Associations between nonparental care experience and preschooler's emotion regulation in the...

JOURNAL OF APPLIED DEVELOPMENTAL PSYCHOLOGY 17,577~596 (1996)

Associations Between Nonparental Care Experience and Presdhooler’s

Emotion degulation in the Presence of the Mother

MICHAEL MORALES

LISA J. BRIDGES

University of California, Riverside

This study tested predictions concerning the associations between children’s non- parentol care experiences and emotion regulation. It involved 53 participants (21 girls and 32 boys) of preschool age and their mothers. The children’s care experiences ranged from those who were raised exclusively at home to those who had nonparental care experience beginning in early infancy. The participants were observed in a laboratory playroom, and their mothers were asked to complete questionnaires regarding their children’s care histories and their current care situations. The children with extensive nonporental care backgrounds were found to be more likely to use self-directed emotion regulation behaviors when faced with a mildly frustrating situation. In the same situation, the children who had experienced minimal nonparental core were found to be more likely to use other-directed emotion regulo- tion behaviors. Additionally, the children who had experienced care settings with

large numbers of children in each care group or who had experienced many caregiver changes were more likely to use self-directed emotion regulation behaviors, whereas those who had experienced settin s with smaller care groups and fewer caregiver changes were more likely to use ot er-directed emotion regulation behav- I

iors.

Considerable debate has arisen over the past several years regarding the impact of nonparental care on children’s socioemotional development. Some researchers have reported negative effects of nonparental care (e.g., Bar-glow, Vaughn, & Molitor, 1987; Belsky, 1988; Belsky & Rovine, 1988; Sroufe, 1988). Schwartz, Strickland, and Krolick (1974), for example, found evidence that preschoolers with day care experience prior to 2 years of age were more aggressive, more noncompliant, and showed a lower tolerance of frustration than did children who entered day care after age 2.

In contrast, however, other researchers have argued that evidence for negative effects of nonparental care per se is inconclusive or nonexistent when the quality of care is taken into account (Scan-, Phillips, & McCartney, 1990; Thompson, 1988). In some contradiction to the Schwartz et al. (1974) paper, for example,

Correspondence and requests for reprints should be sent to Michael Morales, The University of

California, Riverside, Department of Psychology, Riverside, CA 925214426.

577

578 MORALES AND BRIDGES

Howes and Olenick (1986) found that children who attended high-quality day care centers were likely to display more behavioral self-regulation, opera- tionalized as a child’s ability to refrain from touching forbidden toys while in the presence of the mother, and to resist eating during the mother’s absence from the laboratory, than were children without such day care experience.

One area of particular controversy has involved the issue of nonparental care and the quality of attachment. It has been suggested that entry into nonmaternal care in the 1st year of life is a risk factor for a child’s socioemotional development because it may affect the quality of the infant-mother attachment. More specifically, infants who experience frequent separations from their mothers may develop insecure-avoidant attachment relationships with them because of their frequent unavailability (Belsky, 1988; Bretherton & Waters, 1985; Sroufe, 1988). Belsky (1988; Belsky & Rovine, 1988) has suggested that there appears to be an emerging pattern in which 20 or more hr per week of nonmaternal care, particularly if initiated during the 1st year, is associated with the tendency of the infant to avoid or maintain a distance from the mother following separation. The validity of this contention has been vigorously debated (Belsky, 1988; Belsky & Braungart, 1991; Belsky & Rovine, 1988; Clarke-Stewart, 1988, 1989; Hoff- man, 1989; Richters & Zahn-Waxler, 1988; Starr, Phillips, & McCartney, 1990; Sroufe, 1988; Thompson, 1988).

One area that has received less attention concerns the associations between nonparental care and emotion regulation. A number of researchers have suggested that a child’s ability to regulate emotion is based, in part, in early infant- parent interactions (Kopp, 1989; Thompson, 1990; Tronick, 1989). Those interactions, in turn, may be affected by other experiences, such as nonparental care (Field, 1994; Thompson, 1990). The study presented here was designed to examine how the duration of, and timing of entry into, varying child care situations affect the development of emotion regulation during the preschool period.

A general definition of emotion regulation is that it involves the monitoring of emotion and the implementation of various techniques or strategies to control one’s emotional state (Bridges & Grolnick, 1995; Calkins, 1994; Kopp, 1989; Thompson, 1994). More specifically, Bridges and Grolnick (1995) have defined emotion regulation as the set of processes involved in initiating, maintaining, and modulating emotional expression and experience. These researchers have further suggested that emotional responsiveness and the use of particular behavioral strategies are two interdependent component processes involved in emotion regulation .

Emotional responsiveness refers to the extent to which individuals express and experience various emotions. Individual differences in emotional responsiveness are demonstrated by differences in the intensity of emotional expressions and in temporal properties of emotion such as the latency of response to, and recovery from, an arousing stimulus (Bridges & Grolnick, 1995). Moreover, individual

NONPARENTAL CARE AND EMOTION REGULATION 579

differences in emotional responsiveness are related to differences in physiologi- cal responsiveness that may affect the experience and expression of emotional arousal (Rothbart & Posner, 1985).

Strategy use refers to behavioral strategies deployed by the individual to regulate emotional experience (Gianino & Tronick, 1988; Kopp, 1989). Strate- gies may involve a number of different behavioral tendencies, such as approach or withdrawal from a stimulus, orientation of attention toward or away from a stimulus, and physical self-stimulation, as well as cognitive and symbolic manipulations such as when a young child repeats “She’s coming back” while waiting for the return of an absent parent.

