Preschool Emotional Competence: Pathway to Social Competence?

Child Development, January/February 2003, Volume 74, Number 1, Pages 238–256

Preschool Emotional Competence: Pathway to Social Competence? Susanne A. Denham, Kimberly A. Blair, Elizabeth DeMulder, Jennifer Levitas, Katherine Sawyer, Sharon Auerbach-Major, and Patrick Queenan Preschoolers’ (N 5 143) patterns of emotional expressiveness, emotion regulation, and emotion knowledge were assessed. Their contributions to social competence, as evidenced by sociometric likability and teacher ratings, were evaluated via latent variable modeling, both concurrently and across time. Moderation of key results by age and sex was also explored. Emotional competence assessed at 3 to 4 years of age contributed to both concurrent and kindergarten social competence. Even early in the preschool period, contributions of emotional competence to social competence have long-term implications.

Emotional competence is crucial to children’s ability to interact and form relationships with others (Parke, 1994; Saarni, 1990). As Saarni (1990) stated, ‘‘We are talking about how [children] can respond emotionally, yet simultaneously and strategically apply their knowledge about emotions and their expression to relationships with others, so that they can negotiate interpersonal exchanges and regulate their emotional experiences’’ (p. 116). Although these aspects of emotional competence and others continue to develop throughout the lifespan, preschool–aged children already are adept at several component skills of emotional competence (Dunn, 1994). They begin to express a variety of vivid, but not incapacitating, emotions. As well, preschoolers are becoming able to discern their own and others’ emotional states, and to talk about them fluently. Finally, they are beginning to ‘‘up-’’or ‘‘down’’-

Susanne A. Denham, Department of Psychology, George Mason University; Kimberly Blair, Department of Counseling, Psychology, and Special Education, Duquesne University; Elizabeth DeMulder, Initiatives in Educational Transformation, George Mason University; Jennifer Levitas, Katherine Sawyer, and Sharon Auerbach-Major, Department of Psychology, George Mason University; Patrick Queenan, Devereux Foundation. The National Institutes of Mental Health grant R01MH54019 supported the research discussed here while all authors were at George Mason University. A preliminary version was presented at the 1997 meetings of the Society for Research in Child Development. The authors thank the many children, as well as nursery school and daycare directors and teachers, who gave so much time and goodwill so that we could learn about emotional competence, as well as Krysti Batt, Cameron Caswell, Sarah Caverly, Rene Hackney, Teresa Mason, Rebecca Sears, and Meredith Vickery, who assisted in gathering data presented here. Wynne Chin was an invaluable source of support with respect to LVPLS analyses. Correspondence concerning this article should be addressed to Susanne A. Denham, Department of Psychology, George Mason University, 4400 University Drive, Fairfax VA 22030-4444 (703-9934081). Electronic mail may be sent to [email protected].

regulate emotions, depending on their goals (Denham, 1998; Saarni, 1990). Each of the constituent elements of preschoolers’ emotional competence contributes to the crucial task of social competence from 2 to 5 years: successful initiation of peer relationships (Howes, 1987; Parker & Gottman, 1989). During this period, children become more skilled at interacting and at managing the emotions so prominent during these initial forays into the peer arena. The young child who negotiates this developmental task is in good position to continue thriving in a social world. In fact, successful, independent interaction with agemates is a central predictor of later mental health and well-being, beginning during preschool and continuing through during the grade school years when peer reputations solidify (Denham & Holt, 1993; Parker & Asher, 1987; Robins & Rutter, 1990). It is important to specify the ways emotional and social competence are highly related but still separable constructs. Rose-Krasnor’s (1997) recent theorizing is useful in this regard. At the most abstract level, she defined the construct of social competence as effectiveness in interaction, the result of organized behaviors that meet short- and long-term developmental needs. In the case of preschoolers, socially competent behaviors would be organized, as already noted, around the central developmental tasks of positive engagement and self-regulation during peer interaction. Within this theoretical view of social competence, it also is necessary to decide whether to focus on self or otherFAre we interested in accessing the child’s success in meeting personal goals, or their interpersonal connectedness? Our focus here is on adaptation with peers and in the school environment, the ongoing social relations and interactions of children in their social group (Rubin r 2003 by the Society for Research in Child Development, Inc. All rights reserved. 0009-3920/2003/7401-0017

Preschool Emotional Competence

& Ross, 1988). Finally, at the most detailed level, Rose-Krasnor’s model of social competence includes specific social, emotional, and cognitive abilities, behaviors, and motivations that are primarily individual. It is at this level that specific components of emotional competence may contribute to more general social competence. The importance of these social competence outcomes should not be underestimated. Along with marking successful development and predicting later well-being, social competence is increasingly recognized as vital to school readiness (Carlton & Winsler, 1999). For example, socially competent kindergartners are more successful than their less competent counterparts in developing positive attitudes about and adjusting to school, and they get better grades and achieve more (Birch & Ladd, 1997; Ladd, 1990; Ladd, Birch, & Buhs, 1999; Ladd, Kochenderfer, & Coleman, 1996). More specifically, social-emotional indicators, including positive interactions with teachers, positive representations of self derived from attachment relationships, emotion knowledge, emotion regulatory abilities, social skills, and nonrejected peer status, often uniquely predict academic success when other pertinent variables, even earlier academic success, are already taken into account (e.g., Carlton, 2000; Howes & Smith, 1995; Izard et al., 2001; Jacobsen & Hofmann, 1997; O’Neil, Welsh, Parke, Wang, & Strand, 1997; Pianta, 1997, Pianta, Steinberg, & Rollins, 1995; Shields et al., 2001). For example, when children enter school with friends, are well liked, are able to make and sustain new friendships, and are able to initiate positive relationships with their teachers, they also feel more positive about school, participate in school more, and achieve more than children who are not described this way. Kindergartners who are victimized by peers or are aggressive, in contrast, have more school-adjustment problems and are at risk for a potential cascade of problems, including school difficulties, delinquency, and drug abuse (Gagnon, Craig, Tremblay, Zhou, & Vitaro, 1995; Haapasalo & Tremblay 1994; Kochenderfer & Ladd, 1996; Tremblay, Pagani-Kurtz, Masse, Vitaro, & Pihl, 1995). Because social competence is so important, and apparently so intertwined with aspects of emotional competence, we need to better understand its emotional prerequisites. To maximize social competence, researchers and others must carefully scrutinize how elements of emotional competence work together to allow preschoolers to mobilize personal and environmental resources. First, children’s emotional expressiveness is a central aspect of their emotional competence. For example, positive affect

239

is important in the initiation and regulation of social exchanges; sharing positive affect may facilitate the formation of friendships and render one more likable (Denham, McKinley, Couchoud, & Holt, 1990; Sroufe, Schork, Motti, Lawroski, & LaFreniere, 1984). Conversely, negative affect, especially anger, can be problematic in social interaction (Denham et al., 1990; Rubin & Clark, 1983; Rubin & Daniels– Byrness, 1983). Children who express and experience relatively more positive than negative emotions are rated higher by teachers on dimensions such as friendliness and assertiveness, and lower on dimensions such as aggressiveness and sadness. Such children also respond more prosocially to peers’ emotions and are seen as more likable by their peers (Bower, 1985; Denham, 1986; Denham et al., 1990; Denham, Renwick, & Holt, 1991; Eisenberg, Fabes, Murphy et al., 1996; Sroufe et al., 1984; Strayer, 1980). From a functionalist perspective (Campos & Barrett, 1984), it is easy to envision why children’s enduring patterns of emotional expressiveness are such potent intrapersonal supports for, or roadblocks to, interacting with age-mates. If a child is often emotionally negative, especially angry, it is no wonder when his peers flatly assert, as did one of our 3-year-old participants, ‘‘He hits. He bites. He kicked me this morning. I don’t like him.’’ But the happier child’s smile and body language are like beacons signaling, ‘‘Come join me’’ to adults and age-mates alike. Emotion knowledge is the second key component of young children’s emotional competence. Children who understand emotions also are more prosocially responsive to their peers, and rated as more socially competent by teachers, and more likable by their peers (Denham, 1986; Denham et al., 1990; Strayer, 1980). Children who can identify the expression on a peer’s face or comprehend the emotions elicited by common social situations are more likely to react prosocially to their peers’ displays of emotion. For example, if a preschooler sees one peer bickering with another and correctly deduces that the peer suddenly experiences sadness or fear, rather than intensified anger, she may comfort her friend rather than retreat or enter the fray. Interactions with such an emotionally knowledgeable age-mate would likely be viewed as satisfying, rendering that playmate more likable. Similarly, teachers are likely to be attuned to the behavioral evidence of such emotion knowledgeFthe use of emotion language, the sympathetic reactionFand to evaluate it positively. Emotion knowledge allows preschoolers to react appropriately to others, thus bolstering their relationships. In earlier work, we found links

