Feeding Practices and Expectations Among Middle-Class Anglo and Puerto Rican Mothers of 12-Month Old Infants

Beliefs and practices regarding infant feeding were examined among middle-class Puerto Rican and Anglo mothers of 12-month old infants. In particular, Anglo (n=32) mothers in Connecticut and Puerto Rican (n=28) mothers in San Juan were interviewed regarding their expectations for the attainment of selected feeding-related milestones, and observed in an actual feeding situation with their infants. Results indicate that the two groups of mothers demonstrate coherent clusters of beliefs and practices with regard to infant feeding at 12 months. In particular, Anglo mothers reported earlier attainment of self-feeding milestones and a greater emphasis on childrearing goals related to self-maximization, whereas Puerto Rican mothers reported later attainment of self-feeding and a greater emphasis on childrearing goals related to proper demeanor; videotaped observations supported mothers’ self-report of current primary feeding method. Findings are interpreted as providing evidence for the existence of coherent patterns of beliefs and behaviors within each group, with Puerto Rican mothers' patterns differing from those more commonly found among Anglo, middle-class mothers. Feeding Practices and Expectations 3 Feeding Practices and Expectations Among Middle-Class Anglo and Puerto Rican Mothers of 12-Month Old Infants The constructs of egocentrism/independence/individualism versus sociocentrism /interdependence/collectivism have proven to be robust and useful heuristics for understanding broad cultural differences. Briefly, American culture is often described as _individualistic_ in that it conceives of the individual as an _independent, self-contained, autonomous entity who (a) comprises a unique configuration of internal attributes. . . and (b) behaves primarily as a consequence of those internal attributes” (Markus & Kitayama, 1991, p. 224). In contrast, many other cultures are described as _sociocentric_ or _interdependent_ in that they emphasize the fundamental connectedness of human beings to one another. These differing construals of the relation between self and other are frequently described as key components of the beliefs and practices which organize caretakers’ perceptions of and interactions with children, thus constituting a primary aspect of the cultural context of childhood in a variety of cultural settings (Greenfield, 1994; Harwood, J. Miller, & Lucca Irizarry, 1995; Kagitcibasi, 1997). Increasingly, however, researchers have expressed dissatisfaction with the individualism/sociocentrism construct. In particular, a growing debate exists regarding the extent to which parental beliefs and practices do indeed form patterns which are shared at the larger group level, versus the extent to which intragroup variation renders the search for larger group patterns meaningless (Nucci, 1994; Wainryb, 1995). At issue is the extent to which we can speak of children as inhabiting shared developmental niches (Super & Harkness, 1986), versus the extent to which the context of childhood is unique to each individual, thus rendering fruitless the search for patterned variation. Feeding Practices and Expectations 4 The purpose of this paper is to examine withinand between-group variation among Anglo and Puerto Rican middle-class mothers’ beliefs and practices regarding one aspect of childhood’s context: the attainment of developmental milestones regarding feeding at the age of 12 months. The domain is narrowly rather than broadly selected. Because childhood’s context is inevitably complex and multifaceted, one could reasonably argue that children grow up within a matrix of beliefs and practices. By focusing the issue of withinversus between-group variation on a selected domain, we improve our chances of controlling for group differences in the conceptualization and enumeration of different aspects of childhood’s context. The domain we have selected has at least two advantages for research. First, the lives of infants tend to be more under the direct control of the parents or other caregivers than the lives of older children, who often are actively involved with selecting their own environments through contacts with peers, teachers, and other adults (Corsaro, 1997; Scarr, 1992). Although the child’s influence on the parent is present in infancy as well as at later ages, the special demands of infancy require extensive and specific daily caretaking routines across cultures. Second, infant feeding in particular is a universal and age-salient activity, thus making it a relevant, naturally-occurring, and clearly definable context for study across diverse populations. At 12 months, infants have matured enough physically for culturally patterned variation in the attainment of self-feeding milestones to be evident. Contributions of Present Investigation The present study examines Anglo and Puerto Rican middle-class mothers’ beliefs and