to clarify how extensive multiple caregiving systems affect parent child attachment patterns. Several SES differences in infant behavior were also apparent. Infants in upper SES families engaged in more self-play but vocalized less than did infants in middle- and lower SES families. This was not predicted because previous researchers have shown that infants and children from upper SES families tend to vocalize more than infants and children in lower SES contexts. For example, Feiring and Lewis (1981) reported that 3-month-old infants in uppermiddle-class families vocalized more than their middleclass counterparts. However, Feiring and Lewiss study only involved European Americans, and observations were restricted to motherinfant dyads. Our ndings are also not consistent with those obtained in studies of older children; Hart and Risley (1995) reported that preschool-age children from upper SES families had larger vocabularies and talked more frequently than children from lower SES families regardless of ethnicity. Many researchers have tended to focus on vocal interactions with mothers or when mothers are present (Feiring & Lewis, 1981; Hoff, 2003; Richman, Miller, & LeVine, 1992). Perhaps our ndings differ from theirs because we recorded all vocalizations and social interactions throughout the day, not only those to or in the presence of mothers. Similarly, whereas other researchers have suggested that children are spoken to more often in upper SES contexts than are children in lower SES contexts (Hart & Risley, 1995; Hoff, 2003), we found that infants in all three groups were vocalized to at similar levels, although there were some SES differences in styles of social interactions. Specically, infants in upper SES families received more verbal affection than did infants in middle- and lower SES families. Fussing and crying infants in upper SES African Amer-

SOCIAL EXPERIENCES OF AFRICAN AMERICAN INFANTS

663

ican families elicited soothing more often than did infants in middle- and lower SES families, even though infants in the lower SES contexts fussed more than infants in the other groups. This result has implications for possible variations in socioemotional development because maternal responsivity to fussing and crying is a predictor of motherinfant attachment quality among African American and European American infants (Bakersman-Kranenburg et al., 2004). Bakersman-Kranenburg and colleagues (2004) suggested that SES inuences attachment security because poverty appears to impede maternal sensitivity. The elevated levels of infant fussing may be related to different patterns of caregiver responsiveness to fussing and crying. Alternatively, the higher levels of fussing may be related to stress in the household or stress during pregnancy because stress and anxiety during some stages of pregnancy have been associated with elevated levels of infant crying (St. JamesRoberts & Conroy, 2005). Further studies of infants in low-SES contexts, infant crying, and stress during pregnancy are needed to assess this interpretation of our results. Even with the small sample sizes, SES was clearly a predictive variable with medium to large effects on many aspects of infant behavior and experience. Overall, it is apparent that the everyday experiences and behavior of African American infants from different SES backgrounds were quite different, suggesting that SES has a powerful effect on behavior and perhaps development from a very young age. In most cases, signicant SES effects involved differences between infants in the upper as opposed to the middle- and lower SES groups, suggesting the need to further focus on intraethnic group variability. In addition, distinct SES differences in infants social networks suggest that researchers should avoid focusing too narrowly on motherinfant interactions because they may not fully represent the everyday experiences of infants in lower SES contexts where extended kin often play prominent roles. Lastly, although many socioeconomic differences were apparent, these results do not replicate ndings from studies of SES among European Americans, suggesting that SES effects may be mediated by cultural and ethnic backgrounds. It is clear that more studies of non-European American families from a full range of SES backgrounds are needed in order to understand both the relationship between SES and ethnicity and the inuence of SES on infant experience and behavior. Of course, this study had several limitations. First, our sample size was small. We encourage researchers to conduct further studies with larger sample sizes examining minority group families from a broad range of SES backgrounds because this would certainly help scholars disentangle the inuences of ethnicity and SES on infant care and development. Even though we view our focusing on one ethnic group as a strength (Ogbu, 1981), it is important to emphasize that strong conclusions are not warranted. This point is underscored by the heterogeneous nature of our sample. Lastly, our ability to draw conclusions with respect to later development was limited by a cross-sectional design. We hope that our results will stimulate longitudinal

research on the effects of early experience in diverse ethnic and socioeconomic contexts, however.

