Individual differences in infant still-face response at 6 months

Individual differences in infant still-face response at 6 months

Infant Behavior & Development 26 (2003) 421–438 Individual differences in infant still-face response at 6 months George M. Tarabulsy a,∗ , Marc A. Pr...

190KB Sizes 7 Downloads 111 Views

Infant Behavior & Development 26 (2003) 421–438

Individual differences in infant still-face response at 6 months George M. Tarabulsy a,∗ , Marc A. Provost b , Julie Deslandes b , Diane St-Laurent b , Ellen Moss c , Jean-Pascal Lemelin b , Annie Bernier d , Jean-François Dassylva b a

Groupe de recherche sur l’inadaptation psychosociale chez l’enfant, Département d’études sur l’enseignement et l’apprentissage, Faculté des sciences de l’éducation, Université Laval, Laval, Quebec, Canada G1K 7P4 b Groupe de recherche sur le développement de l’enfant et de la famille, Département de psychologie, Université du Québec à Trois-Rivières, Trois-Rivières, Quebec, Canada c Département de psychologie, Université du Québec à Montréal, Montréal, Quebec, Canada d Département de psychologie, Université de Montréal, Montréal, Quebec, Canada Received 5 September 2002; received in revised form 3 March 2003; accepted 4 March 2003

Abstract This study examined the association between infant still-face response and assessments of maternal behavior taken outside the still-face procedure (SFP). We also addressed the contributions of risk status and infant difficultness. Forty-one adolescent mother–infant dyads (high risk), and thirty-five adult mother–infant dyads (low risk) were seen when infants were 6 months old. Home visits were carried out to obtain maternal ratings of infant difficultness and to conduct observations of maternal interactive behavior. The SFP was conducted at the university two weeks later. Infant still-face response was coded for positive affect, negative affect and self-soothing behavior. Regression analyses revealed that maternal behavior was associated with negative affect and self-soothing behavior. In both cases risk status significantly moderated these effects. Infant difficultness significantly moderated the association between maternal behavior and self-soothing behavior, and marginally moderated the link between maternal behavior and negative affect. © 2003 Elsevier Science Inc. All rights reserved. Keywords: Still-face procedure; Maternal behavior; Temperament; Social risk; Infant emotional response



Corresponding author. E-mail address: [email protected] (G.M. Tarabulsy). 0163-6383/$ – see front matter © 2003 Elsevier Science Inc. All rights reserved. doi:10.1016/S0163-6383(03)00039-0

422

G.M. Tarabulsy et al. / Infant Behavior & Development 26 (2003) 421–438

1. Introduction The still-face procedure (SFP; Tronick, Als, Adamson, Wise, & Brazelton, 1978) has been extensively used to illustrate the patterns and cycles of mother–infant interaction. The procedure is composed of three brief phases: during the first phase, mothers are asked to engage their infant in face-to-face interaction in a playful manner. In the second phase, mothers take on a neutral, expressionless face where they do not respond to infant signals, before returning, during the third phase, to a playful, interactive mode. The SFP has been used to gain information on the patterns of social contingencies that characterize mother–infant interaction and the emotional responses of infants when these patterns are violated during the second phase of the procedure (Cohn & Tronick, 1987, 1988; Ellsworth, Muir, & Hains, 1993; Shapiro, Fagen, Prigot, Carroll, & Shalan, 1998; Weinberg, Tronick, Cohn, & Olson, 1999). The second phase of the SFP, where mothers take on a neutral expression and do not respond to infant signals, elicits a certain level of discomfort and sometimes distress in infants. Typically, infants will emit some positive, interactive behavior, followed by a reduction in smiles, and an increase in gaze aversion (looks away from mother) as well as fusses, protests and other manifestations of negative affect. This “still-face effect” (SFE) is perceived by researchers as an attempt by infants to re-establish the reciprocal patterns of interaction of the first phase and a response to their failure to do so. Infants are perceived as reacting to mother’s neutral expression and lack of responsiveness because these are unexpected (Carter, Mayes, & Pajer, 1990; Tronick et al., 1978). Much of the research that documents the SFE has focused on describing normative infant responses at different ages during the first year of life (e.g., Cohn & Tronick, 1988; Toda & Fogel, 1993). Increasingly, however, researchers have devoted more attention to the meaning of individual differences in infant still-face response. These authors have attempted to identify factors in the environmental ecology of the child that may account for variations in infant still-face response. By far, most of the explanations for such differences have focused on variations in maternal behavior in the first phase of the SFP. Several studies have shown that greater levels of maternal positive behavior and affect (Carter et al., 1990), maternal sensitivity (Braungart-Rieker, Garwood, Powers, & Notaro, 1998), the manner in which mothers initiate cycles of interaction (Cohn & Tronick, 1988) and greater overall maternal involvement in this first phase of the procedure, are all systematically associated with lower levels of distress shown by infants in the still-face phase. Other reports by Stack and her colleagues (e.g., Stack & Muir, 1990) and Pelàez-Nogueras, Field, Hossain, and Pickens (1996) have shown that maternal touching of infants during the procedure significantly reduces the amount of negative affect manifested by infants. Both sets of results support the notion that the SFE is systematically associated with the quality of maternal interactive behavior immediately prior to, and during, the still-face phase of the SFP. Such findings have led researchers to speculate that individual differences in the affective responses of infants during the still-face phase may be a function of the quality of mother–infant interaction taken in a broader context, outside of the SFP, and that differences in emotional response to the maternal still face may be a marker of the quality of the developing mother–infant relationship (Braungart-Rieker et al., 1998; Cohn, Campbell, & Ross, 1991; Tarabulsy, Tessier, & Kappas, 1996). The general objective of this study is to examine the proposition that infant

