International Journal of Behavioral Development http://jbd.sagepub.com

Reciprocity of prosocial behavior in Japanese preschool children Keiko K. Fujisawa, Nobuyuki Kutsukake and Toshikazu Hasegawa International Journal of Behavioral Development 2008; 32; 89 DOI: 10.1177/0165025407084055 The online version of this article can be found at: http://jbd.sagepub.com/cgi/content/abstract/32/2/89

Published by: http://www.sagepublications.com

On behalf of:

International Society for the Study of Behavioral Development

Additional services and information for International Journal of Behavioral Development can be found at: Email Alerts: http://jbd.sagepub.com/cgi/alerts Subscriptions: http://jbd.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations (this article cites 34 articles hosted on the SAGE Journals Online and HighWire Press platforms): http://jbd.sagepub.com/cgi/content/refs/32/2/89

Downloaded from http://jbd.sagepub.com at Graduate Uni for Advanced Stu on March 7, 2008 © 2008 International Society for the Study of Behavioral Development. All rights reserved. Not for commercial use or unauthorized distribution.

089-097 084055 Fujisawa (HO)

15/2/08

12:38

Page 89

International Journal of Behavioral Development 2008, 32 (2), 89–97 http://www.sagepublications.com

© 2008 The International Society for the Study of Behavioural Development DOI: 10.1177/0165025407084055

Reciprocity of prosocial behavior in Japanese preschool children Keiko K. Fujisawa

Nobuyuki Kutsukake and Toshikazu Hasegawa

Keio University, Japan and University of Tokyo, Japan

University of Tokyo, Japan

This study investigated the reciprocity of prosocial behavior among 3- and 4-year-old Japanese preschool children during free-play time. Matrix correlation tests revealed positive correlations between the frequencies of object offering given and received within dyads and between the frequencies of helping given and received within dyads. These results suggest that young children reciprocate prosocial behavior spontaneously. Positive correlations were also found between the frequencies of object offering and helping behavior exchanged within dyads, suggesting that children exchanged the two types of prosocial behaviors (i.e., “interchanged”). The interchange was independent of both reciprocity within object offering and reciprocity within helping behavior in 4-year-olds. Friends reciprocated object offerings more frequently than non-friends, suggesting that friendship affects the quantitative aspect of reciprocity. These data provide refined evidence of reciprocity among children and also suggest that reciprocity becomes more complicated as children grow older. Keywords: friendship; naturalistic observation; preschool children; prosocial behavior; reciprocity

Analyses of social interactions between peers provide important insights into children’s social development, as peer interactions are associated with enhanced sociocognitive abilities, emotional understanding, and social skills (Dunn, 2004; Dunn & Slomkowski, 1992; Rubin, Bukowski, & Parker, 1998). Analyses of social interactions also provide direct information about social relationships between peers because the accumulation of previous social interactions between two children determines the characteristics of their dyadic relationship (and conversely, the characteristics of social relationships are reflected in future interactions between them; Hinde, 1979). The exchange of similar prosocial actions between individuals (hereafter defined as “reciprocity;” note that this definition does not include exchanges of negative behavior such as aggression) has been regarded as having an important role in children’s social development (Hinde, 1979; Ross, Cheyne, & Lollis, 1988; Youniss, 1986). Therefore, studies of reciprocity may provide insight into social development in children for the following reasons. First, peer interaction involving reciprocity provides an optimal context for fostering the development of prosocial behavior (reviewed in Eisenberg & Fabes, 1998). This is because preschoolers have other-oriented or relational concerns when responding prosocially toward their peers, whereas they have authority- or punishment-related motives when responding to adults’ requests (Eisenberg, Lundy, Shell, & Roth, 1985). In addition, children are reinforced for being prosocial by exchanges of positive responses among peers. Preschool children respond to peers’ prosocial actions frequently in a reinforcing manner (e.g., smiles, approving of

the act, showing thanks, and sustaining interaction) compared with the responses by their teachers in classrooms (Eisenberg, Cameron, Tryon, & Dodez, 1981). Furthermore, as observed by Eisenberg et al. (1981), children who reacted positively to others’ spontaneous prosocial behaviors frequently received positive feedback for their own prosocial behaviors. Second, reciprocity plays an important role in the cultivation and regulation of social relationships. For example, Ross and Lollis (1989) reported that reciprocity of positive interactions gradually emerges among toddlers, which suggests the gradual development of social relationships. Reciprocal interaction was also observed in dyads of stable friends (Howes, 1983). Reciprocity is regarded as an important component of friendship (Rubin et al., 1998) because balanced social contributions are emphasized at all ages when children describe friendship (Bigelow & LaGaipa, 1975; Goodnow & Burns, 1985; Hartup, 1995; Laursen & Bukowski, 1997; Laursen & Hartup, 2002). Finally, studying reciprocity in children contributes to our understanding of children’s cognitive development because reciprocity is based on complex sociocognitive abilities such as detecting another’s cheating behavior, recalling past interactions, and perceiving others’ needs (Trivers, 1971). The important role of reciprocity in forming and maintaining social relations, which maximizes the coordination in social interaction, seems to favor cognitive development (Hartup, 1995). As described earlier, compared with other types of social interaction, reciprocity offers unique perspectives for understanding social development because reciprocity inevitably involves two interactants. As a result, researchers need to

The authors would like to thank R. Kohata, Y. Usui, and the teachers who participated in this study for their extensive cooperation. We also thank all the children and their parents for their participation. C. Hemelrijk kindly provided the Matrixtester program, for which we are grateful. K. Ozaki kindly provided us with statistical comments. This study was supported by Japan Society for the Promotion of Science (JSPS) Research Fellowships for Young Scientists which were awarded

to the first and the second authors and by the JSPS 21st century COE program: Center for Evolutionary Cognitive Sciences at the University of Tokyo. Correspondence should be sent to Keiko K. Fujisawa, Dept of Humanities and Social Science, Faculty of Letters, Keio University, Mita 2–15–45, Minato-ku, Tokyo 108–8345, Japan; e-mail: fujisawa@flet.keio.ac.jp

Downloaded from http://jbd.sagepub.com at Graduate Uni for Advanced Stu on March 7, 2008 © 2008 International Society for the Study of Behavioral Development. All rights reserved. Not for commercial use or unauthorized distribution.