240

Denham et al.

between emotion knowledge and expressiveness and prosocial behavior and peer status (Denham, 1986; Denham et al., 1990; Denham & McKinley, 1993). Emotion regulation is the third vital aspect of emotional competence. When the intensity, duration, or other parameters of the experience and expression of emotion are ‘‘too much’’ or ‘‘too little’’ to meet goals and expectations of the child or social partners, emotion regulation is needed. During the preschool period, emotion regulation becomes both necessary, because of the increasing complexity of children’s emotionality and the demands of their social world, and possible, because of their increased comprehension and control of their emotionality (Denham, 1998; Lewis, Sullivan, & Vasen, 1987). Thompson (1994) suggested the following definition: ‘‘Emotion regulation consists of the extrinsic and intrinsic processes responsible for monitoring, evaluating, and modifying emotional reactions y to accomplish one’s goals’’ (pp. 27–28). Given our focus on emotion regulation’s contribution to evaluations of preschoolers’ social competence made by potential social partners, we focused on children’s ability to overtly modify their emotional reactions: that is, their coping (Denham, 1998; Gross, 1998; Losoya, Eisenberg, & Fabes, 1998). Young preschoolers often need external support to become skilled at such coping regulation; caregivers’ support allows their strategies to be maximally effective (Kopp, 1989). As they become more autonomous and more capable of cooperation, older preschoolers can collaborate with caregivers’ efforts to regulate their emotions. But, failures of emotion regulation can still be seen throughout the preschool period, in outbursts of temper and distress that may impede social interaction (Bridges & Grolnick, 1994; Cole, Michel, & Teti, 1994). Maternal and teacher reports of such emotion regulatory coping are associated with social competence (e.g., Eisenberg et al., 1995). For example, some investigators have discovered negative associations of ineffective emotion regulation and high emotional negativity with preschoolers’ and older children’s adult-rated social competence (Eisenberg et al., 1995; Eisenberg, Fabes, Murphy, et al., 1996, Eisenberg, Fabes, Guthrie, et al., 1996; Eisenberg, Fabes, et al., 1997; Eisenberg, Guthrie, et al., 1997). In particular, sympathetic responsiveness is related to both dysregulation and negativity, and their interaction (Eisenberg, Fabes, Murphy, et al., 1996; Murphy, Shepard, Eisenberg, Fabes, & Guthrie, 1999). Because of earlier research findings, we focused on lapses of emotion regulation via emotional venting and the interaction

of such venting and children’s expressed anger. We reasoned that social competence would be most evident when venting and expressed anger were low, both separately and in combination. Extant research has examined the constituents of emotional competence piecemeal, one or two at a time. None that we know have looked at all three components of emotional competence outlined here, to assess their unique contributions to multiple assessments of social competence. We need to better specify how these critical aspects of emotional competence contribute, in concert, to social competence, particularly to make findings useful for parents and early childhood educators who wish to foster social competence in their youngsters. Accordingly, our main goal in this study was to examine a model of the prediction of social competence by young children’s own emotional competence, to both replicate and extend our earlier work (see Figure 1 for the overall hypothesized model). Enduring patterns of emotional expressiveness, emotion situation knowledge, and emotion regulatory coping (i.e., the absence of dysregulated venting in response to experienced problem situations, as well as the absence of dysregulated negative reactions to the emotional expression of peers) were used to predict indices of social competence. We attempted to gain a full picture of each participant’s social competence by assessing multiple dimensions, including likability, oppositionality, isolation, and sensitive cooperation, from the differing perspectives of peers and teachers, and across periods from preschool to kindergarten. Both concurrent prediction of social competence at age 3 to 4 and longitudinal prediction of social competence in kindergarten were assessed, with age 3 to 4 emotional competence components as predictors. The longitudinal analyses allow us to make clearer statements about direction of effect in the emotional competence and social competence linkage. We evaluated our model via latent variable partial least squares (LVPLS) modeling so that we could not only examine the separate contributions of latent variables for each aspect of emotional competence, but also their combined contributions. Finally, we evaluated a competing model, in which age 3 to 4 social competence predicted aspects of emotional competence, which in turn predicted kindergarten social competence; perhaps early capabilities to interact with teachers and peers concurrently promote aspects of emotional competence. We also examined possible age and sex differences in the contribution of preschoolers’ emotional competence to their social competence. Thus, we

Preschool Emotional Competence

241

Emotional Expressiv.

Emotion Knowledge

Age 3 to 4 Social Competence

Emotion Regulation

Sex × Emotional Expressiveness

Sex × Emotion Knowledge Sex × Emotion Regulation

Emotional Expressiv. × Emotion Regulation

Kindergarten Social Competence

Age × Emotion Knowledge

Age × Emotion Regulation Figure 1. Overall hypothesized model, including cross-time relations for social competence, with main effect contributions of emotional competence to social competence, as well as contributions moderated by sex, age, and, for the contribution of emotion regulation, emotional expressiveness.

evaluated moderation of the contribution of elements of emotional competence by child sex and age. Regarding gender moderation, the development of boys’ social competence seems particularly sensitive to individual differences in emotional competence (Murphy, 1999; Denham, Mitchell-Copeland, Strandberg, Auerbach, & Blair, 1997; Eisenberg, Fabes, Guthrie, et al., 1996). Also, based on our supposition that sex of child would moderate those aspects of emotional competence that differ most in terms of gender, we were interested especially in how child sex moderated the contributions of indices of emotion regulation. However, it could also be that patterns of expressiveness would contribute to others’ evaluations of boys’ social competence differently from those of girls; perhaps

boys are expected to exhibit some level of dominant negative emotion, so that such expressiveness patterns would be most notable in evaluations of girls’ social competence. Along the same lines, perhaps strengths in emotion knowledge particularly bolster the social competence of boys, whose pathways to social competence appear less multiply determined. As well, based on our supposition that age would moderate those aspects of emotional competence on which it exerts effects, we were interested in how it moderated the contributions of emotion knowledge and emotion regulation indices. The nascent social competence of younger, as opposed to older, preschoolers also might be supported or compromised by unique aspects of emotional competence or

242

Denham et al.

its lack (see Denham, 1998, for age- and genderdependent patterns development for emotional competence). In particular, because both emotion knowledge and emotion regulation become more sophisticated with age (with emotional expressiveness changing more in quality than quantity; Denham, 1998), we could expect that the contributions of these two aspects of emotional competence might afford the younger preschooler with special ‘‘boosts’’ toward effective social interaction. Method Participants We examined the emotional and social competence of 143 predominantly Caucasian, middleincome 3- and 4-year-olds (75 boys) in a multisetting, multimethod design (age: M 5 46 months, SD 5 4.88 months; range 5 32–59 months) at the time of observation during preschool. The median annual income range for these families was $30,000 to $50,000. Of the total sample, 74% were described as Caucasian. The modal level of education for mothers was graduation from college, and 88% of the children lived in two-parent homes. By the kindergarten year, we were able to reconnect with only 104 of these children, but there were no differences on any study variables measured when children were age 3 to 4 between children who continued the study through kindergarten and those who did not. Procedure To obtain the data discussed here we both interviewed and observed children, and administered questionnaires to their parents at home and to teachers in preschool or daycare settings. We thus sought to view various aspects of the children’s emotional and social competence at two time points, during preschool and kindergarten. Various preschool and daycare centers in the Washington, D.C., metropolitan area were targeted based on past liaison relationships and director willingness to participate. Recruitment was a multitiered process. First, graduate research assistants and their research advisor met with center directors and teachers to inform them of the role researchers were to play in their classrooms during the year, as well as to build rapport. Second, letters briefly explaining the research study and encouraging parents to be involved were sent home to the families of each child in the appropriate age range. Third, team members attended parent meetings and

open houses to further personalize the research efforts. Approximately 25% of families recruited decided to participate in the study. More than 95% of those evidencing interest in this family component continued on to complete participation, including the elements reported on here. Because of this low initial interest rate, we compared these families’ with families from two earlier samples, one of which also involved heavy family participation (Denham et al., 1997) and one of which did not involve the burden of home visits, with a greater than 75% participation (Denham et al., 1990). T tests comparing the current sample and these two earlier samples on demographic and study variables showed virtually no significant differences in mean levels across the samples. Thus, we concluded that the low participation rate for this study did not result in biases. Active informed consent was obtained from the parents of all children in the study. Mothers received a packet of questionnaires during a home visit in which observations unrelated to this study were performed. Home visitors collected the questionnaires in person later, providing clarification as needed. Children were observed in their preschool for their expression of emotions and reactions to peer emotions. Teachers also completed questionnaires, one of which was germane to this study, and classmates provided sociometric ratings for each participant. Emotional competence measures were administered when children were 3 to 4 years old, with social competence measured both at 3 to 4 years and 5 to 6 years (i.e., kindergarten).

