Unilateral Friendship Outcomes and Preschool Friendship Formation

Unilateral friendships provide a powerful mechanism to study friendship formation, because these friendships contain interest in friendship, but have not yet become full, reciprocated friendship. This longitudinal study investigated whether the conversations of unilateral dyads identified in the fall were predictive of that dyad’s relationship status in the winter. I hypothesized that the conversations of unilateral dyads would vary widely. I also hypothesized that unilateral dyads that develop into reciprocal friendships will speak more to each other, make more mutual/partner play statements, make more noncompliant/disapproving statements, and more commands/requests than unilateral friendships that remained unilateral or became nonfriend relationships. Seventy-four preschool-aged children (37 boys, 37 girls, Mage = 49.26 months, SD = 7.29 months) from 5 classrooms participated in sociometric interviews in the fall. Unilateral dyads were identified from the sociometric data, and the dyads were observed during a 10-minute free play session. The conversations of the dyads were analyzed, and the outcomes of each of the unilateral dyads were determined using data from a second round of sociometric interviews conducted in the winter. The conversations of the dyads had wide variation in the amount of conversation and the number of the different conversational statements, although, conversations of unilateral dyads in the fall did not predict relationship status in the winter. Future research that examines how a child speaks to the other member of a dyad, and the intent behind their speech, may be more relevant than the content of the conversation when studying preschool friendship formation. PRESCHOOL UNILATERAL FRIENDSHIP OUTCOMES 4 Unilateral Friendship Outcomes and Preschool Friendship Formation The ability to make and maintain friendships in childhood is critical, as it has been linked with later academic achievement, self-image, and mental health (Berndt, 2002). The preschool years are characterized by the rapid development of social skills needed to make these important friendships. Historically, work in this area has focused on characteristics and social skills of individual children as predictors of success in friendship formation (Drewry & Clark, 1985). However, these approaches ignore the fact that any relationship is dependent not just upon the characteristics of the individuals, but also upon the properties of the dyad – properties that exist independently of either dyad member alone. Consequently, this study examines friendship formation during the preschool years by focusing on dyadic behaviors in relation to later friendship outcomes. Many studies have explored the characteristics of friendship, and this research reveals the qualities associated with positive, reciprocal friendship. Less work has focused on processes that lead to this outcome, particularly during the preschool years. One feature of preschool friendship that has received some attention and might be related to friendship formation at this age, is the quality of unilateral friendships, or a relationship in which one child has nominated the other as a friend, but that nomination is not reciprocated. Unilateral relationships are powerful mechanisms to study friendship formation, because in unilateral friendship there is interest in friendship, but not yet reciprocated friendship. In particular, the mechanisms by which unilateral friendships become reciprocal friendships have yet to be thoroughly explored. Examination of the insights provided by unilateral relationships requires understanding the nature of preschool friendships generally, the qualities of unilateral friendships in particular, and the process of friendship formation in early childhood. PRESCHOOL UNILATERAL FRIENDSHIP OUTCOMES 5 Friendship in Preschool For many children, preschool presents the first significant opportunity to interact with children of their own age (Kerns, 2000). Preschool friendships share some characteristics with older friendships; as in adults, friendship in children is defined as a close relationship between two peers (Bukowski & Hoza, 1989). Unlike friendships in older children and adults, however, which are characterized by high levels of prosocial behavior, intimacy, authenticity, and other positive features, while maintaining low levels of conflict, rivalry and other negative features (Berndt, 2002), preschool friendships tend to be characterized by more overt properties of friendship. Thus, elements of friendship such as common activities, propinquity, and sharing secrets are key elements of preschool friendship. Children place increasing value on dispositional characteristics of friendship (e.g., consideration of others’ feelings) as they grow older (Furman & Bierman, 1984). Despite their young age and relative inexperience with peers, preschoolers’ behaviors with friends are distinct from those with non-friends or acquaintances. When placed in a social situation with both a friend and a non-friend, children consistently share toys more equitably, play in closer physical proximity, verbalize more frequently, and express greater satisfaction with friends than with