When Girls Versus Boys Play Alone: Nonsocial Play and ... - PsycNET

11 downloads 29 Views 1MB Size Report
The reticent child may want to engage in social interactions with other peers. 464. This document is copyrighted by the American Psychological Association or ...
Developmental Psychology 2001, Vol. 37. No. 4, 464-474

Copyright 2001 by the American Psychological Association, Inc. 0012-1649/01/S5.00 DOI: 10.1037//0012-1649.37.4.464

When Girls Versus Boys Play Alone: Nonsocial Play and Adjustment in Kindergarten Robert J. Coplan, Marie-Helene Gavinski-Molina, Daniel G. Lagace-Seguin, and Cherami Wichmann

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Carleton University The goal of the present study was to examine the relations between different forms of children's nonsocial play behaviors and adjustment in kindergarten. The participants in this study were 77 kindergarten children (38 boys, 39 girls; mean age = 66.16 months, SD = 4.11 months). Mothers completed ratings of child shyness and emotion dysregulation. Children's nonsocial play behaviors (reticent, solitary-passive, solitary-active) were observed during free play. In addition, teachers rated child behavior problems (internalizing and externalizing) and social competence; academic achievement was assessed through child interviews. Results from regression analyses revealed that different types of nonsocial play were differentially associated with child characteristics and indices of adjustment. For some forms of nonsocial play, the nature of these associations differed significantly for boys and girls.

A defining characteristic attributed to most shy or withdrawn children is a tendency to "play alone" in the company of peers. It is widely accepted that socially noninteractive children are at risk for maladjustment difficulties in later childhood and adolescence (see Rubin, Bukowski, & Parker, 1998, for a recent review). For example, the lack of social interaction in childhood has been concurrently and predictively associated with negative selfperceptions and internalizing difficulties (e.g., Rubin, Chen, McDougall, Bowker, & McKinnon, 1995; Rubin & Mills, 1988). There are two potentially important caveats associated with these findings. First, different forms of nonsocial play seem to be differentially associated with indices of maladjustment in early childhood. Second, although there is some preliminary evidence to suggest that social withdrawal may be more of a risk factor for boys than for girls, potential gender differences in the relations between nonsocial play and adjustment remain largely unexplored. Our goals in the present study were to explore the associates of different subtypes of nonsocial play in early childhood and to investigate potential gender differences in this regard.

Reznick, & Gibbons, 1989; Kagan, Reznick, & Snidman, 1988) have extensively explored the biological underpinnings of characteristically wary behavior in the face of novelty, which they have labeled inhibited behavior. Rubin and colleagues (e.g., Rubin, Hymel, & Mills, 1989; Rubin, Stewart, & Coplan, 1995) have outlined pathways to the development of social withdrawal that describe the interplay between dispositional and biologically based characteristics, parental socialization practices, the quality of relationships within and outside of the family, and macrosystemic forces (i.e., stress, social support, and culture). Further, Asendorpf (1990) has speculated that shyness, social avoidance, and social disinterest can be differentiated in terms of children's motivations to approach and avoid others. In recent years, however, it has become clear that a lack of social interaction is not a sufficient criterion for characterizing children as socially withdrawn and thus at risk for social maladaptation (see Coplan, 2000, for a recent review). Nonsocial play in early childhood is a complex and multidimensional phenomenon (Harrist, Zaia, Bates, Dodge, & Pettit, 1997; Hinde, Tamplin, & Barrett, 1993). Children may play alone for many different reasons (Rubin & Asendorpf, 1993) and may demonstrate many different behaviors while playing alone (Coplan, Rubin, Fox, Calkins, & Stewart, 1994). Researchers have identified three distinct forms of nonsocial play behaviors, which appear to differ in terms of their behavioral, emotional, and motivational underpinnings.