Measure of Emotional Competence: Preschool, Ages 3 to 4 Previously developed observational methodologies were used to examine (a) the emotions expressed by children during free play, and (b) their reactions to their peers’ emotions during free play. Semistructured interviews were used to assess children’s knowledge about their emotions. Care was taken for assistants familiar to the children (but not those who observed them in the home) to administer all measures in the classroom. A previously developed questionnaire was used for maternal report of children’s typical styles of emotion regulation. Naturalistic observation of emotions in the preschool Children were observed in their classroom setting, during free play, by coders using laptop

Preschool Emotional Competence

computers and software designed by Roberts (2002). Researchers made event-based recordings for 5-min trials. Participants were each observed for twelve sessions of two 5-min trials over approximately 6 weeks. Observers obtained twenty-four 5-min observations for each child (12 with the child as focal, counting his or her emotions, and 12 with the child as target, counting his or her reactions to peers’ emotions). Each participating child was viewed as the focal child in the first 5-min interval of each session. During these intervals, the focal child’s happy, sad, angry, afraid, tender, hurt, other, and neutral emotional displays were recorded. These emotions were operationally defined by facial, vocal, and motor indices (see Denham, 1986; Denham et al., 1990). The average kappa for categorization of emotion displays, across raters, was .85 (see also Denham et al., 1997). Each emotion was quantified as its percentage of total emotion displays. Because afraid, hurt, tender, and ‘‘other’’ emotions each accounted for less than 3% of emotion displays, these emotions were not considered further. Understanding of emotion. Children’s understanding of emotion (identification of emotions unequivocally appropriate to certain situations, and inferences of emotions in equivocal situations) was assessed using puppets with detachable faces that depicted happy, sad, angry, and afraid expressions. Because these measures have been described in detail elsewhere (Denham, 1986; Denham & Couchoud, 1990a, 1990b; Denham, Zoller, & Couchoud, 1994), they are summarized as follows: In the emotional situation identification tasks, the puppeteer made standard facial and vocal expressions of emotions while enacting an emotion–laden story. Each child was asked to place on the puppet the face that depicted the puppet’s feeling in the situation. The first emotional situation identification task explored how well children know others’ feelings in eight common situations that elicited unequivocal emotional reactions, such as happiness at being given an ice cream cone or fear at having a nightmare (Borke, 1971; Denham, 1986). The second emotional situation identification task measured how well children identified others’ feelings in situations where the other feels differently from the child under similar circumstances. All situations could elicit one of two emotions, as in feeling happy or afraid to get into a swimming pool (Denham, 1986; Denham & Couchoud, 1990b). Children’s mothers had reported, via a forced–choice questionnaire, how their children would feel in 12 such vignettes. For each vignette, their response determined the emotion expressed by

243

the puppet. That is, the emotion expressed by the puppet did not match what the mother had reported the child would feel. Emotion situation items were summed for the unequivocal emotion situations aggregate (a 5 .79 for eight items), equivocal emotion situations pitting positive and negative emotions (a 5 .84 for six items), and equivocal emotion situations pitting two negative emotions (a 5 .69 for six items). Lapses of emotion regulation: Maternal reports of coping behavior via venting. Children’s coping behavior when faced with emotionally difficult situations with peers was assessed with coping items developed by Eisenberg, Fabes, Nyman, Bernzweig, and Pinuelas (1994). Mothers indicated on a 7-point scale, with 1 indicating never and 7 indicating usually, how often the child would engage in each of several types of coping behavior when confronted with an emotionally arousing problem situation. As already noted, for this study we were interested only in dysregulated emotion regulationFthat is, coping via venting emotional expressiveness or behavior. For this index, item content reflected: instrumental aggression (e.g., hitting another child to obtain a goal), emotional intervention (e.g., crying to elicit help from others), emotional outbursts (e.g., crying to release frustration), and emotional aggression (using aggression to release frustration). Cronbach’s alpha for the four-item summed venting scale equaled .67. Lapses of emotion regulation: Naturalistic observation of coping behavior via negative reactions to peers’ emotions. Our rationale in including the following index was that others’ emotions, particularly at this age, can be dysregulating and thus often require coping efforts. Accordingly, in the second 5-min interval of each observational period described earlier, children within approximately 3 ft of the study child were observed as the focal entity while the study child was observed for his or her reactions. Again, the context of these observations was always free play. Target reactions that denoted dysregulated coping included: opposite affect match (e.g., happy when focal is sad), matching negative affect, displaying hurt feelings, and antisocial (e.g., volitionally exacerbating focal child’s problem; Denham, 1986; Denham et al., 1990). For indices of children’s reactions to emotions, the total for each reaction was divided by the number of emotions reacted to, to generate a proportion of occurrence for each. These reactions were then summed to produce an aggregate parallel to the maternally reported data on venting coping. Average kappa for categorization of reactions to peers’

244

Denham et al.

emotion displays, across raters, was .82 (see also Denham et al., 1997). Cronbach’s alpha for this fouritem rationally derived aggregate equaled .44. Measures of Social Competence: Preschool (Ages 3 to 4) and Kindergarten (Ages 5 to 6) The present study assessed social competence via three methods, each addressing unique aspects of this multifaceted construct (Rose-Krasnor, 1997). All of the children in the class were asked to rate their peers on a continuum of likability to obtain a measure of popularity. Teachers also rated each child’s social competence in the classroom, as viewed from their perspective. All measures were administered both in preschool and in kindergarten years. Teacher-rated social competence. Teachers and daycare providers rated children’s social competence in the classroom via questionnaires. They were uninformed of the study’s hypotheses and completed the questionnaires after they had known children for at least 2 months. The Social Competence and Behavior Evaluation Short Form (SCBE) was used (LaFreniere & Dumas, 1996). On the 30-item version of this wellvalidated measure, teachers rate children on 6-point scales that address different social behaviors associated with the child’s capacity to modulate affect. Three scale scores may be computed from the teacher ratings on specific items: (a) an angryaggressive scale (e.g., often angry, defiant when reprimanded), (b) an anxious-withdrawn scale (e.g., avoids new situations; sad, unhappy, depressed), and (c) a sensitive-cooperative scale (e.g., comforts or assists children in difficulty, takes other children’s viewpoints into account). As seen in the previous examples, both the angryaggressive scale and the anxious-withdrawn scale contain items that overlap conceptually with our observed emotion and dysregulated emotion variables. Thus, because of possible problems of including items referring to anger and aggressive venting in measurement of both emotional and social competence, we first created a new oppositional subscale using only items from the angry-aggressive scale that did not refer directly or solely to anger and aggression (e.g., the first angry-aggressive item listed earlier). Our means of limiting items was derived from our work on emotion language (Denham, Cook, & Zoller, 1992); wherever an item contained a word that we would code as an emotion word, it was deleted. Using this reasoning, we also created an isolated subscale using only items from the anxious-withdrawn scale that did not refer directly

or solely to sadness and anxiety (e.g., the second anxious-withdrawn item referred to earlier). Cronbach’s alphas for the oppositional subscale equaled .83 and .78 for preschool and kindergarten, respectively. Cronbach’s alphas for the isolated subscale equaled .78 and .77 for preschool and kindergarten, respectively. Cronbach’s alphas for the unaltered sensitive-cooperative scale equaled .90 and .82 for preschool and kindergarten, respectively. Peer-rated social competence. All children in the classroom with parental consent (study participants and nonparticipants) were taken out of the classroom by a familiar adult and presented with a set of photographs of all of their classmates (Asher, Singleton, Tinsley, & Hymel, 1979). Children first named all classmates’ photographs, to assure recognition. Then, using ample vocal and facial emotion cues to model the task, the researcher placed one plastic play person in each of three small boxes marked with sketched drawings. The like-a-lot box was marked with a smiling face, the kinda like box was marked with a flat-mouthed face, and the do not like box was marked with a frowning face. After this demonstration rating the likability of the plastic play people, each child was asked to place each of his or her peers’ pictures into the happy face, neutral face, or unhappy face boxes, according to how much he or she liked to play with each peer. At the completion of picture placement, each child chose a sticker reward and was asked to make up his or her own stories with puppets. Concluding with this neutral, constructive activity ensured that the peer rating had no negative effect on the child’s social interactions after returning to the classroom. Based on the sociometric component of the task, each child in the study received scores for both the number of positive ratings and the number of negative ratings from their classmates. One-year test–retest reliabilities for these scores were .38 and .34, respectively, pso.01 (df 5 141). A peer likability score for each child was computed by subtracting the number of negative ratings from the number of positive ratings and dividing by the total number of ratings (see also Denham & Holt, 1993; Denham et al., 1990; Denham, et al., 1997; a 5 .75). In kindergarten, the procedure was changed slightly. Because of the new cognitive abilities of their age, children were presented with 1  5 in. cards with the name of each classmate, one per card. Children matched these cards to the appropriate placement on a 1- to 5-point graphic ‘‘smiley face’’ liking scale. Thus, a 5-point Likert rating scale resulted; we recorded the number of times each participant child’s name was placed on each of the