References
Ainsworth, M. D., Blehar, M., Waters, E., & Wall, S. (1978). Patterns of attachment. Hillsdale, NJ: Erlbaum. Bakersman-Kranenburg, M. J., van IJzendoorn, M. H., & Kroonenberg, P. M. (2004). Differences in attachment security between African-American and white children: Ethnicity or socio-economic status? Infant Behavior & Development, 27, 417 433. Bartz, K. W., & Levine, B. S. (1978). Child rearing by Black parents: A description and comparison to Anglo and Chicano parents. Journal of Marriage and the Family, 40, 709 720. Belsky, J., Gilstrap, B., & Rovine, M. (1984). The Pennsylvania Infant and Family Development Project 1: Stability and change in mother-infant and father-infant interaction in a family setting at one, three, and nine months. Child Development, 55, 692705. Brody, G. H., & Flor, D. L. (1997). Maternal psychological functioning, family processes, and child adjustment in rural, singleparent, African American families. Developmental Psychology, 33, 1000 1011. Caldera, Y. M., & Lindsay, E. W. (2006). Coparenting, motherinfant interaction, and infant-parent attachment relationships in two-parent families. Journal of Family Psychology, 20, 275283. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). New York: Academic Press. Cowan, P. A. (1997). Beyond meta-analysis: A plea for a family systems view of attachment. Child Development, 68, 601 603. de Wolff, M. S., & van IJzendoorn, M. H. (1997). Sensitivity and attachment: A meta-analysis on parental antecedents of infant attachment. Child Development, 68, 571591. Early, D. M., & Burchinal, M. R. (2001). Early childhood care: Relations with family characteristics and preferred care characteristics. Early Childhood Research Quarterly, 16, 475 497. Entwisle, D. R., & Astone, N. M. (1994). Some critical guidelines for measuring youths race/ethnicity and socioeconomic status. Child Development, 65, 15211540. Feiring, C., & Lewis, M. (1981). Middle class differences in mother-child interaction and the childs cognitive development. In T. Field, A. Sostek, P. Vietze, & P. H. Leiderman (Eds.), Culture and early interactions (pp. 6391). Hillsdale, NJ: Erlbaum. Fracasso, M. P., Lamb, M. E., Scholmerich, A., & Leyendecker, B. (1997). The ecology of mother-infant interaction in EuroAmerican and immigrant Central American families living in the United States. International Journal of Behavioral Development, 20, 207217. Fraley, R. C. (2002). Attachment stability from infancy to adulthood: Meta-analysis and dynamic modeling of developmental mechanisms. Personality and Social Psychology Review, 6, 123 151. Gottfried, A. W. (1984). Home environment and early cognitive development: Integration, meta-analyses, and conclusions. In A. W. Gottfried (Ed.), Home environment and early cognitive development: Longitudinal research (pp. 329 342). New York: Academic Press. Graham, S. (1992). Most of the subjects were White and middle class: Trends in published research on African Americans in selected APA journals, 1970 1989. American Psychologist, 47, 629 639. Hamer, J., & Marchioro, K. (2002). Becoming custodial dads: Exploring low-income and working-class African American fathers. Journal of Marriage and the Family, 64, 116 129.