practices regarding the attainment of developmental milestones in relation to infant feeding at 12 months. The rationale for choosing middle-class Anglo and island Puerto Rican mothers as the Feeding Practices and Expectations 5 subject population was as follows: (1) Middle-class Anglo mothers were expected to provide an index of the mainstream, dominant U.S. culture that has shaped the articulation of most normative models of child development, thus providing an appropriate comparison point for the study of other groups; and (2) middle-class island Puerto Rican mothers were expected to represent an educated, professional group in a society that, due to its commonwealth status, shares the technological advances of America. Nonetheless, it was anticipated that these mothers would demonstrate behaviors and perceptions concordant with a cultural meaning system that is different from that of the middle-class Anglo mothers. In particular, Puerto Rican culture is generally considered to emphasize a more sociocentric, and Anglo culture a more individualistic, view of the person and of interpersonal relationships (Diaz Royo, 1974; Harwood, et al., 1995; Lauria, 1982). In keeping with this previous research, it was hypothesized that Puerto Rican mothers will be more likely to organize their beliefs and practices regarding feeding in such a way as to emphasize the infant’s interdependence (e.g., through emphasis on a longer period of being primarily fed by the mother), whereas Anglo mothers will be more likely to organize their beliefs and practices in such a way as to emphasize the infant’s independence (e.g., through emphasis on the early attainment of self-feeding). We also also anticipated that these patterns will be discernible despite within-group variation. Method Population Middle-class Anglo (n=32) and Puerto Rican (n=28) mothers of 12-month old infants participated in the study. The mothers were comprised of women with at least some college education who were from families in which the household head was employed in a white collar Feeding Practices and Expectations 6 or professional occupation (Levels I and II on the Hollingshead index [Hollingshead, 1975]). Anglo mothers were born, reared, and educated in the mainland United States, and spoke American English as their first language; Puerto Rican mothers were reared and educated in Puerto Rico, and spoke Puerto Rican Spanish as their first language. The gender distribution of each mother's target child was roughly equal across the two groups. Similar subject recruitment procedures were employed in the two cultural settings. Specifically, the mothers were obtained through invitational letters distributed through pediatric practices servicing middle-class populations. As can be seen in Table 1, the resulting samples of mothers were well matched on multiple sociodemographic indices. In particular, mothers differed on just 3 demographic variables: Puerto Rican mothers were more likely to be employed outside the home, worked on average more hours per week if employed, and were more likely to report their religious background as Catholic. ---------------------------Insert Table 1 about here ---------------------------Puerto Rican mothers. In addition to general sociodemographic information, data was obtained regarding specific aspects of American influence among the Puerto Rican mothers. In particular, 25% of the Puerto Rican mothers reported that they had spent time in the U.S., and 45% considered themselves bilingual (Spanish-English). The Puerto Rican mothers were also administered the Psychological Acculturation Scale (Tropp, Erkut, Alarcon, Garcia Coll, & Vazquez, 1994), which examines five aspects of cultural identity among Latinos. As can be seen in Table 2, the results of this acculturation measure indicated that the Puerto Rican mothers in Feeding Practices and Expectations 7 our sample had on average a strongly Puerto Rican cultural identity. ---------------------------Insert Table 2 about here ---------------------------Procedure The data for this study were taken from a larger longitudinal study examining culture and mother-infant interactions among middle-class Anglo and Puerto Rican mothers. Mothers were observed and interviewed at home in their native language when their infants were 4, 8, and 12 months of age by ethnically matched, trained interviewers. To examine mothers’ developmental expectations and practices related to feeding, information was obtained regarding their: longterm socialization goals; expectations for the typical attainment of feeding-related milestones; reports of the age at which their own infants attained specific feeding milestones; and primary feeding methods as obtained through both observation and maternal report. Long-Term Socialization Goals To provide an index of cultural beliefs, mothers were administered the Socialization Goals Interview (SGI), a semi-structured interview consisting of four open-ended questions in which parents are