G.M. Tarabulsy et al. / Infant Behavior & Development 26 (2003) 421–438

423

still-face response is linked to the quality of maternal interactive behavior assessed outside the SFP. Additionally, this study addresses the specific contribution of infant temperament on infant response in the SFP. Temperamental features of infant responses have been emphasized in a variety of research procedures but surprisingly, the SFP has not been one of them. Finally, we will also address the possibility that maternal interactive behavior and infant temperament may relate to infant still-face response differently, as a function of social risk status. 1.1. Maternal interactive behavior Two sets of studies support the inference that, as in other structured interaction tasks, individual differences in infant still-face response is a function of the quality of maternal behavior the child is exposed to on a regular basis, outside of the SFP. The first set concerns studies conducted by Field and her colleagues (Field, 1984; Field, Healy, Goldstein, & Guthertz, 1990; Pelàez-Nogueras et al., 1996) and Cohn, Tronick and their colleagues (Cohn et al., 1991; Cohn, Matias, Tronick, Connell, & Lyons-Ruth, 1986) with high-risk, clinically depressed mothers and their infants. The study of this group of dyads is revealing in that the interactive difficulties of depressed mothers are well documented. Depressed mothers engage their infants less frequently in face-to-face interaction, and may be overly passive or aggressive with them. Their interactions are described as lacking in predictability and coherence and infants of depressed mothers more often develop insecure attachment relationships (Cohn et al., 1986; Field, 1994; van Ijzendoorn, Goldberg, Kroonenberg, & Frankel, 1992; Zuravin & DiBlasio, 1992). If the quality of interactions outside the immediate context of the SFP plays a role in individual differences in still-face response, infants of depressed vs. non-depressed mothers should show different responses to the maternal still face. This is in fact what Field and her coworkers found (Field, 1984; see Field, 1994, for a review). When compared to the infants of non-depressed mothers, 3-month-old infants of depressed mothers showed lower levels of both positive and negative affect throughout the SFP, and less negative affect during the still-face phase. Field (1984) and Pelàez-Nogueras et al. (1996) argues that for infants of depressed mothers, the still-face phase may not represent a significant violation of habitual mother–infant interaction, as it does for the infants of non-depressed mothers, and as such, these infants may not be experiencing the same degree of violation of interactive contingency. Similar findings have also been reported by Cohn et al. (1986). Thus, infants who are exposed to mothers with expected problems in caregiving manifest differences in their still-face response, supporting the hypothesis that individual differences in the SFE are associated with the quality of ongoing care. The second set of results that support the idea that infant still-face response is associated with more general aspects of maternal behavior comes from studies linking the SFE and the security of the attachment relationship, assessed in the Ainsworth Strange Situation (Ainsworth, Blehar, Waters, & Wall, 1978). A considerable number of studies in normative and high-risk samples of dyads have demonstrated that the strange situation reflects important aspects of maternal sensitivity and responsiveness during the first 12–18 months of life (see de Wolff & van Ijzendoorn, 1997). Thus, the presence of a relation between infant still-face response and later attachment classification may be taken as support for the suggestion that infant still-face response is meaningfully linked to maternal interactive behavior in broader contexts. Four

424

G.M. Tarabulsy et al. / Infant Behavior & Development 26 (2003) 421–438

studies (Braungart-Rieker, Garwood, Powers, & Wang, 2001; Cohn et al., 1991; Kiser, Bates, Maslin, & Bayles, 1986; Tronick, Ricks, & Cohn, 1982) have shown such an association. The results of these studies converge, indicating that infants who are later classified at 12 or 13 months as having a secure attachment relationship with their mothers, are those who at 4 or 6 months showed lower levels of distress and/or greater levels of positive affect in the still-face phase of the SFP. To the degree that the development of a secure attachment reflects exposure to predictable, warm and coherent caregiving, these studies provide further support for the suggestion that infant still-face response is an indicator of the quality of mother–infant interaction taken in a larger context. In this respect, Cohn et al. (1991) have suggested that the infant’s response during the still face represents a kind of “probe” (p. 373) into the evolving mother–child relationship and overall infant socioemotional development (see also, Moore, Cohn, & Campbell, 2001). 1.2. Inconsistencies in results Studies concerned with the infants of clinically depressed mothers and those that have predicted mother–infant attachment from infant still-face response have provided support for the notion that individual differences in still-face response are associated with mother–infant interaction outside the SFP. However, there is a degree of divergence between these two sets of studies that requires further attention. While Field’s work (1994) suggests that infants exposed to more problematic interactions (infants of depressed mothers) show lower levels of distress, the work linking the SFE with later attachment security shows that infants who will eventually develop a secure attachment (who are exposed to positive mother–infant interactions) also show less distress and/or more positive affect during the still-face phase of the SFP at 6 months. The inference of problematic interactions drawn from studies of depressed mothers is associated with the SFE in a way that is opposite from the inference of problematic interactions drawn from studies that have focused on later attachment. Some of these variations in results may be attributed to the different groups that are involved in the studies. Some studies (Field, 1984; Pelàez-Nogueras et al., 1996) have focused on high-risk mother–infant dyads (clinically depressed mothers, adolescent mothers), while others have examined low-risk dyads only (Braungart-Rieker et al., 2001; Kiser et al., 1986; Tronick et al., 1982). One study, that of Cohn and colleagues (1991), included both depressed and non-depressed mothers and their infants. However, maternal clinical status was not included in analyses that focused on infant still-face response. Developmental process associated with infant still-face response may be different for high- and low-risk dyads and some authors have suggested that responding to the maternal still face may represent a different kind of developmental task for high- and low-risk infants (Field, 1995; Fogel, 2000). In addition, in both sets of studies, maternal interactive behavior is inferred on the basis of two different criteria, clinical status and attachment security. Not only are both criteria imperfectly associated with variations in maternal behavior, but there may also be differences in the kinds of maternal interactive behaviors that each refers to. Thus, while both sets of studies point toward an association between maternal interactive behavior outside the SFP and infant still-face response, the nature of the association remains somewhat uncertain. To address this issue, and to gain a more accurate portrait of the relation between maternal interactive behavior and infant still-face response, it