Subtypes of Nonsocial Play Researchers have approached the study of socially noninteractive children from a number of different theoretical perspectives. For example, Kagan and colleagues (e.g., Kagan, 1997; Kagan,

Reticent Behavior Robert J. Coplan, Marie-Helene Gavinski-Molina, Daniel G. LagaceSeguin, and Cherami Wichmann, Department of Psychology, Carleton University, Ottawa, Ontario, Canada. This research was supported by the Ottawa-Carleton District School Board and a Social Science Research Council of Canada grant. We thank Janis Huygens, Amanda Lee, Fiona Macleod, Angie Mattka, Catherine Mills, Megan Patterson, and Dana Stimpson for their help in the collection and coding of data. We are also grateful to the teachers, principals, children, and families who participated in this study. Correspondence concerning this article should be addressed to Robert J. Coplan, Department of Psychology, Carleton University, 1125 Colonel By Drive, Ottawa, Ontario K1S 5B6, Canada. Electronic mail may be sent to [email protected].

Children engaging in reticent behavior tend to watch others play while making no attempt to join in (onlooking behavior), stare into space, or wander around aimlessly (unoccupied behavior), with no specific purpose in mind (Coplan et al., 1994). On average, children have been observed to engage in reticent behavior about 15% of the time during free play with peers in preschool (Coplan, 2000; Rubin, Maioni, & Hornung, 1976) and about 10% of the time in kindergarten (Rubin, Watson, & Jambor, 1978). It has been suggested that this form of nonsocial play reflects an approachavoidance conflict within the child (Asendorpf, 1990). The reticent child may want to engage in social interactions with other peers 464

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

WHEN GIRLS VERSUS BOYS PLAY ALONE (high social approach motivation) but is inhibited by a fear of social interaction, causing a simultaneous desire to avoid others (high social avoidance motivation). There is converging evidence to suggest that reticent behavior is a marker variable for social anxiety when displayed in both novel (Coplan et al., 1994) and familiar (Coplan & Rubin, 1998) settings. Moreover, results from several studies have indicated concurrent associations between reticent behavior and temperamental shyness in early childhood (e.g., Coplan, 2000; Coplan et al., 1994; Coplan & Rubin, 1998). Reticent behavior has also been associated with the overt display of anxious behaviors during free play (Coplan, 1998) and poor performance during a series of tasks (show-and-tell speeches, ticket sorting) set within social settings (Coplan et al., 1994). In addition, Fox and colleagues (1995) reported that children who exhibited a high frequency of reticent behaviors demonstrated greater right-frontal activation asymmetry on electroencephalograms, compared with their more sociable counterparts. Given this conceptualization, the display of reticent behavior in preschool might be expected to be associated with specific forms of social and emotional maladaptation. Thus, it is not surprising to note that preschool children who more frequently engage in reticent behavior during peer play are rated by mothers and teachers as displaying more internalizing problems (Coplan et al., 1994; Coplan & Rubin, 1998).-Moreover, a significant pathway of associations has recently been reported (Coplan, 2000) to lead from shy temperament, to the display of reticent behavior on the first day of starting a new preschool, to reticent behavior after 6 weeks of school, to internalizing problems 4 months into the preschool year.

Solitary-Passive Behavior Solitary-passive behavior consists of quiet exploration and constructive behavior (i.e., building with blocks, drawing with crayons) while playing alone in the company of peers (Rubin, 1982). This form of nonsocial play is quite common during free play (25-45% of the time) in early childhood (Coplan, 2000; Rubin et al., 1976, 1978). It has been argued that in contrast to children who engage in other forms of nonsocial play, children who more frequently display solitary-passive behavior often prefer solitude to socializing with their peers but do not have a problem when they want to interact with others (Asendorpf, 1993). This tendency may reflect the combination of low social approach and low social avoidance motivations (Asendorpf, 1990). Accordingly, it has been proposed that solitary-passive play reflects social disinterest (Rubin & Asendorpf, 1993). Consistent with this notion, Rubin, Coplan, Fox, and Calkins (1995) reported that solitary-passive behavior was the preferred "play modality" of preschoolers who demonstrated a low frequency of social interactions but who were also emotionally well-regulated. In general, solitary-passive play in early childhood has not been associated with indices of maladjustment (e.g., Coplan et al., 1994; Coplan & Rubin, 1998; Rubin, 1982). In a recent study (Coplan, 2000), significant associations between the display of solitarypassive behavior in preschool and behavior problems (internalizing or externalizing) were not found. In fact, Rubin (1982) has argued that this form of nonsocial play is positively reinforced by peers, parents, and teachers, as it maintains order in the earlychildhood classroom. Moreover, children who more frequently demonstrate solitary-passive behavior are described as being competent problem solvers (Rubin, 1982), and they perform better on