Preschool Emotional Competence

five way points. Then, a weighted sum (with weights of 5 to 1 for each way point, from liked most to liked least) was created for kindergarten peer likability. Results Preliminary Analyses First, Kendall nonparametric correlations examining the relations between individual demographic variables (i.e., family income, race, parental education, perinatal complications, and child’s past or current use of medications) and the predictor and criterion variables were conducted. Because the number of significant results was less than the number expected by chance, none of these demographic variables were used as covariates in later analyses. Descriptive statistics. The group of participants as a whole appeared positive in terms of both emotional and social competence (see Table 1). The children’s emotional profiles showed a predominance of happiness, with some anger and little

Table 1 Descriptions of Measures Measure

Emotional competence Expressiveness Happy displays Sad displays Angry displays Understanding Emotion knowledge: unequivocal Emotion knowledge: Equivocal Positive vs. negative Emotion knowledge: equivocal Negative vs. negative Emotion regulation: maternal report Emotional venting Emotion regulation: observed Emotional venting 3- to 4-year-old social competence SCBE oppositional SCBE isolated SCBE sensitive-cooperative Sociometric likability aggregate Kindergarten social competence SCBE oppositional SCBE isolated SCBE sensitive-cooperative Sociometric likability aggregate

M

SD

Possible range

.79 .05 .11

.16 .12 .11

0–1 ‘‘ ‘‘

14.24 10.41

2.40 2.53

0–16 0–12

9.75

2.63

0–12

15.06

3.70

4–28

.10

.11

0–1

14.71 16.05 42.10 .33

7.95 7.61 9.48 .37

10–42 10–42 10–60 1–1

11.44 15.16 44.87 3.41

6.16 6.19 7.63 0.44

10–42 10–42 10–60 1–5

245

sadness in the preschool. They showed good levels of emotion understanding for their ages. It is not surprising that mothers reported some emotional venting. In terms of social competence indices, teachers rated children at both ages as relatively sensitive/cooperative, and not very oppositional or isolated; children were moderately well liked. There appeared to be adequate variation for correlational analyses. Correlational analyses. Associations of manifest variables for predictors and criteria are shown in Table 2. These preliminary analyses are followed by reports of associations among latent variables calculated by LVPLS. Taking relations among predictors first, aspects of emotional expressiveness were related as expected, as were elements of emotion understanding. The two indices of emotion regulation were, however, unrelated. Next, examining relations between sets of predictors, previously noted associations between emotional expressiveness and understanding of emotion were replicated (Denham, 1986), and observed regulation was related to emotional expressiveness (happiness and anger). Moving onto correlations among criteria, associations among the SCBE scales and sociometric likability were mostly significant both at ages 3 to 4 and in kindergarten. Cross-time correlations for two of the four indices also were significant, albeit small in magnitudeFprobably due to differences in ratersFacross the 2-year period. We next examined correlations between predictors and criteria. Emotional expressiveness patterns were largely unrelated to the social competence evaluations of important others. In contrast, ability to understand equivocal negative emotional situations was related to teachers’ evaluations of social competence during kindergarten. Thus, one type of emotion understanding was related in an interpretable way to later social competence evaluations. Regarding emotion regulation, 3-year-olds who were depicted by mothers as resorting to dysregulated venting were seen by teachers as more oppositional and by peers as less likable at age 3 to 4. The same associations held true for observed emotion regulation. The same observational index predicted concurrent teachers’ ratings of oppositionality, as well as lower peer ratings of likability. Evaluation of Longitudinal Models To continue exploring the links between aspects of emotional competence and social competence, we sought to best summarize the patterns of covariation

246 Denham et al.

Table 2 Intercorrelations Among Elements of Longitudinal Model Variable Emotional competence Expressiveness 1. Happy 2. Sad 3. Angry Understanding 4. Emotion knowledge: unequivocal 5. Emotion knowledge: equivocal, pos./neg. 6. Emotion knowledge: equivocal, neg./neg. Emotion regulation: maternal report 7. Venting Emotion regulation: observed 8. Venting 3 to 4 year-old social competence 9. SCBE oppositional 10. SCBE isolated 11. SCBE sensitive-cooperative 12. Sociometric likability Kindergarten social competence 13. SCBE oppositional 14. SCBE isolated 15. SCBE sensitive-cooperative 16. Sociometric likability n

po.05. po.01. po.001. (two-tailed)

nn

nnn

1

2

y .64nnn y .53nnn .16

3

4

5

6

7

8

9

10

11

12

13

14

15

y

.31nn .18 .20n

.26nn .20n .19n

.07 .01 .02

.02

.01

.01

.04

.04

.05

.31nnn

.05

.49nnn

.00

.01

.01

.06

.05 .15 .01 .16

.02 .00 .03 .13

.16 .21n .04 .05

.11 .07 .11 .04

.03 .06 .04 .03

.08 .15 .14 .07

.20n .03 .18 .22n

.30nn y .11 .16 y .19 .58nnn .42nnn y nn .29 .18 .25n .20n y

.09 .12 .09 .12

.05 .01 .16 .11

.11 .12 .08 .08

.14 .05 .14 .12

.16 .04 .11 .06

.22n .11 .21n .08

.19 .08 .18 .08

.11 .05 .02 .14

y .58nnn y .60nnn .63nnn y y y

.33nnn .12 .28nn .17

.14 .15 .14 .01

.17 .18 .23n .13

.18 y .10 .28nn y .10 .48nnn .33 .13 .12 .12 .29nnn

Preschool Emotional Competence

in a structural model, in which the three aspects of emotional competence (along with moderators of sex and age, and the interaction of emotional expressiveness and emotion regulation) predicted social competence concurrently and at each subsequent year. We conducted LVPLS analysis (Falk & Miller, 1992; Lohmo¨ller, 1989). We chose this method, which is becoming more widely known, because it allows us to explore hypothesized relations among constructs without some of the restrictions of LISREL structural modeling techniques. In particular, LVPLS is appropriate for use with relatively small groups of participants, although it does require a reasonable latent variable: participant ratio. Given the correlation matrix of interest here, a special advantage is that the prediction of latent variables is based on shared variance of the manifest variables. Individual variable residuals and the unreliability associated with measurement error are minimized. Because of its limited assumptions, however, LVPLS yields only approximate estimates of invariant structural parameters. Measurement model. LVPLS allows specification of a measurement model as well as a structural model. Using multiple measures (manifest variables), constructs (latent variables) are created, with the variance shared by the measures taken as a single index of the latent variable. That is, scores for the composite latent variables are computed from principal components weights, derived from analyses of the manifest variables. Thus, a smaller set of theoretical variables is created, whose relations can be investigated without sacrificing information from the larger group of manifest variables. For our study, 11 latent variables were created, including emotional expressiveness, emotion understanding, emotion regulation (i.e., dysregulated coping indices negatively weighted), and social competence at each age (see Table 3). Before analysis, we centered all manifest variables because this procedure helps avoid computational errors by lowering the correlations between the interaction terms and their individual components. Without such a procedure, multicollinearity could lead to an inability for the LVPLS procedure to accurately estimate the underlying interaction construct. Chin, Marcolin, and Newsted (1996) outlined a method to include moderator variables in LVPLS models via multiplicative-term manifest variables (e.g., Sex  Emotional Expressiveness), as also seen in Table 3. This advantageous method allowed us to examine moderation as part of the full model, before decomposition into separate regression equations,

247

with proportionally smaller ns, for each level of the moderator. Examination of the latent to manifest variable loadings in Table 3 reveals none with an absolute value less than .4, one benchmark for noting a manifest variable with very little in common with other measures of a particular construct (before these final results, however, we did have to eliminate the variables for Age  Emotion Regulation reported by mother and Expressed Anger  Emotion Regulation observed because of low loadings). Twenty-four of 29 retained manifest variables (excluding Age  Emotion Regulation observed, and Expressed Anger  Emotion Regulation reported by mother, both of which were single indicators of their respective latent variable) loaded significantly on their latent variables, with the remaining three loading at the borderline significant level. Finally, to assess how accurate the path estimates were to the ‘‘true’’ effects, composite reliabilities based on latent variables’ communalities were calculated (Werts, Linn, & Jo¨reskog, 1974). These reliabilities ranged from .46 to .75. Given the small number of items (i.e., manifest variables) per latent variable, these reliabilities were acceptable according to Nunnally’s (1967) guidelines; that is, those that didn’t meet or exceed a .70 criterion appeared to do so because of a small number of items (i.e., if number of items for each scale were multiplied by a factor of 1 to 3, resulting in 4 to 12 items, the .70 criterion would be met). Correlations among latent variables. Associations among latent variables are shown in Table 4. Onetailed probability levels were used because directions of associations were hypothesized. Emotional expressiveness was related to emotion knowledge and emotion regulation. Every predictor was related to social competence at one age (if not both), except for Age  Emotion Knowledge. Finally, despite the manifest variables’ low cross-time associations (see Table 2), there was a highly significant correlation between ages 3 to 4 and kindergarten social competence. Overview of structural path model. We estimated the latent variables as an exact linear combination of their indicators with the goal of maximizing explained variance for both indicators and latent variables, using PLSGraph software (Chin, 2001). Figure 2 depicts the final structural model, in which path coefficients can be interpreted as standardized beta weights, each estimated after all other paths’ effects have been controlled. To assess whether the paths for interaction and main effects were significant, a jackknife resampling procedure (Efron &

248

Denham et al.