G.M. Tarabulsy et al. / Infant Behavior & Development 26 (2003) 421–438

425

appears necessary to examine maternal behavior directly, outside the SFP, in both a high- and a low-risk group. 1.3. Temperament Another important contributor to individual differences in infant still-face response, and one that has been largely ignored in SFP studies, concerns the domain of temperament. The dearth of studies addressing the link between temperament and still-face response is surprising in that temperament theories have long postulated that individual differences in infant expressions of affect are associated with child-based characteristics, particularly in circumstances that elicit a certain level of distress as does the still-face phase of the SFP (Fox, Henderson, Rubin, Calkins, & Schmidt, 2001). Moreover, many conceptions of temperament and emotionality in children propose that part of the assessment procedures include parental or observational assessments of infant responses to violation of interactive contingencies by strangers and/or parents (Goldsmith & Rothbart, 1994). Viewed in this context, the mother’s still face, which violates expected interactive contingencies, may be conceived as eliciting temperamentally based emotional and behavioral responses. Furthermore, several studies have underlined the role of infant temperament in the quality of mother–infant interactions. For example, van den Boom and Hoeksma (1994) found that mothers of irritable infants have a more passive interactive style, engage their infants less often in interaction and offer less stimulation than mothers of non-irritable infants. The interactions of mothers and irritable infants are reminiscent of the interactions of high-risk dyads and may have a bearing on infant still-face response. It is conceivable that some of the discrepancies noted earlier linking infant still-face response to inferences of maternal behavior outside the SFP may be associated with variations in temperamental characteristics of infants in the different samples, or, perhaps more significantly, to variations in the action of infant temperament as a function of social risk or difficulties in mother–infant interactions. Reactive infants evolving in high-risk settings may be engaged on a developmental trajectory yet uncharted within the domain of infant responses to the maternal still face. Certainly, in other contexts, such children have been viewed as being at greater risk for experiencing problems in emotional development (Crockenberg, 1981; van den Boom, 1994). Only two studies have been found where the issue of temperament has been examined in relation to the SFE. Cohn and coworkers (1991) found that maternal reports of irritability were unrelated to infant still-face response at 2, 4 and 6 months. More recently, Braungart-Rieker and her colleagues (1998) found that maternal ratings of their infant’s negative, reactive temperament were inversely associated with self-comforting behavior, while observer ratings of negative temperament (based on infant behavior during a Bayley developmental assessment) were inversely associated with infant attention to objects during the maternal still face. These authors also conducted father–infant SFPs, where it was found that observer ratings of negative temperament were inversely related to infant positive affect during the paternal still face. Thus, while there is some discrepancy between the two studies that have included temperament-based variables in their analyses, Braungart-Rieker and her colleagues appear to have found some measure of support that part of the response of infants to the maternal still face may be attributable to child-based emotional characteristics. Differences in the results of both studies

426

G.M. Tarabulsy et al. / Infant Behavior & Development 26 (2003) 421–438

may be attributable to the differences in populations studied. Temperament may interact with risk status. Braungart-Rieker and her colleagues focused on a low-risk group while a high proportion of Cohn et al.’s sample was made up of clinically depressed mothers. Moreover, certain potential effects were not examined by either study. For example, while both studies assessed maternal behavior in the first phase of the SFP, neither examined whether temperament and maternal behavior interacted to account for variations in infant still-face response. Within the current study, we attempt to address some of the questions raised by the reviewed research on the individual differences in infant still-face response. Specifically, we examine the associations between observations of maternal behavior taken outside the SFP, maternal reports of infant temperament, and infant still-face response, in both high- and low-risk mother–infant dyads. All dyads in this study were selected from a larger study of child development in high and low social risk contexts. The high-risk group was made up of adolescent mothers and their infants. This specific group was included in this report because the children of adolescent mothers are known to be at risk for a number of social and emotional difficulties in development, thought to be related to problems in interactions and the developing mother–infant relationship. Like mothers from other socially high-risk groups, adolescent mothers vocalize less with their infants, engage them less frequently in eye-to-eye contact and interaction, are more susceptible to neglect and ignore them, and also more frequently develop insecure attachment relationships (Dukewich, Borkowski, & Whitman, 1996; Spieker & Bensley, 1994; van Ijzendoorn et al., 1992; Ward & Carlson, 1995). The low-risk group is made up of middle-class adult mothers and their infants. By including both a high- and low-risk group in this study it is possible to examine whether the associations between infant still-face response and maternal behavior, infant temperament and their interactions are the same within each developmental context.

2. Method 2.1. Subjects Seventy-six mothers agreed to participate in this study with their 6-month-old infants. Mothers were recruited with the help of maternity ward nurses of the major birthing hospital of a small Canadian city in the province of Québec (population 100,000) and with the help of visiting nurses of local community health centers. Adult mothers (N = 35) had a mean age of 29 years (SD = 4.4) and 15 years of schooling (SD = 3.1) at the birth of their child. Mean annual family income was between 30,000 and 45,000$ (Canadian), and all but one of the mothers was married or was living with the biological father of the infant. All but one mother were Caucasian and French-speaking. Inclusion criteria for infants were that they were born weighing more than 2,500 g, without physical or congenital anomalies. Only one of the infants of adult mothers were born before 37 weeks of pregnancy, at 36 weeks. Mean birth weight for this group was 3,533 g (SD = 472 g). There were 15 girls and 20 boys. Adolescent mothers (N = 41) had a mean age of 17 years (SD = 1.56) and had spent 10 (SD = 1.7) years in school. Mean annual income was between 0 and 15,000$ (Canadian). Eleven of the mothers lived alone, 5 with their parents and 25 lived with the biological father

