Symposium
Chair: Erri C. Hewitt
Discussant: Kathryn E. Barnard
Parent-child interactions are a fundamental aspect of early childdevelopment. Substantial research has focused on identifying andquantifying these interactions (see Munson & Odom, 1996 for a review ofscales measuring parent-child interactions). Two of the most widely usedparent-child assessment instruments are Barnard's (1978; 1994) NursingChild Assessment Teaching Scale (NCATS) and Nursing Child AssessmentFeeding Scale (NCAFS). Their stems from the careful scale construction andsubsequent rigorous reliability and validity testing they have undergone(Sumner & Spietz, 1996). This symposium focuses on parent-childinteractions as measured by NCATS and NCAFS and brings together the mostcurrent research from different programs in distinct areas: nursing,developmental psychology, child health institutes, and Early Head StartPrograms. The first talk will examine the relationships among parent-childinteractions, cognitive development, and temperament in a sample of twinsfrom three months to three years of age. This study is unique in itsexamination of the role of temperament not only in parent-childinteractions, but also in later cognitive development. Findings from thisstudy suggest that temperament and cognitive development are both relatedto parent-child interactions. These findings indicate a need to includethe influence of temperament in models identifying the effects ofparent-child interactions on development. The second talk builds on research exploring how mothers' adultattachment classifications affect their engagement in interventionprograms. The talk examines how the mother's attachment classification isrelated to parent-child interactions. Exploratory analysis demonstratedthat mothers with a predominantly dismissing attachment style showed morepositive interaction scores and reduced variability in interactions withtheir children. Mothers with predominantly unresolved attachment stylesdemonstrated lower scores in responding to their child's distress, lesscontingent interactions, and wide variability in their interactions withtheir children. The findings in this study have generated new insightsinto how attachment strategies of mothers affect both their interventionprogram engagement and their interactions with their children. The third talk examines interactions of very low income mothers andtheir infants, and compares them to norms for non-low income mothers.Findings from this study indicate that very low income mothers haveteaching interactions with their children that are quantitatively differentfrom non-low income mothers. Many low income mothers, particularlyyounger, single, and less educated women, demonstrated deficits in verbalaspects of interactions with their children. These findings have importantimplications for understanding how differences in economic status mayaffect parent-child interactions. They also inform future interventionprograms targeted at low income mothers and children. The last talk examines cross-cultural variations in maternalresponsiveness to infant cues of hunger, satiety and early infant growth.This study finds significant differences in interaction styles betweenmothers on the basis of ethnicity, but little difference between infantcues and interaction patterns across cultural groups. The results suggestthat although the behaviors indicating hunger and satiation may be commonto infants, culture influences maternal responsiveness to these infantcues. Exploring cross-cultural differences in parent-child interactions isespecially important because it provides insights about how cultureinfluences child development. Finally, the discussant for this symposium, the creator of theNCATS/NCAFS scales, will share her insights on the current talks and theirrole in the future of parent-child interaction research.
Details of individual items:
paper
A large body of evidence supports the belief that children's earlyinteractions with parents impact subsequent development (Ainsworth & Bell,1974; Bakeman & Brown, 1980). Positive parent-child interactions have beenlinked to later measures of increased cognitive competence (Barnard andEyres, 1979; Beckwith, 1971) and language development (Olson, Bates, &Bayles, 1984). Effects of parent-child interactions on development extendbeyond the cognitive and linguistic domains; affect and emotion regulationhave also been linked to parent-child interactions. Parents' emotionsaffect the quality of children's play (Termine & Izard, 1988). Despitelinks between parent-child interactions and cognitive and emotionaldevelopment, few studies have examined the role of temperament inparent-child interactions and subsequent child development. The currenttalk provides information on the relationships between parent-childinteractions, cognitive development, and temperament in a sample of twinsfrom three months to three years of age. Participants were recruited for participation in an ongoinglongitudinal study examining many domains of development includingsocial-emotional, cognitive, language, and temperament in twins threemonths to three years of age. A sub-sample participated in parent-childinteractions coded using the Nursing Child Teaching and Feeding Scales (Barnard, 1994) at three, nine, and thirty-six months. Cognitivedevelopment was assessed using the Bayley Scales