Journal of Comparative Psychology 2006, Vol. 120, No. 1, 48 –57

Copyright 2006 by the American Psychological Association 0735-7036/06/$12.00 DOI: 10.1037/0735-7036.120.1.48

Peacemaking and Consolation in Japanese Preschoolers Witnessing Peer Aggression Keiko K. Fujisawa, Nobuyuki Kutsukake, and Toshikazu Hasegawa University of Tokyo This article reports developmental changes relating to reconciliation and bystanders’ affiliation with victims of aggression (i.e., consolation) among 3- to 5-year-old Japanese preschool children. Use of the postconflict–matched control (PC–MC) method revealed that the frequency with which reconciliation and consolation were offered to a victim increased steeply in 5-year-olds, compared with 3- and 4-year-olds. The complexity of contextual factors affecting the occurrence of reconciliation and the form of consolation increased with age. Consolation occurred more often before reconciliation than after among all but the 3-year-olds and occurred more often when no reconciliation occurred than when it did occur among all classes. These findings support the view that consolation functions as a substitute for reconciliation, lessening the tension experienced by the victim of aggression. Keywords: reconciliation, consolation, preschoolers, naturalistic observation, PC–MC method

caca mulatta). This observational method systematically compares behaviors observed during the postaggression condition (i.e., PC data) with that of the baseline condition (MC data). Note that primatologists have conventionally used the term “postconflict” to indicate the situation after aggression. Following the usage in these past studies, we also used postconflict (PC) to refer to postaggression. This method allows researchers to compare behaviors during a given time period after aggression with those of a control condition. The PC–MC comparison method has revealed that former opponents affiliate soon after aggression, which is labeled “reconciliation” (Aureli & de Waal. 2000; Aureli, Cords, & van Schaik, 2002). Reconciliation has been confirmed in more than 30 species of social primates (Aureli et al., 2002). Recently, the PC–MC comparison method has been applied to studies of children and has shown that affiliation frequency is greater between former opponents after aggression than the baseline affiliation frequency; former opponents will reconcile even after a brief separation following aggression (Russia: Butovskaya & Kozintsev, 1999; Sweden: Ljungberg, Westlund, & Forsberg, 1999; Kalmyk, Russia: Butovskaya, 2001; Japan: Fujisawa, Kutsukake, & Hasegawa, 2005); and former opponents will reconcile after conflicts without aggression as well as after aggressive conflicts (United States: Verbeek & de Waal, 2001). Furthermore, these studies have shown that social context affects whether opponents reconcile (see below) and that social context functions to reduce the postaggression stress of individual victims (Ljungberg et al., 1999; Fujisawa et al., 2005), as has repeatedly been confirmed in nonhuman primates (Aureli & van Schaik, 1991; Castles & Whiten, 1998; Das, Penke, & van Hooff, 1998; Kutsukake & Castles, 2001). Ways of conflict resolution in children differ by age. For example, Laursen and Hartup (1989) found that aggressiveness was associated with separate outcomes (i.e., opponents were separated from each other) among younger children but not among older children. Conflict termination strategies (e.g., use of conciliation; Verbeek & de Waal, 2001) also differed between younger and older children (Laursen & Hartup, 1989; Verbeek & de Waal,

In developmental psychology, comparative perspectives and cross-disciplinary approaches have promoted our understanding of developmental trajectories of social behavior and social relationships (Hinde, 1972, 1974; Smith, 1984). One such example can be seen in studies of conflict resolution mechanism after the aggression. Conflict is a situation that arises when individuals act on incompatible goals, interests, or actions; such conflict is not necessarily aggressive. Conflict resolution refers to actions that eliminate the incompatibility of attitudes and goals on the part of the conflicting individuals (Aureli & de Waal, 2000). However, traditional studies on conflict resolution among children may have underestimated postconflict reunion among young children. Researchers in these past studies have generally reported only situations observed immediately after conflict (e.g., Sackin & Thelen, 1984; Strayer & Strayer, 1976) and have failed to report observations of the behavior of opponents once separated. Studies of nonhuman primates have introduced a more systematic observation method. The so called postconflict–matched control, or PC–MC comparison, method was introduced by de Waal & van Roosmalen (1979) and was later refined by de Waal and Yoshihara (1983) in analyses of postaggression behavior among captive chimpanzees (Pan troglodytes) and rhesus macaques (Ma-

Keiko K. Fujisawa, Nobuyuki Kutsukake, and Toshikazu Hasegawa, Department of Cognitive and Behavioral Science, University of Tokyo We would like to thank Ryoko Kohata, Yuri Usui, and the teachers and children at the preschool for their great cooperation. C. Hemelrijk kindly provided the Matrixtester program, for which we are grateful. This study was supported by Japan Society for the Promotion of Science (JSPS) Research Fellowships for Young Scientists, which were awarded to Keiko K. Fujisawa and Nobuyuki Kutsukake, and by the JSPS 21st Century COE program: Center for Evolutionary Cognitive Sciences at the University of Tokyo. Correspondence concerning this article should be addressed to Keiko K. Fujisawa, Department of Cognitive and Behavioral Science, Graduate School of Arts and Sciences, Komaba 3-8-1 Komaba, Meguro-ku, Tokyo, Japan 153-8902. E-mail: [email protected] 48

PEACEMAKING AND CONSOLATION IN PRESCHOOLERS

2001, see also Iskandar, Laursen, Finkelstein, & Fredrickson, 1995, for hypothetical conflict situations). Concerning age-related differences of reconciliation, Verbeek and de Waal (2001) investigated both immediate conflict outcomes as well as reconciliation and reported no association between age and the occurrence of reconciliation. On the other hand, Fujisawa et al. (2005) reported that no contextual factors affected the occurrence of reconciliation among 3-year-olds but that close proximity after aggression between former opponents affected the occurrence of reconciliation among 4-year-olds. However, it is unclear whether the changes associated with positive conflict outcomes increase linearly and gradually as children grow older or whether they increase suddenly and steeply at a specific age. Researchers in most previous studies have examined age effects by deriving younger and older age groups from a wide age range (e.g., Laursen & Hartup, 1989; Iskandar et al.,1995; Verbeek & de Waal, 2001) and have often pooled the age data (e.g., 2- to 5-year-olds in Killen & Turiel, 1991), curtailing analysis of behavioral differences in each age class. Also, those researchers have examined age effects cross-sectionally and therefore have been unable to discuss longitudinal changes. Conflict resolution skills are associated with a number of sociocognitive abilities, including enhanced perspective taking and social understanding (see Dunn & Slomokowski, 1992, for a review). These abilities develop dramatically during the preschool period (Astington, 1993). For example, children’s understanding of others’ mental states, which is typically assessed by false-belief procedures, becomes rapidly more sophisticated around the age of 4 (Wellman, Cross, & Watson, 2001). Given the rapid development of such abilities during the preschool period, reconciliation among children may also change dramatically during the same period. In a previous study, we investigated patterns of reconciliation following aggression and the frequency of reconciliation in 3- and 4-year-old Japanese children (Fujisawa et al., 2005). In this study, we collected data 1 year after the study by Fujisawa et al. (2005) from the same groups; this method provided a unique opportunity to evaluate age-related changes in the pattern and frequency of reconciliation as well as the contextual factors affecting reconciliation among children within the same cohorts. Contextual factors affecting the occurrence of reconciliation, particularly the effects of friendship, were also investigated. Nonhuman primate studies have shown that relationship quality between opponents is an important factor in determining the occurrence of reconciliation. The valuable relationship hypothesis predicts that reconciliation is more frequent following aggression between opponents whose relationship has a high value (de Waal & Yoshihara, 1983; Aureli et al., 2002; Cords & Thurnheer, 1993). If the valuable relationship hypothesis can also be applied to the studies of children, then reconciliation should occur more often among friends than among nonfriends. Although previous studies have reported that friends terminated conflict with positive outcomes (Verbeek & de Waal, 2001; Laursen & Hartup, 1989; Hartup, Laursen, Stewart, & Eastenson, 1988), studies of reconciliation have found no statistical influence of friendship between opponents (Verbeek & de Waal, 2001) and have found evidence that nonfriends reconcile more frequently than friends (Butovskaya & Kozintsev, 1999). Finally, from data collected over 2 years, we investigated the pattern of contextual factors affecting the behavioral characteristics and functions of postaggression affiliation from bystanders

49

(i.e., children who were not involved in an aggressive episode) toward the victim of aggression, which was labeled “consolation” in this study. Bystanders are known to play active roles in mitigating and mediating tense situations during or after aggression and may encourage opponents to reconcile (Butovskaya, Verbeek, Ljungberg, & Lunardini, 2000). For example, schoolchildren in Kalmyk, Russia, acted as moderators of aggression by pushing opponents apart and persuading them to stop quarreling (“intervention”; Butovskaya et al., 2000). The prosocial response to a distressed person, which is commonly seen in young children (e.g., Bar-Tal, Raviv, & Goldberg, 1982; Denham, 1986; Strayer, 1980; Caplan & Hay, 1989), is generally assumed to be fostered by sociocognitive skills, particularly perspective taking (Eisenberg & Fabes, 1998); such perspective-taking skills are critical to differentiating between an individual’s own distress and others’ distress and to accurately understanding the nature of others’ emotional reactions (Hoffman, 1982). The positive association between perspective-taking abilities and children’s prosocial behavior has been found (reviewed in Underwood & Moore, 1982). Prosocial behavior has also been assumed to be linked with empathy (Eisenberg & Fabes, 1998; Zahn-Waxler, Robinson, & Emde, 1992). Given that a victim of aggression is in a state of distress (Fujisawa et al., 2005), it is likely that bystanders behave affiliatively or prosocially toward the victim of aggression if they understand the victim’s distress through their perspective-taking abilities or through empathy. If children’s sociocognitive abilities, such as perspective taking and empathy, become more sophisticated with age, then consolation should occur more often as children become older. Researchers who have conducted studies of nonhuman primates have proposed similar hypotheses regarding the relationship between cognitive ability and the occurrence of consolation. Consolation is a characteristic of the postaggression period in chimpanzees and bonobos (Pan paniscus; de Waal & van Roosmalen, 1979; de Waal & Aureli, 1996; Wittig & Boesch, 2003; Kutsukake & Castles, 2004; Palagi, Paoli, & Tarli, 2004; but see Arnold & Whiten, 2001; reviewed in Watts, Colmenares, & Arnold, 2000). In contrast, evidence of consolation is scarce in macaques (de Waal & Aureli, 1996); only one study reported consolation in stump-tailed macaques (Macaca arctoides) after classifying specific behaviors used in affiliation following aggression (Call, Aureli, & de Waal, 2002). De Waal and Aureli (1996) attributed this interspecific variation in consolation between chimpanzees and macaques to cognitive abilities such as empathy, which has been found in chimpanzees but not in macaques. If most bystanders’ affiliative contacts with a victim of aggression function as consolation, this behavior can be predicted to occur more often before opponents reconcile than after and more frequently when reconciliation does not occur than when reconciliation does take place. In addition, quantitative differences should exist between pre- and postreconciliation conditions. Concerning this last point, we predicted that a bystander would select his or her affiliative behavior depending on the situation. That is, a bystander would perform different forms of affiliative behavior in the period following aggressive events, compared with those performed in a control condition. We defined the consolation that is often seen in the postaggression period, but rarely seen in the control condition, as “explicit” consolation and labeled the consolation seen both in postaggression and control conditions as “implicit” consolation.

FUJISAWA, KUTSUKAKE, AND HASEGAWA

50 Method Participants

The study was conducted with the cooperation of two classes at a public preschool in Tokyo. The younger-age class consisted of 9 boys and 6 girls, and the older age class consisted of 12 boys and 10 girls. Table 1 provides detailed information on the participants. To refer to a child’s age, we use the class-unit label indicating the preschool. Hereafter, we refer to children in the class of 3-year-olds for the first year’s observation (T1) and in the class of 4-year-olds for the second year’s observation (T2) as Cohort 1 and the children in the class of 4-year-olds for T1 and the class of 5-year-olds for T2 as Cohort 2. One girl was not included in Cohort 1 because she had an autistic disorder that might have affected the results. By the beginning of the T2 observation period, 1 boy and 1 girl had left Cohort 1 because their families had moved to another area, and 1 boy and 1 girl had joined T2. There were no changes in Cohort 2 membership. All of the children were Japanese from predominantly middle-income families, and spoke Japanese as their first language.

Observation Methods Observations were conducted between October 2001 and March 2002 for T1 and between September 2002 and March 2003 for T2. Observations were conducted during free-play time in a classroom or on a playground. To minimize the disturbance effect of multiple observers, the first author, who is native Japanese and who had established a good rapport with the children and teachers before starting this study, made all the observations. During the observation time, the children were under the care of at least one female teacher. The observer refrained from intervening unless the children were in danger. Following de Waal and Yoshihara (1983), the observer scanned for aggressive conflict occurring in a dyad and carried out 10-min focal observations of children who had been the victims of aggression immediately after the aggression ended. We called the victim of aggression the “target child.” As explained earlier, we termed these data postconflict (PC) data. We regarded aggression as taking place when either of the following two types of behavior occurred: physical threats or deliberate pushing/ pulling, hitting, kicking, or biting (Strayer & Strayer, 1976) or verbal threats, insults, or derogatory comments (Verbeek & de Waal, 2001). Because polyadic aggressive conflicts, in which more than two children jointly attacked one child, were too rare to analyze systematically in our observations, we investigated only dyadic aggression to control for possible differences between dyadic and polyadic interactions. As a control observation, a 10-min baseline (matched control, or MC) observation of the same child was conducted (de Waal & Yoshihara, 1983). MC observations were collected only when the following conditions were met: (a) The MC observation could be conducted within 5 days of the PC observation; (b) the MC observation began within 15 min (either side) of the start time of the PC on a different day (e.g., if a PC observation was

conducted at 10:00 a.m., then the 10-min MC observation should begin between 9:45 a.m. and 10:15 a.m. on a different day); (c) the target child had not been involved in any aggression for at least 10 min before the start of the MC observation; (d) the former opponent was 1–10 m away from the target child to exclude the possibility that former opponents interacted in the MC period simply because they were close to each other, compared with their proximity in the PC period, or the possibility that they did not interact in the MC because of the distance between them; and (e) the location (classroom or playground) was the same. The following PC and MC data were recorded: aggression type (only in PC); the names of the initiator and the target; the timing of the aggression; the distance between opponents following aggression (together: within 1 m; separate: more than 1 m apart); the presence or absence of intervention by a teacher or bystander(s) (only in PC); any affiliative attempt between former opponents and the timing of its occurrence; the distance between the target child and bystander when recording began (together: within 1 m; separate: more than 1 m apart); and any affiliative attempt by a bystander toward the target child and the timing of this occurrence. Following Sackin and Thelen (1984) and Ljungberg et al., (1999), we recorded eight mutually exclusive affiliative types of behavior: offering an apology (only for the opponents), offering a compromise (only for the opponents), making a symbolic offer, offering an object, giving a gentle touch, inviting the partner to play, joining the partner’s play, or engaging in nonaggressive close and friendly talk (excluding apology).

Reliability After collecting T1 data, we checked the data from 30 PC and 30 MC observations (collected exactly in the same manner as other PC and MC observations) for reliability. Two coders (who, after training, were unaware of the conditions relating to any specific encounter, e.g., PC vs. MC or aggressor vs. victim) and the first author coded the video scripts independently. Interrater agreement was calculated using Cohen’s kappa statistic. The kappa coefficient of agreement is the ratio of the proportion of times that the raters agreed (corrected for chance agreement) to the maximum proportion of times that the raters could agree (corrected for chance agreement; Sigel & Castellan, 1988). If there was complete agreement among raters, then the kappa coefficient equaled 1; if there was no agreement (other than the agreement that could be expected to occur by chance) among the raters, then the kappa coefficient equaled 0. Kappa coefficients were computed to assess the coders’ agreement for each category: the presence or absence of each behavior, aggression type (physical or verbal), and distance (close or separate) in 60 video scripts. The Cohen’s kappa values for each category were as follows: aggression type, .80; distance following aggression, .77; apology, .77; object offer, .88; symbolic offer, .75; compromise, .78; invitation to play, .81; nonaggressive close position and talking, .84; joining in the opponent’s play, .78; bystanders’ intervention, .80; object offer by bystander, .86; symbolic offer by bystander, .75; invitation to play by bystander, .81; joining in the target

Table 1 Description of the Number of Children, Their Ages, and Friendships Time Point of Observation Variable

T1

T2

Cohort

Cohort 1

Cohort 2

Cohort 1

Cohort 2

Class Number of children (boys, girls) Mean (and range) of age (months) Number of boy–boy friends Number of girl–girl friends Number of boy–girl friends

3-year-olds (9, 6) 49 (44⬃54) 4 3 0

4-year-olds (12, 10) 61 (56⬃66) 9 7 1

4-year-olds (9, 6) 61 (55⬃65) 5 3 0

5-year-olds (12, 10) 73 (67⬃77) 10 14 3

PEACEMAKING AND CONSOLATION IN PRESCHOOLERS child’s play by bystander, .78; nonaggressive close and talking by bystander, .79; bystander touching the target child’s body, .82. Because the kappa values of a teacher’s intervention (.60) and touching the opponent’s body (.64) were not reliably high, these variables were excluded from further analysis. Also, variables that coders scored differently in the 60 video scripts were also not included in the analyses. The mean kappa coefficient for all categories, except for those of teacher’s intervention and touching the opponent’s body, was .80.

Definition of Friendship Using the instantaneous scan sampling method (Martin & Bateson, 1985) independently of the PC and MC recording, we recorded, for each child, the partners within 1 m every 15 min during the observation period (Cohort 1: n ⫽ 178 for T1, n ⫽ 305 for T2; Cohort 2: n ⫽ 191 for T1, n ⫽ 310 for T2). Then, we calculated the proximity frequency for every possible pair and regarded pairs with a proximity frequency exceeding the mean plus one standard deviation of the proximity frequency in each age group as friends. The possibility that the higher density of children in the classroom biased the results was excluded because a Wilcoxon signedranks test showed no significant difference between the proximity scores for the classroom and playground (T1: n ⫽ 36, z ⫽ ⫺.41, ns; T2: z ⫽ ⫺.40, ns). Thus, we pooled the data from both the classroom and playground. The results of our behavioral evaluation were consistent with teachers’ friendship nominations, in which the teachers in charge of each age group were asked to list each child’s friends. Significant correlations between the friendship matrices from the results of scan sampling and those from the teachers’ nominations were confirmed by the Kr test, which provides a multivariate version of Kendall’s S statistic (Tau Kr), corrected for sample size and ties (see Hemelrijk, 1990, for details): Cohort 1 at T1: Tau Kr ⫽ 0.51, p ⬍ .0001; Cohort 2 at T1: Tau Kr ⫽ 0.50, p ⬍ .0001; Cohort 1 at T2: Tau Kr ⫽ 0.59, p ⬍ .01; Cohort 2 at T2: Tau Kr ⫽ 0.74, p ⬍ .001. This result suggests that our evaluation of friendship was reliable. Table 1 lists the number of friends in each class.

Data Analysis Patterns of reconciliation and consolation. We used the PC–MC comparison method (de Waal & Yoshihara, 1983) to investigate occurrences of reconciliation and consolation. By this method, a PC–MC pair formed 1 data point. We examined the data from target children (i.e., not aggressors) who had at least 3 qualifying PC–MC data points (T1: n ⫽ 22, mean: 6.8 PC–MC pairs per child, range: 3–11 PC–MC pairs, T2: n ⫽ 25, mean: 8.8 PC–MC pairs per child, range: 3–14 PC–MC pairs) when analyses were conducted at the individual level. A PC–MC pair was labeled “attracted” if affiliation between the former opponents occurred only during the PC or earlier in the PC than in the MC. Similarly, a PC–MC pair was labeled “dispersed” if affiliation occurred only in the MC or earlier in the MC than in the PC. A PC–MC pair was labeled “neutral” if it did not occur in either the PC or MC periods or if it occurred at the same time in both the PC and MC periods. Concerning consolation, we considered the first affiliative contact by any bystander with the target child in both the PC and the MC and labeled those contacts as attracted, dispersed, or neutral according to the timing of the occurrence. We compared the percentage of attracted PC–MC pairs with that of the dispersed PC–MC pairs at an individual level to examine both PC affiliation between former opponents (i.e., reconciliation) and affiliation by a bystander to a target child (i.e., consolation). Timing of reconciliation and of consolation. We used the “time-rule” method (Aureli, van Schaik, & van Hooff, 1989). That is, for each PC and MC period, we determined the minute during which the first affiliation occurred between the former opponents. To determine the timing of consolation, we similarly determined the minute during which the first affiliation occurred by any bystander to the target child. We then compared the frequency distribution of the first affiliative contact between former oppo-

51

nents and that by a bystander to a target child in the PC periods with that in the MC periods using the Kolmogorov–Smirnov test. We also compared the frequency distribution of the first affiliative PC contacts between former opponents during the first-year observation period (T1) with that during the second-year observation (T2), using the Kolmogorov–Smirnov test to examine any age-related differences in reconciliation timing. A similar method was used to examine the timing of consolation (i.e., we compared the frequency distribution of the first affiliative contacts by bystanders with target children during the PC for T1 with that for T2). Frequencies of reconciliation and consolation and age-related changes. We used the revised measure of conciliatory tendency (CT; Veenema, Das, & Aureli, 1994) for reconciliation and the triadic conciliatory tendency (TCT; Call, et al., 2002) for consolation. These indices were calculated as follows: (a-d)/T, where a equaled the number of attracted pairs, d equaled the number of dispersed pairs, and T equaled the total number of PC–MC pairs. These indices equaled the difference between the proportions of attracted and dispersed pairs, if there were no neutral pairs. To investigate age-related changes with respect to reconciliation, we compared CTs reported in Fujisawa et al. (2005) with those of the present study. To examine age-related changes with respect to consolation, we compared the TCTs of T1 with those of T2. Contextual factors affecting the occurrence of reconciliatory attempts. To investigate variables affecting the occurrence of reconciliation, we used a stepwise logistic regression analysis based on all PC data for each cohort. We regarded attracted pairs as cases in which reconciliation occurred and dispersed pairs and neutral pairs as cases in which no reconciliation occurred, following Call, Aureli, and de Waal (1999). The dependent variable was the occurrence of reconciliation (occurred or did not occur). Table 2 describes the independent variables in the stepwise logistic regression analysis. To check the generality of the results at the individual level, we used the Wilcoxon signed-ranks test to compare CTs (Veenema, et al., 1994) during PC events involving target children who had more than three PCs. Behavioral characteristics of consolation and contextual impacts. To examine the behavioral characteristics of affiliative interactions in PC situations, we compared the frequency of six types of bystanders’ affiliative behavior. If a particular behavior occurred more frequently in relation to the PC than to the MC, that behavior could be regarded as specific to PC. Otherwise, the behavior could be regarded as related to both the PC and the MC. Therefore, we regarded bystander affiliative behavior that typically

Table 2 The Independent Variables in the Stepwise Logistic Regression Analyses for the Occurrence of Reconciliation and the Type of Consolation Independent Variable

Type

Occurrence of reconciliation Aggression type Distancea Gender combinationa Relationshipa Bystander’s intervention Location

Physical or verbal Together or separate Same or opposite gender Friend or nonfriend Presence or absence Classroom or playground Type of consolation

Aggression type Location Distanceb Gender combinationb Relationshipb Timing of its occurrence a

Between the opponents.

Physical or verbal Classroom or playground Together or separate Same or opposite gender Friend or nonfriend Before or after reconciliation b

Between a bystander and the target child.

FUJISAWA, KUTSUKAKE, AND HASEGAWA

52

occurred during the PC as explicit consolation, whereas the behavior that occurred during both PC and MC situations was regarded as implicit consolation. Stepwise logistic regression analysis was used to investigate the combination of contextual factors that affected the type of consolation. The dependent variable was the type of consolation (explicit or implicit). The independent variables in the stepwise logistic regression analysis are described in Table 2. The function of consolation. To examine the functional aspect of consolation, we compared the proportion of cases in which consolation occurred before reconciliation with that of cases in which it occurred after reconciliation at an individual level. We also compared the proportion of the cases in which consolation took place when no reconciliation occurred with that of the cases in which it took place when reconciliation did occur. If a bystander’s affiliation with the victim of aggression functions to console the victim, such behavior should occur more often before opponents reconcile than after and more frequently when reconciliation does not occur than when reconciliation does take place. All of the analyses were two-tailed, and the level of significance was set at .05. All statistical analyses were performed with R software, version 1.9.0 (Ihaka & Gentleman, 1996).

Results Patterns of Reconciliation and Consolation For the 2 years of observation, 384 PC–MC pairs were collected (65 PC–MC pairs in Cohort 1 at T1; 92 PC–MC pairs in Cohort 2 at T1; 106 PC–MC pairs in Cohort 1 at T2; 121 PC–MC pairs in Cohort 2 at T2). For reconciliation, the percentage of attracted pairs was significantly higher than that of dispersed pairs for both cohorts at T2 (see Table 3). The results show that victims had friendly contacts with former opponents significantly more often during the PC than during the MC. Affiliation with a former opponent was not caused by a general increase in affiliation frequency on the part of the target children because the number of children with whom target children affiliated did not differ between the PC and MC situations (affiliative partners per minute, Wilcoxon signed-ranks test; Cohort 1 at T2: PC: 3.3 vs. MC: 3.3, n ⫽ 10, V ⫽ 20, ns; Cohort 2 at T2: PC: 3.5 vs. MC: 3.6, n ⫽ 15, V ⫽ 51, ns). Rather, affiliation was selective;

the percentage of cases in which the affiliative partner was the former opponent was greater in the PC conditions than in MC conditions (Wilcoxon signed-ranks test; Cohort 1 at T2: 32.0% vs. 16.6%, n ⫽ 10, V ⫽ 55, p ⬍ .01; Cohort 2 at T2: 27.8% vs. 15.1%, n ⫽ 15, V ⫽ 116, p ⬍ .001). For consolation, the percentage of attracted pairs was significantly higher than that of dispersed pairs in Cohort 2 at T2 but insignificant for Cohort 2 at T1 or for Cohort 1 at T1 and T2 (see Table 3). These results indicate that consolation occurred significantly more often and earlier in the PC than in the MC among 5-year-olds; this pattern did not occur among 3- and 4-year-olds.

Timing of Reconciliation and Consolation The greatest differences in cumulative observations of the first friendly contacts between former opponents, between the PCs and the MCs, varied by cohort, and the differences between the PCs and the MCs were not statistically significant in all cohorts (see Table 4). However, the greatest difference in the first bystander’s affiliation to a target child between the PC and MC observations were found after 1 min in all cohorts for all observation periods, and significant differences were found in Cohort 1 at T2 and in Cohort 2 at T2 (see Table 4). The timing of reconciliation at T2 differed significantly from that at T1 for Cohort 2 (D ⫽ 0.26, p ⬍ .01) but not for Cohort 1 (D ⫽ 0.24, ns). Consolation timing was not significantly different for T1 and T2 in either cohort (Cohort 1: D ⫽ 0.29, ns; Cohort 2: D ⫽ 0.24, ns). These results suggest that the timing of reconciliation changed between the ages of 4 and 5, whereas the timing of consolation did not differ according to age.

Reconciliation and Consolation Frequencies and Age-related Changes The CT of Cohort 1 at T2 did not significantly differ from that of Cohort 1 at T1, whereas the CT of Cohort 2 at T2 was significantly higher than that of Cohort 2 at T1 (see Table 5). The TCT of Cohort 2 at T2 was significantly higher than that of Cohort 2 at T1, whereas the TCT of Cohort 1 at T2 did not significantly

Table 3 Percentages of Attracted vs. Dispersed Pairs Concerning Affiliation Between Opponents and Bystander Affiliation Time Point (Age)

Attracted

Dispersed

Wilcoxon signed-rank tests (V)

Affiliation between former opponents Cohort 1 Cohort 2

T2(4years) T2(5years)

59.8% 64.0%

21.6% 22.8%

36* 115**

33.2% 38.5% 33.6% 27.4%

25 45.5 39 91**

Bystander affiliation Cohort Cohort Cohort Cohort

1 2 1 2

T1(3years) T1(4years) T2(4years) T2(5years)

43.1% 46.3% 46.9% 56.1%

Note. The data for target children who had more than three PC–MC data in each year were examined at individual level (T1, Cohort 1: n ⫽ 8, Cohort 2: n ⫽ 14. T2, Cohort 1: n ⫽ 10, Cohort 2: n ⫽ 15). The data for affiliation between opponents at T1 were reported in Fujisawa, Kutsukake, & Hasegawa (2005). V value represents the summed ranks of the positive ranks. * p ⬍ .05. ** p ⬍ .01.

PEACEMAKING AND CONSOLATION IN PRESCHOOLERS

Table 4 Summary of Kolmogorov-Smirnov Tests for the Timings of Reconciliation and Consolation and the Age-Related Differences in Timing Cohort 1 Observation Period Reconciliation Greatest differences (min) D Consolation Greatest differences (min) D * p ⬍ .05.

Cohort 2

T1

T2

T1

T2

5 0.32

3 0.24

4 0.27

1 0.22

1 0.27

1 0.19*

1 0.22

1 0.22**

** p ⬍ .01.

differ from that of Cohort 1 at T1 (see Table 5). This was not due to the patterning of aggressive conflict according to the children’s ages because the proportions of physical aggression and verbal aggression did not differ (Wilcoxon signed-ranks test; physical vs. verbal; Cohort 1 at T1: 50.9% vs. 49.1%, V ⫽ 14.5; Cohort 2 at T1: 53.5% vs. 46.5%, n ⫽ 14, V ⫽ 56.5; Cohort 1 at T2: 52.1% vs. 47.9%, n ⫽ 10, V ⫽ 24.5; Cohort 2 at T2: 45.8% vs. 54.2%, n ⫽ 15, V ⫽ 64.5, all p values ⬎ .1). These findings were also not affected by unique changes in classroom characteristics because the CT and TCT of Cohort 1 at T2 did not differ from those of Cohort 2 at T1 (see Table 5). These results suggest that the frequency of reconciliation and consolation increased not year by year, but rapidly during a specific age period, that is, between the ages of 4 and 5.

showed no frequency difference between the PCs and the MCs or occurred more often in the MCs than in the PCs (see Table 7). These affiliative behaviors were regarded as implicit consolation. The results suggest that some bystanders’ affiliative behaviors can be distinguished in terms of affiliative contacts, particularly those that occur in the PCs and those that occur in both the PCs and the MCs. As shown in Table 8, no variable predicted the type of consolation in Cohort 1 at T1, but the timing before reconciliation was associated with the occurrence of explicit consolation at T2. Two factors at T1 in Cohort 2 (physical closeness between the victim and the bystander and the playground location) and two factors at T2 in Cohort 2 (close distance between the victim and the bystander and the timing before reconciliation) were associated with the occurrence of explicit consolation (see Table 8).

The Function of Consolation The proportion of cases in which consolation occurred before reconciliation was higher than that of cases in which it occurred after reconciliation among all cohorts (Wilcoxon signed-ranks test; Cohort 1 at T2: n ⫽ 10, V ⫽ 2, p ⬍ .1; Cohort 2 at T1: n ⫽ 14, V ⫽ 0, p ⬍ .01; Cohort 2 at T2: n ⫽ 15, V ⫽ 7, p ⬍ .01), except for Cohort 1 at T1 (3-year-olds; n ⫽ 8, V ⫽ 9, ns). The proportion of the cases in which consolation took place when no reconciliation occurred was higher than that of the cases in which it took place when reconciliation did occur among all cohorts (Wilcoxon signed-ranks test; Cohort 1 at T1: n ⫽ 8, V ⫽ 0, p ⬍ .05; Cohort 1 at T2: n ⫽ 10, V ⫽ 1, p ⬍ .01; Cohort 2 at T1: n ⫽ 14, V ⫽ 14, p ⬍ .05; Cohort 2 at T2: n ⫽ 15, V ⫽ 0, p ⬍ .001).

Contextual Factors Affecting Reconciliatory Attempts The occurrence of reconciliation was predicted by “distance” for Cohort 1 at T2 but predicted by “relation” and “bystanders’ intervention” for Cohort 2 at T2 (see Table 6). When examined at the individual level, the CT for Cohort 1 at T2 was higher when opponents were close together after aggressive conflicts than when they were separated (Wilcoxon signed-ranks test; together vs. separate: 65.7% vs. 31.4%, n ⫽ 10, V ⫽ 33, p ⬍ .05). For Cohort 2 at T2, the CT was higher among nonfriends than among friends (Wilcoxon signed-ranks test; nonfriends vs. friends: 68.1% vs. 21.9%; n ⫽ 14,1 V ⫽ 21, p ⫽ .05), although an equal proportion of aggressive conflicts occurred between friends and nonfriends (Wilcoxon signed-ranks test; with friends vs. with nonfriends: 55.6% vs. 44.4%, n ⫽ 15, V ⫽ 64.5, ns). The CT was also higher when a bystander did not intervene than when intervention occurred (Wilcoxon signed-ranks test; the absence of intervention vs. the presence of intervention: 50.6% vs. 12.6%; n ⫽ 15, V ⫽ 19.5, p ⬍ .05). These results indicate that the contextual factors that affect reconciliation differ depending on children’s ages.

Behavioral Characteristics of Consolation and the Impact of Contextual Factors “Gentle touch” and “symbolic offer” occurred relatively frequently in the PCs, compared with in the MCs. These affiliative behaviors were regarded as explicit consolation. “Nonaggressive close position and friendly talking,” “offering object,” “inviting the target child to play,” and ‘joining the target child’s play” either

53

Discussion Reconciliation During the Later Preschool Period We found that the frequency and the timing of reconciliation and the contextual factors affecting its occurrence changed when the children were approximately 4 and 5 years old. The pattern of reconciliation occurrence, however, did not change. This result contrasts with the findings of our previous studies on reconciliation among children, in which no such changes were seen in children who were 3 and 4 years old (Fujisawa et al., 2005). Intracohort comparisons indicate that the frequency of reconciliation (examined by the index of the CT; Veenema et al., 1994) significantly increased between T1 and T2 in the older age cohort (Cohort 2) but did not significantly increase among younger children (Cohort 1). These findings were not due to unique changes in the characteristics of the particular cohort because the frequency of reconciliation at age 4 was similar in both cohorts. This finding not only supports previous reports that older children resolve their conflicts more frequently than younger children (e.g., Verbeek & de Waal, 2001; Laursen & Hartup, 1989) but also reveals that children’s tendency to reconcile steeply increases around the age of 5. Therefore, our prediction of rapid change in the occurrence of reconciliation was confirmed. Sociocognitive abilities, such as 1

One of the fifteen children who had more than three PC–MC pairs in Cohort 2 at T2 did not experience aggression from his friends. Thus, we conducted Wilcoxon signed-ranks tests for 14 children.

FUJISAWA, KUTSUKAKE, AND HASEGAWA

54

Table 5 Mean Values of Conciliatory Tendency (Between Opponents) and Triadic Conciliatory Tendency Measures (From Bystander to Victim), with Longitudinal and Cross-Sectional Age Comparisons 3 years

4 years

5 years

Wilcoxon signed-rank tests (V)

– 41.2

125, ns 75*

– 28.7

14, ns 78*

Conciliatory tendency Cohort 1 (%) Cohort 2 (%) Mann-Whitney test (W)

38.3 –

38.2 35.2 54, ns

Triadic conciliatory tendency Cohort 1 (%) Cohort 2 (%) Mann-Whitney test (W)

9.9 –

13.3 7.7 51, ns

Note. The data for Cohort 1 at 3 years and Cohort 2 at 4 years are from Fujisawa, Kutsukake, & Hasegawa (2005). For longitudinal comparisons, we examined the data for target children who had more than three PC–MC data in both T1 and T2 (Cohort 1: n ⫽ 7; Cohort 2: n ⫽ 13). For cross-sectional comparisons, we examined the data for target children who had more than three PC–MC data at 4 years (Cohort 1: n ⫽ 10, Cohort 2: n ⫽ 14) * p ⬍ .05

studies on children have found that friends resolve conflicts more often and more equitably than do nonfriends (e.g., Hartup, et al., 1988; Vespo & Caplan, 1993). Verbeek and de Waal (2001) investigated both immediate outcomes of conflicts and postconflict reunion (i.e., reconciliation) and reported that relationships between opponents were associated with the immediate positive outcome but not associated with reconciliation. However, these studies investigated only the immediate outcomes of conflicts and did not consider control conditions. Thus, the high frequency of positive conflict resolution among friends may merely reflect the fact that friends ordinarily experience more mutual contact than do nonfriends (Rubin, Bukowski, & Parker, 1998). Research has found that friends have more frequent preconflict interactions than nonfriends and that preconflict interaction is an important predictor of together outcomes (Verbeek & de Waal, 2001). On the other hand, the index of frequency of reconciliation (CT) used in the present study controlled for baseline affiliation. Given that Butovskaya and Kozintsev (1999) found less of a reconciliation tendency (CT) in friends than in nonfriends among 6- to 7-year-old schoolchildren, the influences of relationship (i.e., higher reconciliatory tendency for nonfriends than friends) may emerge around age 5. However, Verbeek and de Waal (2001) found no association between relationship and reconciliation among children who were between 32 to 71 months old. We cannot directly compare Verbeek and de Waal (2001)’s findings

perspective taking and understanding others’ mental states, are assumed to be critical for conflict resolution (Dunn & Slomokowski, 1992). These abilities develop very rapidly during the later preschool period (Astington, 1993); in particular, the ability to understand others develops very rapidly in children at approximately age 4 (Wellman et al., 2001). Although children’s sociocognitive abilities were not directly examined in the present study, the dramatic change in the occurrence of reconciliation around age 5 may reflect the rapid development of sociocognitive abilities in the same period, as well as accumulated social experiences. A number of contextual factors affecting the occurrence of reconciliation increased in importance by age, suggesting that reconciliation is complicated by age. Although no factor influenced the occurrence of reconciliation among 3-year-olds (Fujisawa et al., 2005), in the two classes of 4-year-olds (Cohort 1 at T2 and Cohort 2 at T1) opponents reconciled more frequently when they were closer together soon after an aggressive conflict than they did when they were separated. Previous studies have also noted the importance of proximity between former opponents for conflict resolution (Laursen & Hartup, 1989; Verbeek & de Waal, 2001). The higher frequency of reconciliation (examined by the index of CT) among nonfriends than among friends in the 5-year-old group suggests that the valuable relationship hypothesis was not supported in terms of reconciliation in the present study. Previous

Table 6 Summary of Stepwise Logistic Regression Analysis for Contextual Factors Influencing the Occurrence of Reconciliation Variable Cohort 1 at T2 Distance Cohort 2 at T2 Relation Bystander’s intervention

Estimate

SE

CI

z

OR

CI for OR

0.72

0.41

[⫺0.08, ⫺1.55]

1.75⫹

2.06

[0.92, 4.63]

⫺0.91 ⫺1.02

0.40 0.40

[⫺1.72, ⫺0.13] [⫺1.81, ⫺0.25]

⫺2.26* ⫺2.57*

0.40 0.36

[0.18, 0.88] [0.17, 0.79]

Note. CI ⫽ 95% confidence interval, OR ⫽ odds ratio. p ⬍ .10. * p ⬍ .05.

⫹

PEACEMAKING AND CONSOLATION IN PRESCHOOLERS

though schoolchildren can effectively intervene in aggression using means such as persuasion (Butovskaya, et al., 2000), effective intervention might still be a demanding task for 5-year-olds. To draw firm conclusions, we need further studies to systematically investigate reconciliation after a bystander’s intervention.

Table 7 The Number of Observations of Each Type of Affiliative Behavior Initiated by Bystander Toward Victim Affiliative behavior

PC

Symbolic offer Inviting the target child to play Offering object Gentle touch Nonaggressive close and friendly talking Joining the target child’s play

MC

19⫹ 40 13 80*** 127 28

55

9 33 10 30 202** 48⫹

Consolation and Its Development We systematically investigated affiliative contact by a bystander toward the victim of aggression (i.e., consolation) and provided supporting evidence for our prediction that consolation would occur more often in older children. Bystanders’ affiliative contacts with victims of aggression occurred more frequently and earlier during the PC situations than during the MC situations only among 5-year-olds (Cohort 2 at T2) but not among 3- and 4-year-olds (Cohort 1 at T1 and T2; Cohort 2 at T1). Moreover, the frequency of consolation (examined by the index of the TCT (Call et al., 2002) significantly increased between T1 and T2 in older children (Cohort 2) but did not increase among younger children (Cohort 1). The frequency of consolation at age 4 was similar in both cohorts. Thus, similar to the situation with reconciliation, the approximate age of 5 may constitute a distinct period for acting affiliatively toward the victim of aggression, compared with such behavior in younger children. This finding agrees with the finding that older children exhibit prosocial behavior more often than younger children (reviewed in Eisenberg & Fabes, 1998; ZahnWaxler, Radke-Yarrow, Wagner, & Chapman, 1992; Zahn-Waxler & Smith, 1992). Similar to the age-related change in reconciliation, the rapid change with respect to consolation around age 5 may be associated with the development of sociocognitive abilities such as perspective taking and empathy. It is commonly assumed that higher perspective-taking skills are associated with an accurate understanding of others’ emotional reactions (Hoffman, 1982) and perspective-taking skills may foster empathy and consequently lead to prosocial behavior (reviewed in Eisenberg & Fabes, 1998). In studies of nonhuman primates, the interspecific differences in consolation occurrence can be attributed to differences in sociocognitive abilities such as empathy and perspective taking (de Waal & Aureli, 1996). This finding may parallel our hypothesis that development of sociocognitive abilities may underlie changes