There are a number of ways in which different emotion regulation strategies can be described. For example, emotion regulation strategies vary in the extent to which they are active (e.g., approach, reorientation of attention away from an arousing aspect of the environment toward another specific object or individual) versus passive (e.g., nondirected focus on the environment), internal (e.g., cognitive or symbolic manipulations of a situation) versus external (e.g., using a blanket or other “special object”), and self-directed (e.g., behaviors designed to regulate the individual’s own emotional state) versus other-directed (e.g., behaviors designed to regulate a caregiver’s behavior).

Early nonparental care experiences may influence developing emotion regulation styles in a number of ways, particularly with respect to the use of self- and other-directed emotion regulation strategies. In home- and center-based child care, children most frequently experience settings with several children of the same or similar age per adult care provider. In such a setting, it may be expected that children would develop more self-directed forms of emotion regulation, because of the competition for attention that may take place and the consequent increased latencies in caregiver response that they may experience. Additionally, children who are in substantial amounts of nonparental care may also experience their mothers as being less consistently available to provide assistance in regulat- ing emotions in potentially frustrating or otherwise distressing situations than do children who have experienced little or no nonparental care. If this is the case, greater reliance on self-directed, rather than other-directed, strategies would be expected even in situations in which the mother is present.

The study presented here was designed to examine the associations between nonparental care and the development of emotion regulation in a sample of preschoolers whose care experience ranged from those who were cared for exclusively by their mothers to those who had nonparental care experience beginning in infancy. The associations were assessed within a laboratory-based procedure involving mildly frustrating delays during which the child was required to wait to receive a clearly visible, desirable object. Brief frustrating delay situations in the presence of the mother or other adults are common occurrences for young children who must often wait for a drink, a snack, or an out-of-reach toy or other


object. This is likely to be similarly true for children in and out of nonparental care. Thus, the psychological equivalence of the setting for the two groups of children should be quite high.

It was expected that children who had experienced extensive nonparental care would be more likely to use self-directed emotion regulation strategies in this situation, and that children who had little or no nonparental care experience would be more likely to use other-directed emotion regulation strategies. Among children with nonparental care backgrounds, it was predicted that children who had experienced care settings with large numbers of children per adult or who had experienced many caregiver changes within their care setting would use more self-directed emotion regulation behaviors, and fewer other-directed behaviors, than would those who had experienced settings with smaller numbers of children per adult and fewer caregiver changes.

In relation to emotional responsiveness, it was expected that children who spent many hours each week in nonparental care would exhibit more positive emotion than would children with little or no nonparental care experience during a period in which the parent could not interact with them. Conversely, it was predicted that children with little or no nonparental care experience would be more positive than would children with extensive nonparental care when the parent was interactive. Among children with nonparental care backgrounds, it was predicted that children who had experienced care settings with fewer children per adult, or who had experienced fewer caregiver changes, would be more positive when their parent could interact with them, and less positive when their parent could not interact with them during the frustrating situations.

METHODS

Participants Participants in this study were 53 (21 girls, 32 boys) preschoolers (M = 48 months, SD = 5.4) and their mothers. The children came from one- (9) and two- parent (44) middle- to upper middle-class families. Maternal education ranged from high school to postgraduate level, with the median educational level being some college (at least 1 year). Of the children, 48 were White (non-Hispanic), two were African American, two were Hispanic, and one was Asian.

Mother-Reported Child Care Experience Mothers were asked to report on the types of nonparental care utilized (i.e., center-based care, home-based care, babysitters, or relatives) and the extent and quality of this care. Extent measures included the number of hours the child was currently in nonparental care per week, the child’s age at first entry into some form of nonparental care, and how long the child had been in his or her current care situation. Several commonly accepted (e.g., Starr, Phillips, & McCartney, 1990; Young & Zigler, 1986) qualitative measures were used, including the


child:adult ratio, the number of children in the child’s group, and the frequency of caregiver turnover within the current care setting. Furthermore, the mothers were asked to describe previous care arrangements, as well, using the same set of questions.

At the time of observation, 16 children had never experienced any form of regular nonparental care. The median age of first entrance for children who had entered nonparental care (N = 37) was 24 months (range, 0 to 40 months). These children had been in their current care situation for a median of 19.5 months (range, 1 to 40 months), and spent a median of 16.5 hr per week (range = 5 to 60 hr) in nonparental care. Ten of these children were cared for by nonrelated babysitters (N = 3) or by relatives (N = 7). In these settings, there was a median child:adult ratio of 1: 1 (range, 1 to 2). The remaining 27 children were in home- and center-based care. In these settings, there was a median child:adult ratio of 7.8:1 (range, 3 to 12 to l), and the median class size was 12.0 (range, 3 to 36). The median number of caregiver changes experienced within the current care situation was 1.0 (range, 0 to 7).

Because the children in this study ranged in age, some adjustments were necessary in variables involving age of entrance and length of time in current care situation. Because this was the age of the youngest child in the sample, 40 months was used as a cap for these variables. Thus, children who had never entered nonparental care, and those who had entered nonparental care after 40 months were assigned a value of 40 months on the variable, age of entry.

Laboratory Procedure

Overview. The laboratory procedures were adapted from those developed by Bridges and Grolnick (1995) for use with younger children. They consisted of placing the child in two mildly frustrating forced-delay situations separated by a 5-min free-play period. Two different situations were used in order to assess the child’s emotional responsiveness and self-directed regulatory strategies under varying levels of external support and assistance. During each situation the child was presented with an attractive object and required to wait 6 min to obtain it. In one situation (parent-active), mothers were told they could interact with their child as they wished, with the only restriction being they should not retrieve the stimulus object for their child. During the other situation (parent-passive), mothers were asked to fill out questionnaires and to interact with their child as little as possible.