of Mental Development(Bayley, 1993) at three and thirty-six months and the Uzgiris-Hunt (1975)object permanence task at nine and twelve months. Finally, temperament wasassessed using the Infant Behavior Questionnaire (IBQ; Rothbart, 1981) atthree and twelve months. In exploratory analyses, children with higher IBQ fear scores hadless positive interactions with their parents. Children with high fear onthe IBQ at three months had feeding interactions with their parents at ninemonths that were characterized by less cognitive growth fostering (r-.51,p<.01, n119) and decreased parent response to child distress (r-.54,p<.01, n119). Children who showed higher fear on the IBQ at twelve monthshad teaching interactions with parents at nine months characterized by lesscognitive (r-.51, p<.01, n.46) and social (r-.32, p<.05, n46) growthfostering, and less parent sensitivity to child cues (r-.31, p<.05, n46). Children rated as laughing and smiling more on the IBQ at three months hadfeeding interactions with their parents at nine months characterized byincreased social growth fostering (r.37, p<.05, n40) and increased parentresponse to child distress (r.38, p<.01, n40). A similar pattern emergedfor children who were rated as laughing and smiling more on the IBQ attwelve months. Their feeding and teaching interactions with parents atnine months were characterized by increased social (Feeding:r.38, p<.05,n48; Teaching:r.32, p<05, n46) and cognitive (Feeding:r.46, p<.01,n48; Teaching:r.30, p<.05, n46) growth fostering, increased parentsensitivity to child cues (Teaching:r.31, p<.05, n46), and increasedparent response to child distress (Teaching:r.32, p<.05, n46). Few relationships were found between temperament and cognitive developmentat three, nine, or thirty-six months. Modest relationships were foundbetween cognitive development and parent-child interactions, whichreplicates previous work by Olson, Bates, & Bayles (1984). These findings, linking temperament and parent-child interactions,suggest that it is necessary to not only include measures of temperament inmodels examining parent-child interactions, but that differentiating amongaspects of temperament is also needed. Our results indicate that fearfuland pleasurable aspects of temperament may have the most salient effects onparent-child interactions.
paper
Throughout the past four years we have been working with an EarlyHead Start program, studying how adult attachment classifications impactthe mother's engagement in the intervention program and the impact of anintervention component designed to enhance the parent-child relationship.A final research goal has been identifying how the child's emergingattachment strategy is related to the parent-child relationshipspecifically dyadic outcomes. Within this longitudinal study we have examined the issue ofclient engagability in an intervention program. After the subjects had beenin the program at least one year we asked program staff to select thoseparents they experienced as easy to engage and those they experienced asdifficult to engage in the EHS program. We have presented the differencesbetween these two groups at SRCD (1999). We found that the best predictorof program engagement was the Adult Attachment Classification. The AAI hadbeen collected at program intake. There were significant differencesbetween major adult attachment classifications, with the difficult toengage (N10) being more unresolved with regard to trauma and the easy toengage being predominately dismissing (N 11) Qualitative differences werein the level of playfulness and nurturing the mothers had had in their ownchildhood. There was also a lower score on the Nursing Child AssessmentFeeding Scale (NCAFS) subscale of Sensitivity to Cues for the difficult toengage mothers. This is our first presentation concentrating on the parent-childinteraction. This data was generated at two time points, ages 14 months athome and 19 months in the laboratory, using the NCAST Teaching Scale. The teaching situation involved the mother teaching the child atask. All interaction episodes were video taped and coded by reliable(>90%) observers with no knowledge of the program or the subjects. Thetapes available were from subjects with both 14 and 19-month data, whichincluded 5 Easy cases and 6 Difficult cases. This exploratory analysis ofa small number of well-documented cases revealed significant differencesbetween groups at both ages and within group differences at the two ages.The Teaching Scales demonstrated mothers with dismissing attachmentstrategies showing more positive scores and reduced variance. We foundthat the difficult to engage mothers, demonstrated lower scores inresponding to their child's distress and in contingent interaction and widevariance between subjects in most subscales and total scores. Thedifficult to engage mothers were less positive and contingent at the 19month lab visit episode; while their children were more responsive andcontingent. We believe this possibly was a condition of the environment,were the unresolved mother became more anxious in our lab and thissubsequently negatively influenced her behavior; whereas for the childrenthe new environment was a positive alerting experience. The currentfindings have generated some new insights into how attachment strategiesand program engagement affects dyadic outcomes.