non-friends (Garcia Werebe & Baudonnière, 1988). The differences between friends and non-friends are apparent in other social contexts as well. When pairs of preschoolers are randomly assigned to be either a teacher or a learner, friends are more involved with their partners, more emotionally expressive, and more competitive than nonfriends. Teachers who are friends are rated as more domineering, and learners as more playful and friendly. Thus, preschool friends engage in distinctly different kinds of interactions than non-friends in a variety of social contexts (Brachfeld-Child & Schiavo, 1990). PRESCHOOL UNILATERAL FRIENDSHIP OUTCOMES 6 Kerns’ (2000) analysis of the behaviors of pairs of preschool best-friends during a 1-hour free-play session revealed substantial variation on measures of positive social orientation, harmony, control, responsiveness, cohesiveness, and coordinated play. While the majority of friendships in this study were characterized by harmonious, responsive, and interactive play, some friendships were characterized by harmonious and independent play, and others were disjointed, marked by low levels of both coordinated play and positive interaction. A majority of the friendships in this study, whether highly harmonious or disjointed, remained stable over the course of one year; however, somewhat unexpectedly, the most stable friendships were those that were characterized by low levels of coordinated play and positive interaction. The generally high rates of friendship stability in this study may be attributed to the fact that only best friendships were recruited. The unexpected results regarding which type of friendships were the most stable may also be related to the manner in which the participants were recruited. Participants in this study were recruited via advertisements and by asking mothers whether her child had a best friend who would be available to participate in the study with her child. That the mothers of the children in the study were not only aware of their children’s friendship, but also had the ability to coordinate with the other child’s parents, indicates that the friendships in this study were likely to have a high degree of parental support, be formed at a relatively young age, and have relatively frequent interaction. Such a relationship may begin to take on the properties of a sibling relationship, which is often characterized by relatively high rates of conflict and disjointed behavior (Recchia & Howe, 2009). If this study were conducted in a preschool context, as is the current study, where children have access to a greater choice of friends, children might be more likely to pursue and maintain friendships with high levels of coordinated play and positive affect. Indeed, PRESCHOOL UNILATERAL FRIENDSHIP OUTCOMES 7 kindergarten friendships characterized by high validation and low rates of conflict were found to be the most stable over a 4-month period (Ladd, Kochenderfer, & Coleman, 1996). The different results regarding the stability of preschool friendship indicate that the children’s environment may be a key variable in the children’s development of friendships over time. Consequently, environment may also play a role when children originally develop, or form, these friendships. Unilateral Friendship A unilateral friendship is one in which an individual nominates another as a friend, but that friendship nomination is not reciprocated. Unilateral relationships differ from other relationships; elementary-aged children report liking unilateral friends less than reciprocal ones but more than non-friends. Elementary-aged children also report liking unilateral-given friends, friends they nominated, more than unilateral-received friends, friends from whom they received (but did not give) nominations (Hundley & Cohen, 1999). Children also tend to know more about their reciprocal friends than their unilateral ones; for example, children are more accurate at predicting shared characteristics of reciprocal friendships than unilateral ones (Ladd & Emerson, 1984). Reciprocal friends are more likely to be similar to each other with regard to interpersonal understanding, proximity, popularity, and self-concept than unilateral friends (Drewry & Clark, 1985; Kurdek & Krile, 1982). While unilateral friendships have characteristics that distinguish them from other peer relationships, unilateral friendships can take on characteristics of both reciprocal and non-friend relationships. For instance, the nature of conflict between unilateral dyads contains some characteristics of reciprocal friends and some of non-friends. The nature and intensity of conflicts resemble reciprocal dyads; however, the manner in which unilateral pairs negotiate their conflict, with both parties sticking to their original positions, is more similar to non-friend PRESCHOOL UNILATERAL FRIENDSHIP OUTCOMES 8 dyads. After conflict, unilateral dyads arrive at outcomes that tend to resemble non-friends, with one winner and one loser (Fraysse, 1994). However, unilateral dyads are more similar to reciprocal pairs in the way interaction tends to resume immediately after conflict (Fraysse, 1994; Hartup, Laursen, Stewart & Eastenson, 1988), perhaps due to the desire of the unilateral-given partner to resume