Table 3 Latent Variable Loading for Manifest Variables in Longitudinal Model Variable

Loading

t value

Composite reliability (communality)

Age 3 to 4 Emotional expressiveness Observed happiness Observed sadness Observed anger Emotion knowledge Emotion knowledge: unequivocal Emotion knowledge: equivocal Positive/negative Equivocal situation knowledge Negative/negative Emotion regulation strategies Maternal report: venting Observed venting Social competence SCBE oppositional SCBE isolated SCBE sensitive-cooperative Sociometric likability Sex  Emotional Expressiveness interactions Sex  Happiness Sex  Sadness Sex  Anger Sex  Emotion Knowledge interactions Sex  Unequivocal Situation Knowledge Sex  Equivocal Situation Knowledge Positive/negative Sex  Equivocal Situation Knowledge Negative/negative Sex  Emotion Regulation Strategies Sex  Maternal Report: Venting Sex  Observed Venting Age  Emotion Knowledge interactions Age  Unequivocal Situation Knowledge Age  Equivocal Situation Knowledge Positive/negative Age  Equivocal Situation Knowledge Negative/negative Age  Emotion Regulation Strategies Age  Observed Venting Emotional Expressiveness  Emotion Regulation Strategies Anger  Maternal Report: Venting Kindergarten Social competence SCBE oppositional SCBE isolated SCBE sensitive-cooperative Sociometric likability po.05. po.01. nnn po.001. + po.10. (one-tailed)

n

nn

.946 .493 .754

23.94nnn 3.72nnn 8.86nnn

.57

.875 .815

16.50nnn 7.12nnn

.73

.874

33.59nnn

.438 .926

2.27n 22.75nnn

.52

.832 .437 .805 .563

1.66n 1.43+ 1.66n 1.69n

.46

.972 .673 .589

2.54nn 1.91n 3.26nnn

.58

.824 .797

12.04nnn 7.12nnn

.70

.882

27.74nnn

.670 .731

2.54nn 2.22n

.49

.761 .845

1.53+ 1.54+

.75

.973

1.76n

1.00

.00

F

1.00

.00

F

.777 .414 .819 .627

12.21nnn 1.84n 22.31nnn 8.81nnn

.46

Preschool Emotional Competence

249

Table 4 Correlations Among Latent Variables Latent variable

1

2

3

4

5

6

1. Emotional expressiveness y 2. Emotion knowledge .22nn y 3. Emotion regulation .38nnn .00 y 4. Sex  Emotionality .94nnn .22n .30nn y n nnn 5. Sex  Emotion Knowledge .20 .95 .02 .23nn y 6. Sex  Emotion Regulation .26nn .02 .90nnn .23nn .02 y 7. Emotional Expressiveness  Emotion Regulation .11 .02 .07 .07 .02 .07 y 8. Age  Emotion Regulation .02 .07 .31nnn .02 .04 .32nnn 9. Age  Emotion Knowledge .03 .00 .08 .07 .05 .07 10. Age 3 to 4 social competence .09 .13+ .39nnn .06 .12 .39nnn n 11. Kindergarten social competence .17 .23nn .20n .16n .23nn .21n

7

8

9

10

.09 y .02 .14+ y .13+ .16n .06 y .14+ .08 .02 .38nnn

po.05. po.01. nnn po.001. + po.10. (one-tailed)

n

nn

Gong, 1983) was performed. In this procedure, the partial least squares parameters of a series of random subsamples of the total sample were iteratively tested, until significance could be estimated based on their convergent findings. Given these procedures, the final model in Figure 2, in which nonsignificant hypothesized paths were deleted, can be summarized by noting that emotion regulation predicted age 3 to 4 social competence. Emotional expressiveness, emotion knowledge, Age  Emotion Regulation, and age 3 to 4 social competence predicted kindergarten social competence. Emotional Expressiveness  Emotion Regulation and Age  Emotion Regulation predicted social competence at both ages. Finally, emotional expressiveness predicted both emotion knowledge and emotion regulationFmore positive children were both more knowledgeable about emotions and more able to regulate them. The overall model can also be evaluated. Two indices are available to specify whether the model accounts for meaningful variance in the latent variables. First is the coefficient root mean square of the covariance (RMS COV[E,U]) between the manifest variable residuals and the latent variable residualsFthe proportion unaccounted for by the model (Falk & Miller, 1992). This coefficient is 0 in a perfect model, and values over .20 indicate poor fit. The path model in our study yielded a RMS COV(E,U) value of .10, indicating good to excellent fit between model and data. Second, the significance of R2s for each of the endogenous latent variables (i.e., emotion knowledge, emotion regulation, and social competence at both ages) were tested and found to be significant at the po.01 level or better.

Finally, it is important to specify the nature of the significant moderation pathways. To do so, we constructed sets of separate regression equations for boys and girls, younger and older children (split at the median age at study inception, 46 months), and low and high negatively expressive children for each pertinent predictor. Thus, emotion regulation predicted age 3 to 4 social competence more strongly for girls than for boys, bs 5 .376 and .255, pso.01 and .05, respectively. Emotion regulation also predicted age 3 to 4 social competence more strongly for younger than for older children, bs 5 .335 and .293, pso.01 and .05, respectively. Emotion knowledge predicted social competence more strongly for children who were younger, as opposed to older, at the study’s inception, bs 5 .229 and .085, pso.05 and ns, respectively, for age 3 to 4 social competence, and bs 5 .251 and .209, pso.05 and .10, respectively, for kindergarten social competence. Finally, emotion regulation predicted age 3 to 4 social competence more strongly for low-negative children than for high-negative children, bs 5 .365 and .253, pso.01 and .05, respectively. Conversely, emotion regulation predicted kindergarten social competence more strongly for high-negative children than for lownegative children, bs 5 .220 and .097, pso.05 and ns, respectively. Alternative model. An alternate model was tested, examining direct paths from age 3 to 4 social competence to concurrent emotional competence, and from age 3 to 4 emotional competence to kindergarten social competence (see Figure 3; because age 3 to 4 social competence was seen as a predictor of emotional competence in this model, links between age 3 to 4 social competence and

250

Denham et al.

R 2emotion knowledge = .048, F (1, 141) = 7.11**. R 2emotion regulation = .145, F (1, 141) = 23.90 ***. R 2age 3 to 4 social competence = .214, F (9, 133) = 36.21 ***. R 2kindergarten

social competence

= .214, F (9, 94) = 27.13 *** ** p < .01. *** p < .001.

Emotional Expressiv.

B = .218, t = 2.64**

Emotion Knowledge

B = .398, t = 7.60***

B = .313, t = 2.26* Age 3 to 4 Social Competence

B = .321, t = 11.50***

Emotion Regulation B = .233, t = 2.05* B = .401, t = 3.19*** B = -.065, t =

-285**

B = .152, t = 3.45** Sex × Emotion Regulation

Emotional Expressiv. ×Emotion Regulation

B = .188, t = 1.64*

B = -.211, t = -2.62**

Age × Emotion Knowledge

Age × Emotion Regulation

Kindergarten Social Competence

B = -.053, t = -1.51+

B = -.216, t = -1.75*

Figure 2. Longitudinal model of age 3 to 4 emotional competence predicting age 3 to 4 and kindergarten social competence, and age 3 to 4 social competence predicting kindergarten social competence, with only significant pathways shown. n po.05. nnpo.01. nnnpo.001. +po.10. (one-tailed)

moderator latent variables were not included in the model). Overall, the RMS COV(E,U) value was again .10, indicating good to excellent fit between model and data In this alternative model, social competence predicted all three latent variables for emotional competence, but only the R2s for emotion knowledge and emotion regulation were significant, as in the main model, suggesting that their interrelation was as important, if not more so, as the pathways from age 3 to 4 social competence. Two, instead of three, moderation paths to kindergarten social competence were significant in this model. Overall, the main difference in the two models appeared to be that the total R2 for emotion

regulation was enhanced by its prediction by age 3 to 4 social competence.