465

object-focused tasks than on people-focused tasks (Coplan et al., 1994). Although solitary-passive behavior does not appear to be an indicator of maladjustment in early childhood, this may change with increasing age. Children who consistently and over time engage in less social interaction than their peers may lag behind in the acquisition of important social skills (Rubin & Asendorpf, 1993). Moreover, as children age, they are more likely to view solitary types of play as being deviant from the norm (Younger & Daniels, 1992). In this regard, Rubin and colleagues (e.g., Rubin & Mills, 1988) have reported that in middle and later childhood, solitary-passive behavior becomes increasingly associated with peer rejection, negative self-perceptions, and internalizing difficulties.

Solitary-Active Behavior Solitary-active behavior involves boisterous, repetitive functional/sensorimotor behaviors (i.e., repeatedly banging two blocks together) and dramatizing (i.e., animating a doll) while playing alone in the company of peers (Rubin, 1982). It has been suggested that this form of nonsocial play is reflective of social immaturity and impulsivity (Coplan & Rubin, 1998). Although it occurs rather infrequently during free play (approximately 3% of the time), Rubin (1982) has speculated that solitary-active behavior is quite negatively salient to the peer group, even in early childhood. As such, children who frequently engage in solitary-active behaviors may be playing alone because their social initiations are rejected by their peers (Bowker, Bukowski, & Zargarpour, 1998; Rubin, 1982; Rubin & Asendorpf, 1993). Thus, although some children play alone because they withdraw from others (e.g., due to social fear or social disinterest), children who engage in solitary-active behaviors are thought to be isolated by others (Asendorpf, 1993). Solitary-active behavior in early childhood has been associated with peer rejection, poor social problem-solving (Rubin, 1982), and maternal ratings of impulsivity (Coplan et al., 1994). Moreover, the display of solitary-active play is positively associated with externalizing problems in the preschool (Coplan, 2000; Coplan & Rubin, 1998; Coplan, Wichmann, & Lagace-Seguin, in press).

Gender Differences and Nonsocial Play Researchers have not reported any significant gender differences in the frequency of occurrence of reticent, solitary-passive, and solitary-active behaviors (e.g., Coplan et al., 1994; Coplan & Rubin, 1998; Rubin, 1982). However, gender differences in the correlates of nonsocial play in early childhood remain largely unexplored. Conceptually, the assumption that different forms of nonsocial play might be differentially associated with indices of adjustment for young boys and girls was derived from the notion that in contemporary society, shyness or withdrawal may be less acceptable for boys than for girls (Sadker & Sadker, 1994). Consistent with this assertion, results from some studies indicate that parents respond differently to shyness from their sons and daughters—shyness in girls is more likely to be rewarded and accepted by parents, whereas shyness in boys is more likely to be discouraged (Engfer, 1993; Stevenson-Hinde, 1989). RadkeYarrow, Richters, and Wilson (1988) reported that mothers were less accepting of their sons who were shy and more affectionate