interaction with the unilateral-received partner. Unilateral Friendships among Preschool-aged Children. Unilateral friendships are more prevalent among preschool-aged children than older children (Gleason & Hohmann, 2006; Guralnick, Gottman & Hammond,1996; Vaughn, Azria, Krysik, Caya, Bost, Newell & Kazura, 2000). Unilateral dyads have been found to have varying levels of stability; reciprocal friends are more likely to be maintained across a 6-month period than unilateral friends (Drewry & Clark, 1985; Gershman & Hayes, 1983). The ages and popularity scores of unilateral dyads are more disparate than those of reciprocal friends (Drewry & Clark, 1985). Clear behavioral distinctions exist between these two types of relationships. In naturalistic observation, reciprocal friend tend to interact more frequently than unilateral friends (Vaughn, Colvin, Azria, Caya & Krzysik, 2001). Although unilateral friendships are common in preschoolers, and comprise a significant portion of preschoolers’ social environments, how preschoolers understand these friendships is still unclear. Preschoolers are just beginning to develop the ability to recognize that others’ mental representations of the world may or may not match their own (Astington, 1993). The ability to recognize the one-sided nature of a unilateral friendship may directly influence preschoolers’ understanding of these friendships. Preschoolers appear to make some cognitive distinctions between these types of friendships. Preschoolers cite different reasons for liking unilateral vs. reciprocal friends; children are more likely to cite common activities and positive PRESCHOOL UNILATERAL FRIENDSHIP OUTCOMES 9 evaluation (e.g., s/he is nice) as reasons for liking reciprocal friends than unilateral friends (Hayes, Gershman, & Bolin, 1980). Additionally, preschool-aged children appear to have differing expectations of reciprocal and unilateral friends with regard to social provisions, or social benefits, of friendship. The specific social provisions children expect of reciprocal vs. unilateral friendships do not differ, but children expect that reciprocal friends will offer these social provisions to a greater degree (Gleason & Hohmann, 2006). Reciprocal dyads’ average social competence also tends to be greater than that of unilateral dyads (Vaughn et al., 2001). Friendship Formation Much research on children’s friendship formation has focused on friendship selection, or identifying likely friendship candidates rather than the social processes involved in friendship formation. Preschool-aged children are highly influenced by similarities in sex, race, age, and common activities when selecting their friends (Aboud & Mendelson, 1998). In addition to these demographic characteristics, preschoolers also appear to be influenced by other, more nuanced, personal characteristics. Temperamental qualities, such as activity level, impulsivity, and soothability have all been shown to influence the likelihood of being selected as a friend. The effects of temperament and gender are intertwined; girls are more likely to choose low activity level friends (i.e., other girls), while boys are more likely to choose high activity level friends (i.e. other boys) (Gleason, Gower, Hohmann & Gleason, 2005). The probability of being chosen as a friend is related to a given child’s general characteristics, but it is also related to specific interactions between a given pair. Less research has focused on the underlying social mechanisms and processes involved in preschool friendship formation. One exception is Masters and Furman’s (1981) study which demonstrated that a given child is more likely to be chosen as a “liked peer” or friend if the PRESCHOOL UNILATERAL FRIENDSHIP OUTCOMES 10 interactions between that child and the chooser are characterized by high rates of reinforcing and neutral acts. Another key exception is a study by Gottman (1983), which performed extensive analysis of conversations between unacquainted dyads during three free-play sessions over a period of 6 weeks at one of the child’s homes and marked their progression towards friendship. This analysis included, among many other variables, the amount of conversation, statements describing mutual play or partner’s play, commands and requests, and noncompliance and disapproval. Gottman theorized that children who “hit it off” were those with the ability to interact in a connected fashion, noted high numbers of statements involving the dyad’s mutual play, or commenting on their partner play. Children who hit it off were also able to exchange information successfully through commands and requests, to manage conflict (although rates of noncompliance and disapproval were high even among children who hit it off), and to establish a common-ground activity. Further, Gottman suggested that play that requires high levels of interaction has greater potential for friendship but also greater risk for conflict. Thus, children’s progression from non-acquaintanceship to friendship is characterized by the balance children must find between playing at a level with high enough social interaction to form friendship, while at the same time successfully managing conflict. While this landmark study chronicled the process of friendship