Discussion We were encouraged by our effort to identify the aspects of emotional competence that contribute to a key developmental task of the preschool periodsuccessful interaction with peers. Simultaneous consideration of multiple indicators for three hypothesized constituents of emotional competence was a step forward in specifying which elements contribute to social competence, both concurrently

Preschool Emotional Competence

251

R 2emotional expressiveness = .014, F (1, 141) = 2.00, ns. R 2emotion knowledge = .057, F (1, 141) = 8.52 **. R 2emotion regulation = .286, F (1, 141) = 56.48***. R 2kindergarten social competence = .214, F (9, 94) = 28.24 ***.

** p < .01. *** p < .001.

Emotional Expressiv. B = .123, t = 5.72***

B = .137, t = 2.70** Emotion Knowledge

B = .130, t = 3.45***

B =.132, t = 2.01* Emotion Regulation

B = .080, t = 2.24*

B = .142, t = 5.59***

Emotional Expressiv. × Emotion Regulation

Age 3 to 4 Social Competence

B = -.175, t = 2.45* Kindergarten Social Competence

B = -.263, t = 3.06 ***

Age × Emotion Regulation

B = -.209, t = = 2.10*

Figure 3. Competing model, showing age 3 to 4 social competence predicting age 3 to 4 emotional competence and kindergarten social competence, with only significant pathways shown. n po.05. nnpo.01. nnnpo.001. +po.10. (one-tailed)

and across time. Emotional competence at ages 3 to 4 made contributions to both age 3 to 4 and kindergarten social competence, suggesting that the abilities of emotional competence acquired by ages 3 to 4 become stable, with continuing import. The patterns of these latent variables’ contributions differ depending on whether each is especially important for preschoolers in general, boys or girls, or relatively younger and older preschoolers. Thus, these findings enhance our basic understanding of the affective foundations of young children’s social competence and could allow for targeted prevention efforts with older and younger preschoolers, boys and girls. The first of three emotional competence latent variables involved emotional expressiveness. This latent variable indexed predominantly happy, not sad or angry, emotions displayed during interaction.

We have argued elsewhere (e.g., Denham, 1986) that ‘‘feeling good’’ in many situations not only ‘‘greases the cogs’’ of ongoing social interaction, but also makes it easier for a child to enter the peer world in the first place. Furthermore, the enduring patterns of clear emotional well-being signaled by a happy child are likely to make a positive impression on a preschool teacher, and happier playmates probably don’t ‘‘make waves’’ in the sometimes-choppy sea of preschool interactions; they are just easier for the other children to like. However, we need to know more about how happiness works within interaction. Thus, we should examine whether peer entry, peer provocation, higher level sociodramatic play, amity with friends, and other hallmarks of ongoing preschool social interaction are fostered by children’s observed happiness (e.g., Gottman & Mettetal, 1986; Marcus, 1987). Furthermore, not all happy displays

252

Denham et al.

are created equal. For example, Arsenio and colleagues (e.g., Arsenio & Lover, 1997) have made it clear that happiness shown during preschool conflict is likely to be a sign of ‘‘happy victimizing,’’ and definitely not well received by others. As well, anger has been reported elsewhere as a detriment to young children’s social competence (Denham et al., 1990; Denham & McKinley, 1993). Chronically angry children are less likely to meet another’s positive overtures in kind, but more likely to respond hostilely. Those who more often show sadness seem to puzzle their peers, who may ultimately shun them. These children’s demeanor makes them more difficult, both by omission and commission, to play with. In fact, dealing with negative emotions, either intra- or interpersonally, makes effective interaction much more difficult across the lifespan; in contrast, more positive individuals are like ‘‘interaction magnets’’ (Carstenson & Turk, 1998; Dua, 1993; Marcus, 1987). So, it is easy to see how a predominance of happy displays as opposed to angry or sad displays could lead to others’ positive evaluations of one’s social competence. Children who showed this pattern of enduring emotional expressiveness with their playmates as early as ages 3 to 4 probably built up positive reputations with teachers and peers alike, and probably encountered more social competencebuilding experiences than their less happy, sadder, more angry counterparts. It was a bit surprising that emotional expressiveness did not predict age 3 to 4 social competence directly. One explanation could be that variations in emotional expressiveness do not yet affect others’ evaluationsFteachers may expect some level of negative emotionality, and peers who are just beginning to make appraisals of others’ likability may not yet focus on their emotional expressiveness per se. Turning to the emotion knowledge latent variable, for the total sample (but particularly for those less than 46 months old at the study’s inception), emotion knowledge assessed at ages 3 to 4 predicted social competence, both contemporaneously and in kindergarten. Older preschoolers (i.e., older than 46 months) typically showed less variation on our measure of emotion understanding; perhaps younger children’s use of such knowledge more uniquely supported their socially competent behavior or particularly stood out to raters. Emotion knowledge as assessed here could give preschoolers information in critical peer situations, such as group entry and provocation, by helping to promote conflict resolution and other positive peer behavior (Dunn, 1995; Dunn & Herrera, 1997; Edwards, Manstead, &

MacDonald, 1984; Goldman, Corsini, & de Urioste, 1980; Hughes Dunn, & White, 1998). Last, the latent variable of emotion regulation may be seen as a second-order outcome of emotional expressiveness, given its link with emotional expressivenessFperhaps a child knows that he or she feels angry, what made him or her angry, how he or she is experiencing and expressing it, and figures out what to do. Given this interpretation, it may be no surprise that indices of lapses in emotion regulation predicted age 3 to 4 social competence even more potently than emotional expressiveness patterns. How one responds to one’s experience of emotion may enhance or hinder social relations more than one’s enduring patterns of expressed emotions. We found that children who were less likely to vent their emotions, especially girls and younger participants, were evaluated as more socially competent at either ages 3 to 4 or in kindergarten. These children showed restrained emotion in response to problematic situationsFthey were relatively unlikely to cry or behave aggressively. Uncontrolled negativity may have been especially conducive to raters’ negative evaluations because it contradicted stereotypical views of girls’ social competence and reinforced notions of the immaturity of children who are ‘‘young in grade.’’ Girls especially may be expected to adhere to the nuances of ‘‘nice’’ behavior toward others, so that their emotional regulation may be an especially important part of the emotional competence that predicts concurrent social competence. For example, Fagot (1984) found that teachers of young preschoolers (i.e., 2 to 3 years old) reacted particularly negatively when girls showed behaviors in two categories that included demanding attention; expressing verbal aggression; taking objects; crying and whining; and hitting, pushing, or shoving. Thus, when a girl or a preschooler who is young in grade behaves less positively, this negativity may be more salient than the negative behavior of boys or more mature children, because girls are expected to be nice and older children’s negative behavior may be less overt (Abramovitch, Corter, & Lando, 1979; Crick, Casas, & Mosher, 1997). In particular, this set of findings highlights the continuing need to shed light on girls’ social worlds across time during the preschool period. For example, the precursors and emergence of girls’ relational aggression in preschool and kindergarten warrant study. As well, abilities to refrain from dysregulated venting particularly benefited children who were more negatively emotional; their age 3 to 4 social

Preschool Emotional Competence

competence, but not that of less negatively emotional children, was predicted by emotion regulation. This pattern of moderation fits well with that also found by Eisenberg and colleagues (Eisenberg, Fabes, Guthrie, et al., 1996; Eisenberg, Fabes, Murphy, et al., 1996; Eisenberg, Fabes, et al., 1997). It makes sense that emotion regulation should aid those who most need to regulate! However, emotion regulation at ages 3 to 4 predicted kindergarten social competence slightly better for children low in negative emotions. It could be that children who experience fewer intensely negative emotions, and avoid unrestrained outbursts when they do feel negative, also are more able, by kindergarten, to use more constructive emotionregulation strategies, such as active problem solving, cognitive restructuring, and avoidance. This longitudinal linkage for low-emotional children needs further study. It is also noteworthy that emotional expressiveness patterns predicted both emotion knowledge and emotion regulation at ages 3 to 4. Firm establishment of a linkage between being emotionally positive and emotion knowledge, already noted in Denham (1986), suggests that it may be profitable to promote children’s positivity so they feel comfortable and able to learn about their emotional world. As noted earlier, the association of emotional expressiveness and emotion regulation also makes intuitive senseFchildren who experience and display more negative emotion probably have more to regulate, but this undertaking may be overwhelming, leading to venting in less experienced 3- and 4year-olds. Finally, we did attempt to assess the adequacy of a model in which age 3 to 4 social competence predicted concurrent indices of emotional competence, which then predicted kindergarten social competence. It is perhaps not surprising that, due to the ubiquitous possibility of bidirectional influences in any concurrent correlations, age 3 to 4 social competence did predict indices of emotional competence measured at the same period, and the model fit well. However, evaluation of the overall model suggested that the main information gleaned was the prediction of emotion regulation by age 3 to 4 social competence. This finding hints at the possibility that the burden of others’ evaluationsFwhether those of teachers who consider a child difficult and let this appraisal leak into their interactions, or those of peers who decide that a dysregulated child is too much trouble to play withFmay add to the child’s difficulties in regulation. Better evaluation of alternative models awaits research in which social