G.M. Tarabulsy et al. / Infant Behavior & Development 26 (2003) 421–438

427

of the child at the time of the study. All were Caucasian and French-speaking. Six infants were born between 34 and 37 weeks of pregnancy, all others after 37 weeks. Mean birth weight was 3,395 g (SD = 467). There were 21 girls and 20 boys. An additional six dyads were originally part of the study but were not included in present analyses because of irregularities in the SFP. One of the infants of an adolescent mother was too fussy for the procedure to be carried out. Three of the mothers interacted with the infant during the still-face phase by laughing, touching or gesturing to the child and two adolescent mothers gave their infants toys or pacifiers during the procedure. These tapes were viewed as inappropriate and were not coded. 2.2. Measures and procedure 2.2.1. Maternal Behavior Q-Set (MBQS) (Pederson & Moran, 1995) The MBQS is a 90-item instrument for measuring the quality of maternal behaviors during mother–infant interactions in the home. Each item describes potential maternal behaviors. Items are first sorted as being most-like, neutral or unlike the mother under observation. Each group of items is then sorted again into three groups, yielding a total of nine clusters of items. Items in the first cluster, perceived as most descriptive of the mother, are assigned a score of 9; items in the second group receive a score of 8, and so on. Correlations, corresponding to the score for maternal behavior score, are calculated between subject sorts and a criterion sort for the prototypically sensitive and responsive mother provided by the developers of the instrument. Thus, scores vary from −1.0 (less sensitive/responsive) to 1.0 (prototypically sensitive/responsive). The development of the MBQS is anchored in attachment theory, and particularly in the descriptions of sensitivity and responsiveness provided by Ainsworth and her colleagues (Ainsworth et al., 1978). Pederson, Moran and their colleagues (Pederson, Gleason, Moran, & Bento, 1998; Pederson & Moran, 1995, 1996; Tarabulsy, Avgoustis, Phillips, Pederson, & Moran, 1997) have provided detailed descriptions of the development of the MBQS, as well as validity and reliability information. The longitudinal studies of these authors have demonstrated that the MBQS is useful in predicting several aspects of social and emotional development. Moreover, the MBQS is significantly correlated with other assessments of maternal behavior, such as those measured with the HOME Inventory (Caldwell & Bradley, 1978) and the Ainsworth scales (Ainsworth et al., 1978). These reports have also shown high levels of inter-rater agreement and test-retest stability. 2.2.2. Infant characteristics questionnaire (ICQ) (Bates, Freeland, & Lounsbury, 1979) The ICQ has often been used in temperament research and is particularly suited for 6-monthold infants. The scale focuses on the basic notion of “difficultness” or “fussiness” in the study of temperament (Bates, 1992). It is composed of four subscales: (1) fussy-difficultness; (2) unadaptedness; (3) dullness of expression; and (4) unpredictability. A summary score of “overall difficultness” is also derived and will be used presently in all analyses. When completing the ICQ, mothers determine to what degree their infant’s behavior corresponds to questions asked in each specific item. Items are rated on a 7-point Likert scale where a score of 4 corresponds to an “average” infant response, scores of 1 are viewed as being not at all irritable, and scores

428

G.M. Tarabulsy et al. / Infant Behavior & Development 26 (2003) 421–438

of 7 are viewed as very difficult responses by the child. Higher summary scores indicate that mothers perceive their infant as being more fussy than others, as being more reactive to novelty and more difficult to soothe. Bates (1992) reports high levels of convergence with various subscales of other maternal reports of temperament that address dimensions similar to the four ICQ scales, such as the Revised Infant Temperament Questionnaire (Carey & McDevitt, 1978) and the Infant Behavior Questionnaire (Rothbart, 1986), as well as high test-retest correlations. In the current report, Chronbach’s internal consistency alpha coefficients were the following: difficultness = 0.68, unadaptedness = 0.82, dullness = 0.73, unpredictability = 0.63. Chronbach’s alpha for the summary difficultness score was 0.90. 2.2.3. Still-face procedure (Tronick et al., 1978) The SFP was conducted in the laboratory, in a manner similar to that described in other reports. Infants were seated in a child seat (similar to a car seat) placed on a table, facing mother, approximately a meter away. A different video camera focused on each of the members of the dyad and images were recorded simultaneously, using a split-screen generator and a digital timer. The procedure was composed of three 2-min episodes. Mothers were instructed to interact with their infants in a playful manner during the first phase, as they would under normal circumstances at home. During the second phase, they were to adopt a neutral expression and not respond to their infant. In the final phase, they were to return to their playful, interactive mode. All phases were separated by light taps on a two-way mirror. The present report focuses on infant behaviors and affect manifested during the still-face phase only. As has been reported in other studies (e.g., Mayes & Carter, 1990), phases 2 or 3 were aborted if infants became highly distressed. If distress in phase 2 persisted for more than 20 s, the experimenter had to move to the next phase—if in phase 3, the experimenter had instructions to end the procedure. 2.3. Coding scheme The coding scheme was devised from a reading of other reports having focused on infant still-face responses (Field, 1984; Kiser et al., 1986; Mayes & Carter, 1990) and on descriptions of the AFFEX facial coding scheme (Izard, 1995). Ten codes were identified, sorted into three broad categories of infant behaviors and affect. Analyses will focus on the three categories only. 2.3.1. Positive affect expression Positive vocalizations—Vocalizations by the infant, directed towards the mother or not, manifested in the absence of an AFFEX expression for sadness (triangle shaped eyebrows and/or raised chin with lowered mouth corners), anger (lowered eyebrows, drawn together and/or open mouth, with corners drawn downward) or fear (raised eyebrows, drawn together and/or open mouth with retracted corners). Positive vocalizations may be accompanied by AFFEX expressions for interest (raised eyebrows), or surprise (raised eyebrows and round, open mouth). Positive affect—AFFEX expressions of joy (squinted eyes with a smile), or smiles that otherwise did not meet AFFEX criteria for joy.

G.M. Tarabulsy et al. / Infant Behavior & Development 26 (2003) 421–438

429

2.3.2. Negative affective expressions This category reflects infant manifestations of mild to intense distress, discomfort and/or fussiness. Fusses—Negative vocalizations (cries, whines or vocal manifestations of discomfort) that did not exceed five seconds in duration and were manifested in the absence of sadness, anger or fear. Frowns—Facial expressions of discomfort that did not clearly meet AFFEX expressions of sadness, anger or fear. Extended fusses—Behaviors coded as fusses but that exceeded five seconds in duration and were manifested in the absence of sadness, anger or fear. Moderate distress—Behaviors coded as fusses or extended fusses accompanied by expressions of sadness, anger or fear. Crying was also coded here. Extreme distress—Behaviors coded as extended fusses or cries but where infants seemed to be more distressed than the descriptions these codes allowed. Infants for whom this code was attributed may have arched their backs and/or cried intensely. 2.3.3. Self-soothing behavior The following behaviors have been described by other researchers as reflecting infant attempts to manage distress during the still-face phase of the procedure (e.g., Field, 1994; Mayes & Carter, 1990). Self-comforting behavior—Behaviors in this code included thumb or finger-sucking, playing with a part of the infant seat or a piece of clothing. Gazes away from mother—Looks away from mother to focus on something on the wall. Coders viewed tapes and noted the occurrence of behaviors and affect expressions, and their duration. The scores used in the analyses are the proportions of time that infants manifested each of the behaviors and affect expressions. Coders were unaware of any other details of the study at the time of coding. Inter-rater agreement (intra-class correlations), established for 17 infants (22.4%), ranged between a low of 0.70 (moderate distress) to 0.95 (fusses, frowns, positive vocalizations). All but two of the codes had agreements superior to 0.80. Agreement for the three categories of infant behaviors and affect ranged were as follows: positive affect— 0.92; negative affect—0.95 and self-soothing behavior—0.75. Correlations between the three categories were the following: negative affect and positive affect—r = −.46 (p < .0001); negative affect and self-soothing—r = −.23 (p < .05); positive affect and self-soothing— r = −.06 (ns). 2.4. Home visits Two trained observers conducted semi-structured home visits of two to three hours when infants were 6 months old. The training of home visitors was composed of the following: (A) seminars on early mother–infant interaction. The seminar included theoretical aspects and the viewing of a number of video-tapes of mother–infant interactions, involving both adolescent and adult mothers; (B) the description of videotaped interactions, followed by actual coding of a number of taped interactions with the MBQS; (C) approximately 10 practice home visits followed by post-visit interviews with the first author. These interviews had as their purpose to