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

466

COPLAN, GAVINSKI-MOLINA, LAGACE-SEGUIN, AND WICHMANN

and tender to their shy daughters. Similarly, shy boys tended to have more negative interactions with parents, whereas shy girls had more positive ones (Simpson & Stevenson-Hinde, 1985). In keeping with the notion that shyness in girls is considered more acceptable, some empirical evidence suggests that withdrawn boys have more adjustment difficulties from early childhood into adulthood than withdrawn girls do. In early childhood, it has been reported that extremely shy preschool-aged boys have more behavior problems than extremely shy girls (Stevenson-Hinde & Glover, 1996). In middle childhood, Rubin, Chen, and Hymel (1993) found that socially withdrawn boys, but not girls, described themselves as more lonely and as having poorer social skills than their average peers. Morison and Masten (1991) reported that shy boys had lower self-esteem in adolescence than shy girls. Finally, Caspi, Elder, and Bern (1988) found that boys who were shy in childhood married, became fathers, and established careers at a later age than their nonshy peers. In comparison, girls who were shy in childhood did not marry or start families later than other women within the same cohort. The Present Study Some empirical support for a conceptual model linking child temperament, nonsocial play, and adjustment in an early school setting has been described and provided in a recent study (Coplan, 2000). Specific child characteristics are thought to predict the display of different forms of nonsocial play in preschool, which are in turn associated with different adjustment outcomes. However, as with other studies in this area, potential gender differences in the associates of nonsocial play behaviors remain predominantly unexplored. In order to explore this possibility, we undertook a short-term longitudinal project with a sample of kindergarten children. Measures of child shyness and emotional dysregulation were collected 2 months into the kindergarten year. After 6 months, observations of nonsocial play were undertaken. Finally, near the end of the school year, teachers rated social adjustment, and we conducted child interview assessments of academic achievement. Hypotheses

Shyness and Emotion Dysregulation Given the results of past research relating nonsocial play to various child characteristics, we expected that (a) shyness would be related to reticent behavior (e.g., Coplan, 2000), (b) emotion regulation (i.e., less emotionality, ease in soothing) would be associated with solitary-passive behavior (e.g., Coplan & Rubin, 1998), and (c) emotional dysregulation (negative emotionality and difficulty in soothing) would be related to solitary-active behavior (e.g., Coplan et al., 1994). It is possible that there are gender differences in the underlying psychological meanings of one or all of the different forms of nonsocial play (reticent, solitary-passive, and solitary-active behaviors). In this regard, we explored gender differences in the child characteristics that may be associated with different forms of nonsocial play behaviors. However, given the lack of previous research in the area, hypotheses regarding possible sex differences in these relations were largely explorative.

havior problems) and cognitive (academic achievement) domains. Drawing on the extant literature relating nonsocial play and adjustment, we hypothesized that, overall, (a) reticent behaviors would be negatively related to social competence (e.g., Stewart & Rubin, 1995) but associated with maladjustment along the internalizing dimension (e.g., Coplan & Rubin, 1998), (b) solitaryactive behaviors would also be negatively related to social competence (e.g., Rubin, 1982) but related to maladaptation along the externalizing dimension (e.g., Coplan et al., in press), and (c) solitary-passive play would not be associated with indices of maladjustment (e.g., Coplan et al., 1994). Because no previous research has explored the associations between the different forms of nonsocial play and indices of academic research, hypotheses in this regard were exploratory in nature. It is also possible that there are gender differences in the relations between one or all of the different forms of nonsocial play and indices of adjustment. However, to date these associations have yet to be explored empirically. In this regard, gender differences in the relations between reticent, solitary-passive, and solitary-active behaviors and indices of adjustment in the kindergarten classroom were also explored. Given the premise that shyness-withdrawal (broadly defined) is seen as more acceptable in girls than in boys, we speculated that nonsocial play (in all its forms) would be more strongly associated with maladjustment among kindergarten-aged boys than girls.

Method Participants The participants in this study were 77 children (38 boys, 39 girls; mean age = 66.16 months, SD = 4.11 months) attending half-day public kindergarten classes in seven schools in a mid-sized city in Ontario, Canada. The sample was dispersed across many different schools because the children were part of an ongoing longitudinal research project (following the children from public junior kindergarten classes in a number of different schools) conducted in conjunction with a local school board. This project concerned the relations between teacher training and child outcomes (see Coplan, Wichmann, Lagace-Seguin, Rachlis, & McVey, 1999). All children from the original longitudinal sample (originally recruited 16 months earlier) for whom current data were available were included in the present sample except those children who did not speak English as a first language. The 77 children in the current sample represented approximately 60% of the original longitudinal sample. The drop in sample size was primarily due to children's moving away or changing schools during the intervening span of 16 months and from parents' not granting consent for this follow-up investigation. Children were recruited through information and consent letters sent home with them from school. The participating schools were selected as a function of their locations in diverse neighborhoods throughout the city. Thus, the current sample was representative of a wide range of socioeconomic status and ethnicity. Just less than 50% of fathers and about 63% of mothers in the sample did not possess a university degree. Thirty-two percent of fathers and 22% of mothers had graduated from a university, and 18% of fathers and 15% of mothers reported at least some postgraduate education. School board policy did not permit the collection of racial or ethnicity data.

Procedure and Measures Indices of Adjustment For the purposes of the present study, indices of adjustment included variables from both the social (social competence, be-

Maternal ratings. Three months into the school year (November), mothers completed the Colorado Child Temperament Inventory (CCTI; Buss & Plomin, 1984; Rowe & Plomin, 1977). This scale is used to assess

467

WHEN GIRLS VERSUS BOYS PLAY ALONE

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

parental perceptions of various child characteristics. The CCTI has wellestablished psychometric properties (Rowe & Plomin, 1977). For the purposes of the present study, we were particularly interested in the subscales of Shyness (5 items—e.g., "Child takes a long time to warm up to strangers"), Emotionality (5 items—e.g., "Child often fusses and cries"), and Soothability (5 items—e.g., "When upset by an unexpected situation, child quickly calms down"). In the current sample, Cronbach's alphas for these subscales were .80 for Shyness, .77 for Soothability, and .81 for Emotionality. Following procedures outlined by Rubin, Coplan et al. (1995), the factors of Emotionality and Soothability (reversed) were standardized, aggregated, and then restandardized to create a composite variable in which high scores would be representative of emotion dysregulation. Behavioral observations. Six months into the school year (February), children's behaviors during free play were observed. Children were provided with between 45 min and an hour (depending upon the school) of undisturbed free play each day. Behaviors were coded over a 3-4-week period; we used an adapted version of the Play Observation Scale (Rubin, 1989). Each child was observed for a series of 10-s intervals for 4-5 min per day on at least three separate occasions. In total, each child was observed for 12 min, yielding 72 coding intervals per child. For each 10-s interval, the child's predominant free-play behavior was recorded. Of particular interest for the present study were the time-sampled codes of reticent behavior (e.g., unoccupied and onlooking behaviors), solitary-passive play (e.g., solitary-constructive and solitary-exploratory behaviors), and solitary-active play (e.g., solitary-functional and solitarydramatic behaviors). These raw scores were proportionalized by dividing by the total number of coding intervals. The observational data were collected by six trained research assistants. Interrater reliability was computed for pairs of researchers on the basis of 720 codes of data (120 min) collected before the start of the study. For a complete variable matrix, Cohen's kappa between pairs of observers ranged from .80 to .86. Academic achievement. Near the end of the school year (May/June), a multimethod assessment of children's academic achievement was undertaken. Each child was assessed individually and asked a series of questions designed to examine early academic skill (e.g., basic concepts, literacy, and numeracy skills). Interviews were conducted by three trained female research assistants. The interview protocol was designed by the school board and included 28 items (correct-incorrect) derived from the primary battery of the Metropolitan Readiness Test (Nurss & McGauvran, 1986). Results from factor analyses revealed that all 28 items loaded onto a single factor. Cronbach's alpha for this scale was .88. An aggregate score representing interview academic achievement was thus created. Teachers provided a second measure of children's academic achievement. On a 5-point Likert scale, teachers rated children in terms of their improvement over the course of the school year on a number of different academic skills, including understanding language, using language, reading, math skills, science and technology, ability to print, fine and gross motor skills, and reasoning. Results from factor analyses revealed that all 9 items loaded onto a single factor. Cronbach's alpha for this scale was .94. The child's score for each of these items was aggregated to represent teacher-rated academic achievement. The two measures of academic achievement were significantly correlated with each other (r = .54, p < .001). Thus, they were standardized and aggregated to create a global measure of children's academic achievement. Social adjustment. Also near the end of the school year (May/June), teachers rated children's behavior problems and social competence by means of the Preschool Behavior Questionnaire (Behar & Stringfield, 1974). This scale has well-established psychometric properties (e.g., Hoge, Meginbir, Khan, & Weatherall, 1985; Rubin & Clark, 1983). On the basis of results from recent research (e.g., Hoge et al., 1985; Moller & Rubin, 1988), we used the broader two-factor solution. This resulted in subscales pertaining to internalizing problems (9 items—e.g., "Is worried—worries about many things") and externalizing problems (15 items—e.g., "De-

stroys own or other's belongings"). A subscale with items pertaining to social competence (9 items—e.g., "Will invite bystanders to join a game") was also added, following the procedures outlined by Tremblay, Vitaro, Gagnon, Piche, and Royer (1992). In the current sample, Cronbach's alphas were .92 for the Externalizing, .81 for the Internalizing, and .91 for the Social Competence subscales.

Results Preliminary Analyses Means, standard deviations, and ranges for the three forms of nonsocial play are displayed in Table 1. Results from a series of t tests indicated no significant differences between boys and girls in terms of the observed display of reticent, solitary-passive, and solitary-active behaviors. In addition, results from correlational analyses revealed no significant associations between child age or maternal-paternal education and the three nonsocial play forms. As a result, these variables were not controlled for statistically in subsequent analyses.

Regression Analyses Overview. To explore the relations between the nonsocial play forms, child characteristics, and indices of social and academic adjustment in kindergarten, we performed a series of regression analyses. The goal of these analyses was to assess the main effects and any interactions with gender. We used Cohen's partialed products technique (Aiken & West, 1991; Cohen, 1978; Cohen & Cohen, 1983). Following procedures outlined by Aikin and West (1991), we standardized gender, child characteristics (shyness and emotion dysregulation), and measures of nonsocial play (reticent and solitary-passive behaviors). Given the extremely low frequency of occurrence of solitary-active behavior, the analysis of this variable (using an alternate statistical technique) is presented in a later section. The first set of equations concerned the prediction of nonsocial play (reticent, solitary-passive) from gender and child characteristics. With reticent behavior serving as the dependent variable, the main effects variables of child characteristics (shyness and emotion dysregulation) and gender were entered into the regression equation as a block. In the next step, the two-way interaction terms were entered, represented by multiplicative products (Shyness X Gender, Emotion Dysregulation X Gender, and Shyness X Emotion Dysregulation). In the third step, the three-way interaction term was entered (Shyness X Emotion Dysregulation X Gender). The procedure was then repeated with solitary-passive behavior serving as the dependent variable.

Table 1 Frequency of Nonsocial Play Behaviors Nonsocial play form Variable

Reticent

Solitary-passive

Solitary-active

M SD Range

.09 .09 0-.51

.31 .18 0-.78

.02 .02 0-.10

Note.

Values represent proportions of observed intervals.

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

468

COPLAN, GAVINSKI-MOLINA, LAGACE-SEGUIN, AND WICHMANN

The next set of equations involved the prediction of indices of adjustment from measures of nonsocial play and gender. Separate regression equations were computed for each of the indices of adjustment (internalizing problems, externalizing problems, social competence, and academic achievement) where nonsocial play (reticent or solitary-passive behavior) and gender were first entered into the regression equation. The interaction term for gender and nonsocial play was then entered (reticent or solitary-passive behavior). Reticent behavior. Results from regression analyses revealed that shyness was significantly and positively associated with reticent behavior (/3 = .25), ?(74) = 2.13, p < .05. Thus, consistent with predictions, children who were rated by their mothers as being more shy tended to display more reticent behavior during free play with peers. No significant main effect of gender was found, and there were no significant two-way or three-way interactions between shyness, emotional dysregulation, and gender in the prediction of reticent behavior. Results are displayed in Table 2. In terms of indices of adjustment, as hypothesized, reticent behavior was significantly and negatively associated with social competence (j3 = -.28), /(74) = -2.59, p < .05. Contrary to predictions, the relation between reticent behavior and internalizing problems did not reach significance. However, reticent behavior was significantly and negatively related to academic achievement (/3 = -.28), r(74) = -2.51, p < .05, and externalizing problems (/3 = -.24), t(74) = -2.18, p < .05. As we had found previously, there were no significant main effects of gender or interactions between reticent behavior and gender in the prediction of indices of adjustment. Results are displayed in Table 3.

Table 2 Prediction of Reticent and Solitary-Passive Behaviors From Child Characteristics and Gender Variable

P

Unique variance accounted for

Dependent variable: Reticent behavior Step 1 Shyness Dysregulation Gender Step 2 Shyness X Gender Dysregulation X Gender Shyness X Dysregulation Step 3 Shyness x Dysregulation X Gender

.25* -.09

.02 .01 .01 .02 .01

Dependent variable: Solitary-passive behavior Step 1 Shyness Dysregulation Gender Step 2 Shyness x Gender Dysregulation x Gender Shyness x Dysregulation Step 3 Shyness x Dysregulation X Gender p < .05.

-.07 -.24* .16 .06* .00 .01 .01

Solitary-passive behavior. With regard to child characteristics, results indicated a significant main effect of emotional dysregulation in the prediction of solitary-passive behavior. As hypothesized, emotion dysregulation was negatively related to the display of solitary-passive behavior (j3 = -.24), r(74) = -2.10, p < .05, indicating that children rated as being less negatively emotional and easier to soothe were more likely to display solitary-passive behaviors during free play. No significant main effect relation with shyness was indicated. A significant interaction was found between shyness and gender in the prediction of solitary-passive behavior (sr2 = .06), r(73) = 2.28, p < .05. No other significant two-way or three-way interactions were indicated. Results are displayed in Table 2. In terms of relations with indices of adjustment, consistent with expectations, no significant main effect relations were found between solitary-passive play and any of the indices of adjustment. However, significant interaction effects between solitary-passive behavior and gender were indicated for internalizing problems (Art2 = .05), F(l, 73) = 3.84, p < .05, academic achievement (A/?2 = .10), F(l, 73) = 7.65, p < .01, and a trend for social competence (Art2 = .04), F(l, 73) = 3.45, p < .07. Results are displayed in Table 3. Interactions were explored by recomputing the regression analyses separately for boys and girls. Results are displayed in Table 4. Shyness was positively associated with solitary-passive behavior in boys but significantly negatively associated with solitarypassive behavior in girls. For indices of adjustment, a consistent pattern of results emerged. Solitary-passive behavior was negatively associated with adjustment in boys and positively associated with adjustment in girls. More specifically, for boys, solitarypassive behavior was positively associated with internalizing problems and negatively associated with social competence and academic achievement. The opposite pattern of results was observed for girls; solitary-passive behavior was negatively related to internalizing problems, positively related to academic achievement, and relatively unrelated to social competence. Solitary-active behavior. Solitary-active play was observed to occur in less than 2% of coding intervals. Subsequently, this variable was recoded in order to create two groups: The first group consisted of children who did not demonstrate any solitary-active behaviors during free play (comparison group, n = 50), and the second group contained children who demonstrated any number of solitary-active behaviors during free play (solitary-active group, n = 27). Results from chi-square analyses revealed that gender distribution of the comparison (22 boys, 28 girls) and solitaryactive (16 boys, 11 girls) groups did not differ from expected values. To predict membership in the solitary-active group from measures of child characteristics, gender, and their interactions, we computed a logistic regression, with shyness, dysregulation, and gender entered in Step 1, followed by all two-way interactions entered in Step 2, and the three-way interaction between gender, shyness, and dysregulation entered in Step 3. Results indicated a main effect of emotion dysregulation only (B = .80, Wald = 3.70, p < .05), such that membership in the solitary-active group was associated with greater emotion dysregulation (negative emotionality, difficult to soothe) than in the comparison group. A series of 2 (gender) X 2 (solitary-active group) analyses of variance was then performed for each of the indices of adjustment. Relevant means and standard deviations are displayed in Table 5.

469

WHEN GIRLS VERSUS BOYS PLAY ALONE

Table 3 Prediction of Indices of Adjustment From Measures of Nonsocial Play and Gender

Reticent behavior

Gender

.07 -.24* -.28* -.28*

-.04 -.18 .16 .01

Solitary-passive

Gender

Solitary-Passive X Gender

.01 -.02 -.12 .05

-.04 -.18 .17 -.01

.05* .02 .04t .10**

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Internalizing problems Externalizing problems Social competence Academic achievement

Internalizing problems Externalizing problems Social competence Academic achievement tp