paper
There are multiple factors which influence the development ofchildren. Because caregivers, especially mothers, have a significant rolein their children's development, assessments and interventions must includethe transactions between the mother-child dyad. In this dyadicrelationship each of the partners brings their own characteristics to theinteraction. The mother must be sensitive to her infant's cues and respondto the infant in ways that will alleviate distress and promotesocial-emotional and cognitive growth. The infant, in turn, must beresponsive to the mother and be clear in the cues demonstrated during theinteraction. Some mothers, because of their personal characteristics andexperiences, may find it difficult to relate to their infants; some infantsmay have characteristics that make it difficult for their mothers to respond. Interactions of very low income mothers and their infants wereobserved in a mid-sized community in the mid-western United States (n172). The women were primarily white, single, and in their early twenties.Trained data collectors observed mothers using the Nursing Child AssessmentTeaching Scale (NCAST; Barnard, 1994) during a planned teaching situationin the home. The Teaching Scale has Cronbach's alphas ranging from .64 to.94 for the subscales and from .87 to .90 for the total scale. On average, mothers scored at least one standard deviation belowthe NCAST mean for most of the parent subscales and the total scale.Almost 45% scored below the NCAST 10th percentile on parent subscales aswell as total scales. Older mothers scored significantly higher thanyounger mothers for cognitive growth fostering. Mothers with more thanhigh school education scored significantly higher than mothers with lessthan high school education for social-emotional growth fostering. Motherswho were married scored significantly higher than single mothers in thisarea as well. Most infants (74%) scored above the 10th percentile cut-off. Many mothers, particularly younger, single, and less educatedwomen, demonstrated deficits in multiple areas of interactions. Most ofthe weak areas were in the verbal areas. These mothers may be unaware thattheir infants can engage in meaningful interactions. The infants in thissample were generally responsive to their mothers and were able to engagein the interaction. Providers working with families similar to these couldhelp mothers understand their role in teaching their infants and canencourage mothers to talk to and smile at their infants and to praise theirchildren's efforts. Providers can guide mothers to choose appropriatetasks for their infants to learn and can model how their infants areresponsive to them.
paper
The aim of the current study was to examine variations in maternalresponsiveness to infant cues of hunger, satiety and early infant growth inthree culturally diverse groups living in London. The hypothesis drew ontwo distinct areas of developmental psychology: the study of themother-infant interaction during feeding and the study of culturalvariations in the mother-infant interaction. The former provides evidenceof differences in the behavior of mothers that breast and bottle feed toinfant cues of hunger and satiety; the latter, that culture is one of thefactors shown to guide maternal patters of responsiveness to infants. Three groups of infants were recruited: one group known to show, asearly as six weeks, a decline in weight gain compared to normativepatterns, one group known to exceed growth rates in the first weeks oflife, and a third demonstrating convergence with normal growth rates.Bangladeshi infants (N18) typified the declining weight group, Nigerian(N21) the accelerated group and White UK (N23) the normal weight grouprespectively. Primiparous mothers born and educated in Bangladesh andNigeria, who had recently moved to London, were recruited during the lasttrimester of pregnancy. A group of White UK mothers were recruited from thesame hospitals. Mothers and infants were seen postnatally at home when infants weresix weeks, three months and six months old. Video recordings of the feedinginteractions at three and six months were coded using a standardizedobservation scale (NCAST Feeding Scale, Barnard 1978). The mean 'Total' NCAST scores- the sum of the 76 individual itemsthat comprise the Scale- differed significantly between groups at boththree and six months. At both time points, the White UK consistently hadthe highest mean score, the Nigerian the mid-position and the Bangladeshithe lowest. Analyses of the 'Mother-only' and 'Infant-only' subscalesidentified significant differences between all three groups of mothers onthe basis of ethnicity, but fewer differences between the infants. The likelihood of scoring 'yes' to individual items pertaining tothe initiation, pacing and termination of the feed differed significantlybetween the White UK and Bangladeshi mothers, but not between the White UKand Nigerian mothers. In contrast there were few differences between theinfants from all three groups. The results suggest that although the behaviors indicating hungerand satiation may be common to infants, culture is an influence on maternalresponsiveness to these infant cues. The complex picture of similaritiesand differences between and within the three ethnic groups pose furtherquestions of the research to date, and for future research.