formation to a great degree, examination of friendship formation in a preschool context has greater ecological validity as children acquire much of their social experiences with peers in preschool. Study of friendship formation in a social context in which interest in friendship is established, rather than one in which the children are unacquainted, may better isolate the social processes related to friendship formation from those related to initial attraction and compatibility. Unilateral friendships provide such a social context. PRESCHOOL UNILATERAL FRIENDSHIP OUTCOMES 11 Studying Friendship Formation via Unilateral Dyads While previous research investigating unilateral friendships has helped to identify the ways in which unilateral relationships are similar to and different from reciprocal and non-friend relationships, these categorical descriptions of unilateral relationships may be too broad. Unilateral friendships may be as varied as the individuals that comprise them; therefore these descriptions fail to capture the range and variety of unilateral relationships during the preschool years. The relative instability of unilateral friendships suggests dynamic social development. What remains unexplored is the potential role of unilateral friendships in preschoolers’ friendship formation. This longitudinal study seeks to determine why some preschool friendships succeed while others do not by identifying the dyadic behaviors of unilateral friendships that are related to different unilateral friendship outcomes. Unlike other studies, which have monitored the interactions of unacquainted pairs and marked their progression towards friendship (Gottman, 1983), this study has the ability to identify what behaviors are associated with specifically transitioning (or not) from initial interest to reciprocal friendship. In unilateral friendship there is interest in friendship, but not yet full-reciprocated friendship. This study has greater ecological validity than Gottman’s study because it takes place in a preschool where most early childhood peer interactions occur. Most importantly, this study differs from Gottman (1983) in that this study focuses on individual differences between dyads rather than commonalities in the friendship formation process. By examining differences in the conversation between unilateral dyads that go on to become reciprocal, become non-friend, or remain unilateral, it becomes possible to begin to ascertain why some dyads become friends while others do not. PRESCHOOL UNILATERAL FRIENDSHIP OUTCOMES 12 First, I hypothesize that the conversations of the unilateral dyads will vary widely. Second, I hypothesize that conversations between unilateral dyads observed at Time 1 will be related to unilateral friendship outcomes at Time 2. Third, I hypothesize that unilateral friendships at Time 1 that become reciprocal friendships at Time 2 will 1) generally speak more with each other 2) speak more about the mutual play they are engaging in or in the play that their partner is engaging in 3) be more noncompliant to partner requests 4) disapprove of their partners more and 5) be more likely to command their partner or request certain behaviors from their partners than those dyads that become non-friends or remain unilateral at Time 2. Methods Participants Seventy-four preschool-aged children (37 boys, 37 girls, Mage = 49.26 months, SD = 7.29 months) participated. Children were recruited from two classrooms at a laboratory preschool affiliated with a liberal arts college in New England, and three classrooms at a nearby day care facility. The majority came from a predominantly white, upper-middle class suburban area. General parental consent at the laboratory preschool was obtained upon children’s enrollment; before any sessions were run, parents had access to a summary of the study and retained the right to withdraw their child from this particular investigation. Parental consent was obtained individually for children from the day care. Procedure Sociometric interviews were conducted twice (fall and winter). The interview protocol was based on previous studies in which the friendships identified via the interview had high correlations with both observational data and teacher report (Brachfeld-Child & Schiavo, 1990; Lindsey, 2002; Rosen, Furman, & Hartup, 1988), and differed only in that children were allowed PRESCHOOL UNILATERAL FRIENDSHIP OUTCOMES 13 to make unlimited peer nominations, rather than being limited to three nominations. Unlimited nominations prevent the total number of unilateral nominations in the data set from being artificially increased. At the start of the sociometric interview, participants were asked to identify the names of all of their classmates. Next, participants were asked to select the pictures of the children they “like to play with the most.” From those nominations, children were asked to rank their nominations in order of whom they “like to play with the very most.” Next, children rated all of their classmates by sorting them into three groups: “play with a lot,” “play with sometimes,” and “play with not so much.” From the sociometric interviews conducted in the fall, unilateral pairs were identified (see Coding and Reliability below). Between the fall and