asked to describe the qualities they (a) would and (b) would not like their children to possess as adults, and to describe children they know who possess at least the beginnings of those (c) positive and (d) negative qualities (Harwood, 1992). Previous research (Harwood, Schoelmerich, Ventura-Cook, Schulze, & Wilson, 1996) has indicated that the first two questions, which probe for adult socialization goals, may provide a better index of broad cultural belief systems than the two questions regarding desirable and undesirable child Feeding Practices and Expectations 8 behavior; this study will therefore focus on mothers' open-ended responses to the first two questions. Mothers’ responses to the first two questions of the SGI were coded at the level of individual word and phrase descriptors into five mutually exclusive content categories; these five categories were created on the basis of open-ended responses given by samples of Anglo and Puerto Rican mothers to the SGI in previous research (Harwood, 1992; Harwood, et al., 1995), and are as follows: (1) Self-Maximization, or concern that the child become self-confident and independent, and fully develop his or her talents and abilities as an individual; (2) Self-Control, or concern that the child learn to control negative impulses towards greed, aggression, and egocentrism; (3) Lovingness, or concern that the child develop the capacity for emotional intimacy; (4) Decency, or concern that the child grow to meet basic societal expectations such as being hard-working, law-abiding, and honest; and (5) Proper Demeanor, or concern that the child behave respectfully, get along well with others, and fulfill reciprocal role obligations, particularly within the family. Less than five percent of mothers' responses failed to fit into one of these five content categories, attesting to their broadly encompassing nature. Because previous research has indicated that the categories of Self-Maximization and Proper Demeanor capture culturally central dimensions consistent with independent and interdependent socialization goals among Puerto Rican and Anglo mothers (Harwood, et al., 1995, 1996), these two categories were deemed of particular importance in this study. Coding of the SGI was performed by a trained graduate assistant blind to the study hypotheses as well as to each mothers’ sociocultural group. Reliability in coding mothers’ responses was calculated on 50% of the sample with a second blind judge. Overall agreement Feeding Practices and Expectations 9 reached a level of .85 (Cohen_s kappa, range .80-.89). Developmental Milestones To examine group differences in developmental expectations, mothers were asked to complete a checklist indicating their expectations of the age ranges at which most children attain five feeding-related developmental milestones: able to eat solid food; able to drink from an infant cup by self; able to eat finger foods; able to be weaned from the breast or bottle; able to feed self using utensils. For each milestone, mothers chose from the following age ranges: 0-3 months, 3-6 months, 6-9 months, 9-12 months, 12-18 months, 18-24 months, and 2-3 years. Because mothers may feel that there is a range of variability in the attainment of milestones, age ranges were used instead of asking mothers to give specific ages. Maternal Report Regarding Own Child Preliminary questioning indicated that, at 12 months, all infants in both groups had already tried finger foods and drinking from an infant cup; mothers were therefore asked to report the age in months that their own child had first tried finger foods and drinking from a cup. In addition, mothers were asked whether or not their infant had tried using a spoon. Current Feeding Practices Maternal report. To examine group differences in feeding practices, mothers were asked to report the primary means by which they currently fed their infants: bottle/breast, spoonfeeding, or infant self-feeding. Videotaped observation. Mothers were asked to feed their infant as they normally would, and were then videotaped for 8 minutes; to ensure a normal feeding situation, interviews were Feeding Practices and Expectations 10 scheduled around infants’ typical feeding times. To avoid potential bias, all videotaped observations were conducted before interviewing mothers about their beliefs and practices. Videotaped observations of infant feeding were coded by an undergraduate research assistant blind to the study hypotheses into one of three categories: Mother spoon-feed, mother bottlefeed, or infant self-feed (e.g., finger foods or attempt to use utensils). No mothers breast-fed at the 12-month observation. Results Long-Term Socialization Goals To determine the relative proportion of each mothers’ responses codeable into the category of Self-Maximization as opposed to the category of Proper Demeanor, a proportional measure was created by computing the natural logarithm of the number of statements each mother made codeable into the Proper Demeanor category divided by the number of statements she made codeable into the Self-Maximization category. This measure yielded a value along which mothers varied, from a low of -8.4 (the relative number of Self-Maximization responses outweighing the relative number of Proper Demeanor responses) to a high of 2.4 (the number of Proper Demeanor responses outweighing the number of Self-Maximization responses). A one-way analysis of variance on each mothers’ proportion of Proper Demeanor to SelfMaximization responses yielded a significant main effect for Group, F(1,55) = 7.1, p < .05. In particular, compared to Anglo mothers, Puerto Rican mothers generated a greater proportion of Proper Demeanor compared to Self-Maximization responses (respective means = -1.5, .19). Within-group variability. To examine within-group variability in the proportion of Proper Demeanor to Self-Maximization responses generated by mothers, separate analyses of Feeding Practices and Expectations 11 variance were conducted within-group examining whether the Anglo and Puerto Rican mothers differed among themselves with regard to the relative use of Proper Demeanor versus SelfMaximization responses according to the following five factors: child's gender; child’s birth order (first, second, laterborn); maternal employment status; marital status; and mother’s religious background. A sixth factor, bilingual status, was included with the analyses for the Puerto Rican mothers. Stepwise regression analyses were then used to examine the following seven withingroup predictors of proportion of Proper Demeanor to Self-Maximization responses: socioeconomic status as indexed by the Hollingshead scale; maternal and paternal education; maternal and paternal age; total number children; and number hours worked per week by mother. Two additional variables, number years spent in the U.S., and psychological acculturation score, were included for the analysis with the Puerto Rican mothers. The p-level for entry into the model was set at .05, whereas the p-level for exclusion from the model was set at .10. These analyses yielded a few significant predictors. In particular, among the Puerto Rican mothers, number years in the U.S. predicted relatively greater use of the SelfMaximization category in the stepwise regression analysis (R-squared = .29, p < .05). Among the Anglo mothers, having more children and working more hours per week predicted relatively greater use of the Proper Demeanor category (R-squared = .40, p < .01). In addition, the analyses of variance indicated that, among the Anglo mothers, birth order (respective means first/second/laterborn = -8.35, -1.61, -.41. p < .01), maternal employment (respective means employed/not employed = -.67, -2.88, p < .05), and religious background (respective means None/Protestant/Jewish/Catholic = -1.77, -1.44, -7.88, -.24) were significant factors. However, Feeding Practices and Expectations 12 group differences remained significant even when first-borns, unemployed mothers, and Jewish subjects were removed from the analyses, indicating that the group differences obtained were not simply the result of within-group skewing by selected subpopulations. Developmental Milestones Chi-square analyses were conducted by group membership for each of the five feedingrelated milestones. The results of these analyses indicated significant group differences (p < .01) in the age at which infants were typically expected to attain two of the five milestones. In particular, compared to the Anglo mothers, Puerto Rican mothers expected later typical attainment for being able to drink from an infant cup, and to use utensils by self (62.5% of Anglo but only 14.3% of Puerto Rican mothers expected that the typical infant should be able to drink from a cup by 9 months of age, p < .01, and begin using utensils by 12 months of age, p < .01). Maternal Report Regarding Own Child To examine group differences in maternal report regarding their own infant’s attainment of feeding-related milestones, a multivariate analysis was performed by group membership on mothers’ reports of the age finger foods were introduced, and the age a cup was introduced. This analysis yielded a main effect for Group, F(2,54) = 3.2, p < .05. Follow-up univariate analyses indicated that, compared to the Anglo mothers, Puerto Rican mothers reported that their infants first drank from a cup at a significantly later age (respective means = 8.5, 9.8 months, p < .05). To examine group differences in whether infants had tried using a spoon, a chi-square analysis was performed. This analysis indicated that 84% of the Anglo but only 52% of the Puerto Rican infants had tried self-feeding with a spoon by 12 months of age (p < .01). Current Feeding Practices Feeding Practices and Expectations 13 Maternal report. A chi-square analysis was performed by group membership on mothers’ reports of primary method of feeding at 12 months: bottle; spoon-feeding; or infant self-feeding with finger