Arts & Architecture

Mother-infant teaching interactions and attachment security in Euro-American and Central-American immigrant families

Description
Mother-infant teaching interactions and attachment security in Euro-American and Central-American immigrant families
Published
of 11
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  See discussions, stats, and author profiles for this publication at:https://www.researchgate.net/publication/222275859 Mother-infant teachinginteractions and attachmentsecurity in Euro-American andCentral-American...  Article   in  Infant Behavior and Development · April 1997 DOI: 10.1016/S0163-6383(97)90019-9 CITATIONS 23 READS 69 4 authors , including: Some of the authors of this publication are also working on these relatedprojects: SEIKA-NRW (Early language education in preschools)   View projectMichael E LambUniversity of Cambridge 521   PUBLICATIONS   18,408 CITATIONS   SEE PROFILE M p FracassoTowson University 14   PUBLICATIONS   279   CITATIONS   SEE PROFILE All content following this page was uploaded by Michael E Lamb on 15 September 2014. The user has requested enhancement of the downloaded file.  INFANT BEHAVIOR AND DEVELOPMENT 20 2), 1997, pp. 165-l 74 ISSN 0163-6383 Copyright 0 1997 ABLEX Publishing Corporation All rights of reproduction in any form resewed. zyxwvut Mother-Infant Teaching Interactions and Attachment Security in Euro-American and Central-American Immigrant Families AXEL SCH~LMERICH MICHAEL E. LAMB BIRGIT LEYENDECKER MARIA P. FRACASSO National Institute of Child Health and Human Development Forty mothers who had recently immigrated from Central America and 42 mothers from upper- middle-class Euro-American families were videotaped at home teaching their 4.. 8-, and 12. month-old infants to perform three tasks at each age. Overall, didactic maternal behavior decreased as the infants grew older, whereas task-oriented infant behavior increased with age. Fre- quency and duration measures of maternal and infant behavior in the teaching sessions were largely unrelated to attachment status, as assessed in the Strange Situation when the infants were 13 months old. When the timing of maternal behavior relative to infant behavior was used as a measure of maternal sensitivity, some of the expected relations were evident, however. Dyads who were later classified as disorganized had negative scores on a measure of joint attention to objects. Resistant dyads were characterized by high levels of coordination of social attention. while disor- ganized dyads had very low or negative coordination scores. Mutual coordination of maternal teaching was highest in dyads involving securely attached infants. attachment security mother-infant interaction didactic interaction Hisoanic families Maternal sensitivity and responsiveness are reli- ably associated with the security of infant- mother attachment (Ainsworth, Blehar, Waters, & Wall, 1978; Lamb, Thompson, Gardner, & Charnov, 1985) yet although maternal sensitivity is typically viewed as a characteristic evident in a variety of situations (Lamb & Easterbrooks, 1981), cross-situational stability has not been widely studied. As a result, we do not know which behaviors in which contexts are forma- tively important, particularly as variations in measures, samples, and methods hinder compar- isons across studies. The goal of the present study was to examine the association between attach- ment security and mother-infant interaction in a teaching situation, using data drawn from two quite different samples in order to assess the gen- eralizabiity of any association we discerned. Although mother-infant interaction is usu- ally studied in unstructured or free play situa- Direct all correspondence to: Michael E. Lamb, Section on Social and Emotional Development, National Institute of Child Health and Human Development, 9190 Rockville Pike, Bethesda, MD 20814. Axe1 Schtilmerich is now at the Martin-Luther Universit&t of Halle, and Maria Fracasso at Towson State University. tions (Ainsworth, 1979; Ainsworth, Blehar, Waters, & Wall, 1978, Belsky, Rovine, & Taylor, 1984; Egeland & Farber, 1984), Cairns (1979) and Maccoby and Martin (1983) have suggested that maternal responsiveness is best assessed in somewhat structured situations, Few researchers have assessed maternal sensitivity (Bee et al., 1982) or responsivity (Levitt, Guacci, & Coffman, 1993) in teaching situa- tions, which are structured by a behavioral goal the mother is instructed to pursue. We expected that the mothers’ behavior would be shaped by these demands and that the infants would become increasingly task-oriented with age. As in other contexts, however, we expected to find that sensitive and responsive didactic behavior would promote well-coordinated interaction and the development of secure infant-mother attach- ment. In two earlier studies, researchers have shown that indices of behavioral coordination based on interval level data obtained in unstruc- tured home observations (“synchrony”: Isabella, Belsky, & von Eye, 1989; “harmony”: Schtilmerich, Fracasso, Lamb, & Broberg, 1995) predict subsequent attachment status. In the current study, we attempted to develop com- parable indices of mutual coordination from 165  166 Schiilmerich Lamb Leyendecker and Fracasso shorter observations of interaction in more structured settings. In comparison with the mea- sures developed by Isabella et al. (1989) and Scholmerich et al. (1995), our measures were designed to achieve a higher level of resolution using continuously observed data. In addition, because so much less is known about the ante- cedents of “disorganized” attachment behavior (Main & Solomon, 1986) than about the ante- cedents of secure (B), avoidant (A), and resis- tant (C) attachment behavior, we made a special effort to explore the antecedents of disorganized behavior in normative samples. Very little is known about the antecedents of attachment behavior in different sociocul- tural contexts. In a large meta-analysis, van IJzendoom and Kroonenberg (1988) reported a relatively small but statistically significant effect of culture on the distribution of attach- ment classifications, and Lamb, Thompson, Gardner, and Charnov (1985) similarly sug- gested that socioeconomic, ethnic, and cultural factors affect behavior in the Strange Situation. To further explore the universality and general- izability of current conceptions of attachment formation, we included two very different sam- ples in the present study. The first sample com- prised recent immigrants from Central America, and the second sample was chosen to represent upper-middle class families from Euro-American backgrounds. Central Ameri- cans were selected because they are the fastest growing segment of the U.S. population, and the second sample was selected to represent the groups most frequently studied in the past. A secondary goal of our study was thus to test the cross-cultural generality of the association between mother-infant interaction in a struc- tured situation and the security of infant- mother attachment. Two research teams have previously studied Spanish-speaking mothers teaching their infants. De Cubas and Field (1984) reported several differences between Black and Cuban teenage and adult mothers observed interacting with their 12-month-old infants. Cuban mothers simultaneously demonstrated and verbalized more often than black mothers did. The authors failed to find the expected differences in the frequency of contingent and meaningful feed- back, however, perhaps because contingent feedback was rare in both groups. Garcia-Co11 et al. (1992) later reported that Puerto Rican mothers were more verbally directive, asked fewer questions, and presented more toys dur- ing play than did English-speaking Caucasian mothers on the mainland. In the teaching situa- tion, the Puerto Rican mothers praised and questioned their infants less often than the Anglo mothers did. The teaching behavior of middle-class North Americans and members of other cultural groups have been compared in three other stud- ies. Dixon, Levine, Richman, and Brazelton (1984) studied mothers teaching 6- to 30-month- old infants, and found that Gusii (African) moth- ers verbalized less than American mothers did. Bomstein and his colleagues (1992) found that mothers from Japan and France were more responsive in social and dyadic contexts, whereas American mothers were especially responsive in interactions involving object- related infant activities. Field and Pawlby (1980) reported that lower-class dyads in both the United States and the United Kingdom engaged in fewer distal interactions (including verbalization) than middle-class dyads did. Instructional maternal behaviors-especially verbalization-thus appear to vary depending on socioeconomic status, cultural background, and level of education. In sum, we expected (a) mothers to organize their behavior according to the tasks, and infants to show increasingly task-oriented behavior over the first year of life, and (b) behavioral coordination, as an index of maternal sensitivity, to be a significant predictor of attachment secu- rity. In addition, we expected to find (c) similar associations between mother-infant interaction and attachment security in both samples, with (d) differences between the samples most prom- inent in the patterning and frequency of mater- nal vocalization. METHODS Participants The Central-American sample consisted of 40 mothers and their firstborn infants (22 boys, 18 girls). The participants were recruited through advertisements in local newspapers, radio stations, churches, and hospitals. The families had been in the United States for an average of 27.43 months and 97% indicated that they only spoke Spanish at home. Eighty percent of them were from El Salvador, 15% were from Guatemala, and 5% were from Nicaragua. Annual house- hold income in 1990 averaged $13,800.00, and the mothers’  leaching Interactions and Attachment 167 Hollingshead scores ranged from 8 to 37 with an average of 20.65 (Hollingshead, 1975). Mothers averaged 22.48 years of age when their babies were born, and only one third had completed high school. The Euro-American mothers had sharply different socio-economic backgrounds. All mothers were born in the United States, and had an average Hollingshead score of 56.11 (Range: 38-66). Annual family income averaged $78,110.00. Ninety-three percent of the mothers had com- pleted college, and 44% had some graduate or professional training. Additionally, mothers were significantly older at the time of birth of the first child (M = 30.41 years). There were 20 boys and 22 girls in the sample. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIH Procedure All families were videotaped in their homes when the infants were 4, 8, and 12 months of age. The families were visited by a culturally matched experimenter, who asked mothers in their native language to teach their infants three tasks. The tasks were: mouthing a ring, picking up a cube, and flipping a bucket (4 months); ringing a bell, retrieving a toy from under a cloth, and clapping hands (8 months); stacking two blocks, pushing blocks with a hoe, and drawing a line on paper (12 months). The tasks were designed to range from easy to difficult at each age, and each lasted between two and three minutes. Mothers were asked to terminate each individual task session when they thought the infant had learned the task. If mothers did not terminate the task within three minutes, the experimenter interrupted the activity and introduced the next task. The standard Ainsworth Strange Situation procedure was administered in our laboratory when the infants were 13 months old. Measures The videotapes were analyzed in real time by recording the onset and offset of discrete behaviors. We could thus obtain exact information on the timing, frequency, and mean dura- tion of each behavior and so identify behavioral co- occurences and sequences. The mutually exclusive and exhaustive codes were designed to provide detailed descrip- tions of the interaction and permit the creation of composite or molar codes (Bakeman & Gottman, 1986). Eight passes were made through each videotape to code behavior in eight domains: mother’s visual atfention (infant’s face, infant’s body, environment, camera/observer, other people or TV, no attention, not visible); infant’s visual attenrion (mother’s face, mother’s body, environment, camera/observer, other people or TV, no attention, not visible, and drowsy/sleepy); mother’s vocalization (negative, neutral, babytalk, vocalizes to observer, vocalizes to other people, sings, play vocaliza- tion, no vocalization); infant’s vocalizution (positive, neu- tral, negative, laugh/squeal, no vocalization); distance and position of the dyad relative to one another (body contact, arm reach, more than arm reach, mother leaves room, infant moves, en face, en face with body contact, not visible); infant’s activities (cuddle, reach for object, social play, tac- tile exploration of teaching object, tactile exploration of other object, visual exploration, hands or shows object, not visible); mofher’s activities (signals for infants’ attention, structure task, manipulates object, take object, demonstrate task, provide variation, plays off task, not visible), and mother’s other activities (guides infant’s hands, restrain, moves infant, social play, soothe, caress, caregive). The codes for distance and position were dyadic by definition, as was the code for social play. All the other codes, though, were independent in that they did not require observers to judge whether the behavior occurred in response to the part- ner’s behavior. (A copy of the codebook can be obtained from the authors.) Scores from the three tasks were highly intercorrelated and were thus combined into one record. The frequencies and durations of the behaviors were then prorated to stan- dardize for differences in total observation time, and the overlap as well as the sequential organization of the differ- ent behaviors were then analyzed. In addition, several “metacodes” were defined by combining single codes. Three types of measures were then created: (a) basic observed teaching behaviors; (b) vocalization patterns; and (c) maternal sensitivity as measured by the degree of coor- dination between the mother’s behaviors and those of the child. The basic teaching behaviors were: (a) infanr urtempts rusk (infant’s attention on environment while hav- ing tactile contact with the teaching object), (b) morher reaches (mother demonstrates teaching object or orients infant’s attention toward teaching task), and (c) mother inrrusive reaching (mother teaches while infant attempts task). Vocalization patterns were examined using (d) over- all mufernul vocalization rate, e) rate of responsive mater- nal vocalizations. and (f) infant vocalization rate. The vocalization rates were computed as number of vocaliza- tions divided by interaction time in minutes, and the rate of responsive maternal vocalization was computed as the number of maternal vocalizations produced three seconds or less after termination of an infant vocalization divided by the total length of these observations. Mutual coordination was analyzed by computing the log-odds ratios for (g) social attention (mother directs her attention to the infant while infant directs attention to mother), (h) joint uftention (both mother and infant direct their attention to the teaching object), (i) mother reaches while infant attends MTIA), and (j) infant attempts the tusk while mother observes IATMO). The first two scores repre- sent more traditional dimensions of attention in social inter- actions, and the latter two attempt to capture specific behavior elicited by the instructions. Mutual coordination is by definition different from the accumulated time spent in a joint state, since the degree to which one partner shows the behavior in question while the other does not needs to be represented in the score as well. Stem, Jaffe, Beebe, and Bennett (1975) suggested that cross-product ratios (Fleiss, 1973) be used to determine whether vocalizations between infants and mothers were coactional or represented a conver- sation-structure. Their approach was criticized by Elias, Broerse, Hayes, and Jackson (1984) mainly because they pooled data across subjects and used frequency counts based on sampling intervals. We avoided these problems by com- puting the scores for each individual dyad based on the con- tinuously coded behaviors. To make the scores usable in subsequent statistical analyses, we followed Wickens’ (1993) suggestion that log-odds-ratios be computed using the following formula: a+ S) * c+ S) LoR =I” d+ S) * b+ S)  168 Schiilmerich lamb Leyendecker and Fracasso where a represents the time both mother and infant showed the appropriate behavior, c represents the time both