winter sociometric interviews, each unilateral pair was escorted from their classroom to a separate room equipped with a pretend kitchen, a dramatic play area similar to those found in the children’s classrooms. The experimenter explained to the children that the children were going to play together in the pretend kitchen, and briefly showed the children the available toys. While the children played, the experimenter sat in a chair beside the play area and pretended to work. The experimenter did not engage in the children’s play except in the case of conflict or when asked a direct question. The children played for 10 minutes and were told when they had 1 minute left to play in order to prepare them to transition back to their classrooms. All sessions were videotaped. Coding and Reliability Friendship Pairings. Following the first round of sociometric interviews, children were sorted into unilateral pairs. Unilateral matches were first drawn from nominations, and the lowest ranked (1 stood for 1 nomination) unreciprocated nominations were used. In the case that a child had no unreciprocated nominations, or the lowest unreciprocated nomination was PRESCHOOL UNILATERAL FRIENDSHIP OUTCOMES 14 ranked higher than five, ratings were used to match unilateral dyads. In unilateral pairs composed from the ratings data, the target child was matched with a child whom s/he rated as “plays with a lot” and from whom s/he received a “plays with not so much” rating. In the case that such a pair was not possible, a pair was used in which the target child was matched with a child whom s/he rated as “plays with a lot” and from whom s/he received a “plays with sometimes” rating. In unilateral pairs made using ratings data, preference was given to same-sex dyads and those whom the target child had listed in their nominations. Some children were chosen as a unilateral-given match from multiple target children, but no child was more than three children’s unilateral-given match. If multiple choices existed, preference was given to children who were not already chosen as a unilateral-given match in an effort to limit the number of times any one child was represented in the data. In order to check inter-rater reliability, a second researcher used the rules to make pairs. Coders agreed on 94.8% of pairs and all discrepancies were resolved by discussion. Based on the second (winter) round of sociometric interviews, all unilateral pairs identified in the fall were categorized as either still unilateral, changed to reciprocal, changed to non-friend, or changed to reverse-unilateral (which member of the dyad was the unilateral-giver and which was the unilateral-receiver reversed at Time 2). Unilateral friends were identified using the same criteria as in the fall. Reciprocal friendships were identified using children’s nominations. I determined which nominations were reciprocated, and averaged the nomination the target child gave with the nomination s/he received (an average of 1 indicated that both children nominated each other first in the nominations process). All reciprocated nominations with averaged scores of less than 5 were considered reciprocal friend relationships. Non-friend relationships were defined as pairs who neither nominated each other nor had a rating PRESCHOOL UNILATERAL FRIENDSHIP OUTCOMES 15 combination of “play with a lot-play with a lot” or “play with a lot-play with sometimes.” Coders agreed on 87.0% of pair outcomes and all discrepancies were resolved by discussion. Conversational Variables. The conversations of the children during the observations were coded using the procedure described by Furman (1987). First, all conversations were transcribed from the video recordings of the observations. Next, a second transcriber checked and revised the transcript where necessary. Last, each transcript was re-checked by the original transcriber who determined which edits should be kept for conversational analysis. After transcription, all transcripts were then divided into “thought units,” or an expressed idea or fragment. This unit could be one utterance or several, and it could be either a phrase or a sentence (Furman, 1987; Gottman, 1983). Thought units were categorized into five categories: 1) mutual/partner play statements, or statements that refer to the children’s joint play or the partner’s play, 2) noncompliance, or refusals to obey the partner’s command or request, 3) disapproval, insults or other forms of disapproval of the partner’s behavior, 4) commands/requests, or statements intended to direct the partner’s or dyad’s behavior, or 5) other, statements that did not fall into any of the other four categories. For example, the statement, “The babies aren’t sleeping, but yours is awake,” contains two thought units, “The babies aren’t sleeping,” and “but yours is awake,” and contains two mutual/partner play statements. Coders used both transcripts and videos to code the conversational variables. Inter-rater reliability was obtained with the help of coders unfamiliar with the hypotheses. Two coders coded one-third of the transcripts. With the exception of noncompliance (r = 0.532), inter-rater reliability was high; Pearson-r values ranged from 0.910 to 0.994. Discrepancies were resolved using the author’s