089-097 084055 Fujisawa (HO)

90

15/2/08

12:38

Page 90

FUJISAWA ET AL. / JAPANESE PRESCHOOLERS

consider the prosocial acts of two children simultaneously, which is not always required in studies of social behavior (Hinde & Stevenson-Hinde, 1987). Many experimental studies (Birch & Billman, 1986; Dreman & Greenbaum, 1973; Levitt, Weber, Clark, & McDonnell, 1985; Staub & Sherk, 1970) have revealed the existence of “temporal reciprocity” (i.e., an immediate exchange of similar acts). Reciprocity of prosocial behavior has been confirmed in children of at least 2 years of age (Levitt et al., 1985). In addition, it has been reported that even 6-monthold infants show a prototype of reciprocity (responses to peers’ distress, Hay, Nash, & Pedersen, 1981; and responses to peers’ touch, Hay, Nash, & Pedersen, 1983), even though these results did not necessarily suggest that these very young children were capable of having specific social relations with peers (Hay et al., 1983). Ross et al. (1988) proposed that reciprocation in the longterm reflects the quality of relationships between young children. Although temporal reciprocity is typically examined in experimental studies, reciprocity examined in naturalistic observational studies (Howes, 1983, 1987; Strayer, 1980) can represent “dyadic reciprocity,” that is, balanced exchanges of similar acts between individuals in the long-term. If the dyad members have a good relationship, a child does not require an immediate return of his or her favor because of the expectation that the partner will respond in the future when needed. Temporal reciprocity may, therefore, be unrealistic in natural situations because it is likely that a child who is kind to a peer does not necessarily need reciprocation at that moment. Based on this argument, the reciprocity referred to in this article is dyadic reciprocity, not temporal reciprocity. Although reciprocity among children was studied intensively a few decades ago, more recent studies are rare. Also, previous studies on reciprocity have encountered both theoretical and statistical problems. For example, Strayer (1980) used naturalistic observation to investigate reciprocity in groups of preschool children and reported significant positive associations between giving and receiving prosocial behaviors within dyads. However, it is likely that children had particular classmates with whom they spent long (or short) periods and with whom they affiliated frequently (or less frequently). For example, friends among preschoolers are typically characterized as individuals who frequently share positive activities (Howes, 1983). Because Strayer (1980) did not control for variations in affiliation frequency with different classmates, reciprocity might have been statistically detected even if children did not reciprocate selectively to classmates from whom they had received prosocial behavior, but simply directed prosocial behavior toward one with whom they were frequently affiliated. In other words, it is possible that the reciprocity of prosocial behavior was confounded by the frequency of affiliative interaction among peers. In addition, previous studies focused on only one form of prosocial behavior (Birch & Billman, 1986; Levitt et al., 1985) and did not examine the possibility that children repaid one form of prosocial behavior with another form of prosocial behavior (hereafter, “interchange”). This possibility is highly likely because preschool-age children exhibit various kinds of prosocial behaviors and can choose behavior appropriate for their partner (Eisenberg & Fabes, 1998). Finally, a previous approach that used a simple correlation test to investigate reciprocity (Strayer, 1980) is problematic from a statistical standpoint because data points from the same children were not independent. That is, such an

approach violates the statistical assumption that each data point is independent. The problem is exaggerated when there are significant individual differences in behavioral capabilities or tendencies. In this study, we analyzed reciprocity among 3- and 4-yearold Japanese children during free-play time. In the first part of this study, we used an innovative statistical method (the matrix correlation test or Kr test; Hemelrijk, 1990a, 1990b; see Methods for details) that overcame the problems in the previous studies mentioned above. This test has advantages over simple and ordinary correlation tests. First, this test includes a random permutation procedure, which resolves the problem of data dependency within a social interaction matrix (Hemelrijk, 1990a; Manly, 1997; Sokal & Rohlf, 1995). Second, the individual differences in behavior and the variations between behaviors are taken into account in this test, making the test results robust to the effect of outliers (e.g., children who are extremely prosocial or a behavior with low occurrence compared with other behaviors). We specifically examined: (a) whether peers reciprocated prosocial behavior, (b) whether the reciprocity of prosocial behavior was independent of the frequency of affiliation in a dyad, and (c) whether one type of prosocial behavior was interchanged for a different type of prosocial behavior. Based on these analyses, we aimed to provide sophisticated evidence of reciprocity among preschoolers. In the second part of this study, we investigated the association between reciprocity and friendship. Although reciprocity is assumed to be a common characteristic of friendships (Bukowski & Sippola, 1996; Rubin et al., 1998), few studies have empirically examined differences in reciprocity between friends and non-friends in young children (see Youniss, 1986 for a study of elementary school children), and most experimental studies of prosocial reciprocity have not considered the effect of friendship itself (Dreman & Greenbaum, 1973; Leiter, 1977; Levitt et al., 1985). Previous studies have tended to classify dyadic relationships dichotomously according to whether the prosocial behavior was reciprocated (Dreman & Greenbaum, 1973; Leiter, 1977; Levitt et al., 1985). This dichotomy, however, may underestimate occurrences of reciprocity among non-friends, since it is unlikely that children are entirely non-reciprocal to their non-friends’ prosocial behavior under natural conditions. For example, observational reports have noted that reciprocity occurred even in dyads that did not necessarily meet friendship criteria (Howes, 1983). Thus, the non-friends may reciprocate to some extent, but the degree of the reciprocation may be less than the degree of reciprocation between friends. To test this possibility, a measure that enables us to examine the “quantity” of reciprocity is necessary. To use a quantitative proxy measure of dyadic reciprocity, we improved Silk, Seyfarth, and Cheney’s (1999) Reciprocity Index so that we could examine the balance in exchanges of social behaviors within dyads, accounting for the frequency of the social behavior that occurred within the dyads (see Methods). Using this measure, we investigated whether friends reciprocated more in quantity than nonfriends. This study has two broad implications for understanding social development in preschoolers. First, previous studies rarely investigated the developmental process of reciprocity in young children (Strayer, 1980; Strayer, Wareing, & Rushton, 1979; see Youniss, 1986 for developmental changes in elementary school children). We predicted that reciprocity or

Downloaded from http://jbd.sagepub.com at Graduate Uni for Advanced Stu on March 7, 2008 © 2008 International Society for the Study of Behavioral Development. All rights reserved. Not for commercial use or unauthorized distribution.

089-097 084055 Fujisawa (HO)

15/2/08

12:38

Page 91

INTERNATIONAL JOURNAL OF BEHAVIORAL DEVELOPMENT, 2008, 32 (2), 89–97

interchange would be more apparent among 4-year-olds than among 3-year-olds for the following two reasons. First, sociocognitive abilities increase as young children become older (Astington, 1993), and these abilities constitute the basis for reciprocal exchanges. Second, prosocial behavior changes both qualitatively and quantitatively as children become older (Eisenberg & Fabes, 1998), with older children engaging in a greater variety of prosocial behaviors than younger children. This change enables older children to choose a given type of prosocial behavior that is appropriate to the social context when reciprocating or interchanging. We also predicted that the effects of friendship on the degree of reciprocity would increase in 4-year-olds relative to 3-year-olds. This is because reciprocity is assumed to be a common characteristic of friendships (Bukowski & Sippola, 1996; Rubin et al., 1998), and friendships among 4-year-olds are more stable than those among 3-year-olds (Guralnick & Groom, 1988; Hymel, 1983; Uehara, 2004). The second implication relates to cultural differences. Participants of most previous studies on reciprocity of prosocial behavior have been from western cultures (Birch & Billman, 1986; Levitt et al., 1985; Staub & Sherk, 1970; Strayer, 1980; Strayer et al., 1979); understanding of prosocial behavior in eastern cultures is thus limited. In contrast to western societies in which children are expected to learn prosocial behavior incidentally, Japanese society encourages children to demonstrate prosocial behavior beginning in their early years (Stevenson, 1991). For example, although North American mothers often make their children share with peers when intervening in object conflicts (Ross, Tesla, Kenyon, & Lollis, 1990), teachers in America often do not respond in a reinforcing manner to children’s prosocial behaviors (Eisenberg et al., 1981). In contrast, Japanese teachers create various types of settings that reinforce children’s prosocial behavior (Stevenson, 1991). This study was not designed to test the cultural differences between western and eastern cultures directly; however, it may help to fill gaps in our knowledge of social development in preschoolers.