Each mother-child pair was observed in the parent-active and the parent- passive situation. The order of these two situations was counterbalanced, as was which of two attractive items (either food or a wrapped gift) was presented during each delay. Thus, the combination of the delay object order and the parent instructional set order led to four possible delay conditions (e.g., parent-active/food and parent-passive/gift; parent-passive/food and parent-active/gift). A


single male experimenter conducted all assessments, and the laboratory procedures were videotaped for later coding.

Procedure for Wrapped Gift Dehy. During this episode, the experimenter brought a colorfully wrapped gift into the laboratory playroom. He allowed the child to get a close look at it, and said, “This present is for you, but I have to leave the room for a few minutes. When I come back, you can have it.” The experimenter then put the gift on top of a cabinet in the child’s view, but out of reach, and left the room. After 6 min, he returned and presented the child with the gift (a set of coloring pens and a coloring book).

Procedure for Food Delay. The food delay episode was the same as the gift delay, except that the stimulus was a single-serving bag of graham cookies. One additional different was that, after 3 mins, the experimenter entered, gave the child a cookie, and told the child that he or she would receive the rest of the cookies in a few more minutes. He then left for another 3 mins, after which he returned and gave the child the cookies. The break in the middle of the food delay episode was included to maintain consistency with the procedures described by Grolnick, Bridges, and Connell (1996).

Behavioral Coding Child behaviors conceptualized as emotion regulation strategies were coded during frustration situations. The particular strategies examined were selected to be consistent with those examined in earlier studies with younger children using the same laboratory situations (Bridges & Grolnick, 1995; Grolnick, Bridges, & Connell, 1996). These behaviors are presented in Table 1.

Two coders worked simultaneously but independently to note strategy use within lo-set intervals across each 6-min delay situation. Multiple strategies were possible in the same interval. Disagreements between coders were resolved by discussion and, when necessary, replaying videotaped segments. Summary scores used in data analyses consisted of the percentage of intervals (of a total of 36 intervals per situation per subject) containing each specific regulatory strategy.

Emotionality Coding. Child emotionality was coded similarly by two independent raters in the same lo-set intervals as strategy use. As with strategy coding, disagreements were resolved with discussion and, when necessary, replaying the relevant videotaped segments. To maximize independence of emotionality and strategy coding, all emotionality coding was conducted following the completion of all strategy coding. The affect scale designed for use in this study assessed expressed emotion on a positive to negative continuum, taking into account both facial and vocal expressions of affect. Ratings ranged from 1 to 7, with 1 signifying intense positive emotion denoted by full smile and sustained positive vocalizations; 4 indicating neutral, or no expression of either positive or


TABLE 1 Description of Emotion Regulation Strategies (ESRs)

ESR Child Behaviors

Self-initiated play (Act.play) Active play with parent (Act.engage) Low-level play

Using toy for intended purpose, other task directed play.

Low-level engagement of parent (Low-level engage) Self-soothing physical (Self-soothe physical) Self-soothing symbolic (Self-soothe symbolic)

Attempts to get the parent to play, ongoing reciprocal activity with parent.

Looking at toys or objects in room, wander- ing around the room.

Looking at toys held by parent, responding briefly to parent.

Mouthing hands, clothing. Hugging, mouthing a toy or object.

Talking about the experimenter. Enacting a pretend scenario where the frustration object is obtained.

Passive focus on frustration object Looks at frustration object from a distance; (Pass.focus) points at object. Active focus on frustration object (Actfocus)

Tries to retrieve frustration object; calls for parental assistance.

Playful focus on frustration object (Play.focus) Other-directed proximal (Other-dir.proximal)

Engages in playful banter regarding the frustration object.

Other-directed distal (Other-dir.distal)

Primarily include proximal forms of comfort seeking behaviors directed toward parent (e.g., touching parent).

Primarily include distal forms of comfort seeking behaviors directed toward parent (e.g., looking at parent).

negative emotion; and 7 denoting intense negative emotion, indicated by angry cryface, cryface, or frown, and sustained negative vocalizations. Intervals in which a child’s face was not in view, and he or she was not vocalizing, were not coded.

Summary scores used for later analyses included a mean emotion score across intervals, the peak or most intense negative emotion expressed, a latency to distress score based on the number of intervals before the onset of distress, and the minimum (most positive or least negative) emotion expressed.

Reliability of Observational Coding. Cohen’s Kappas (Cohen, 1960) were used to compute interrater agreement, prior to discussion, for both emotionality and behavioral strategy ratings. Each reliability rating was based on a random subsample of 20 subjects. Kappa coefficients for emotionality codings were: parent-active = .93, and parent-passive = .99. Kappas for strategy codings were computed for each strategy separately and across strategies for each situation. Kappa coefficients for each strategy coding ranged from .96 to .99 for parent-


active (median, .99), and .97 to .99 for parent-passive (median, .98). Overall Kappa coefficients for strategy codings were .92 for parent-active, and .96 for parent-passive.

RESULTS

Preliminary Analyses

Creation of Other-directed and Self-directed Emotion Regulation Strategy Composites. Because of the expectation that the self- versus other-directed di- mension might be most influenced by nonparental care, two composite variables, a self-directed emotion regulation variable and an other-directed emotion regulation variable, were created for use in our analyses, in addition to the individual emotion regulation strategy variables. Also, many of these strategies occur relatively infrequently with this situation. Thus, combining conceptually related strategies together served to preserve behavioral information while reducing problems with skewed distributions.

The self-directed composite variable consisted of the percentage of intervals containing emotion regulation strategies that could be defined as self-directed strategies. These included self-initiated play, low-level play, self-soothing physical, self-soothing symbolic, and passive focus on the object of frustration. The other-directed composite variable consisted of the percentage of intervals containing emotion regulation strategies that could be defined as other-directed strategies. These included active play with the parent, active focus on the object of frustration (because these most frequently involved requests for assistance from the mother), playful focus on the object of frustration, other-directed con- tact (proximal), and other-directed distal.