foods (at 12 months, no mother reported that she was primarily breast-feeding). This analysis yielded a significant group difference, with 74.2% of the Anglo mothers, but only 14.3% of the Puerto Rican mothers reporting that their infants primarily self-fed by 12 months of age (Pearson chi-square, 2df, = 21.3, p < .01). Videotaped observations. The 8-minute videotaped observations of a typical feeding situation confirmed mothers’ self-reports regarding primary method of feeding. In particular, a chi-square analysis by group membership of observed feeding method (mother bottle-feed, mother spoon-feed, infant self-feed), indicated that 81.3% of the Anglo but only 3.6% of the Puerto Rican infants primarily self-fed during the videotaped observations (Pearson Chi-Square (2df) = 38.1, p < .01). The correlation (Kendall’s tau) between maternal self-report of primary feeding method and observed feeding behavior was .66 (p < .01). Within-group variation. The lack of within-group variation in observed method of feeding was striking. In particular, only six Anglo infants did not primarily self-feed, whereas only one Puerto Rican infant did. This near-uniformity of practice rendered within-group analyses on this variable problematic. Discussion This study provides evidence for the existence of clearly distinct patterns of beliefs and practices for middle-class Anglo and Puerto Rican 12-month old infants with regard to feeding. In particular, with remarkable consistency Anglo mothers were encouraging their infants at this age to self-feed, whereas Puerto Rican mothers were spoonor bottle-feeding. Consistent with Feeding Practices and Expectations 14 this practice, Anglo mothers placed relatively greater emphasis on long-term socialization goals related to maximizing the self as an autonomous agent, whereas Puerto Rican mothers placed relatively greater emphasis on goals related to maintaining respectful and appropriate interpersonal relationships. Group differences in maternal expectations for the attainment of feeding-related milestones were also consonant with these findings; specifically, when significant differences arose, they were in the direction of Anglo mothers expressing earlier expectations and reporting earlier attainment by their own child. Examination of within-group differences suggested greater variability in mothers’ longterm socialization goals than in their observed practices. However, this variability was patterned in distinct ways across the two groups. Among the Puerto Rican mothers, having spent time in the U.S. predicted greater use of goals related to Self-Maximization, suggesting that more intensive exposure to Anglo-American culture may be associated with greater adoption of some of its cultural goals and ideals; alternatively, Puerto Rican mothers who hold goals and ideals more consonant with Anglo-American culture may seek to spend time in the U.S. Among Anglo mothers, working more hours, having had more children, and being Catholic predicted greater use of goals related to Proper Demeanor. One possible explanation for these findings is that Anglo mothers who work more hours or have more children are busier, and thus may adopt a more pragmatic approach to childrearing, emphasizing the importance of cooperative behavior over individual agency; alternatively, mothers who are Catholic may represent a distinct ethnic enclave (e.g., Italian American) within the middle-class Anglo mothers in this study, and variations in beliefs and practices may reflect differing patterns among different ethnic groups. Together, the findings of this study suggest that middle-class Puerto Rican and Anglo Feeding Practices and Expectations 15 mothers demonstrate coherent clusters of beliefs and practices. For Anglo mothers, this cluster includes: long-term socialization goals that emphasize the importance of maximizing the self as an autonomous agent; maternal expectations for early attainment of selected feeding-related milestones; and practices which encourage early self-feeding. For Puerto Rican mothers, this cluster includes: long-term socialization goals which stress the importance of appropriate relatedness; maternal expectations for later attainment of selected feeding-related milestones; and practices which extend the period of time that the infant is fed by the mother rather than encouraged to self-feed. Within-group differences were also found, particularly with regard to mothers’ long-term goals. However, typical patterns emerged in each group despite the presence of internal variation. With regard to potential bias in either mothers’ self-reports of current feeding practices or observed feeding method, it should be noted that the correlation between observed practice and mothers’ reports of current primary feeding method was .66. This suggests that mothers did not simply report that their typical method of feeding was whatever they had just been observed doing; indeed, they tended to report a somewhat broader range of