Method Participants The study was conducted at a public nursery school in Tokyo. Two classes of 3-year-olds (Class 3A: nine boys and five girls; mean age = 44.1 months, SD = 3.0; Class 3B: eight boys and eight girls; mean age = 43.4 months, SD = 3.4) and two classes of 4-year-olds (Class 4A: nine boys and three girls; mean age = 55.1 months, SD = 2.5; Class 4B: seven boys and nine girls; mean age = 54.8 months, SD = 3.3) participated in this study. Class 3A and Class 4A consisted of entirely different children, whereas Class 3B and Class 4B consisted of mostly the same children. This means that this study included both crosssectional (3A and 4A; 3A and 4B; 3B and 4A) and longitudinal (3B and 4B) designs in terms of age-difference comparisons. This was because classes marked with a B had participated in our 2-year longitudinal project. One boy had left Class 3B, and one girl had joined Class 4B when observation of the 4B group began. All of the children were Japanese and spoke Japanese as their first language.

91

Procedure Observational data were collected during morning free-play time for a school year (except for summer and winter holidays). All observations were conducted in a classroom. During the study period, one or two teachers were present during classroom observation. Twenty sessions of 5-minute focal observations of each child (i.e., 100 minutes per child in total) were videotaped. Focal observations for each child in each class were conducted in a random order in each session. It took about 2 weeks to observe all children in a class for a session. If a child to be filmed was absent on the scheduled observation day, he or she was filmed as soon as possible afterward for an even accumulation of observational data. Observations were not made when a focal child was engaging in teacher-structured activities. Children and teachers were accustomed to having their activities filmed before this study commenced. One observer, who had established a good rapport with the children and teachers before starting this study, moved around freely in the classroom and filmed all the children with a portable video camera.

Coding Frequencies of affiliative and prosocial behaviors that occurred between a focal child and peers were coded. The initiator and the recipient of the social behavior were also identified. Affiliative behavior. Six kinds of affiliative behaviors were coded: hand to body, body to body, talking, show, approach, and look (see Table 1 for descriptions). These behaviors were mostly based on Strayer’s (1980) coding but “talking” and “show” were added. Prosocial behavior. Two kinds of prosocial behavior were coded: object offering and helping (see Table 1 for descriptions). These behaviors were chosen for this study because of their relatively frequent occurrence during the preschool period (Eisenberg & Fabes, 1998). We did not code prosocial behaviors that appeared to benefit the initiator as well as the recipient, such as behaviors that occurred as part of a play ritual or as cooperation. Also, we did not code the behaviors if their initiator acted forcefully or aggressively toward the recipient or if these behaviors were not welcomed by the recipient (e.g., an initiator helped to place a puzzle piece, but the recipient wanted to complete the puzzle by him- or herself without help).

Reliability After extensive training, the first author and one research assistant independently coded 75 randomly selected sessions (i.e., 75  5 minutes) to assess coding reliabilities. Kappa coefficients were as follows: hand to body, .71; body to body, .70; talking, .81; show, .75; approach, .78; look, .73; object offering, .84; helping, .75. We calculated the sum of the frequencies of all six affiliative behaviors that occurred in dyads and used this sum in further analyses because internal consistency was moderately good (Cronbach’s alpha = .75 [3A], .73 [3B], .79 [4A], .72 [4B]).

Downloaded from http://jbd.sagepub.com at Graduate Uni for Advanced Stu on March 7, 2008 © 2008 International Society for the Study of Behavioral Development. All rights reserved. Not for commercial use or unauthorized distribution.

089-097 084055 Fujisawa (HO)

92

15/2/08

12:38

Page 92

FUJISAWA ET AL. / JAPANESE PRESCHOOLERS

Table 1 Descriptions of coded affiliative and prosocial behaviors Affiliative behaviors Hand to body A child spontaneously and gently touched his/her partner’s body, including the partner’s hands, and then released the partner. Body to body A child spontaneously and gently clung to his/her partner’s body and then let go of the partner’s body. Talking A child spontaneously talked to his/her partner, excluding verbal aggression (e.g., insults, derogatory comments). An instance of a child’s talking was considered to end when the child stopped talking and started doing something else. Show A child spontaneously showed an object or his/her actions (e.g., a toy, a book, or his/her own dancing) to his/her partner. Approach A child spontaneously approached within 1 m of a partner, excluding cases in which the child accidentally approached his/her partner. Look A child looked at his/her partner, excluding staring in an unfriendly or angry manner. Prosocial behaviors Object offering A child spontaneously gave objects (e.g., a toy) to his/her partner. Helping A child spontaneously assisted his/her partner to accomplish the partner’s goal (e.g., putting on a smock, doing up buttons, holding a ramp for another child during a racing car game, or pulling the partner in a wagon while playing a firefighter game).

Friendship The head teacher of each class was asked to nominate each child’s friend(s) because sociometric measures are generally less reliable with children younger than 4 years (Hymel, 1983). The teacher could nominate multiple numbers of friends and was allowed to nominate no child when the teacher believed that a given child did not have any friends. We defined two children as a friend pair only when each was nominated as being a friend to the other by the teacher. That is, we did not define a unidirectional dyad, in which the teacher nominated child X as a friend of child Y but did not nominate child Y as a friend of child X, as a friend pair. Based on this definition, a child could be a member of more than one friend pair (range of the number of friends identified per child: 0–6 [3A]; 2–7 [3B]; 1–5 [4A]; 1–6 [4B]). Nineteen, 24, 20, and 21 pairs were reciprocally nominated in Classes 3A, 3B, 4A, and 4B, respectively (16 boy–boy pairs and 3 girl–girl pairs [3A]; 10 boy–boy pairs, 9 girl–girl pairs, and 5 boy–girl pairs [3B]; 17 boy–boy pairs and 3 girl–girl pairs [4A]; and 3 boy–boy pairs, 11 girl–girl pairs, and 7 boy–girl pairs [4B]).

Analysis Descriptive statistics. To investigate the age differences in frequencies of affiliative and prosocial behaviors that occurred

within dyads, we used a generalized linear mixed model (GLMM; Schall, 1991). Mixed models allow both fixed and random terms to be fitted to the model. Fixed terms examine its effects on the dependent variable. Random terms are used to represent subject-specific (i.e., a dyad in the present study) random variation and take into consideration repeated sampling within the same individuals or groups (Crawley, 2002; Schall, 1991). In the model, we included frequencies of affiliative or prosocial behaviors that occurred within dyads as a dependent variable with a Poisson error structure, the age of dyads as an independent variable, and dyads themselves as a random term. By fitting the random term to the model, we could resolve the problems of data dependency caused by the longitudinal sample of 3B and 4B (i.e., the same dyads were observed in 3B and in 4B) and the fact that the same child could be a member of more than one dyad in each class. Reciprocity and interchange. We used the matrix correlation test (hereafter, Kr test) to examine reciprocity and interchange. The Kr test enabled us to analyze whether individuals direct a given type of behavior relatively more (or less) toward social partners according to the extent that they receive more (or less) of that behavior from the partners. The Kr test was developed from Kendall’s tau approach and was a special version of Mantel’s randomization test (Hemelrijk, 1990a). We used Hemelrijk’s (1990a) permutation computer program (MATSQUAR software) for the Kr test. First, we constructed two actor–receiver matrices of social interactions. The values of cells in the matrices represent the frequencies of behaviors from an actor to a receiver. All cells in one matrix corresponded to cells in the other matrix. The Kr test statistic (hereafter, Tau Kr) is calculated from within-row comparisons between corresponding cells in two matrices (Hemelrijk, 1990a, 1990b). This statistic considers whether the change between the two cells within one row in one matrix is in the same direction as the change between the corresponding cells in the other matrix. Tau Kr is the sum of the number of pairs of cells whose change is in the same direction in two matrices minus the number of pairs of cells whose change is in the opposite direction in two matrices, while taking into account the pairs of cells in which values neither increased nor decreased (see Hemelrijk, 1990a, 1990b for details and example computation). We calculated the Tau Kr between the two matrices of the original (i.e., observed) data. Next, MATSQUAR randomly permutated one of the two matrices, while keeping the other matrix unchanged. The order of the individuals in the rows was kept identical to that in the columns to preserve the diagonal. We then calculated Tau Kr from these two matrices. This permutation procedure was repeated to produce the distribution of Tau Kr, which enabled reliable judgment of the p-value of the Tau Kr from the original data. In this article, we conducted 5000 permutations, which is a realistic minimum for estimating a significance level of about .01 (Manly, 1997). Finally, the one-tailed p-value of Tau Kr from the original data was calculated by computing the percentage of all permutations that yielded a Tau Kr equal to or greater than the Tau Kr calculated from the original data. Based on this percentage, we assessed the significance of the results; if the percentage of the Tau Kr from the original data fell within the rightmost 5% of the distribution, we regarded the association between the two matrices as statistically significant. In Kr tests, only the right-sided one-tailed probability value was used; the left-sided p-value was meaningless because

Downloaded from http://jbd.sagepub.com at Graduate Uni for Advanced Stu on March 7, 2008 © 2008 International Society for the Study of Behavioral Development. All rights reserved. Not for commercial use or unauthorized distribution.

089-097 084055 Fujisawa (HO)

15/2/08

12:38

Page 93

INTERNATIONAL JOURNAL OF BEHAVIORAL DEVELOPMENT, 2008, 32 (2), 89–97

it indicated the opposite of reciprocity/interchange (Hemelrijk, 1990a). The Kr test has many advantages beyond intracorrelation tests. First, the Kr test is robust to data in which there are considerable individual differences in the frequencies of a behavior or a considerable variation in frequencies between behaviors (Hemelrijk, 1990a). This is because the Kr test consists of within-row (i.e., within-individual) comparisons that consider only the direction of the changes between two cells in one row in one matrix and the direction change between its corresponding cells in the other matrix. Because of this advantage, the Kr test is also robust to a few extreme values (i.e., outliers), which may have strongly biased the results of other tests. The Kr test procedure (i.e., randomly permuting one of the two matrices a considerable number of times and computing the Tau Kr statistic each time) could resolve the problem of non-independence of dyads that might arise when analyzing dyadic interactions within a social interaction matrix (Hemelrijk, 1990a; Manly, 1997; Sokal & Rohlf, 1995). In addition, the random permutation procedure in the Kr test could resolve the problems of unknown sampling distributions of dyadic values because a sufficiently large number of random permutations can generate a distribution of Tau Kr with which the significance of the Tau Kr from the original observed data can be assessed (Hemelrijk, 1990a; Manly, 1997). Using the Kr test, we first examined whether children reciprocated and interchanged prosocial behavior. We examined: (a) the correlation between object offering from child X to child Y and object offering from child Y to child X; (b) the correlation between helping from child X to child Y and helping from child Y to child X; (c) the correlation between object offering from child X to child Y and helping from child Y to child X; and (d) the correlation between helping from child X to child Y and object offering from child Y to child X. MATSQUAR enabled us to examine relationships between two matrices after controlling for the effect of the third matrix (Hemelrijk, 1990b). Using this, we examined whether children reciprocate the partner’s prosocial behavior regardless of the frequency of affiliation with the partner. We examined: (a) the reciprocity of object offering, partialling out the frequency of the affiliation in a dyad; and (b) the reciprocity of helping, partialling out the frequency of the affiliation in a dyad. If we found a significantly positive association between the frequencies of the two types of prosocial behavior exchanged within a dyad, this would not necessarily mean that the children interchanged the different types of prosocial behavior because it is not clear whether the received prosocial behavior was reciprocation for the same behavior previously performed by the recipient (i.e., reciprocity) or reciprocation for a different type of prosocial behavior previously performed by the recipient (i.e., interchange), unless the frequency of the same prosocial behavior by the recipient was controlled for. For example, child X and child Y could have offered each other objects with similar frequency, and child X could have helped child Y with a predominant frequency. In such a case, the matrix correlation test might detect a positive relationship between the frequency of helping from child X to child Y and the frequency of object offering from child Y to child X. However, this correlation would likely be a by-product of the reciprocity of the object offering between child X and child Y. Thus, to investigate the interchanges between different types of prosocial behavior, it was necessary to partial out the frequency of one type of prosocial behavior and see whether

93

interchanges were independent from reciprocity between each dyad. To do so, we examined: (a) the interchange between object offering from child X to child Y and helping from child Y to child X, controlling for helping from child X to child Y; and (b) the interchange between helping from child X to child Y and object offering from child Y to child X, controlling for object offering from child X to child Y. To test whether the interchanges were independent of the affiliation frequency of dyads, we ran separate analyses in which the frequency of affiliation was partialled out by placing it within a third matrix when investigating the correlations between object offering and helping. Reciprocity between friends versus non-friends. The Kr test evaluates the existence of reciprocity at classroom level but it does not investigate the reciprocity at dyadic level (Silk et al., 1999). To quantify the degree of reciprocity within dyads, we used an improved version of Silk et al.’s (1999) Reciprocity Index. Consider a dyad of two individuals in which one initiated a behavior s times and the other d times (s > d). In total, the behavior within the dyad occurred N times (s + d = N). We refer to the cumulative binomial probability of s in N events as S and that for d in N events as D. Silk et al.’s (1999) Reciprocity Index was calculated as D divided by S. The reciprocity index approached 0 when interactions within a dyad were due primarily to one of the two individuals; it approached 1 when interactions were due evenly to the two participants. A weakness of Silk et al.’s (1999) Reciprocity Index is that it may fraudulently detect a high level of dyadic reciprocity when examining dyads in which very few or no social exchanges occurred. Therefore, we modified the Reciprocity Index of Silk et al. (1999) as follows: (0.5/S – 0.5)  2. This formula is based on the fact that the cumulative binomial probability of one individual in a dyad is 0.75 if two children exchange a prosocial behavior with each other, one time. The cumulative binomial probability gradually approaches 0.5 if two children exchange prosocial behaviors with each other exactly the same number of times which is close to an infinite number. Therefore, we could incorporate the frequency of prosocial behavior by dividing 0.5 by S. In addition, we subtracted 0.5 and multiplied by 2, so that this value ranges from 0 (unbalanced exchanges) to 1 (complete reciprocation). The value for a dyad in which balanced exchanges occurred is higher than for a dyad in which less balanced exchanges occurred. This value has an advantage over the Silk et al. (1999) Reciprocity Index because it reflects the degree of reciprocity taking account of the total frequency of a behavior that occurred in a dyad. For example, the value is 0.33 when two individuals in a dyad did a behavior to each other one time (i.e., the behavior occurred two times in total); it is 0.52 when two individuals in a dyad did a behavior to each other three times (i.e., the behavior occurred six times in total). This indicates that the latter dyad was more reciprocal than the former dyad. Hereafter, we call this value the Modified Reciprocity Index (MRI). We excluded MRIs of dyads in which offering/helping occurred one or less times from the analyses because it is meaningless to examine the degree of reciprocity in those dyads. Means and standard deviations of MRIs of offering and of helping per dyad in each class were calculated (Table 2). We used the GLMM to investigate the age differences in MRIs. As in testing the age differences in frequencies of affiliative or prosocial behaviors, we included MRIs of dyads as a

Downloaded from http://jbd.sagepub.com at Graduate Uni for Advanced Stu on March 7, 2008 © 2008 International Society for the Study of Behavioral Development. All rights reserved. Not for commercial use or unauthorized distribution.

089-097 084055 Fujisawa (HO)

15/2/08

12:38

94

Page 94

FUJISAWA ET AL. / JAPANESE PRESCHOOLERS

dependent variable with a binomial error structure, age of dyads as an independent variable, and dyads themselves as a random term. We examined the age difference in MRIs of offering and of helping in separate analyses. To examine the hypothesis that preschool children adjust the quantity of reciprocation toward a partner’s object offering or helping behavior according to their friendship, we used a GLMM that included MRIs of dyads as a dependent variable with a binomial error structure, friendship status (friend or non-friend) of dyads as an independent variable, and dyads themselves as a random term. Children could have had more than one friend, creating non-independency in the MRI data. However, as stated earlier, the problems of pseudoreplication and data dependency were resolved by fitting the dyads themselves as a random term in the GLMM.

Results Descriptive data Table 2 presents the means and standard deviations of affiliative or prosocial behaviors within dyads by class (the number of the possible dyads in a class: 3A, n = 91; 3B, n = 120; 4A, n = 66; 4B, n = 120). GLMM showed no age differences in the frequencies of affiliative or prosocial behaviors occurring in dyads (Table 3).

Reciprocity and interchange Table 4 presents Kr test statistics for the reciprocity of object offering and reciprocity of helping in all classes. The relative amounts of object offering were significantly correlated with the relative amounts of object offering received in all classes. Similarly, the relative amounts of helping were significantly correlated with the relative amounts of helping received in all classes, apart from Class 3A, for which the result approached significance. Moreover, the reciprocity of object offering for all classes was confirmed when the number of affiliative acts was partialled out, which suggests that reciprocity of object offering

Table 3 Results of generalized linear mixed modeling for age comparisons of the frequencies of affiliative or prosocial behaviors within dyads, and Modified Reciprocity Indexes Beta

SE

df

t

–0.02

0.08

395

–0.30+

Prosocial behaviors Object offering Age (3 years vs. 4 years) Helping Age (3 years vs. 4 years)

–0.04

0.13

395

–0.33+

0.23

0.22

395

1.03+

Modified Reciprocity Index Object offering Age (3 years vs. 4 years) Helping Age (3 years vs. 4 years)

0.11

0.20

196

0.52+

0.50

0.27

87

1.84+

Affiliative behaviors Age (3 years vs. 4 years)

Note. Separate analyses were conducted. +p < .1; **p < .01; ***p < .001.

was independent of the affiliative interactions in a dyad. Reciprocity of helping, with the frequency of the affiliative interaction in a dyad partialled out, was confirmed in Class 3B and Class 4B. This trend in Class 4A class approached significance, but was not significant in Class 3A (Table 4). Kr tests demonstrated the interchange between object offering and helping in all classes (Table 4). Interchange between object offering and helping was independent of the reciprocity of each prosocial behavior in 4-year-olds but not in 3-year-olds. The correlation between the relative amounts of

Table 4 Results of Kr tests: reciprocity of object offering, reciprocity of helping, and those with controlling affiliation frequency in a dyad; interchange between object offering and helping, interchange with controlling initiations of each prosocial behavior, and interchange with controlling affiliation frequency in a dyad.The Kr test statistics (Tau Kr) were reported

Table 2 Mean and standard deviations for the frequencies of total affiliative or prosocial behaviors within dyads and Modified Reciprocity Indexes 3A Mean (SD)

3B Mean (SD)

4A Mean (SD)

4B Mean (SD)

50.55 (43.99)

49.67 (35.18)

53.76 (48.38)

57.31 (63.07)

2.03 (3.40) 0.73 (1.19)

3.16 (3.97) 1.13 (2.07)

2.55 (3.16) 0.73 (1.38)

2.61 (3.82) 1.54 (3.28)

Modified Reciprocity Index Object offering 0.12 (0.16) Helping 0.05 (0.09)

0.12 (0.16) 0.08 (0.12)

0.14 (0.14) 0.12 (0.20)

0.13 (0.17) 0.11 (0.15)

Affiliative behaviors Prosocial behaviors Object offering Helping

Note. Standard deviations are in parentheses.

Reciprocity Object offering Affiliation controlled Helping Affiliation controlled Interchangea Object offering/helping received Initiation of helping controlled Affiliation frequency controlled Helping/object offering received Initiation of object offering controlled Affiliation frequency controlled

3A

3B

4A

4B

0.25*** 0.16* 0.12+ 0.03

0.24*** 0.14** 0.21*** 0.16**

0.42*** 0.25** 0.26*** 0.16+

0.43*** 0.26*** 0.28*** 0.15*

0.15*

0.14*

0.23**

0.29***

0.12+

0.08

0.15*

0.21***

0.05

0.06

0.08

0.12*

0.16*

0.16**

0.28**

0.27**

0.06

0.09+

0.15*

0.11*

0.07

0.09

0.16*

0.12*

Note. +p < .1; *p < .05; **p < .01; ***p < .001. a Initiation of each prosocial behavior and the affiliation frequency were controlled in separate analyses.

Downloaded from http://jbd.sagepub.com at Graduate Uni for Advanced Stu on March 7, 2008 © 2008 International Society for the Study of Behavioral Development. All rights reserved. Not for commercial use or unauthorized distribution.

089-097 084055 Fujisawa (HO)

15/2/08

12:38

Page 95

INTERNATIONAL JOURNAL OF BEHAVIORAL DEVELOPMENT, 2008, 32 (2), 89–97

object offering and the relative amounts of helping received was significant when the relative amounts of helping were partialled out in classes of 4-year-olds. This correlation approached significance in Class 3A, but not in 3B. The correlation between the relative amounts of helping and the relative amounts of object offering received was significant when the relative amounts of object offering were partialled out in classes of 4-year-olds. This correlation approached significance in Class 3B, but not in 3A. When affiliation frequency in a dyad was partialled out, there was a significant correlation between object offering and helping received in one class of 4-year-olds (4B). Interchange between helping and object offering received, with affiliation frequency in a dyad partialled out, was confirmed in two classes of 4-year-olds but was not in the 3-year-olds (Table 4).

Reciprocity between friends versus non-friends Table 2 presents means and standard deviations of the MRIs of dyads for offering and helping by class. The GLMM showed that MRIs of object offering did not differ according to age, whereas the MRI of helping in 4-year-olds tended to be higher than for 3-year-olds (Table 3). The GLMM also indicated that the MRI of offering was significantly higher for friends, mean (SD) = 0.19 (0.18), than for non-friends, mean (SD) = 0.10 (0.14). However, the MRI of helping was not different between friends, mean (SD) = 0.11 (0.14) and non-friends, mean (SD) = 0.08 (0.14) (Table 5).

Discussion In this study, we conducted naturalistic observations in sameage peer groups of 3- and 4-year-olds and found positive correlations between the frequencies of helping given and received and the frequencies of object-offering given and received within dyads. This suggests that Japanese preschool children reciprocate prosocial behavior. This result is consistent with previous experimental studies reporting that children who received more prosocial behavior from a partner in previous sessions directed more prosocial behavior toward that partner in subsequent sessions, while children who received less from a partner in previous sessions directed less to their partner in ongoing sessions (Birch & Billman, 1986; Peterson, Hartmann, & Gelfand, 1977; Staub & Sherk, 1970). Our results reinforce the findings from observational studies that used simple correlation tests (Strayer, 1980). Because the Kr test has several important advantages over ordinary correlation

Table 5 Results of generalized linear mixed modeling for friend–nonfriend comparisons of the Modified Reciprocity Indexes Beta

SE

df

t

Object offering Friend versus non-friend

–0.51

0.20

196

–2.55*

Helping Friend versus non-friend

–0.28

0.36

87

–0.77*

Note. Separate analyses were conducted. *p < .05; ***p < .001.

95

tests when using dependent data with unknown sampling distributions and data with variations in individual behavioral frequency (Hemelrijk, 1990a), our study also provides more precise evidence of reciprocity among preschoolers. The social relations model (Kenny, 1994; Kenny & La Voie, 1984), another statistical method that has been used to examine dyadic reciprocity (Card, Hodges, Little, & Hawley, 2005; Cook, 2001; Ross & Lollis, 1989; Ross, Stein, Trabasso, Woody, & Ross, 2005; Stevenson, Leavitt, Thompson, & Roach, 1988), also considers the data dependency of social interaction within dyads and individual differences in behavior. The Kr test has a compensatory function to the social relations model as the Kr test allows examination of the association between two matrices controlling for the effect of the third matrix (e.g., the frequency of affiliation within dyads; see Hemelrijk, 1990b). It is possible that reciprocity of prosocial behavior was affected by affiliative interaction among peers, a concern that was not addressed in previous studies. In our study, we found that reciprocity of object offering was confirmed in all classes, even after controlling for the frequency of affiliative interaction. Reciprocity of helping was also confirmed in all classes when controlling for affiliation, except in one class of 3-yearolds. These results suggest that preschool-age children do not simply direct prosocial behavior toward those with whom they are frequently affiliated, but that they reciprocate selectively to peers from whom they receive prosocial behavior. It is known that infants and toddlers show a prototype of reciprocity. For example, even 6-month-old infants reciprocate physical contact (Hay et al., 1983), while 2-year-olds show reciprocity of contributions to positive interactive sequences (Ross & Lollis, 1989). These findings suggest that even very young children may socially influence their peers. However, the age around which young children begin to show spontaneous reciprocity of prosocial behaviors has remained unknown. For example, it has been reported that 2-year-old children showed reciprocity of offering when their mothers requested that they do so (Levitt et al., 1985). However, it is unknown whether 2-year-olds show this reciprocity without their mothers’ encouragement. Naturalistic observations of mixed-age groups of 3-year-olds have revealed reciprocity in 3- and 4-year-olds (Strayer, 1980; Strayer et al., 1979). However, it was unclear whether the 3-year-olds spontaneously reciprocated in those studies because elder children might have encouraged them to do so. Because this study focused on natural situations within same-age classes, it can now be said conclusively that 3-year-old children show spontaneous reciprocity of prosocial behavior. This study demonstrates that children interchange different types of prosocial behavior; the Kr test also revealed a significant correlation between object offering and the receipt of helping and a correlation between helping and the receipt of object offering for all classes. This suggests that preschool children reciprocated their partner’s prosocial behavior with not only a similar action, but also by using a different form of prosocial act. Based on the fact that: (a) the correlation between object offering from child X to child Y and helping from child Y to child X was significant after controlling for helping from child X to child Y, and (b) the correlation between helping from child X to child Y and object offering from child Y to child X was also significant when object offering from child X to child Y was controlled in 4-year-olds, we can say that the interchange between offering and helping was independent of the reciprocity within object offering and

Downloaded from http://jbd.sagepub.com at Graduate Uni for Advanced Stu on March 7, 2008 © 2008 International Society for the Study of Behavioral Development. All rights reserved. Not for commercial use or unauthorized distribution.

089-097 084055 Fujisawa (HO)

96

15/2/08

12:38

Page 96

FUJISAWA ET AL. / JAPANESE PRESCHOOLERS

the reciprocity within helping in 4-year-olds. The interchanges were not confirmed in 3-year-olds after controlling for other types of prosocial behavior. This suggests that the recipients reciprocated a partner’s prosocial behavior regardless of the type of prosocial behavior in 4-year-olds. The original recipient of prosocial behavior may not need to return the favor in the same way as it was received. Rather, they may return partners’ favors using different forms of prosocial behavior that might be more appropriate for their partner in a different situation (Eisenberg & Fabes, 1998). Gouldner (1960) argued that individuals do not necessarily reciprocate the favor of the social partner with a favor of the same form, but may reciprocate with a favor of a different form that has similar “value” to the received favor. Likewise, 4-year-olds may interchange the partner’s former favor with a different form of prosocial behavior regarded as having a similar value. These results showed more complex types of social exchange occurring than were previously known because past studies simplified social exchange among children by pooling all types of prosocial behavior (Strayer, 1980; Strayer et al., 1979). Analyses of reciprocity are fundamentally different from analyses of individual differences in social behavior or on differential allocation of social behavior according to the identity of social partners. This is because the former analyses consider the balance between the frequencies of social behavior exchanged in a dyad, whereas later analyses do not. Therefore, studies of reciprocity have the potential to provide novel knowledge about social relationships in children. For example, analyses of reciprocity may enable us to distinguish friendships with mutual preference from those with unilateral preference (Rubin et al., 1998). The present study quantified the degree of reciprocation using the MRI based on Silk et al.’s (1999) Reciprocity Index and found that the MRI for object offering was higher for dyads of friends than for dyads of nonfriends. This result means that preschool children reciprocate with their friends in a more balanced way than with nonfriends, suggesting that preschool children adjust the degree of reciprocity according to friendship.This may be one of the ways in which preschoolers regulate or maintain their relationships (Hartup, 1995). Despite awareness that reciprocity is central within children’s friendships (Bukowski & Sippola, 1996; Rubin et al., 1998), few studies have investigated the link empirically (Dreman & Greenbaum, 1973; Leiter, 1977; Levitt et al., 1985). Our findings empirically support the link between reciprocity and friendship. In addition, our findings emphasize the importance of the quantitative analysis of reciprocity, such as by using the MRI, although previous studies treated reciprocity dichotomously (i.e., the occurrence of reciprocity vs. nonoccurrence of reciprocity), which could oversimplify exchanges of prosocial behavior (Dreman & Greenbaum, 1973; Howes, 1983; Leiter, 1977; Levitt et al., 1985). We did not find a difference between friends and non-friends in the MRI for helping. This might be due to the fact that children may sometimes help peers regardless of their friendship (Berndt, 1981). Because children are particularly likely to assist people who will clearly benefit from their assistance (Eisenberg & Fabes, 1998), a difference between friends and non-friends in terms of helping might not be observed. The present study examined the developmental process of reciprocity during the preschool period, an area that had not been thoroughly examined in previous studies (see Youniss, 1986, for the developmental process during the elementary school period). The 3- and 4-year-olds differed in terms of

interchanges that were independent of reciprocity of the same behavior. While 4-year-olds interchanged between object offering and helping independent of the reciprocity of each prosocial behavior, this was not apparent in 3-year-olds. This suggests that reciprocity becomes more complicated as children grow older. In addition to the extent of accumulated social experiences, it is possible that the age-related differences in interchange between 3-and 4-year-olds run parallel with the rapid development of sociocognitive abilities occurring around this age period (Astington, 1993; Wellman, Cross, & Watson, 2001). That is, interchanges that are independent of reciprocity would require reasoning to judge whether a child’s favor to his or her partner has similar “value” to the favor received from the partner (Gouldner, 1960); they would also require the ability to choose an appropriate behavior from a variety of prosocial behaviors to satisfy the partner’s need and the social context. We had predicted that the effects of friendship on the degree of reciprocity would increase in 4-year-olds relative to 3-yearolds. However, this prediction was unsupported because neither the MRI of offering nor that of helping was significantly affected by age, and friendships affected only the MRI of offering. This suggests that the adjustment for the degree of reciprocation according to friendships did not differ between 3-year-olds and 4-year-olds, although studies on relationship stability among preschoolers have reported that children over age 4 show stable relationships, but children under this age fail to show stability (Guralnick & Groom, 1988; Hymel, 1983; Uehara, 2004). Perhaps the interval between age 3 and age 4 is too small to detect differential reciprocation based on friendships according to age. Furthermore, although there have been many studies of reciprocity in western cultures (Birch & Billman, 1986; Levitt et al., 1985; Staub & Sherk, 1970; Strayer, 1980; Strayer et al., 1979), few studies have been conducted in eastern cultures. Although comparing western and eastern reciprocity results is beyond the scope of this study, our analysis may help to fill the gap and provide an opportunity to discuss existing cultural differences. For example, the results of this study may not be universally found in western cultures. It is known that prosocial behavior of children is associated with socializing experiences (Eisenberg & Fabes, 1998; Hay, 1994). Compared with western cultures, Japanese culture is characterized by extensive efforts to enhance the development of children’s prosocial behavior (Stevenson, 1991). As a result, Japanese children are given many more opportunities to learn prosocial behaviors than children in western cultures. For example, the presence of interchanges, which were not investigated in previous studies from western cultures, may be attributable to cultural differences. To examine whether there are any cultural influences on prosocial behavior, it would be interesting to investigate whether similar results would be obtained in western societies, by using identical methodology and the new statistical tests.

References Astington, J.W. (1993). The child’s discovery of the mind. Cambridge, MA: Harvard University Press. Berndt, T.J. (1981). Effects of friendship on prosocial intentions and behavior. Child Development, 52, 636–643. Bigelow, B.J., & LaGaipa, J.J. (1975). Children’s written descriptions of friendship: A multidimensional analysis. Developmental Psychology, 11, 857–858.

Downloaded from http://jbd.sagepub.com at Graduate Uni for Advanced Stu on March 7, 2008 © 2008 International Society for the Study of Behavioral Development. All rights reserved. Not for commercial use or unauthorized distribution.

089-097 084055 Fujisawa (HO)