Mean Diflerences in Strategy Use-During Food and Gift Situations. Distinc- tions between parent-passive and parent-active episodes were of conceptual inter- est for this study, whereas differences between cookie and gift episodes were not. To determine whether significant differences existed in strategy use or emotionality during food and gift situations, independent sample t tests were conducted comparing behavior and emotionality scores for food and gift presentations within parent-active and parent-passive situations, separately. Of the 25 t tests that were conducted, only two were significant at p < .05, a number that may have occurred by chance. Therefore, food and gift presentation situations were combined in all subsequent analyses.

Mean Differences in Strategy Use During Parent-Active and Parent-Passive Situations. The means and standard deviations of the percentage of intervals in which the entire sample of children exhibited each of the emotion regulation strategies for each type of strategy are shown in Table 2.


TABLE 2 Means, Standard Deviations, and t-Values for Strategy Use

During Parent-Active and Parent-Passive Episodes

Emotion Regulation Stragegy

Parent-Active Parent-Passive

M SD M SD t

Act.play 6.3 14.0 Act.engage 88.4 18.5 Low-level play 0.6 1.9 Low-level engage 0.8 2.5 Self-soothing physical 0.7 2.7 Self-soothing symbolic 1.3 2.5 Passive focus 5.2 6.8 Active focus 1.1 4.0 Playful focus 3.3 4.7 Other-directed proximal 2.2 8.1 Other-directed distal 0.6 2.1

58.6 27.3 -11.5”” 28.1 18.0 15.7‘“”

6.4 9.1 -4.9”“”

2:7 4.1 4.7 1.9 2.1 1.5 1.6

6.5 -2.6’” 4.8 -3.5”“” 5.4 0.5 3.3 -1.5 4.3 1.4 6.3 0.5 3.7 -1.8+

Note. Active engagement during the parent-passive episode was possible because although mothers were instructed not to initiate interactions with their children, children attempted to initiate interactions with their mothers. It was not possible to code low-level engagement of parent during the parent-passive episode because part of the definition of this strategy includes cases in which the parent initiates interactions with the child.

+p < .lO. l p < .05. **p < .Ol. “‘p < .OOl. ****p < .OOOl.

As can be seen from Table 2, children’s strategy use differed as a function of parent availability. During the parent-active episode the vast majority of events involved active play with the mother, and self-initiated active was relatively infrequent. During the parent-passive episode, the most common strategy was self-initiated play, although attempts to interact with the otherwise-occupied mother were quite frequent as well. Similarly, children’s use of self-soothing behaviors decreased during the parent-active episode and increased during the parent-passive episode. Finally, distal forms of comfort-seeking behavior were marginally more prevalent during the parent-passive episode.

Factor Analyses of Emotionality Indices. Means and standard deviations for individual emotion indices were as follows. Parent-active episode: mean emotion score, M = 3.5, SD = 0.4; peak negative emotion, M = 4.6, SD = 0.7; latency to distress, M = 267.1 s, SD = 120.4 s; and minimum negative, M = 2.1, SD = 0.6. Parent-passive episode: mean emotion score, A4 = 3.8, SD = 0.3; peak negative emotion, M = 4.5, SD = 0.8; latency to distress, M = 298.7 s, SD = 113.4 s; and minimum negative, M = 2.4, SD = 0.7. To examine whether it would be feasible to combine emotion indices, thus reducing the total number of emotion variables, principal components factor analyses using promax oblique rotation were conducted with the four indices: mean emotion score,


peak negative emotion, latency to distress, and minimum negative (or most positive) emotion. Based on the mineigen criterion, two factors were extracted for each episode. In all cases all loadings were greater than .79, and no cross- loadings were greater than .22. Together these factors accounted for 81% (parent-active episode) and 89% (parent-passive episode) of the total variance in the observed variables. Interfactor correlations were .09 for parent-active episode and .28 for parent-passive episode. The factors were the same for both types of episode. The first factor included the mean level of emotion across the episode and the minimum (most positive) code given. Thus, this factor may be interpreted as a positive emotionality factor. The second factor included the peak negative emotion code given, as well as the latency to distress onset, and thus may be interpreted as a negative emotionality factor. Based on these analyses, composites were created by summing the standardized (and in the case of latency to onset of distress, reversed) scores on the appropriate variables.

Age Differences. Given that the children observed in this study ranged in age from 38 to 60 months, it was thought that age may be associated with measures of emotion regulation behavior and emotionality. Pearson product-moment correlations were used to examine associations among emotion regulation composite variables, individual emotion regulation strategies, emotionality composite variables, and child age. Only two significant associations were found between child age and emotion regulation measures. Specifically, older children engaged in more self-directed emotion regulation behaviors in general, r = .33, p < .05, and in more self-initiated play, r = .31, p < .05, during the parent-passive situation than did younger children.

Gender Differences. Gender differences in emotion regulation behavior and emotionality were examined using t tests. Only two significant differences were found in use of emotion regulation strategy. Specifically, the boys exhibited more passive focus, c = 2.89, p < .Ol, and more playful focus, t = 2.48, p -=c .05 on the frustration object during the parent-active situation than did the girls.

Associations Between Child Care Variables. As noted earlier, mothers were asked to report on several qualitative measures both for the child’s current nonparental care situation and for previous situations. The correlations between measures of current and previous situations (if any) were all high and significant: number of caregiver changes, r = .79; adult:child ratio, r = .65; group size, r = .59; all p < .OOOl. Indeed, 74% of the children had not changed care setting since entering into nonparental care. Because of this, and because reports of current conditions were likely to be the most accurate (Henry, Moffitt, Caspi, Langley, & Silva, 1994), only the current situation variables were used in the analyses. Furthermore, there was a strong correlation between child:adult ratio and group size within the current care situation, r = .79, p < .OOOl. Because


group size has been shown to be an important determinant of child outcomes (Belsky, 1985; Howes, 1983) adult:child ratio was dropped, and the number of caregiver changes and group size were retained as indices of the quality of the current nonparental care situation.

Child Care Characteristics as Predictors of Emotion Regulation and Emotionality A series of hierarchical multiple regression analyses was conducted to test predictions regarding the relations between child care characteristics and children’s emotion regulation behavior while in the laboratory. Each emotion regulation strategy, the two emotion regulation composites, and each emotionality factor (positive and negative emotion factors) was examined separately.

Because the two indices of nonparental care quality, number of caregiver changes and group size, were only applicable to children with nonparental care experience, some analyses were conducted for the entire sample (i.e., children with and without nonparental care experience), whereas others included only the children who had experienced regular nonparental care. Furthermore, because 10 parents failed to report the number of caregiver changes and group size within the current care setting, a subsample of 26 children was used in the analyses of nonparental care quality. The 10 children not included in these analyses were distributed across the sitter care, family care, and center care groups. Thus, predictor variables for the entire sample included the age at entrance into nonparental care and the number of hours per week a child was in nonparental care. Predictor variables for the nonparental care sample included the number of caregiver changes, and the number of children in a child’s care group, in the current care situation.

Age at Entrance and Nonparental Care Hours as Predictors of Emotion Regulation During Delay Situations. The first set of regression analyses used the entire sample to examine the relations between extent of nonparental care experience and children’s emotion regulation behavior while in the laboratory. It was predicted that children who entered nonparental care early in life, and who spent many hours per week in nonparental care would use more self-directed regulatory strategies while in the laboratory than those children who had been raised exclusively in the home, or had minimal nonparental care experience.

A series of hierarchical multiple regression procedures were used to predict the two emotion regulation composites (other- and self-directed strategy use), as well as each of the separate emotion regulation strategy variables. Age at entrance was entered first in each model, followed by the number of hours each week a child spent in nonparental care. This ordering was decided upon because of the temporal precedence of age at entry, as well as the proposal made by researchers such as Belsky (1988; Belsky & Rovine, 1988) that earlier entry into nonparental care would have a more deleterious effect than later entry. The


TABLE 3 Results of Hierarchical Multiple Regression Analyses Associating Age

at Entrance and Care Hours per Week to Emotion Regulation

Age at Care Hours Entrance per Week

Outcome Variable P P RZ

Parent-Active Self-directed composite -0.35’ - 0.11

-0.21 0.24 0.15 Other-directed composite 0.39”’ - 0.14

0.25 -0.25 0.18 A&play -0.37” - 0.12

-0.27+ 0.16 0.14 Act.engage 0.39’” - 0.14

0.19 -0.35’ 0.21

Parent-Passive Low-level play -0.28 - 0.08

-0.13 0.26 0.13 Self-soothe.symbolic 0.23+ - 0.06

Note. Dashes are entered when variable is not considered. +p < .lO. ‘p < .05. **p < .Ol.

df F

1,47 6.03” 2,46 4.04 1,47 7.78’” 2,46 4.97’ I,47 6.64’” 2.46 3.72’ I,47 7.78” 2,46 6.27’”

1,47 4.13” 2,46 3.30’ 1.47 2.82+

results of the analyses, which resulted in significant or marginally significant F values for the overall model, are presented in Table 3.

As can be seen in Table 3, the analyses involving composite strategy measures were significant only for the parent-active episode. Age at entrance into nonparental care was predictive of both self-directed and other-directed emotion regulation behavior in the parent-active situation. The regression models maintained significance when the second variable, number of hours in nonparental care each week, was included, although the increment in prediction of strategy use attributable to care hours was nonsignificant for either self- or other-directed strategies, sr2 = .04 and .04, respectively, both NS. Consistent with predictions, the more experience a child had with nonparental care, the more that child used nonsocial or self-directed regulatory strategies to regulate emotion during the parent-active episode.

In examining models predicting use of discrete strategies, analyses suggested that the differences in the use of self- and other-directed strategies in the parent- active episode were primarily attributable to differences in the use of self-initiated or active play alone versus engaging in play with the mother. In these analyses, there did appear to be a unique contribution of the number of hours the child was currently in nonparental care each week to the prediction of active play with the mother, over and above the association with age at entrance into nonparental care, u-2 = .07, p < .05. Children who spent more time each week in

NONPARENTAL CARE AND EMOTION REGUlAT!ON 589

nonparental care were less likely to have high levels of engagement with the mother during the parent-active delay situation.

Age at entrance and number of hours in nonparental care each week were not predictive of overall use of self- or other-directed strategies within the parent- passive delay situation. However, low-level play during the parent-passive episode was predicted by these variables. The earlier a child began nonparental care, the more he or she would engage in low-level independent play when the parent was not actively participatory. Once again, the model maintained significance when the number of nonparental care hours per week was added, although the increment in R2 was not significant. Finally, a marginal association was found between age at entrance and self-soothing symbolic regulatory behaviors during this episode. Thus, the later in life a child entered nonparental care, the more likely it was that child would talk about the experimenter or engage in pretend play involving themes of waiting or obtaining the stimulus object.

In sum, consistent with predictions, children who entered nonparental care early in life, and/or who spent many hours each week in nonparental care, used more self-directed or nonsocial emotion regulation strategies while in the laboratory. Conversely, children who entered nonparental care later in life, or not at all, and who spent fewer hours per week in nonparental care, used more other- directed or social strategies while in the laboratory.

Group Size and Number of Caregiver Changes as Predictors of Emotion Regulation. This set of regression analyses used the nonparental care sample (N = 26) to test the effects of the quality of nonparental care on children’s emotion regulation behavior while in the laboratory. It was predicted that children who were in care settings with many children per adult, or who had experienced many caregiver changes, would use more nonsocial or self-directed means to regulate emotion. No a priori ordering of these variables was specified. Thus, simultaneous multiple regression procedures were used to predict the emotion regulation composites, as well as each individual emotion regulation strategy, from the two dependent variables. The results of these analyses are presented in Table 4. Only those models that reached significant or marginal levels of significance, out of a total of 18 analyses conducted are reported in Table 4.

The results of these analyses were mixed, and overall there were few significant results. In the mother-active episode, neither of the two composite emotion regulation variables were significantly predicted by the two qualitative care measures, although there was a marginally significant association between the number of caregiver changes experienced by the child and the tendency to engage in other-directed regulatory strategies, suggesting that as the number of caregiver changes experienced by the child increased, the tendency to use other- directed strategies for emotion regulation decreased, r = - .33, p < _ 10. One multiple regression involving a specific strategy within the parent-active episode did result in a significant R? As predicted, children who had experienced many


TABLE 4 Results of Simultaneous Multiple Regressfon Analyses Associating Group Size

and Number of Caregiver Changes to Emotion Regulation

Outcome Variable

Group # Caregiver Size Changes

B B R2 ‘df F

Parent-Active Act.play

Parent-Passive Self-directed composite Other-directed composite Plav.focus

0.15 0.55” 0.25 2,24 3.93’

0.14 -0.34” 0.18 2.24 2.70+ -0.11 0.42’ 0.23 2.24 3.50’

0.05 0.61” 0.27 2.24 4.36’

+p < .lO. ‘p < .05. l *p < .Ol. ***p < .OOl.

caregiver changes in their current care situation and whose group size in their child care setting was relatively large were more likely to engage in self-initiated independent play.

During the parent-passive episode, the use of other-directed regulatory strategies was significantly predicted, and self-directed regulatory strategy use was marginally predicted, by the regression model including both group size and the number of caregiver changes experienced by the child. Because of the association found for the full sample between child age and self-directed regulatory strategy use during the parent-passive episode, the regression (involving the subsample of 26 children) was run again with child age entered first, followed by group size and then number of caregiver ‘changes. As with the two-predictor model, the overall model was only marginally significant, F = 2.48, p -=c .lO, and this marginal significance was achieved only upon entry of the number of caregiver changes variable, t = -2.07, p < .05.

The direction of the significant and marginal effects were in the opposite direction of our hypothesis. Specifically, children tiho had experienced many caregiver changes were more likely to use other-directed emotion regulation strategies, and less likely to use self-directed strategies, than were children with fewer changes. In examining the individual emotion regulation strategies, it appears that children with relatively many caregiver changes were likely to engage in more playful focus on the object of frustration (an other-directed strategy) than were children with fewer changes.

Age al Entrance and Nonparental Care Hours as Predictors of Emo- tionality. This series of regression analyses again used the entire sample to test the associations between age at entrance and number of hours per week in nonparental care and children’s emotionality while in the laboratory. It was predicted that children who had experienced little or no nonparental care would


be less negative during the parent-active episode than children with extensive nonparental care experience. Conversely, it was expected that children with extensive nonparental care experience would be less negative during the parent- passive episode. To test this hypothesis, age at entrance was once again entered first in each model, followed by the number of hours each week a child spent in nonparental care.

These analyses resulted in one significant model, but in the opposite direction of predictions. Age at entrance and number of care hours each week a child spent in nonparental care predicted negative emotionality during the parent-passive episode, F = 5.84, p < .Ol. Controlling for age of entrance into nonparental care, negative emotionality during the parent-passive episode was related to the number of hours each week a child spent in nonparental care, n-2 = . 11, P < .Ol . That is, the more hours per week a childwas in nonparental care, the more negative he or she was during the parent-passive episode.

Group Size and Number of Caregiver Changes as Predictors of Emo- tional&y. This series of regression analyses used the nonparental care sample to test the effects of group size and number of caregiver changes in the current care situation on children’s emotionality while in the laboratory. It was predicted that children who had experienced care settings with fewer children per adult, or who had experienced fewer caregiver changes, would express more positive emotionality during the parent-active episode than during the parent-passive episode. However, analyses resulted in no significant models associating group size or number of caregiver changes with children’s expressed emotionality.

DISCUSSION

This study was designed to examine associations between children’s nonparental care experiences and emotion regulation within mildly frustrating situations with the mother. To summarize our hypotheses: We expected that children with extensive nonparental care experience would show more self-directed, and less other- directed emotion regulation strategy use during frustrations with the mother present than would children with less nonparental care experience. Furthermore, we expected that the children with more out-of-home care would be less dis- tressed by frustrations, particularly those in which the mother was relatively unresponsive to them, than would the children with little or no out-of-home care.

Our findings regarding strategy use were fairly consistent with these hypotheses, although associations were much stronger during the parent-active delay than during the parent-passive delay. Our findings regarding differences in expression of negative emotion, however, were in the opposite direction from our hypotheses. Specifically, children with more out-of-home care, beginning earlier in life, tended to be more negative within the parent-passive delay situation.


A second set of hypotheses suggested that qualitative measures of child care, including group size and the number of caregiver changes experienced by the child, would be associated with emotion regulation strategy use. Specifically, we expected that among children with nonparental care experience, those whose situations involved more children and more caregiver changes would be likely to engage in more self-directed strategies for emotion regulation than would those children with more caregiver consistency and (potentially) more individual attention.

The results of these analyses were consistent with our hypotheses within the parent-active situation, but were in contradiction to our hypotheses in the parent- passive situation, although there were few significant associations overall. Chil- dren with experiences of low-quality care involving large numbers of children in their care group, and those who had experienced many caregiver changes, were more likely to play alone in the parent-active situation than were children with smaller group sizes and fewer caregiver changes. In the parent-passive situation, however, the reverse was true: Children who had experienced many caregiver changes and large numbers of children within their groups actually engaged in fewer self-directed emotion regulation behaviors, and more other-directed behaviors, while in the laboratory than did those children who had smaller care groups and fewer caregiver changes. The unique contribution of the number of caregiver changes to emotion regulation strategy use appeared to be responsible for all of the significant or marginally significant effects found in these analyses.

Each of the differences in emotion regulation behavior and emotionality among children with and without extensive backgrounds in nonparental care are noteworthy in light of current debates regarding the effects of nonparental care on emotional development. As has been suggested by a number of researchers (e.g., Field, 1987; Kopp, 1989; Thompson, 1990; Tronick, 1989), the origins of the ability to regulate emotion are thought to lie in children’s early experiences. Differences in early experiences, including earlier or later exposure to routine nonparental care and variations in the weekly amounts of such care, should result in individual differences in how children regulate emotion.

Our results may be open to a number of interpretations. On the one hand, they may suggest that children with extensive nonparental care experience, as well as those with less stable care situations involving larger numbers of children and more caregiver changes, are developing more independent ways of coping with emotionally arousing situations. This is certainly to be expected when one con- siders the inherent nature of most nonparental care environments in which many children per adult is generally the norm, and individual adult attention is some- times hard to come by. Indeed, increased reliance on self-directed emotion regulation strategies could be quite adaptive within that situation.

That the significant differences in emotion regulation strategy use associated with age at entry and number of hours in nonparental care each week were found primarily in the parent-active situation, and not in the parent-passive situation,


may suggest another possibility, however. Mothers in the parent-active situation were not restricted in their activities with their preschoolers, and yet the children who experienced extensive nonparental care, and who entered nonparental care at a relatively early age, were significantly less likely to play or otherwise interact with them than were the children with later, and less extensive, nonparental care. This could suggest that our results are not demonstrating a mature and generally adaptive independence of emotion regulation. Rather, children with relatively extensive nonparental care could be somewhat avoidant of their mothers, preferring to focus their attention elsewhere even when she is readily available.

Such a conclusion may be supported by the finding that age of entry into nonparental care and number of hours in care each week were associated with more, rather than less, negative emotion during the mother-passive delay situation where independent, self-directed emotion regulation would be most useful to the child. Support may also be suggested from the finding that the same qualitative measures of out-of-home care (particularly the number of caregiver changes experienced by the child) that were positively associated with playing alone in the mother-active situation were positively associated with other-directed regulatory behaviors (and negatively associated with self-directed emotion regulation strategies) during the parent-passive situation.

Children’s avoidance of their mothers was not assessed directly in this study. Neither was maternal behavior or emotion during the two delay situations. Be- cause of this, the extent to which we can compare and contrast the two possible explanations for our pattern of results is limited. It is interesting to note how our results may relate to the empirical and theoretical debates regarding the effects of early nonparental care on infant-mother attachment (e.g., Belsky, 1988; Belsky & Rovine, 1988; Clarke-Stewart, 1989). Despite pronounced differences in the assessment situation, coding procedures, age of subjects, and outcomes of inter- est, it does appear that we are observing the same lessened tendency among children with early and extensive nonparental care experience to interact with their mothers under conditions of mild distress that has been noted by attachment researchers.

This possible consistency between our findings and those of attachment researchers, most notably Belsky (1988; Belsky & Rovine, 1988), may be relevant to one other area of disagreement among researchers examining the effects of early nonparental care on children’s socioemotional development. Clarke-Stew- art (1989) and others (e.g., Hoffman, 1989; Starr, Phillips, & McCartney, 1990) have suggested that the usual assessment procedure for attachment, the Ainsworth and Wittig (1969) Strange Situation, may not be psychologically equivalent for infants whose mothers are employed, because features of the Strange Situation designed to stress the infant are similar to events routinely experienced by infants in nonparental care. As such, these infants may be less stressed by the Strange Situation, and thus may not display attachment behaviors


to the same degree as home-reared infants. However, as previously mentioned, brief frustrating delay situations in the presence of the mother or other adults are common occurrences for young children, and the psychological equivalence of the setting for the two groups of children may be greater than is the case for the Strange Situation.

Our results provide additional support for the importance of high-quality child care, especially with regard to caregiver stability. The number of caregiver changes experienced by children in the nonparental care setting was particularly influential in how these children regulated emotion while in the laboratory. Consequently, child care environments characterized by low turnover of care- givers and small group sizes would seem most conducive to producing positive child outcomes.

In summary, the findings of the study presented here revealed that there are clear associations between children’s emotional behavior and their care experiences. As children’s experiences with nonparental care become more extensive, so too does their reliance on nonsocial or self-directed emotion regulation behaviors to deal with emotionally arousing situations, most notably in situations where the mother is available to interact actively with her child. In addition, these data have shown that the quality of nonparental care situations can have an impact on children’s emotional behavior. When children are exposed to care settings with large numbers of children in each care group, or when they experience many caregiver changes within their care setting, their use of self-directed emotion regulation strategies also increases. It may be that this reflects the child’s ability to adapt to a caregiving environment with less immediately available external supports.