430

G.M. Tarabulsy et al. / Infant Behavior & Development 26 (2003) 421–438

help observers focus on relevant mother–infant interaction during their visits. In total, training observers for home visits took about 2 months. Each observation visit was led by one of the observers. This observer had most of the contact with the mother and completed the MBQS after the visit, whereas the other observer prepared and manipulated different materials. Visits included an interview with the mother, a developmental assessment of the infant, a brief, 5-min free-play period with toys brought by the home visitors, and a series of questionnaires that the mother had to complete. Tasks were chosen and conducted such that, apart from the free-play sequence, research procedures and infant demands competed for maternal attention (for a detailed description, see Pederson & Moran, 1995). Home visitors kept running notes of maternal and infant behaviors and interactions during the visit. Following the visits, observers completed their notes independently and the visit leader completed the MBQS. In the case of 16 dyads (21%), both observers completed the MBQS with an agreement rate of 0.91 (intra-class correlation). Following the home visit, an appointment was made with the mother to visit the university laboratory, approximately two weeks later, to carry out the SFP.

3. Results Table 1 presents descriptive statistics for the major variables in this study. All scores were transformed into z-scores prior to further analyses to ensure greater respect of primary distribution assumptions of regression analyses. To address study objectives, three multiple regression analyses were carried out, one for each category of infant responses during the maternal still-face: positive affect, negative affect and self-soothing behavior. The three predictive variables were inserted in the regression equation: risk status (dummy coded as 0 for adult mother dyads and 1 for adolescent mother dyads), infant difficultness and maternal behavior. Interactions between risk and difficultness, risk and maternal behavior and maternal behavior and infant difficultness were examined. Regressions are summarized in Table 2.

Table 1 Means and standard deviations for maternal sensitivity, ICQ summary score and infant still-face response variables (n = 76) Variable

Maternal behavior Infant difficultness Negative affect in SFP Positive affect in SFP Self-soothing in SFP

Adolescent mother dyads

Adult mother dyads

Means (SD)

Means (SD)

0.29 (0.43) 22.35 (8.49) 48.63 (40.32) 17.80 (20.30) 30.49 (25.61)

0.52 (0.39) 23.46 (8.56) 51.00 (45.19) 20.39 (18.53) 39.94 (31.48)

Note A: Infant response in SFP scores are percentages of total time. Total percentage may exceed 100% because infants sometimes showed more than one behavior/expression of affect at the same time. Conversely, total percentage may not attain 100% in cases where infants did not show codable behaviors. Note B: The only variable that was different for adolescent and adult dyads was maternal behaviour (p < .01).

G.M. Tarabulsy et al. / Infant Behavior & Development 26 (2003) 421–438

431

Table 2 Summary of linear multiple regressions Variable

b

Standard error

Beta

T

p

Dependent variable: positive affect expressiona Intercept 0.08 Group −0.26 Infant difficultness 0.66 Maternal behavior 0.39 Maternal behavior × difficultness 0.12 Group × difficultness 0.44 Group × maternal behavior −0.09

0.28 0.39 0.72 0.61 0.22 0.44 0.38

0.00 −0.08 0.38 0.24 0.07 −0.41 −0.09

0.30 −0.67 0.91 0.64 0.56 −0.99 −0.24

ns ns ns ns ns ns ns

Dependent variable: negative affect expressionb Intercept −0.11 Group −0.07 Infant difficultness 0.85 Maternal behavior −2.27 Maternal behavior × difficultness −0.54 Group × difficultness −0.29 Group × maternal behavior 1.59

0.40 0.54 1.01 0.86 0.30 0.62 0.54

0.00 −0.01 0.32 −0.92 −0.20 −0.18 1.03

−0.28 −0.13 0.84 −2.64 −1.76 −0.47 2.96

ns ns ns .01 .10 ns .01

Dependent variable: self-soothing behaviorc Intercept 0.06 Group −0.17 Infant difficultness −0.22 Maternal behavior 0.84 Maternal behavior × difficultness −0.35 Group × difficultness 0.18 Group × maternal behavior −0.52

0.16 0.21 0.40 0.34 0.12 0.24 0.21

0.00 −0.08 −0.21 0.86 −0.34 0.28 −0.86

0.37 −0.79 −0.55 2.50 2.94 0.73 −2.49

ns ns ns .05 .01 ns .05

(F(6, 75) = 0.46, p > .83) model R2 = 3.82%. (F(6, 75) = 2.59, p < .05) model R2 = 18.37%. c (F(6, 75) = 2.83, p < .05) model R2 = 19.76%. a

b

Significant models emerged for infant negative affect and self-soothing behavior but not for infant positive affect. 3.1. Negative affect Results indicate the presence of a main effect for maternal behavior. Overall, infant negative affect was inversely related to the quality of maternal interactive behavior. However, this effect must be interpreted in terms of a significant interaction observed between maternal behavior and risk status, as well as a marginally significant (p = .08) interaction between maternal behavior and infant difficultness. Separate regressions were carried out for the high-risk and the low-risk groups to break down the risk by maternal behavior interaction. The direction of the main effect for maternal behavior was confirmed for the low-risk group (beta = −0.26) indicating that the infants of more sensitive mothers showed lower levels of negative affect during the still-face phase. For the high-risk group of adolescent mothers, however, this association was direct (b = 0.38), suggesting that infants of more sensitive mothers showed greater amounts of negative affect. This interaction was graphically depicted in Fig. 1 by breaking down the entire sample of dyads