mother and infant did not show the respective behaviors, while h and indicate those times when one of the partners showed the behavior while the other did not. The constant is added to avoid dividing by zero when either zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCB or b was zero. (For a basic introduction to cross-product ratios see Fienberg, 1982). Note that these scores do not merely reflect the amount of time spent in a joint state, but the amount of time spent in the joint state re[arive to the amount of time spent in dissociative states. The log-odds ratio is negative if the prod- uct of the dissociative states is larger than the product of the joint states, zero if they are equal, and positive if there is more coordination than dissociation among the interactants’ behaviors. Intercoder Reliability Four different assistants coded the tapes. All four repeatedly coded ten tapes and compared the reliability of their coding for individual codes in one-second and five-second intervals using Cohen’s Kappa. Once coders were reliable with the training-tapes K > .60), coding began. About 15% of the tapes were coded by two observers to assess reliability: Half when coding began and the remainder throughout the coding process, which lasted approximately one year. The coders did not know which tapes would later be used to assess reli- ability. Across all codes used in the analyses, averaged K was .79, with a range of .68 to .90. Several codes were extremely infrequent and therefore dropped from further analysis (mother negative vocaliza- tion, mother’s vocalization inaudible, infant drowsy/sleepy. infant social play, mother not visible, and mother leaves room), and some codes were not sufficiently reliable (tc < .60). Inspection of the confusion-matrices revealed that the codes for infant positive and neutral vocalizations were fre- quently interchanged, as were the codes for mother’s atten- tion on infant’s face and mother’s attention on infant’s body. Therefore, we created two new codes combining the inter- changed behaviors (infant vocalization and mother attention on infant). These codes were sufficiently reliable K for attention: 31, for vocalizations: .74), and they were used in the analyses reported here. Infant-Mother Attachment Security of attachment in the Strange Situation was assessed using the procedures described by Ainsworth et al. (1978). Videotapes of these sessions were sent to an inde- pendent facility (University of Washington, Seattle) for coding by three raters whose reliability had previously been established in training sessions at the University of Minne- sota (S. Paris, J. Moon, and B. Conley), and who were completely blind to all other information on our subjects and the goals of the study. Each tape was coded by two rat- ers, who agreed on 79.5 % of the tapes from the Central- American sample and on 81.5 % of the tapes from the Euro-American sample when the four-group classification including D was considered. Agreement for the forced clas- sification without the D-group was practically identical: the coders agreed on just one additional case from the Central American sample. All disagreements were resolved in a conference with a fourth rater (S. Spieker) who reviewed the videotapes with the srcinal coders. Mother-infant dyads were classified as secure (B), insecure-avoidant (A), insecure-resistant (C), or disorganized (D), using the crite- ria developed by Ainsworth et al. (1978) and Main and Solomon 1986, 1990). Missing Data At each age, data for one or several dyads could not be gath- ered, either for technical reasons or because the infant was sick or out of town. At four months, data for 5 of the 82 dyads were missing; at 8 and I2 months, only one dyad could not be filmed. The Strange Situation was completed for 74 out of the 82 infants. For all longitudinal analyses of data from the teaching sessions, we used the 74 infants with complete data at all three ages. and in analyses involving the attachment classifications we used data from 69 infants for whom we had complete observational and Strange Situation data. RESULTS (a) Teaching Behaviors and Vocalization Patterns Unexpectedly, the Central-American mothers spent 40% to 60% more time teaching their infants than the Euro-American mothers did. All scores included in the following analyses were prorated to account for differences in the obser- vation times. In both samples, mothers and infants were frequently engaged in didactic behaviors and exploration of the teaching objects, respec- tively. For between 16% and 26% of the total time, infants held or touched the objects that were used in the teaching tasks, and for an additional 17% to 22% of the time they looked at the teaching objects as they were held or demonstrated by the mothers. Mothers were engaged in active teaching between 23% and 40% of the total time. Repeated-measures MANOVAs with group (2) and age of infant (3) as the independent variables revealed no effects for group but significant effects for age on infant attempts and mother teaches (see Table 1). As expected, infants’ activity increased with age in both samples, whereas maternal teaching activity declined over time in the Central-American sample. Mothers’ teach- ing activity in the Euro-American sample decreased from 4 to 8 months, but increased again between 8 and 12 months. This is reflected in the significant group by age inter- action. Maternal teaching activities varied systemat- ically depending on the infants’ state. When the infants were occupied with the teaching objects,
Search
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks