International Journal of Behavioral Development # 2004 The International Society for the
2004, 28 (3), 204209 Study of Behavioural Development
http://www.tandf.co.uk/journals/pp/01650254.html DOI: 10.1080/01650250344000415 Imaginary companions and peer acceptance Tracy R. Gleason Wellesley College, MA, USA Early research on imaginary companions suggests that children who create them do so to compensate for poor social relationships. Consequently, the peer acceptance of children with imaginary companions was compared to that of their peers. Sociometrics were conducted on 88 preschool-aged children; 11 had invisible companions, 16 had personied objects (e.g., stuffed animals animated by the child) and 65 had no imaginary companion. The three groups were compared on positive and negative nominations, social preference, social impact, and total number of reciprocal friends. Given the positive correlation between pretend play and social competence, fantasy predisposition was used as a covariate. The groups did not differ on number of positive nominations by peers, total number of reciprocal friends, or social preference scores. However, compared to their peers, children with personied objects had higher social impact scores, largely as a result of negative nominations. Attention is thus called to the differences between personied object and invisible imaginary companions, and to the underlying social cognition that may be involved in their creation. The creation of an imaginary companion is simultaneously a typical and an atypical pathway through development in early childhood. The phenomenon is typical in that an imaginary companion who is controlled by the child and usually friendly is not considered a form of psychopathology (Silberg, 1998; Trujillo, Lewis, Yeager, & Gidlow, 1996). Moreover, some estimates suggest that imaginary companions are relatively common; incidence can be as high as 65% (D. Singer & Singer, 1990) when they are dened as including invisible compa- nions, personied objects (i.e., objects such as stuffed animals or dolls that are animated by the child), and extensive impersonation of a character. However, the phenomenon is atypical in that the emergence of an imaginary companion is not universal to preschool-aged children. Indeed, if only invisible friends are included in the denition, estimates of incidence can be as low as 6% (e.g., Harvey, 1918). Perhaps because many but not nearly all children engage in this kind of intense fantasy, researchers have searched for correlates of imaginary companion formation within several domains, including perception (Harvey, 1918), social cognition (Glea- son, 2002; Taylor & Carlson, 1997), and creativity (Schaefer, 1969). Although each of these areas has enjoyed some success in distinguishing between children with and without imaginary companions, the correlates of the phenomenon within social development have not been fully explored. Examination of the relation between the formation of imaginary companions and social development seems promis- ing given that many hypotheses regarding the origins of pretend friends relate to social conditions. In particular, imaginary companions have been related to childrens relation- ships with and acceptance by peers. For example, the creation of an imaginary companion has been explained as an effort to relieve loneliness or compensate for poor social relationships (Bender & Vogel, 1941; Harvey, 1918; Manosevitz, Prentice, & Wilson, 1973). In fact, the suggestion implied by some studies is that children with imaginary companions may be creating them precisely because they lack the social skills or opportunities to create real friends (Ames & Learned, 1946; Nagera, 1969; Svendsen, 1934). Some support for this theory comes from research by Harter and Chao (1992). Using the written version of the Pictorial Scale for Perceived Competence and Social Acceptance (Harter & Pike, 1984), these authors asked teachers to rate the competence of children in their classrooms who did and did not have imaginary companions. Children with imaginary companions were rated as less competent than their peers on the overall measure, which included a component of social acceptance by peers. The nature of relationships with imaginary companions seems to support the notion that pretend friends are created by socially incompetent children who may have been rejected by their peers. For example, parents often explain childrens creation of an imaginary companion in terms of the childs need for a relationship, and some pretend friends are imaginary versions of real people (Gleason, Sebanc, & Hartup, 2000). Unlike real friends, who will not always conform to a childs desires, an imaginary companion is subject to the childs whim (Taylor, 1999) and thus does not require sophisticated social skills for successful interaction. Moreover, children claim that their relationships with imaginary companions are equally as important as their relationships with real people (Mauro, 1991), suggesting that pretend friends occupy a status in the childs world that would typically be occupied by a real friend. Taken together, these elements of pretend friends existence seem to suggest that they are created to compensate for poor Correspondence should be addressed to Tracy R. Gleason, Depart- ment of Psychology, Wellesley College, Wellesley, Massachusetts 02481-8203, USA; e-mail: tgleason@wellesley.edu. This research was supported in part by a Wellesley College Faculty Award. Thanks are due to the children and teachers who participated in the study, as well as to Jean Gleason, Bill Hartup, and David Kaiser for their comments on earlier versions of this manuscript. The author is also grateful to Anneliese Bass, Kate Fiske, Amy Gower, and Lisa Hohmann for help with data collection and coding and especially to Bill Bukowski for his assistance with data analyses. peer relationships, but this question has never been tested empirically. The goal of this study was thus to test the connection between the creation of imaginary companions and childrens peer acceptance. Not all of the literature supports the notion that children who create imaginary companions have poor peer relation- ships. In fact, some evidence indicates that children who create imaginary companions are more sociable than other children. For example, compared to their peers, children with imaginary companions are more cooperative with adults and other children and demonstrate more positive affect in their play (Partington & Grant, 1984; D. Singer & Singer, 1990). In addition, parental reports of childrens shyness have suggested that children with imaginary companions are less shy than their peers, at least during the preschool period (Mauro, 1991). Children with pretend friends also have real friends and playmates at rates similar to children without imaginary friends (Gleason et al., 2000; Manosevitz et al., 1973; Masih, 1978). These ndings run directly counter to the idea that children with imaginary friends are rejected by their peers. If children who create imaginary companions are not doing so to compensate for their relationships with real peers, then they may develop pretend friends simply because they enjoy the companionship available from such a friend (Taylor, 1999)after all, most imaginary companions are abandoned when real playmates are around (Manosevitz et al., 1973). As with real friends, children play with their imaginary compa- nions, pretend with them, involve them in their daily routines, and occasionally argue with them (Gleason et al., 2000; Taylor, 1999). In fact, rather than being the result of rejection by peers, imaginary companions may enhance childrens peer relationships by providing a forum for children to practice interacting with a peer. After all, children with imaginary companions appear to conceptualise their pretend friends as relationship partners, and appear to attribute similar qualities to these quasi-relationships as they do to relationships with best friends (Gleason, 2002). This extra social rehearsal may even mean that children with imaginary companions are particularly adept at forming and maintaining peer relation- ships in comparison to other children. If so, imaginary companion creators might be attractive playmates, as socially competent children often enjoy high acceptance by peers (Connolly & Doyle, 1981; Howes, 1988). The idea that children with imaginary companions may have heightened social skills is supported by the positive relation between social competence and fantasy play more generally (Connolly & Doyle, 1984; Doyle & Connolly, 1989; Garvey, 1977; Rubin & Maioni, 1975). Formulating an imaginary companion constitutes the creation of a fantasy (Taylor, 1999), and just as pretend play is described as an elaboration of elements of real experiences (Harris, 2000), the creation of an imaginary friend might be an elaboration of elements of real relationships. After all, fantasy play and the creation of an imaginary companion appear to be related. In a longitudinal study, Acredolo and colleagues (Acredolo, Good- wyn, & Fulmer, 1995) demonstrated that 4-year-olds with imaginary companions had shown a pattern of signicantly more interest in fantasy play than their peers since infancy. Accordingly, the positive relationship between fantasy play and social competence may extend to children who create imaginary companions as well, even if just as a function of involvement in pretence and not as a function of the existence of pretend friends. The positive relation between fantasy play and social competence raises another issue. Children with imaginary companions are considered high-fantasy (e.g., Taylor & Carlson, 1997), meaning that their predilections for pretence, and not their imaginary companions per se, might be responsible for some of the ndings relating the creation of imaginary companions to sociability. In other words, being a high-fantasy child may be connected to social competence and therefore to acceptance by peers (Connolly & Doyle, 1984), and having an imaginary companion may not contribute to that social success over and above ones involvement in fantasy. Therefore, the creation of an imaginary companion and predilection for fantasy ought to be examined as separate contributors to childrens acceptance by their peers. Another factor that might inuence the relation between imaginary companions, fantasy, and peer acceptance is the denition of imaginary companion used to identify children with and without pretend friends. This distinction is impor- tant, as systematic differences have been found between different types of imaginary companions. In particular, parents frequently describe invisible imaginary companions as provid- ing egalitarian relationships, akin to friendships, whereas children appear to nurture their personied objects much as a parent nurtures a child (Gleason et al., 2000). If interactions in these quasi-relationships inuence interactions with real peers, then children with invisible companions may be at an advantage, in that invisible friends may offer social rehearsal that is more directly applicable to peer relationships than that of children with personied objects. This investigation was designed to test directly the notion that having an imaginary companion is related to acceptance by peers. One possible hypothesis was that children may create imaginary companions in order to compensate for inadequate social relationships, or as an alternative to interactions with real peers. Presumably, then, children with either invisible imagin- ary companions or personied objects would be less accepted by their peers than children without imaginary companions. However, the bulk of the literature suggests that acceptance by peers should not be lower for children with imaginary companions because of their sociability (Taylor, 1999) and their extensive involvement in fantasy (Connolly & Doyle, 1984). Conceptualising the imaginary companion as a venue for social rehearsal also suggests that the peer acceptance of children with imaginary companions should not be lower than that of their peers, and may even be higher. Moreover, acceptance by peers might be higher for children with invisible companions over those with personied objects given the differential nature of the relationships that children often form with these different types of companions. Sociometric inter- views thus were conducted with a large group of preschool- aged children to examine their social status in the classroom as a function of whether they had imaginary companions. Childrens scores on a measure of fantasy predisposition were factored out of these analyses to more clearly indicate the relative contributions of imaginary companions to a childs acceptance by peers independently of his or her fantasy level. Method Participants A total of 105 preschool children were invited to participate in INTERNATIONAL JOURNAL OF BEHAVIORAL DEVELOPMENT, 2004, 28 (3), 204209 205 206 GLEASON / IMAGINARY COMPANIONS AND PEER ACCEPTANCE the study and 92 accepted. Four children had to be dropped from the study because they were missing fantasy interviews owing to absences or the end of the school year. The remaining 88 children (43 boys, 45 girls) ranged in age from 43 to 69 months (M = 54.8 months, SD = 7.43 months) and were enrolled in six different classrooms in two private preschools located in an upper-middle class suburban area. Approxi- mately 84% of the children were white. Materials This investigation was part of a larger study of childrens fantasy play and imaginary companions. Materials relevant to this portion of the study are described below. Sociometric interview. Childrens preferences for particular classmates were assessed using a sociometric interview. Participants were shown individual photographs of the children in their classrooms whose parents had given permission for participation in the study. Participants were asked to name the child pictured, after which the photo was placed face up on the table. Once all the pictures were out, children were asked to point to pictures of three children with whom they most liked to play. The researcher then removed all of the photos except those the child had chosen, and asked the child to choose the one friend of the three with whom he or she liked to play with the very most. Teacher nominations. While other researchers have asked children to specify who they do not like to play with or simply do not like in the classroom to acquire negative nominations (e.g., Asher & Dodge, 1986; Coie, Dodge, & Coppotelli, 1982), the schools where we collected data were reluctant to use this method. Instead, each childs primary teacher nominated a peer with whom the child rarely or never played from among a list of children in the class who were participating in the study. A single negative nomination was obtained as data were collected toward the end of the year when children knew each other well and were likely to have played with most of their classmates at some point. Although obviously childrens nominations of each other would have been ideal, teacher nominations seemed to be a viable alternative given the overlap of their responses with childrens responses on such measures in past research (Wu, Hart, Draper, & Olsen, 2001) and the relation of teacher-based rankings of popularity to childrens social competence (Con- nolly & Doyle, 1981). Teachers were blind to the hypotheses of the study and to childrens imaginary companion status. Fantasy interview. A fantasy interview developed by Singer and Singer (1990) was administered to the children. This interview has been used to categorise children into high and low fantasy groups (J. Singer, 1961; J. Singer & Streiner, 1966), and correlates with other measures of fantasy such as parents reports of the childs impersonation of characters, the ability to use an imaginary object (versus a body part) in a pretend action, and creation of imaginary companions (Taylor & Carlson, 1997). Children were asked to name their favourite story, toy, TV show, and game. In addition, they were asked to tell the researcher what they liked to do alone, what they thought about when lying in bed at night, and what they liked to play with others. The responses were coded as either fantasy oriented (e.g., Cinderella, playing house), which received a score of 1, or reality oriented (e.g., puzzle, tag), which received a score of 0. Finally, they were asked if they ever pretended to be an animal or another person. Afrmative responses received a score of 1, and negative responses a zero. Childrens fantasy scores were the sum of their responses to the 10 questions. Parent report of imaginary companions. Parents were asked to complete a short questionnaire regarding whether their child had an imaginary companion. If so, the parent was asked to specify if the companion was invisible or a personied object and, regardless of type, to report the companions name and give a general description. Imaginary companion interview. Each child was asked whether or not he had a pretend friend using the procedure described by Taylor, Cartwright, and Carlson (1993). This procedure has been used successfully to identify children with imaginary companions, as children understand that these friends are not real (Taylor, 1999). If the child responded afrmatively to the question, then the imaginary companion interview was con- ducted. If the child responded that he or she did not have an imaginary companion though the parent had reported one, the child was asked What about X? Your (Mom/Dad) tells me that you have a pretend friend named X, is X your pretend friend? Although seven children initially reported no imagin- ary companion (four with invisible friends, three with personied objects), no child whose parent had named an imaginary companion responded negatively to this question. Children with imaginary companions were asked some general questions about the companion to help ensure that the children were not creating it on the spot. Interviewers also ascertained whether the companion was invisible or a personied object. Procedure Because they were combined with other measures not included here, the sociometric interview was administered in one session, and the fantasy and imaginary companion interviews in a separate session. Before the fantasy interview, the imaginary companion questionnaire was distributed to parents, and teachers were asked for the negative peer nominations. Children were asked about imaginary companions regardless of their parents reports, as children do not always share their companions with their parents. Approximately 2 weeks after the imaginary companion interview, children were questioned a second time about imaginary companions to be sure their companions were consistent. Parents were also asked to clarify that any companion mentioned by a child was not a real friend, and that any personied objects mentioned by either child or parent were animated by the child and not just carried around or used for comfort. Children were classied as having an imaginary companion if they named the same companion in both interviews, or if the companion they named once was corroborated by the parent. Coding Fantasy interview. Of the 10 items on the fantasy interview, 3 correlated either weakly or not at all with the other items. These threefavourite story, television show, and gamewere thus dropped from the total scores. Without these items, reliability on the scale was low but acceptable (Cronbachs alpha = .59). Ratings of social status. Three different ratings of social status were obtained. The number of positive and negative nomina- tions each child received from the other children (positive) or the teacher (negative) was calculated. The number of reciprocal friends (out of the three choices) each child had was also computed. All scores except the number of reciprocal friends were converted to z-scores so that the ratings could be compared across classrooms. Results Imaginary companions and fantasy interviews Of the 88 children in the study, 18.2% (n = 16; 7 girls, 9 boys) had personied object imaginary companions and 12.5% (n = 11; 9 girls, 2 boys) had invisible imaginary companions. These proportions are somewhat lower than those found in other studies (e.g., Gleason et al., 2000; Taylor & Carlson, 1997), but are within the typical range (Taylor, 1999). A chi-square examination of gender by companion type was signicant, w 2 (1, N = 27) = 3.91, p 5 .05, in that more girls than boys had invisible imaginary companions. Because gender and companion type were related, an independent samples t-test was conducted comparing boys and girls scores on the fantasy interview. No gender difference emerged. A one-way ANOVA on the fantasy scores using imaginary companion status as the independent factor was marginally signicant, F(2, 87) = 2.71, p = .07. Children with personied objects had the highest fantasy scores, followed by children with invisible companions and children with no imaginary companions. (See Table 1 for means and standard deviations.) Measures of social status The teacher-provided negative nominations were rst tested to verify that teachers chose largely opposite-sex peers for these measures. Although the validity of teacher-based nominations has not been examined, verifying that teachers primarily chose opposite-sex peers for the negative nominations provided some support that their choices might be a reasonable proxy for the childrens nominations, given that preschool-aged children typically nominate opposite-sex peers when asked who they rarely play with or do not like (Ramsey, 1995). For girls, teachers in this sample chose boys for the negative nominations in 90.7% of the cases, and for boys, girls were nominated in 89.2% of cases. A chi-square analysis conducted on gender of child by gender of child nominated by the teacher was highly signicant, w 2 (1, N = 84) = 45.92, p 5 .001. Children with invisible companions, personied objects, and no imaginary companion were compared according to their standardised scores on the positive and negative nomina- tions (see Table 1 for means and standard deviations). Fantasy score was used as a covariate to ensure that any effects could be attributed to the creation of an imaginary companion per se and not simply to high involvement in fantasy, especially since the three groups differed marginally on the fantasy interview. A 3 (imaginary companion status) 2 (gender) MANCOVA resulted in a signicant difference only for negative nomina- tions, F(5, 82) = 2.18, p = .05, Z 2 = .14. No interaction effects occurred, but the main effect for imaginary companion status was marginally signicant, F(2, 82) = 2.35, p = .10, Z 2 = .06. The standardised scores on the negative nominations of the children with personied objects were high and positive, those of the children with invisible friends high and negative, and those of the children with no imaginary companions were close to zero. Reciprocity of friendships To examine the hypothesis that the number of reciprocal friendships a child had might differ by imaginary companion status, once again a 3 2 ANOVA was run with gender and imaginary companions as xed factors and fantasy score as a covariate. No signicant differences were found. (See Table 1 for means and standard deviations.) Discussion In general, children with and without imaginary companions appear to be more similar than different in terms of their acceptance by peers. This claim is supported particularly by the effect sizes of the ndings, which were small even where statistically signicant. The results suggest that having or not having an imaginary companion per se does not correspond to differences in childrens peer acceptance, although possibly the type of companion a child creates may have some small bearing on his or her popularity within the classroom. Specically, compared to their peers, children who create personied object imaginary companions appeared to have slightly elevated negative nominations (as provided by teachers). Moreover, these results were obtained while controlling for the childrens predisposition for fantasy as measured by the fantasy interview, INTERNATIONAL JOURNAL OF BEHAVIORAL DEVELOPMENT, 2004, 28 (3), 204209 207 Table 1 Means and standard deviations for fantasy interview scores, social status measures, and number of reciprocal friends (maximum 3) by imaginary companion status Imaginary companion status Invisible friend (n = 11) Personied object (n = 16) No imaginary companion (n = 61) Mean SD Mean SD Mean SD Fantasy interview 4.27 1.95 4.44 1.09 3.44 1.82 Positive nominations 0.0182 0.98 0.3142 1.08 0.1055 0.96 Negative nominations 0.5537 0.63 0.4905 1.20 0.0143 0.93 No. of reciprocal friends 1.45 1.04 1.80 0.94 1.44 0.89 208 GLEASON / IMAGINARY COMPANIONS AND PEER ACCEPTANCE meaning that for children with personied objects, their imaginary companions weakly relate to their social standing above and beyond any inuence derived from their status as high fantasy children. These ndings have several implications for understanding the relation between imaginary companions and peer acceptance. Imaginary companions as compensation for poor relationships The hypothesis that children create imaginary companions in order to compensate for poor relationships is unsupported by these data. Children with imaginary companions received statistically equivalent positive nominations from their peers to those of children without imaginary companions, and the groups did not differ signicantly in number of reciprocal friends. Children with imaginary companions thus are not creating them because they have few or no real friends. They appear to be as well-liked as children without imaginary companions. Concerns, therefore, that children might be creating imaginary companions because they cannot create satisfactory relationships with real peers seem unfounded. Initially, the fact that children with personied objects received slightly more negative nominations than their peers suggests that they might be rejected children. However, given that they did not differ from their peers on positive nomina- tions or number of reciprocal friends, these children appear to have friends. Of course, many rejected children have friends (Hartup, 1983), but despite its correlational nature, the pattern of the data does not support the notion that children create a personied object as a result of rejection from the peer group. The effect was only marginally signicant and quite weak, suggesting that overall, children with personied objects are probably not signicantly more likely than their peers to receive negative nominations. Imaginary companions as a venue for social rehearsal The idea that the creation of an imaginary companion might enhance a childs social competence and thus peer acceptance by providing a forum for practising social interactions with peers received little support. The marginal differences between the imaginary companion groups on the negative nominations suggested at best that invisible companions, but not personi- ed objects, kept children from receiving negative nomina- tions. This nding may indicate that the creation of an invisible imaginary companion enhances the social competence of children to a small but largely unremarkable degree. Similarly, the hypothesis that children with invisible imaginary compa- nions would receive higher ratings of peer acceptance than the children with personied objects received some weak support but, of course, only in terms of negative nominations. However, these ndings contribute to a growing body of literature suggesting that invisible companions and personied objects are subtly distinct phenomena with separate underlying social processes. The vertical nature of childrens relationships with personied objects (Gleason et al., 2000) implies that the focus of this fantasy is framed by the hierarchical nature of the relationship. Ostensibly, this emphasis on the self as the competent partner might be reected in childrens interactions with their peers, meaning that their concepts of friendship may also not be as structurally egalitarian as those of their peers. Thus, these children may be inclined to take a leadership role both in instances where such behaviour is warranted and where it is not. Such behaviour would be consistent with the ndings Harter and Chao (1992) reported for girls with imaginary companions. The girls in their study created less competent imaginary companions and were rated as less socially competent than their peers. Teachers also rated the boys as less socially competent, but the companions they created were superordinate rather than subordinate to themselves. Although Harter and Chao did not specify if the imaginary companions in their study were invisible or personied, one interpretation of these ndings and those of the present study is that children with personied objects have concepts of peer relationships that are more similar to the schemas used for parentchild relationships than to those of egalitarian friendships (Gleason, 2002), and these schemas may occasionally hamper their interactions with peers. Imaginary companions and fantasy predisposition The creation of an imaginary companion does not seem to contribute to peer acceptance over and above what is contributed by fantasy for positive nominations by peers. These results suggest that the robust relationship between engagement in pretend play and socially competent behaviour (Connolly & Doyle, 1984; Doyle & Connolly, 1989; Rubin & Maioni, 1975) is not facilitated by the creation of an imaginary companion. Nor, of course, are positive aspects of peer acceptance hindered by the existence of a pretend friend. Once again, the lesson here appears to be that children who create imaginary companions are more similar to than different from their peers in terms of positive peer acceptance. For negative nominations, the slight between-group differ- ences are clearly connected to having or not having a personied object over and above any group differences in fantasy predisposition. Although the groups differed marginally on their fantasy scores, factoring these scores out of the analyses ensured that the differences that emerged were not inuenced by a general pattern of involvement in fantasy on the part of the children with personied objects. In fact, given their generally high predilection for fantasy, children with personied objects would be expected to be among the more socially competent children in the classrooms (Connolly & Doyle, 1984). Limitations The use of teacher rather than peer ratings and only a single rating for the negative nominations may have inuenced the results. Teacher ratings of childrens sociometric status are not perfectly correlated with childrens ratings; however, nor are they completely independent (Wu et al., 2001). Another possible inuence may have been the presence of the childrens personied objects at school even if the children did not play with them there. Many children, especially those in full-day care, bring a precious object to school, and approximately two- thirds of personied objects have been to school at some point (Gleason et al., 2000). The presence of these objects may have inuenced teachers ratings of these children in systematic ways, although presumably children with transitional objects may have brought them to day care as well, and teachers may not have been able to distinguish children with transitional versus personied objects. Last, condence in the results is limited by the small number of children who had imaginary companions and by the gender disparity of the invisible imaginary companion group. The imbalance in this group may have prevented the detection of interactions between gender and imaginary companion status on the standardised nomina- tion scores and, furthermore, group differences in imaginary companion status could have been owing to gender. Replica- tion would certainly enhance condence in the results. Future directions and conclusions The results of this investigation suggest that concerns about childrens creation of an imaginary companion as a way to compensate for poor or nonexistent friendships should be put to rest. Although the slightly elevated negative nominations of the children with personied objects invite further investigation into the differences between these children and those with invisible friends or no imaginary companions, the ndings do not suggest that children who create imaginary companions have lower rates of acceptance than their peers. Instead, future research could usefully focus on two sets of questions with respect to the links between real and imaginary friendships. First, concepts of real and imaginary friendships appear to be similar (Gleason, 2002), but the concepts underlying these relationships in early childhood have not been fully described. The relation between the qualities of real and imaginary friendships may provide interesting insights into how children view their friendships, what they expect from them, and how they might like them to function. Second, although not reected in childrens acceptance by peers, the creation of imaginary companions may be connected to childrens social competence within early peer relationships. Specically, behaviours observable in relationships with imaginary compa- nions may be mirrored in relationships with real friends. A child who successfully negotiates a conict with an imaginary companion may be able to apply this skill with real children, meaning that fantasy or role-play might be a way to help children who need to work on their social skills. Another interesting topic for future research concerns the nature of the personied object imaginary companion and its correlates in social experience and social cognition. Many personied objects emerge from transitional objects (D. Singer & Singer, 1990) but this process has not been explored, nor has any research examined why this transformation occurs to some but not all transitional objects. In addition, if children with personied objects are conceptualising their friendships differ- ently from other children, the nature of these representations deserves some attention. Last, further description of the variation in types of imaginary companions may continue to shed light on the functions of these fantastical friends and the roles that they play in childrens early social development. Manuscript received November 2002 Revised manuscript received May 2003 PrEview publication February 2004 References Acredolo, L., Goodwyn, S., & Fulmer, A. (1995, April). Why some children create imaginary companions: Clues from infant and toddler play preferences. Poster presented at the biennial meetings of the Society for Research in Child Development, Indianapolis, IN. Ames, L., & Learned, J. (1946). Imaginary companions and related phenomena. Journal of Genetic Psychology, 69, 147167. Asher, S., & Dodge, K. (1986). Identifying children who are rejected by their peers. Developmental Psychology, 22, 444449. Bender, L., & Vogel, F. (1941). Imaginary companions of children. American Journal of Orthopsychiatry, 11, 5665. Coie, J., Dodge, K., & Coppotelli, H. (1982). Dimensions and types of social status: A cross-age perspective. Developmental Psychology, 18, 557570. Connolly, J., & Doyle, A.