253

competence and emotional competence are each evaluated at more than two time points. Because of the ever-present nature of emotion at the core of social interaction and well-being, discerning the nature of linkages from emotional to social competence is a vital task for developmentalists (Sroufe et al., 1984). In the present study, lack of positive affect, inability to regulate affect and cope behaviorally, and stunted abilities to use emotion understanding were clear risk factors for the children who participated in our project (see also Denham, Blair, Schmidt, & DeMulder, 2002). To flesh out our knowledge to make a difference in children’s lives both contemporaneously and predictively, we need even more details such as those uncovered here, and we need to overcome some of the limitations of our current workFfor example, we need information of the sort gathered here for other cultural and ethnic groups to generalize our findings and make them more useful. However, some specific suggestions can be made for preschool and daycare teachers and parents to aid their charges in attaining particular aspects of emotional competence that have long-term utility. For example, teaching about feelings may be especially helpful for children 4 years old and under. All preschoolers, but especially those who are more emotional, might benefit from learning means to avoid dysregulated coping, to respond to peers’ emotions prosocially instead of antisocially. In sum, we have focused on the separate and joint prediction of indices of social competence via emotional expressiveness, emotion knowledge, and emotion regulation. We have begun delineating boundary conditions for these relations (e.g., age, sex, and level of expressiveness) that ultimately may have important applied implications.

References Abramovitch, R., Corter, C., & Lando, B. (1979). Sibling interaction in the home. Child Development, 50, 997–1003. Arsenio, W., & Lover, A. (1997). Emotions, conflicts, and aggression during preschoolers’ freeplay. British Journal of Developmental Psychology, 15, 531–542. Asher, S. R., Singleton, L. C., Tinsley, B. P., & Hymel, S. (1979). A reliable sociometric measure for preschool children. Developmental Psychology, 15, 443–444. Birch, S. H., & Ladd, G. W. (1997). The teacher-child relationship and children’s early school adjustment. Journal of School Psychology, 35, 61–79. Borke, H. (1971). Interpersonal perception of young children: Egocentrism or empathy? Developmental Psychology, 5, 263–269.

254

Denham et al.

Bower, B. (1985). Caution: Emotions at play. Science News, 127, 226–267. Bridges, L. J., & Grolnick, W. S. (1994). The development of emotional self-regulation in infancy and early childhood. In N. Eisenberg (Ed.), Review of personality and social psychology. Vol. 15. Social development (pp. 185–211). Thousand Oaks, CA, Sage. Campos, J. J., & Barrett, K. C. (1984). Toward a new understanding of emotions and their development. In C. E. Izard, J. Kagan, & R. B. Zajonc (Eds.), Emotions, cognition, & behavior (pp. 229–263). New York: Cambridge University Press. Carlton, M. P. (2000). Motivation and school readiness in kindergarten children. Dissertation Abstracts International Section A Human and Social Sciences, 60(11-A), 3899. Carlton, M. P., & Winsler, A. (1999). School readiness: The need for a paradigm shift. School Psychology Review, 28, 338–352. Carstenson, L. L., & Turk, C. S. (1998). Emotion in the second half of life. Current Directions in Psychological Science, 7, 144–149. Chin, W. W. (2001). PLS Graph. Version 03.00 build 279. Houston, TX: Soft Modeling. Retrieved October 29, 2002, from www.plsgraph.com. Chin, W. W., Marcolin, B. L., & Newsted, P. R. (1996). A partial least squares latent variable modeling approach for measuring interaction effects: Results from a Monte Carlo simulation study and voice mail emotion/adoption study. Proceedings of the Seventeenth International Conference on Information Systems, 21–41. Cole, P. M., Michel, M. K., & Teti, L. O. (1994). The development of emotion regulation and dysregulation: A clinical perspective. Monographs of the Society for Research in Child Development, 59 (2-3, Serial No. 240), 73–102. Crick, N. R., Casas, J. F., & Mosher, M. (1997). Relational and overt aggression in preschool. Developmental Psychology, 33, 579–588. Denham, S. A. (1986). Social cognition, social behavior, and emotion in preschoolers: Contextual validation. Child Development, 57, 194–201. Denham, S. A. (1998). Emotional development in young children. New York: Guilford Press. Denham, S. A., Blair, K. A., Schmidt, M., & DeMulder, E. (2002). Compromised emotional competence: Seeds of violence sown early? American Journal of Orthopsychiatry, 72, 70–82. Denham, S. A., Cook, M. C., & Zoller, D. (1992). Baby looks very sad: Discussions about emotions between mother and preschooler. British Journal of Developmental Psychology, 10, 301–315. Denham, S. A., & Couchoud, E. A. (1990a). Young preschoolers’ understanding of emotion. Child Study Journal, 20, 171–192. Denham, S. A., & Couchoud, E. A. (1990b). Young preschoolers’ understanding of equivocal emotion situations. Child Study Journal, 20, 193–202.

Denham, S. A., & Holt, R. W. (1993). Preschoolers’ likability as cause or consequence of their social behavior. Developmental Psychology, 29, 271–275. Denham, S. A., & McKinley, M. J. (1993). Sociometric nominations of preschoolers: A psychometric analysis. Early Education and Development, 4, 109–122. Denham, S. A., McKinley, M., Couchoud, E. A., & Holt, R. (1990). Emotional and behavioral predictors of peer status in young preschoolers. Child Development, 61, 1145–1152. Denham, S. A., Mitchell-Copeland, J., Strandberg, K., Auerbach, S., & Blair, K. (1997). Parental contributions to preschoolers’ emotional competence: Direct and indirect effects. Motivation and Emotion, 21, 65–86. Denham, S. A., Renwick, S., & Holt, R. (1991). Working and playing together: Prediction of preschool social–emotional competence from mother–child interaction. Child Development, 62, 242–249. Denham, S. A., Zoller, D., & Couchoud, E. A. (1994). Preschoolers’ causal understanding of emotion and its socialization. Developmental Psychology, 30, 928–936. Dua, J. K. (1993). The role of negative affect and positive affect in stress, depression, self-esteem, assertiveness, Type A behaviors, psychological health, and physical health. Genetic, Social, & General Psychology Monographs, 119, 515–552. Dunn, J. (1994). Understanding others and the social world: Current issues in developmental research and their relation to preschool experiences and practice. Journal of Applied Developmental Psychology, 15, 571–583. Dunn, J. (1995). Children as psychologists: The later correlates of individual differences in understanding of emotions and other minds. Cognition and Emotion, 9, 187–201. Dunn, J., & Herrera, C. (1997). Conflict resolution with friends, siblings, and mothers: A developmental perspective. Aggressive Behavior, 23, 343–357. Edwards, R., Manstead, A. S. R., & MacDonald, C. J. (1984). The relationship between children’s sociometric status and ability to recognize facial expressions of emotion. European Journal of Social Psychology, 14, 235–238. Efron, B., & Gong, G. (1983). A leisurely look at the bootstrap, the jackknife, and cross-validation. The American Statistician, 37, 36–48. Eisenberg, N., Fabes, R. A., Guthrie, I. K., Murphy, B. C., Maszk, P., & Holmgren, R., et al. (1996). The relations of regulation and emotionality to problem behavior in elementary school children. Development & Psychopathology, 8, 141–162. Eisenberg, N., Fabes, R. A., Murphy, B. C., Karbon, M., Smith, M., & Maszk, P. (1996). The relations of children’s dispositional empathy-related responding to their emotionality, regulation, and social functioning. Developmental Psychology, 32, 195–209. Eisenberg, N., Fabes, R. A., Murphy, B. C., Maszk, P., Smith, M., & Karbon, M. (1995). The role of emotionality