15/2/08

12:38

Page 97

INTERNATIONAL JOURNAL OF BEHAVIORAL DEVELOPMENT, 2008, 32 (2), 89–97 Birch, L.L., & Billman, J. (1986). Preschool children’s food sharing with friends and acquaintances. Child Development, 57, 387–395. Bukowski, W.M., & Sippola, L.K. (1996). Friendship and morality: (How) are they related? In W.M. Bukowski, A.F. Newcomb, & W.W. Hartup (Eds.), The company they keep. Friendship in childhood and adolescence (pp. 238–261). Cambridge: Cambridge University Press. Card, N.A., Hodges, E.V.E., Little, T.D., & Hawley, P.H. (2005). Gender effects in peer nominations for aggression and social status. International Journal of Behavioral Development, 29, 146–155. Cook, W.L. (2001). Interpersonal influence in family systems: A social relations model analysis. Child Development, 72, 1179–1197. Crawley, M.J. (2002). Statistical computing: An introduction to data analysis using S-Plus. Chichester: Wiley. Dreman, S.B., & Greenbaum, C.W. (1973). Altruism or reciprocity: Sharing behavior in Israeli kindergarten children. Child Development, 44, 61–68. Dunn, J. (2004). Children’s friendships. The beginning of intimacy. Oxford: Blackwell. Dunn, J., & Slomkowski, C. (1992). Conflict and the development of social understanding. In C.U. Shantz & W.W. Hartup (Eds.), Conflict in child and adolescent development (pp. 70–92). Cambridge: Cambridge University Press. Eisenberg, N., Cameron, E., Tryon, K., & Dodez, R. (1981). Socialization of prosocial behavior in the preschool classroom. Developmental Psychology, 17, 773–782. Eisenberg, N., & Fabes, R.A. (1998). Prosocial development. In W. Darmon & N. Eisenberg (Vol. Eds.), Handbook of child psychology:Vol. 3. Social, emotional, and personality development (5th ed., pp. 701–778). New York: Wiley. Eisenberg, N., Lundy, T., Shell, R., & Roth, K. (1985). Children’s justifications for their adult and peer-directed compliant (prosocial and nonprosocial) behaviors. Developmental Psychology, 21, 325–331. Goodnow, J.J., & Burns, A. (1985). A child’s eye. Sydney: Alwin. Gouldner, A.W. (1960). The norm of reciprocity: A preliminary statement. American Sociological Review, 25, 161–178. Guralnick, M.J., & Groom, J.M. (1988). Friendships of preschool children in mainstreamed playgroups. Developmental Psychology, 24, 595–604. Hartup, W.W. (1995). Cooperation, close relationships, and cognitive development. In W.M. Bukowski, A.F. Newcomb, & W.W. Hartup (Eds.), The company they keep. Friendship in childhood and adolescence (pp. 213–237). Cambridge: Cambridge University Press. Hay, D.F. (1994). Prosocial development. Journal of Child Psychology and Psychiatry, 35, 29–71. Hay, D., Nash, A., & Pederson, J. (1981). Responses of six-month-olds to the distress of their peers. Child Development, 52, 1071–1075. Hay, D., Nash, A., & Pederson, J. (1983). Interaction between six-month-old peers. Child Development, 54, 557–562. Hemelrijk, C. (1990a). Models of, and tests for, reciprocity, unidirectionality, and other social interaction patterns at group level. Animal Behaviour, 39, 1023–1029. Hemelrijk, C. (1990b). Matrix partial correlation test used in investigations of reciprocity and other social interaction patterns at a group level. Journal of Theoretical Biology, 143, 405–420. Hinde, R.A. (1979). Towards understanding relationships. London: Academic Press. Hinde, R.A., & Stevenson-Hinde, J. (1987). Interpersonal relationships and child development. Developmental Review, 7, 1–21. Howes, C. (1983). Patterns of friendship. Child Development, 54, 1041–1053. Howes, C. (1987). Peer interaction of young children. Monographs of the Society for Research in Child Development, 53 (1, Serial No. 217). Hymel, S. (1983). Preschool children’s peer relations: Issues in sociometric assessment. Merrill-Palmer Quarterly, 29, 237–260. Kenny, D.A. (1994). Interpersonal perception: A social relations analysis. New York: Guilford Press. Kenny, D.A., & La Voie, L.J. (1984). The social relations model. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 18, pp. 142–182). San Diego, CA: Academic Press.

97

Laursen, B., & Bukowski, W.M. (1997). A developmental guide to the organisation of close relationships. International Journal of Behavioral Development, 21, 747–770. Laursen, B., & Hartup, W.W. (2002). The origins of reciprocity and social exchange in friendships. New Directions for Child and Adolescent Development, 95, 27–40. Leiter, M.P. (1977). A study of reciprocity in preschool play groups. Child Development, 48, 1288–1295. Levitt, M.J., Weber, R.A., Clark, M.C., & McDonnell, P. (1985). Reciprocity of exchange in toddler sharing behavior. Developmental Psychology, 21, 122–123. Manly, B.F. (1997). Randomization, bootstrap and Monte Carlo methods in biology. Texts in statistical science (2nd ed.). London: Chapman & Hall. Peterson, L., Hartmann, D.P., & Gelfand, D.M. (1977). Developmental changes in the effects of dependency and reciprocity cues on children’s moral judgments and donation rates. Child Development, 48, 1331–1339. Ross, H.S., Cheyne, A., & Lollis, S.P. (1988). Defining and studying reciprocity in young children. In S.W. Duck (Ed.), Handbook of personal relationships (pp. 143–160). Chichester: Wiley. Ross, H., & Lollis, S. (1989). A social relations analysis of toddler peer relationships. Child Development, 60, 1082–1091. Ross, H., Stein, N., Trabasso, T., Woody, E., & Ross, M. (2005). The quality of family relationships within and across generations: A social relations analysis. International Journal of Behavioral Development, 29, 110–119. Ross, H., Tesla, C., Kenyon, B., & Lollis, S. (1990). Maternal intervention in toddler peer conflict: The socialization of principles of justice. Developmental Psychology, 26, 994–1003. Rubin, K.H., Bukowski, W., & Parker, J.G. (1998). Peer interactions, relationships, and groups. In W. Darmon & N. Eisenberg (Vol. Eds.), Handbook of child psychology:Vol. 3. Social, emotional, and personality development (5th ed., pp. 619–700). New York: Wiley. Schall, R. (1991) Estimation in generalized linear models with random effects. Biometrika, 78, 719–727. Silk, J.B., Seyfarth, R.M., & Cheney, D.L. (1999). The structure of social relationships among female savanna baboons in Moremi Reserve, Botswana. Behaviour, 136, 679–703. Sokal, R.R., & Rohlf, F.J. (1995). Biometry:The principles and practice of statistics in biological research (3rd ed.). New York: Freeman. Staub, E., & Sherk, L. (1970). Need for approval, children’s sharing behaviors and reciprocity in sharing. Child Development, 41, 243–252. Stevenson, H.W. (1991). The development of prosocial behaviour in large-scale collective societies: China and Japan. In R.A. Hinde & J. Groebel (Eds.), Cooperation and prosocial behaviour (pp. 89–105). Cambridge: Cambridge University Press. Stevenson, M.B., Leavitt, L.A., Thompson, R.H., & Roach, M.A. (1988). A social relations model analysis of parent and child play. Developmental Psychology, 24, 101–108. Strayer, F.F. (1980). Child ethology and the study of preschool social relations. In H.C. Foot, A.J. Chapman, & J.R. Smith (Eds.), Friendship and social relations in children (pp. 235–265). New York: Wiley. Strayer, F.F., Wareing, S., & Rushton, J.P. (1979). Social constraints on naturally occurring preschool altruism. Ethology and Sociobiology, 1, 3–11. Trivers, R.L. (1971). The evolution of reciprocal altruism. Quarterly Review of Biology, 46, 35–57. Uehara, I. (2004). Developmental changes in consistency of preferential feeling for peers and objects around age four. Psychological Reports, 94, 335–347. Wellman, H.M., Cross, D., & Watson, J. (2001). Meta-analysis of theoryof-mind development: The truth about false belief. Child Development, 72, 655–684. Youniss, J. (1986). Development in reciprocity through friendship. In C. ZahnWaxler, M. Cummings, & R. Iannotti (Eds.), Altruism and aggression (pp. 88–106). Cambridge: Cambridge University Press.

Downloaded from http://jbd.sagepub.com at Graduate Uni for Advanced Stu on March 7, 2008 © 2008 International Society for the Study of Behavioral Development. All rights reserved. Not for commercial use or unauthorized distribution.