664

FOUTS, ROOPNARINE, AND LAMB M. H. Bornstein (Ed.), Handbook of parenting: Vol. 4. Applied parenting (2nd ed., pp. 95121). Mahwah, NJ: Erlbaum. Ogbu, J. U. (1981). Origins of human competence: A culturalecological perspective. Child Development, 52, 413 429. Quintana, S. M., Aboud, F. E., Chao, R. K., Contrera-Grau, J., Cross, W. E. J., Hudley, C., et al. (2006). Race, ethnicity, and culture in child development: Contemporary research and future directions. Child Development, 77, 1129 1141. Richman, A. L., Miller, P. M., & LeVine, R. A. (1992). Cultural and educational variations in maternal responsiveness. Developmental Psychology, 28, 614 621. Rogoff, B., Mistry, J., Goncu, A., & Mosier, C. (1991). Cultural variation in the role relations of toddlers and their families. In M. H. Bornstein (Ed.), Cultural approaches to parenting (pp. 173184). Hillsdale, NJ: Erlbaum. Roopnarine, J. L., Fouts, H. N., Lamb, M. E., & Lewis-Elligan, T. E. (2005). Mothers and fathers behaviors toward their 3- to 4-month-old infants in lower, middle, and upper socioeconomic African American families. Developmental Psychology, 41, 723 732. St. James-Roberts, I., & Conroy, S. (2005). Do pregnancy and childbirth adversities predict infant crying and colic? Findings and recommendations. Neuroscience and Biobehavioral Reviews, 29, 313320. Super, C., & Harkness, S. (1982). The infants niche in rural Kenya and metropolitan America. In L. L. Adler (Ed.), Crosscultural research at issue (pp. 4755). New York: Academic Press. Super, C., & Harkness, S. (1997). The cultural structuring of child development. In J. Berry, P. Dasen, & T. S. Saraswathi (Eds.), Handbook of cross-cultural psychology: Basic processes and human development (pp. 139). Needham, MA: Allyn & Bacon. Thornton, M. C., Chatters, L. M., Taylor, R. J., & Allen, W. R. (1990). Sociodemographic and environmental correlates of racial socialization by Black parents. Child Development, 61, 401 409. Uttal, L. (1999). Using kin for child care: Embedment in the socioeconomic networks of extended families. Journal of Marriage and the Family, 61, 845 857.

Harkness, S., & Super, C. M. (1992). Childcare in Kenya. In M. E. Lamb, K. J. Sternberg, C. P. Hwang, & A. G. Broberg (Eds.), Childcare in context: Cross-cultural perspectives (pp. 419 476). Hillsdale, NJ: Erlbaum. Hart, B., & Risley, T. R. (1995). Meaningful differences in the everyday experience of young American children. Baltimore: Brookes Publishing. Heath, S. B. (1989). Oral and literate traditions among Black Americans living in poverty. American Psychologist, 44, 367 373. Hewlett, B. S., & Lamb, M. E. (Eds.). (2005). Hunter gatherer childhoods. Somerset, NJ: Transactions/Aldine. Hewlett, B. S., Lamb, M. E., Shannon, D., Leyendecker, B., & Scholmerich, A. (1998). Culture and infancy among Central African foragers and farmers. Developmental Psychology, 334, 653 661. Hoff, E. (2003). Causes and consequences of SES-related differences in parent-child speech. In M. C. Bornstein & R. H. Bradley (Eds.), Socioeconomic status, parenting, and child development (pp. 147160). Mahwah, NJ: Erlbaum. Jackson, J. F. (1993). Multiple caregiving among African Americans and infant attachment: The need for an emic approach. Human Development, 36, 87102. Lamb, M. E., Bornstein, M. H., & Teti, D. M. (2002). Development in infancy. Mahwah, NJ: Erlbaum. Lareau, A. (2003). Unequal childhoods. Berkeley: University of California Press. Leyendecker, B., Lamb, M. E., Scholmerich, A., & Fracasso, M. P. (1995). The social worlds of 8-month-old and 12-month-old infants: Early experiences in two subcultural contexts. Social Development, 4, 194 208. Leyendecker, B., Lamb, M. E., Scholmerich, A., & Fricke, D. M. (1997). Contexts as moderators of observed interactions: A study of Costa Rican mothers and infants from differing socioeconomic backgrounds. International Journal of Behavioral Development, 21, 1534. Low, B. (2005). Families: An evolutionary anthropological perspective. In J. L. Roopnarine & U. Gielen (Eds.), Families in global perspectives (pp. 14 32). Boston: Allyn & Bacon. Magnusson, K. A., & Duncan, C. J. (2002). Parents in poverty. In

Received May 27, 2006 Revision received March 14, 2007 Accepted March 15, 2007