-B. (1981). Assessment of social competence in preschoolers: Teachers versus peers. Developmental Psychology, 17, 454462. Connolly, J., & Doyle, A.-B. (1984). Relation of social fantasy play to social competence in preschoolers. Developmental Psychology, 20, 797806. Doyle, A.-B., & Connolly, J. (1989). Negotiation and enactment in social pretend play: Relations to social acceptance and social cognition. Early Childhood Research Quarterly, 4, 289302. Garvey, C. (1977). Play. Cambridge, MA: Harvard University Press. Gleason, T. (2002). Social provisions of real and imaginary relationships in early childhood. Developmental Psychology, 38, 979992. Gleason, T., Sebanc, A., & Hartup, W. (2000). Imaginary companions of preschool children. Developmental Psychology, 36, 419428. Harris, P. (2000). The work of the imagination. Oxford: Blackwell. Harter, S., & Chao, C. (1992). The role of competence in childrens creation of imaginary friends. Merrill-Palmer Quarterly, 38, 350363. Harter, S., & Pike, R. (1984). The pictorial scale of perceived competence and social acceptance for young children. Child Development, 55, 19691982. Hartup, W.W. (1983). Peer relations. In P. Mussen (Series Ed.), & E.M. Hetherington (Ed.), Handbook of child psychology: Vol. 4. Socialization, personality, and social development (pp. 103196). New York: Wiley. Harvey, N. (1918). Imaginary playmates and other mental phenomena of children. Ypsilanti, MI: State Normal College. Howes, C. (1988). Same- and cross-sex friends: Implications for interaction and social skills. Early Childhood Research Quarterly, 3, 2137. Manosevitz, M., Prentice, N., & Wilson, F. (1973). Individual and family correlates of imaginary companions in preschool children. Developmental Psychology, 8, 7279. Masih, V. (1978). Imaginary play companions of children. In R. Weizman, R. Brown, P. Levinson, & P. Taylor (Eds.), Piagetian theory and the helping professions (pp. 136144). Los Angeles: University of Southern California Press. Mauro, J. (1991). The friend that only I can see: A longitudinal investigation of childrens imaginary companions (Doctoral dissertation, University of Oregon, Eugene, 1991). Dissertation Abstracts International, 52, 4995. Nagera, H. (1969). The imaginary companion: Its signicance for ego development and conict solution. Psychoanalytic Study of the Child, 24, 165195. Partington, J., & Grant, C. (1984). Imaginary playmates and other useful fantasies. In P. Smith (Ed.), Play in animals and humans (pp. 217240). New York: Basil Blackwell. Ramsey, P. (1995). Changing social dynamics in early childhood classrooms. Child Development, 66, 764773. Rubin, K., & Maioni, T. (1975). Play reference and its relationship to egocentrism, popularity and classication skills in preschoolers. Merrill-Palmer Quarterly, 21, 171179. Schaefer, C. (1969). Imaginary companions and creative adolescents. Develop- mental Psychology, 1, 747749. Silberg, J. (1998). Interviewing strategies for assessing dissociative disorders in children and adolescents. In J. Silberg (Ed.), The dissociative child: Diagnosis, treatment and management (2nd ed., pp. 4768). Lutherville, MD: Sidran Press. Singer, D., & Singer, J. (1990). The house of make-believe. Cambridge, MA: Harvard University Press. Singer, J. (1961). Imagination and waiting ability in young children. Journal of Personality and Social Psychology, 29, 396413. Singer, J., & Streiner, B. (1966). Imaginative content in the dreams and fantasy play of blind and sighted children. Perceptual and Motor Skills, 22, 475482. Svendsen, M. (1934). Childrens imaginary companions. Archives of Neurology and Psychiatry, 32, 985999. Taylor, M. (1999). Imaginary companions and the children who create them. New York: Oxford University Press. Taylor, M., & Carlson, S. (1997). The relation between individual differences in fantasy and theory of mind. Child Development, 68, 436455. Taylor, M., Cartwright, B., & Carlson, S. (1993). A developmental investigation of childrens imaginary companions. Developmental Psychology, 29, 276285. Trujillo, K., Lewis, D.O., Yeager, C., & Gidlow, B. (1996). Imaginary companions of school boys and boys with Dissociative Identity Disorder/ Multiple Personality Disorder. Child and Adolescent Psychiatric Clinics of North America, 5, 375391. Wu, X., Hart, C., Draper, T., & Olsen, J. (2001). Peer and teacher sociometrics for preschool children: Cross-informant concordance, temporal stability, and reliability. Merrill-Palmer Quarterly, 47, 416443. INTERNATIONAL JOURNAL OF BEHAVIORAL DEVELOPMENT, 2004, 28 (3), 204209 209