Preschool Emotional Competence and regulation in children’s social functioning: A longitudinal study. Child Development, 66, 109–128. Eisenberg, N., Fabes, R. A., Nyman, M., Bernzweig, J., & Pinuelas, A. (1994). The relation of emotionality and regulation to preschoolers’ anger-related reactions. Child Development, 65, 1352–1366. Eisenberg, N., Fabes, R. A., Shepard, S. A., Murphy, B. C., Guthrie, I. K., & Jones, S., et al. (1997). Contemporaneous and longitudinal prediction of children’s social functioning from regulation and emotionality. Child Development, 68, 642–664. Eisenberg, N., Guthrie, I., Fabes, R. A., Reiser, M., Murphy, B. C., & Holgren, R., et al. (1997). The relations of regulation and emotionality to resiliency and competent social functioning in elementary school children. Child Development, 68, 295–311. Fagot, B. I. (1984). Teacher and peer reactions to boys’ and girls’ play styles. Sex Roles, 11, 691–702. Falk, R. F., & Miller, N. (1992). A primer for soft modeling. Akron, OH: The University of Akron Press. Gagnon, C., Craig, W. M., Tremblay, R. E., Zhou, R. M., & Vitaro, F. (1995). Kindergarten predictors of boys’ stable behavior problems at the end of elementary school. Journal of Abnormal Child Psychology, 23, 751–766. Goldman, J. A., Corsini, D. A., & de Urioste, R. (1980). Implications of positive and negative sociometric status for assessing social competence of young children. Journal of Applied Developmental Psychology, 1, 209–220. Gottman, J. M., & Mettetal, G. (1986). Speculations about social and affective development of friendship and acquaintanceship through adolescence. In J. M. Gottman & J. Parker (Eds.), Conversations of friends: Speculations on affective development (pp. 192–237). New York: Cambridge University Press. Gross, J. J. (1998). The emerging field of emotion regulation: An integrative review. Review of General Psychology, 2, 271–299. Haapasalo, J., & Tremblay, R. E. (1994). Physically aggressive boys from ages 6 to 12: Family background, parenting behavior, and prediction of delinquency. Journal of Consulting and Clinical Psychology, 62, 1044– 1052. Howes, C. (1987). Social competence with peers in young children: Developmental sequences. Developmental Review, 7, 252–272. Howes, C., & Smith, E. W. (1995). Relations among child care quality, teacher behavior, children’s play activities, emotional security, and cognitive activity in child care. Early Childhood Research Quarterly, 10, 381–404. Hughes, C., Dunn, J., & White, A. (1998). Trick or treat? Uneven understanding of mind and emotion and executive dysfunction in ‘‘hard-to-manage’’ preschoolers. Journal of Child Psychology & Psychiatry & Allied Disciplines, 39, 981–994. Izard, C. E., Fine, S., Schultz, D., Mostow, A., Ackerman, B., & Youngstrom, E. (2001). Emotions knowledge as a predictor of social behavior and academic competence in children at risk. Psychological Science, 12, 18–23.

255

Jacobsen, T., & Hofmann, V. (1997). Children’s attachment representations: Longitudinal relations to school behavior and academic competency in middle childhood and adolescence. Developmental Psychology, 33, 703–710. Jenkins, J. M., & Oatley, K. (2000). Psychopathology and short-term emotion: The balance of affects. Journal of Child Psychology and Psychiatry, 41, 463–472. Kochenderfer, B. J., & Ladd, G. W. (1996). Peer victimization: Cause or consequence of school maladjustment? Child Development, 67, 1305–1317. Kopp, C. B. (1989). Regulation of distress and negative emotions: A developmental view. Developmental Psychology, 25, 343–354. Ladd, G. (1990). Having friends, keeping friends, making friends, and being liked by peers in the classroom: Predictors of children’s early school adjustment? Child Development, 61, 1081–1100. Ladd, G. W., Birch, S. H., & Buhs, E. S. (1999). Children’s social and scholastic lives in kindergarten: Related spheres of influence? Child Development, 70, 1373–1400. Ladd, G. W., Kochenderfer, B. J., & Coleman, C. C. (1996). Friendship quality as a predictor of young children’s early school adjustment. Child Development, 67, 1103– 1118. LaFreniere, P. J., & Dumas, J. E. (1996). Social competence and behavior evaluation in children ages 3 to 6 years: The short form (SCBE-30). Psychological Assessment, 8, 369–377. Lewis, M., Sullivan, M., & Vasen, A. (1987). Making faces: Age and emotion differences in the posing of emotional expression. Developmental Psychology, 23, 690–697. Lohmo¨ller, J. B. (1989). Latent variable path modeling with partial least squares. New York: Springer-Verlag. Losoya, S., Eisenberg, N., & Fabes, R. A. (1998). Developmental issues in the study of coping. International Journal of Behavioral Development, 22, 287–314. Marcus, R. F. (1987). The role of affect in children’s cooperation. Child Study Journal, 17, 153–168. Murphy, B. C., Shepard, S. A., Eisenberg, N., Fabes, R. A., & Guthrie, I. K. (1999). Contemporaneous and longitudinal relations of dispositional sympathy to emotionality, regulation, and social functioning. Journal of Early Adolescence, 19, 66–97. Nunnally, J. (1967). Psychometrics. New York: McGraw-Hill. O’Neil, R., Welsh, M., Parke, R. D., Wang, S., & Strand, C. (1997). A longitudinal assessment of the academic correlates of early peer acceptance and rejection. Journal of Clinical Child Psychology, 26, 290–303. Parke, R. D. (1994). Progress, paradigms, and unresolved problems: A commentary on recent advances in our understanding of children’s emotions. Merrill-Palmer Quarterly, 40, 157–169. Parker, J., & Asher, S. A. (1987). Peer relations and later personal adjustment: Are low accept children ‘‘at risk’’? Psychological Bulletin, 102, 357–389. Parker, J. G., & Gottman, J. M. (1989). Social and emotional development in a relational context: Friendship interaction from early childhood to adolescence. In T. J. Berndt

256

Denham et al.

& G. W. Ladd (Eds.), Peer relationships in child development (pp. 95–131). New York: Wiley. Pianta, R. C. (1997). Adult-child relationship processes and early schooling. Early Education and Development, 8, 11– 26. Pianta, R. C., Steinberg, M., & Rollins, K. (1995). The first two years of school: Teacher-child relationships and deflections in children’s classroom adjustment. Development & Psychopathology, 7, 295–312. Roberts, W. L. (2002). Software for observing behavior in natural settings. Retrieved October 28, 2002, from http:// www.cariboo.bc.ca/ae/psych/roberts/homepage. htm#focal. Robins, L. N., & Rutter, M. (1990). Straight and devious pathways from childhood to adulthood. Cambridge, UK: Cambridge University Press. Rose-Krasnor, L. (1997). The nature of social competence: A theoretical review. Social Development, 6, 111–135. Rubin, K. D., & Clark, M. L. (1983). Preschool teachers’ ratings of behavioral problems: Observational, sociometric, and social–cognitive correlates. Journal of Abnormal Child Psychology, 11, 273–286. Rubin, K. D., & Daniels–Byrness, T. (1983). Concurrent and predictive correlates of sociometric status in kindergarten and grade 1 children. Merrill–Palmer Quarterly, 29, 337–352. Rubin, K. H., & Ross, H. S. (1988). Toward the study of social competence, social status, and social relations. Monographs of the Society for Research in Child Development, 53 (1, Serial No. 217), 79–87.

Saarni, C. (1990). Emotional competence. In Ross Thompson (Ed.), Nebraska symposium: Socioemotional development (pp. 115–161). Lincoln, University of Nebraska Press. Shields, A., Dickstein, S., Seifer, R., Guisti, L., Magee, K. D., & Spritz, B. (2001). Emotional competence and early school adjustment: A study of preschoolers at risk. Early Education and Development, 12, 73–96. Sroufe, L. A., Schork, E., Motti, F., Lawroski, N., & LaFreniere, P. (1984). The role of affect in social competence. In C. E. Izard, J. Kagan, & R. B. Zajonc (Eds.), Emotions, cognition, & behavior (pp. 289–319). Cambridge, UK: Cambridge University Press. Strayer, J. (1980). A naturalistic study of empathic behaviors and their relation to affective states and perspective-taking skills in preschool children. Child Development, 51, 815–822. Thompson, R. A. (1994). Emotional regulation: A theme in search of definition. 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), 25–52. Tremblay, R. E., Pagani-Kurtz, L., Masse, L. C., Vitaro, F., & Pihl, R. O. (1995). A bimodal preventive intervention for disruptive kindergarten boys: Its impact through midadolescence. Journal of Consulting and Clinical Psychology, 63, 560–568. Werts, C. E., Linn, R. L., & Jo¨reskog, K. G. (1974). Intraclass reliability estimates: Testing structural assumptions. Educational and Psychological Measurement, 34, 25–33.