432

G.M. Tarabulsy et al. / Infant Behavior & Development 26 (2003) 421–438

Fig. 1. Relations between infant negative affect during the still face and maternal behavior as a function of risk group.

on the basis of maternal behavior and charting means for infant negative affect as a function of risk group and maternal behavior. All dyads were assigned to one of three maternal behavior groups according to whether they were observed to be in the lower, middle or upper third in terms of the quality of maternal behavior. In the lower third, there were 9 adult (25.7%) and 16 adolescent (39.0%) mothers. The middle third was made up of 8 adult (22.9%) and 18 adolescent (43.9%) mothers, whereas the upper third was composed of 18 adult (51.4%) and 7 adolescent (17.1%) mothers. The marginal interaction between maternal behavior and infant difficultness was broken down following a procedure described by Aiken and West (1991). These authors propose to break down interactions between continuous variables by fixing the level of one of the interactive variables (difficultness) at one standard deviation above the mean and one standard deviation below the mean by examining the association between the other interactive variable (maternal behavior) and the dependent variable (infant negative affect) in each regression. Simple effects for fixed levels of infant difficultness show that with greater levels of difficultness, maternal behavior is more closely associated with negative affect (b = −1.14) than at lower levels of difficultness (b = −0.70). So while the overall direction of the effect is the same at all levels of difficultness, the relation between maternal behavior and negative affect appears to be more important when infants are more irritable. 3.2. Self-soothing behavior Results indicate a significant association between maternal behavior and self-soothing behavior, where infants of sensitive mothers showed greater levels of self-soothing behavior when confronted with the maternal still face. However, this effect must be interpreted in light of two

G.M. Tarabulsy et al. / Infant Behavior & Development 26 (2003) 421–438

433

Fig. 2. Relations between infant self-soothing during the still face and maternal behavior as a function of risk group.

significant interactions, one between maternal behavior and risk status and one between infant difficultness and maternal behavior. Separate regressions were carried out for the high-risk and the low-risk groups to break down the risk by maternal behavior interaction. The marginal tendency observed for the sample as a whole was confirmed only for the low-risk group (beta = 0.33) indicating that the infants of more sensitive mothers showed greater levels of self-soothing during the still-face phase. For the high-risk group of adolescent mothers, however, this association was inverse (beta = −0.23), suggesting that infants of more sensitive mothers showed lower levels of self-soothing. We chose the same strategy to depict this interaction as in the previous risk group by maternal behavior interaction predicting infant negative affect (see Fig. 2). The breakdown of the maternal behavior by infant difficultness interaction was carried out by fixing levels of infant difficultness at one standard deviation above the mean and one standard deviation below the mean. Simple effects for fixed levels of infant difficultness show that with greater levels of difficultness, maternal behavior is less closely associated with negative affect (beta = 0.50) than at lower levels of difficultness (beta = 1.22). Overall, the direction of the association is the same, but the relation between maternal behavior and infant self-soothing is greater when infants are perceived as less fussy.

4. Discussion The purpose of this paper was to investigate the association between the quality of maternal interactive behavior assessed outside the SFP and infant still-face response while considering potential contributions of child temperamental characteristics and social risk. While previous work that had addressed the association between maternal behavior and infant still-face response by studying high-risk dyads or by predicting attachment security both suggested that

434

G.M. Tarabulsy et al. / Infant Behavior & Development 26 (2003) 421–438

the two constructs were linked, certain inconsistencies emerged as to the nature of this association. Moreover, few reports had examined this association using both a high- and a low-risk group while considering possible interactions with infant temperament. 4.1. Maternal interactive behavior The results of the current study support the idea that maternal behavior is associated with infant still-face response and that infant response to the maternal still face is representative of mother–infant interactive experience. Overall, maternal behavior was found to be inversely related to the proportion of time infants engaged in negative affect during the still-face phase, and directly related to the amount of self-soothing behavior emitted by infants. As such, these results support previous findings from studies that predicted attachment security from infant still-face response in that greater levels of maternal sensitivity are linked to less negative affect and more adaptive behavior on the part of the infant (Braungart-Rieker et al., 2001; Cohn et al., 1991; Tronick et al., 1982). These specific effects do not support results obtained from high-risk samples (e.g., Pelàez-Nogueras et al., 1996). In both cases, however, results must be interpreted in light of significant interactions with the level of social risk. For the group of adolescent mothers, infant negative affect increased and self-soothing behavior decreased with greater levels of sensitivity. The pattern of findings is inversed for the low-risk group of adult mother–infant dyads. Here, with greater levels of sensitivity, infants show less negative affect and somewhat greater levels of self-soothing behaviors. This interaction indicates that some of the inconsistencies observed in the literature on the association between maternal behavior and infant still-face response, evoked by previous studies with high-risk dyads or by those that predicted attachment security may actually have tapped into differences in the way that infants respond to mothers who are parenting in different ecological contexts. The responses of infants in high- and low-risk contexts do not appear to follow the same patterns. The responses of infants in the low-risk group are concordant with studies that have mostly used low-risk samples to link still-face response to attachment security. The pattern of infant responses in the high-risk group are similar to findings reported by Field and her colleagues with the infants of depressed mothers (Field et al., 1990) and adolescent mothers (Pelàez-Nogueras et al., 1996). Thus, by examining the interaction between maternal behavior taken outside the SFP and risk status, the current results shed new light on hypotheses linking maternal behavior and infant still-face response. From a theoretical perspective, these results raise the issue of the function and developmental significance of different patterns of infant responses to the maternal still face. When the infants of adolescent mothers exposed to low levels of sensitivity show little or no negative affect in response to the still face, does this have the same function as when the child of a low risk, sensitive mother shows low levels of negative affect? We suggest that these findings imply that different developmental processes may be at work in each of these groups, reflecting the broad developmental paths that each dyad is currently engaged in. Further longitudinal research in this area (e.g., Braungart-Rieker et al., 2001; Moore et al., 2001) linking these factors to later outcome will certainly help in unraveling the meaning and function of infant responses in different caregiving contexts.