This study is one of the first to explore associations between emotion regulation and nonparental care experience in this fashion. Despite the limitations resulting from the relatively small sample size, findings support the importance of examining emotion regulation strategy use in examinations of the possible impact of early nonparental care on children’s so&emotional development. Future studies with larger samples and using longitudinal designs will be required to explore the extent to which results such as ours reflect an adaptation that may be conducive to the development of a flexible repertoire of social and nonsocial emotion regulation strategies versus a less-than-optimal relationship between the child and the mother.

REFERENCES

Ainsworth, M.D.S., & Wittig, M. (1969). Attachment and exploratory behavior of one-year-olds in a

strange situation. In B.M. Foss (Ed.), Dererminanrs ofinfant behavior (Vol4, pp. 11 l- 136).

London: Methuen. Barglow, P., Vaughn, B.E., & Molitor, N. (1987). Effects of maternal absence due to employment on

the quality of infant-mother attachment in a low-risk sample. Child Development, 58, 94%

954.


Belsky, J. (1988). The “effects” of infant day care reconsidered. Early Childhood Research Quar- terly, 3, 235-272.

Belsky, J., & Braungart, J.M. (1991). Are insecure-avoidant infants with extensive day-care experi-

ence less stressed by and more independent in the strange situation. Child Development, 62, 567-571.

Belsky, J., & Rovine, M.J. (1988). Nomnatemal care in the first year of life and the security of

attachment. Child Development, 59. 157-167. Bretherton, I., & Waters, E. (1985). Growing points of attachment theory and research. Monographs

of the Society for Research in Child Development, 50 (l-2, Serial No. 209). Bridges, L.J., & Grolnick, W.S. (1995). The development of emotional self-regulation in infancy and

early childhood. In N. Eisenberg (Ed.), Social Development (pp. 185-211). Newbury Park,

CA: Sage. Calkins, S.D. (1994). Origins and outcomes of individual differences in emotion regulation. The

Development of emotion regulation: Biological and behavioral considerations. Monographs of the Society for Research in Child Development, 59 (2-3, Serial No. 240).

Clarke-Stewart, K.A. (1988). “The ‘effects’ of day care reconsidered” reconsidered: Risks for par-

ents, children, and researchers. Early Childhood Research Quarterly, 3, 293-318. Clarke-Stewart, K.A. (1989). Day care: Maligned or malignant. American Psychologist, 44, 266

273. Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and Psychological

Measurement, 20, 37-46. Field, T. (1987). Affective and interactive disturbances in infants. In I.D. Osofsky (Ed.), Handbook

of infant development, (2nd ed., pp. 972- 1005). New York: Wiley.

Field, T. (1994). The effects of mother’s physical and emotional unavailability on emotion regulation.

The Development of emotion regulation: Biological and behavioral considerations. Mono-

graphs of the Society for Research in Child Development, 59 (2-3, Serial No. 240). Gianino, A., & Tronick, E.Z. (1988). The mutual regulation model: The infant’s self and interactive

regulation coping and defense. In T. Field, P. McCabe, & N. Scheiderman @is.), Stress and coping (pp. 47-68). Hillsdale, NJ: Erlbaum.

Grolnick, W.S., Bridges, L.J., & Connell, J.P. (1996). Emotion regulation in two-year-olds: Strate-

gies and emotional expression in four contexts. Child Development, 67, 928-941. Henry, B., Moffitt, T.E., Caspi, A., Langley, J., & Silva, P.A. (1994). On the “remembrance of

things past”: A longitudinal evaluation of the retrospective method. Psychological Assess- ment, 6, 92- 101.

Hoffman, L.W. (1989). Effects of maternal employment in the two-parent family. American Psychol- ogist, 44, 283-292.

Howes, C. (1983). Caregiver behavior in center and family day care. Journal of Applied Develop- mental Psychology, 4, 99- 107.

Howes, C., & Olenick. M. (1986). Family and child care influences on toddler’s compliance. Child Development, 57, 202-216.

Kopp, C.B. (1989). Regulation of distress and negative emotions: A developmental view. Develop- mental Psychology, 25, 343-354.

Richters, J.E., & Zahn-Waxier, C. (1988). The infant day care controversy: Current status and future directions. Early Childhood Research Quarterly, 3, 319-336.

Rothbart, M.K., & Posner, MI. (1985). Temperament and the development of self-regulation. In

L.C. Hartlage & CF. Telzrow (Eds.), The neuropsychology of individual di$erences, (pp. 93-123). New York: Plenum.

Scan, S., Phillips, D., & McCartney, K. (1990). Facts, fantasies, and the future of child care in the United States. Psychological Science, 1, 26-35.

Schwartz, J.C., Strickland, R.G., & Krolick, G. (1974). Infant day care: Behavioral effects at preschool age. Developmental Psychology, IO, 502-506.


Sroufe, L.A. (1988). A developmental perspective on day care. Early Childhood Research Quarterly, 3, 283-291.

Thompson, R.A. (1988). The effects of infant day care through the prism of attachment theory. Early Childhood Research Quarterly, 3, 213-282.

Thompson, R.A. (1990). Emotion and self-regulation. In R.A. Thompson (Ed.), Socio-emotional development: Nebraska Symposium on Motivation, (Vol. 36, pp. 367-467). Lincoln, NE: University of Nebraska Press.

Thompson, R.A. (1994). Emotion regulation: A theme in search of definition. The Development of

emotion regulation: Biological and behavioral considerations. Monographs offhe Societyfor Research in Child Development, 59 (2-3, Serial No. 240).

Tronick, E.Z. (1989). Emotions and emotional communication in infants. American Psychologist, 44, 112-l 19.

Young, K.T., & Zigler, E. (1986). Infant and toddler day care: Regulations and policy implications.

American Journal of Orthopsychiatry, 56, 43-55.

Associations between nonparental care experience and preschooler's emotion regulation in the...

Documents

Transcript of Associations between nonparental care experience and preschooler's emotion regulation in the...