feeding activities than was actually observed. It is interesting in this regard that the difference between observed and reported practice was for both groups in the direction of what might be considered ‘culturally more ideal’ for the observations. In particular, 14% of Puerto Rican mothers reported that their infants typically self-fed; however, only one Puerto Rican infant was observed to primarily selffeed. Conversely, 74.2% of Anglo mothers reported that their infants typically self-fed; however, 81.3% of Anglo infants were observed to self-feed. This suggests that among both groups of mothers, there may have been a tendency to “put one’s best foot forward” for the Feeding Practices and Expectations 16 observations; that is, although the overwhelming majority of mothers in both groups reported that they typically engaged in feeding activities that were consistent with the cultural norm (selffeeding among Anglo mothers, spoonor bottle-feeding among Puerto Rican mothers), an even higher percentage chose the culturally normative mode of feeding for the actual observation. The fact that the methods chosen to demonstrate “good parenting practice” appear to have been in diametrically opposite directions among mothers in the two groups provides evidence for the constraining power of cultural norms on mothers’ beliefs and behaviors. Although this constraining power may also have led to some inflation of mothers’ reported primary feeding method, it seems unlikely that the majority of these low-risk, middle-class mothers would have reported practices that have no basis in the reality of their infants’ everyday lives. Indeed, preliminary analyses of 24-hour time diaries which mothers provided on their infants’ daily activities furnish additional confirmation that Anglo infants spend significantly more time selffeeding at 12 months than do Puerto Rican infants (Feng, Harwood, Leyendecker, & Schoelmerich, in preparation, 2000). Finally, these findings are noteworthy in light of previous studies which have sought to relate maternal expectations regarding a variety of social and self-care milestones to demographic factors such as socioeconomic status, maternal education, or age of the mother (Fox, Platz, & Bentley, 1995; Fulton, Murphy, & Anderson, 1991). These studies appear to be operating on the premise that there is one universal set of ages at which children should attain certain developmental milestones, and that it is in the best interest of infants to have parents who have these appropriate expectations; mothers who lack these expectations may be viewed as “at risk” for child abuse or neglect. Our findings provide evidence that a different but coherent Feeding Practices and Expectations 17 cluster of normative beliefs, expectations, and practices for the attainment of selected feedingrelated milestones emerges among this highly educated, low-risk group of middle-class Puerto Rican mothers. In particular, the attainment of early autonomy in feeding does not appear to be equally important among all cultural groups, nor should it be considered a risk-marker if parenting practices fail to emphasize it. Instead, these findings point to the extent to which everyday life in infancy is culturally patterned, and highlight the importance of understanding models of normative development alternative to our own. The growth in psychological research on parents’ beliefs has brought about greater recognition that much of parental cognition is itself socially and culturally organized. In addition, there has been increased acknowledgment that childrearing practices can only be evaluated in relation to culturally valued parenting goals. It is not difficult to think of parenting behaviors that are more or less universally recognizable, such as bottle-, spoon, and self-feeding. The present study provides support that such behaviors are integrated into different meaning systems from one culture to the next, and that cultural norms regarding developmentally appropriate practice constrain parental beliefs and behaviors. Conversely, what must not be forgotten in this age of reliance on expert childrearing advice is that cultural goals and ideals influence our views regarding the normative attainment of milestones and consequent desirable parenting behavior. Although we would not expect a 1month old infant to self-feed, it is by no means self-evident or universally desirable that this should be the primary mode of feeding among 12-month olds around the world, or even among diverse populations within our own society. Cultural variation in beliefs and practices regarding Feeding Practices and Expectations 18 the attainment of a wide variety of developmental tasks throughout infancy and childhood deserves further examination, even as our own normative models need greater acknowledgment of the ways in which culture influences our conceptions of what constitutes typical development and consequent appropriate practice. Feeding Practices and Expectations 19