G.M. Tarabulsy et al. / Infant Behavior & Development 26 (2003) 421–438

435

4.2. Temperament Unlike the findings of Braungart-Rieker and her colleagues (1998) and like the findings of Cohn et al. (1991), there were no main effects attributed to infant difficultness for any of the three categories of infant responses during the still-face phase. However, difficultness was found to moderate the association between maternal interactive behavior and infant self-soothing response. When infants were perceived as being less difficult, the positive association between maternal behavior and the amount of self-soothing the infant engaged in was stronger. A statistically marginal interaction was also found between difficultness and maternal behavior in accounting for infant negative affect. The interaction indicates that with more difficult infants, a stronger, inverse association emerged between maternal behavior and negative affect. Thus, infant fussiness relates differently to still-face response, as a function of maternal behavior and the class of responses considered. One interpretation of the findings on temperament is that fussier infants elicit maternal interactive behavior that focuses on signaling behaviors and expected, subsequent interaction, to regulate the emotions elicited in distressing circumstances, such as the maternal still face. Fussier infants overtly depend on mothers to help in managing distress, hence they respond with negative affect to their mother’s still face. Infants who are less fussy may be exposed to interactive behavior where the emphasis is placed on the coping skills that are associated with self-soothing in times of distress. In this interpretation, fussiness, or lack thereof, disposes infants to interact with parents in a way that emphasizes either dyadic processes or self-reliance as a means of dealing with experienced distress. An additional suggestion from this data is that difficult temperament is not a problematic circumstance per se, as it is sometimes portrayed in the developmental literature. The current results suggest that while temperament is important in determining the nature of the child’s response to distressing circumstances, the action of temperament appears to vary as a function of the quality of mother–infant interactions the infant is exposed to on a daily basis. Similar suggestions have been made by Belsky (1999) in the context of the development of attachment security, and de Vries (1989) in the study of cultural differences in infant temperament.

5. Conclusion The results of this study provide support for the association between maternal behavior and infant still-face response. Both infant negative affect and self-soothing behavior during the still-face phase appear to be linked to observations of maternal behavior during a two- to three-hour home visit occurring about two weeks prior to conducting the laboratory procedure. However, the most significant findings from this paper are that maternal behavior interacts with both social risk and infant temperament in its association to infant still-face responses. There is some caution needed in interpreting some of these effects—the sample of infants is small, some of the effects are marginal and clearly there is a need for replication of findings. However, by including both a high- and a low-risk groups, the current findings are helpful in addressing some of the discrepancies observed in previous research. Moreover, by examining the interactions between temperament and maternal behavior, some new hypotheses emerge,

436

G.M. Tarabulsy et al. / Infant Behavior & Development 26 (2003) 421–438

in line with current theories of potential roles for infant temperament in social and emotional development. Clearly, the developmental significance of the pattern of results obtained in this study will depend on future replication and longitudinal investigation of later outcome. It is hoped that the current results will be helpful in stimulating such research and underscore the importance of the SFP in developmental research.

Acknowledgments The authors would like to thank the mothers and infants who generously donated their time, and the following persons who provided invaluable assistance during data collection: Sophie Allaire, Manon Fournier, Marie-France Gagnon, Isabelle Hémond, Marie Larose, and Julie Robitaille. We also want to thank Stéphane Duchesne for helpful comments throughout the writing process. This study was made possible by funding from the Fonds pour les chercheurs et l’avancement de la recherche, the Conseil québécois de la recherche sociale and the Social Sciences and Humanities Research Council of Canada.

References Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage. Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: A psychological study of the strange situation. Hillsdale, NJ: Lawrence Erlbaum Associates. Bates, J. E. (1992). Information on the infant characteristics questionnaire. Unpublished manuscript, Department of Psychology, Indiana University, Bloomington, IN. Bates, J. E., Freeland, C. A. B., & Lounsbury, M. L. (1979). Measurement of infant difficultness. Child Development, 50, 794–803. Belsky, J. (1999). Interactional and contextual determinants of attachment security. In J. Cassidy & P. Shaver (Eds.), Handbook of attachment: Theory, research and clinical applications. New York: Guilford. Braungart-Rieker, J. M., Garwood, M. M., Powers, B. P., & Notaro, P. C. (1998). Infant affect and affect regulation during the still-face paradigm with mothers and fathers: The role of infant characteristics and parental sensitivity. Developmental Psychology, 34, 1428–1437. Braungart-Rieker, J. M., Garwood, M. M., Powers, B. P., & Wang, X. (2001). Parental sensitivity, infant affect and affect regulation: Predictors of later attachment. Child Development, 72, 252–270. Caldwell, B. M., & Bradley, R. H. (1978). Home observation for the measurement of the environment. Unpublished manual, University of Arkansas at Little Rock. Carey, W. B., & McDevitt, S. C. (1978). Revision of the infant temperament questionnaire. Pediatrics, 61, 735–739. Carter, A. S., Mayes, L. C., & Pajer, K. A. (1990). The role of dyadic affect in play and infant sex in predicting infant responses to the still-face situation. Child Development, 61, 764–773. Cohn, J. F., Campbell, S. B., & Ross, S. (1991). Infant response in the still-face paradigm at 6 months predicts avoidant and secure attachment at 12 months. Development and Psychopathology, 3, 367–376. Cohn, J. F., Matias, R., Tronick, E. Z., Connell, D., & Lyons-Ruth, K. (1986). Face-to-face interactions of depressed mothers and their infants. New Directions for Child Development, 34, 31–45. Cohn, J. F., & Tronick, E. Z. (1987). The sequence of dyadic states at 3, 6, and 9 months. Developmental Psychology, 23, 68–77. Cohn, J. F., & Tronick, E. Z. (1988). Mother–infant face-to-face interaction: Influence is bidirectional and related to periodic cycles in either partner’s behavior. Developmental Psychology, 24, 386–392.

G.M. Tarabulsy et al. / Infant Behavior & Development 26 (2003) 421–438

437

Crockenberg, S. B. (1981). Infant irritability, mother responsiveness, and social support influences on the security of infant–mother attachment. Child Development, 52, 857–865. de Vries, M. W. (1989). Difficult temperament: A universal and culturally embedded concept. In W. B. Carey & S. C. McDevitt (Eds.), Clinical and educational applications of temperament research. Berwyn, PA: Swets North America. de Wolff, M. S., & van Ijzendoorn, M. H. (1997). Sensitivity and attachment: A meta-analysis on parental antecedents of infant attachment. Child Development, 68, 571–591. Dukewich, T. L., Borkowski, J. G., & Whitman, T. L. (1996). Adolescent mothers and child abuse: An evaluation of risk factors. Child Abuse & Neglect, 20, 1031–1047. Ellsworth, C. P., Muir, D. W., & Hains, S. M. (1993). Social competence and person-object differentiation: An analysis of the still-face effect. Developmental Psychology, 29, 63–73. Field, T. M. (1984). Early interactions between infants and their postpartum depressed mothers. Infant Behavior and Development, 7, 527–532. Field, T. M. (1994). The effects of mother’s physical and emotional unavailability on emotion regulation. In N. A. Fox (Ed.), The development of emotion regulation: Biological and behavioral considerations. Monographs of the Society for Research in Child Development, 59(2/3, Serial No. 240). Field, T. (1995). Psychologically depressed parents. In M. H. Bornstein (Ed.), Handbook of parenting: Vol. 4. Applied and practical parenting (pp. 85–99). Hillsdale, NJ: Erlbaum. Field, T. M., Healy, B., Goldstein, S., & Guthertz, M. (1990). Behavior-state matching and synchrony in mother–infant interactions of nondepressed versus depressed dyads. Developmental Psychology, 26, 7–14. Fogel, A. (2000). Systems, attachment, and relationships. Human Development, 43, 314–320. Fox, N. A., Henderson, H. A., Rubin, K. H., Calkins, S. D., & Schmidt, L. A. (2001). Continuity an discontinuity of behavioral inhibition and exuberance: Psychophysiological and behavioral influences across the first four years of life. Child Development, 72, 1–21. Goldsmith, H. H., & Rothbart, M. K. (1994). The laboratory temperament assessment battery. Unpublished manual, The University of Wisconsin, Department of Psychology. Izard, C. E. (1995). The maximally discriminative facial movement coding system. Unpublished manuscript, University of Delaware. Kiser, J. L., Bates, J. E., Maslin, C., & Bayles, K. (1986). Mother–infant play at six-months as a predictor of attachment security at thirteen months. Journal of the American Academy of Child Psychiatry, 25, 68–75. Mayes, L. C., & Carter, A. S. (1990). Emerging social regulatory capacities as seen in the still-face situation. Child Development, 61, 754–763. Moore, G. A., Cohn, J. F., & Campbell, S. B. (2001). Infant affective responses to mother’s still face at 6 months differentially predict externalizing and internalizing behaviors at 18 months. Developmental Psychology, 37, 706–714. Pelàez-Nogueras, M., Field, T. M., Hossain, Z., & Pickens, J. (1996). Depressed mothers’ touching increases infants’ positive affect and attention in still-face interactions. Child Development, 67, 1780–1792. Pederson, D. R., Gleason, K., Moran, G., & Bento, S. (1998). Maternal attachment representations, maternal sensitivity, and the infant–mother attachment relationship. Developmental Psychology, 34, 925–933. Pederson, D. R., & Moran, G. (1995). A categorical description of attachment relationships in the home and its relation to q-sort measures of infant–mother interaction. In E. Waters, B. Vaughn, G. Posada, & K. Kondo-Ikemura (Eds.), Caregiving, cultural and cognitive perspectives on secure-base behavior and working models: New growing points of attachment theory and research. Monographs of the Society for Research in Child Development. Pederson, D. R., & Moran, G. (1996). Expressions of the attachment relationship outside of the strange situation. Child Development, 67, 915–927. Rothbart, M. K. (1986). Longitudinal observation of infant temperament. Developmental Psychology, 22, 356–365. Shapiro, B., Fagen, J., Prigot, J., Carroll, M., & Shalan, J. (1998). Infants’ emotional and regulatory behaviors in response to violations of expectancies. Infant Behavior and Development, 21, 299–313. Spieker, S. J., & Bensley, L. (1994). Roles of living arrangements and grandmother support in adolescent mothering and infant attachment. Developmental Psychology, 30, 102–111. Stack, D. M., & Muir, D. W. (1990). Tactile stimulation as a component of social interchange: New interpretations for the still-face effect. British Journal of Developmental Psychology, 8, 131–145.

438

G.M. Tarabulsy et al. / Infant Behavior & Development 26 (2003) 421–438

Tarabulsy, G. M., Avgoustis, E., Phillips, J., Pederson, D. R., & Moran, G. (1997). Similarities and differences in mothers’ and observers’ descriptions of attachment behaviors. International Journal of Behavioral Development, 21, 599–619. Tarabulsy, G. M., Tessier, R., & Kappas, A. (1996). Contingency detection and contingent organization of behavior in interactions: Implications for socioemotional development in infancy. Psychological Bulletin, 119, 25–41. Toda, S., & Fogel, A. (1993). Infant response to the still-face situation at 3 and 6 months. Developmental Psychology, 29, 532–538. Tronick, E. Z., Als, H., Adamson, L., Wise, S., & Brazelton, T. B. (1978). The infants’ response to entrapment between contradictory messages in face-to-face interaction. Journal of American Academy of Child Psychiatry, 17, 1–13. Tronick, E. Z., Ricks, M., & Cohn, J. F. (1982). Maternal and infant affect exchange: Patterns of adaptation. In T. Field & A. Fogel (Eds.), Emotion and interaction: Normal and high risk infants. Hillsdale, NJ: Erlbaum. van den Boom, D. C. (1994). The influence of temperament and mothering on attachment and exploration: An experimental manipulation of sensitive responsiveness among lower-class mothers with irritable infants. Child Development, 65, 1457–1477. van den Boom, D. C., & Hoeksma, J. B. (1994). The effect of infant irritability on mother–infant interaction: A growth-curve analysis. Developmental Psychology, 30, 581–590. van Ijzendoorn, M. H., Goldberg, S., Kroonenberg, P. M., & Frankel, O. J. (1992). The relative effects of maternal and child problems on the quality of attachment: A meta-analysis of attachment in clinical samples. Child Development, 63, 840–858. Ward, M. J., & Carlson, E. A. (1995). Associations among adult attachment, maternal sensitivity, and infant attachment in a sample of adolescent mothers. Child Development, 66, 69–79. Weinberg, M. K., Tronick, E. Z., Cohn, J. F., & Olson, K. L. (1999). Gender differences in emotional expressivity and self-regulation during early infancy. Developmental Psychology, 35, 175–188. Zuravin, S., & DiBlasio, F. (1992). Child-neglecting adolescent mothers. Journal of Interpersonal Violence, 7, 471–489.