Note. To examine whether observed value was higher than expected value, we calculated the probability of the occurrence of adjusted residual in standard normal distribution for each cell. ␹2 ⫽ 48.82, df ⫽ 5, p ⬍ .001; effect size w ⫽ 0.28. PC ⫽ post conflict, MC ⫽ matched control. ⫹ p ⬍ 10. ** p ⬍ .01. *** p ⬍ .001.

with ours because they did not use an index of CT. Verbeek and de Waal (2001) also did not systematically examine the effect of age-related differences in relationships and reconciliation; thus, the finding of no association between relationship and reconciliation might be due to the various age ranges of the children in their study. In 5-year-olds, affiliative activities after aggressive conflicts between nonfriends might be used to restore a relationship damaged by aggression. Such affiliative activities between friends do not carry as much meaning as those among nonfriends, given that the bonds between friends are strong and, thus, their restoration after damage requires relatively little effort (Butovskaya & Kozintsev, 1999). Studies of nonhuman primates have proposed that relationship characteristics other than value, such as low “security” between the opponents (the perceived probability that the relationship with the partner will change), facilitate the occurrence of reconciliation (Cords & Aureli, 2000). For example, Cords and Aureli (1993) found that juvenile long-tailed macaques reconciled more often with “insecure” nonkin than with “secure” kin. Another contextual factor that impacted the occurrence of reconciliation among 5-year-old children was bystanders’ intervention. Because the occurrence of reconciliation after a bystander’s intervention was rare (only 1 case in 39), it is likely that bystanders could not effectively intervene between former opponents. Al-

Table 8 Summary of Stepwise Logistic Regression Analysis for Contextual Factors Influencing the Type of Consolation Variable

Estimate

SE

CI

z

OR

CI for OR

Cohort 1 at T1 Cohort 2 at T1 Location Distancea Cohort 1 at T2 Timing of reconciliation Cohort 2 at T2 Distancea Timing of reconciliation

—

—

—

—

—

—

⫺1.63 2.28

0.86 0.89

0.20 9.79

[0.04, 1.05] [1.71, 56.04]

2.36

0.78

[1.04, 4.23]

3.04**

10.54

[2.31, 48.21]

0.89 2.05

0.44 0.66

[0.03, 1.78] [0.88, 3.56]

2.01* 3.10**

2.43 7.75

[1.02, 5.75] [2.12, 28.34]

[⫺3.61, 0.11] ⫺1.91⫹ [0.70, 4.33] 2.56*

Note. CI ⫽ 95% confidence interval; OR ⫽ odds ratio. a Between a bystander and the target child. ⫹ p ⬍ 10. * p ⬍ .05. ** p ⬍ .01.

56

FUJISAWA, KUTSUKAKE, AND HASEGAWA

in the occurrence of consolation among children around the age of 5. This study also showed that some forms of bystanders’ affiliation (explicit consolation) occurred during the PC and other forms (implicit consolation) occurred in both the PC and the MC. Thus, the prediction that a bystander would display different forms of affiliative behavior in the period following aggression, compared with a control condition, was confirmed. This finding suggests that affiliative acts toward the victim of aggression can be distinguished from those that occur during ordinary playtime. Developmental changes were also found to be related to contextual factors influencing the form of consolation. Explicit consolation is likely to manifest as a clear intent to offer consolation, given that explicit forms of contact occurred after aggression toward the victim. Explicit consolation may have been used by 4- and 5-year-olds to console the victim because it might have been easy for them to notice the victim’s distress when they were near him or her.

The Functions of Consolation The present study provides data relating to the functional aspects of consolation; these aspects have remained largely unexplored because of scant previous research. If much of bystanders’ affiliative contact with victims of aggression functions to console victims in their distress, then such affiliation would be expected to occur more often before reconciliation than after reconciliation and to occur more frequently when reconciliation did not occur than when reconciliation had occurred because victims of aggression experience higher stress before reconciliation than after (Fujisawa et al., 2005). These predictions were confirmed in all cohorts except for the youngest one. Consolation occurred more often before reconciliation among all cohorts except for the youngest cohort (3-year-olds), and it occurred more often in situations in which no reconciliation occurred for both cohorts. These results indicate that consolation that occurs before reconciliation or in cases in which there is no reconciliation has a more pronounced consolatory character than behaviors that occur after reconciliation. Our findings suggest that consolation may be used as a substitute mechanism for reconciliation (de Waal & Aureli, 1996; Palagi et al., 2004). Although reconciliation seems to be the best, consolation might function as a substitute for reconciliation, thus lessening the tension of the victim of aggression. Explicit consolation occurred more often before reconciliation among 4-year-olds in T2 and among 5-year-olds. This suggests that explicit bystanders’ affiliation assumes a consolatory function. Therefore, the prediction that the bystander’s affiliative behavior toward the victim of aggression would quantitatively differ preand postreconciliation was confirmed. It is likely that the 4- and 5-year-olds in the present study chose explicit consolation in which the intention to console was much clearer than in the case of implicit consolation. Considering that the occurrence of consolation was irrelevant to that of the timing of reconciliation and that none of the contextual factors affected the type of consolation offered among 3-year-olds, it is likely that consolation does not function at this age. In summary, the frequency of reconciliation and consolation increased rapidly from ages 4 to 5, compared with the frequency seen from ages 3 to 4, and the contextual factors affecting the occurrence of reconciliation became complicated as children got older. Children used explicit consolation when the former oppo-

nents did not reconcile, suggesting that consolation might act as a substitute for reconciliation and function to lessen the stress experienced by the victim of aggression. These rapid changes in postaggression behavior may be linked to the rapid development of sociocognitive abilities in children around this age (Astington, 1993). It should be noted that these new findings were revealed by the PC–MC observational method used to systematically compare postaggression and control data. Therefore, this study emphasizes the importance of comparative perspectives and interdisciplinary approaches between human developmental psychology and animal studies for understating social behavior among children and child development.

References Arnold, K. & Whiten, A. (2001). Post-conflict behaviour of wild chimpanzees (Pan troglodytes schweinfurthii) in Budongo Forest, Uganda. Behaviour, 138, 649 – 690. Astington, J. W. (1993). The child’s discovery of the mind. Cambridge, MA: Harvard University Press. Aureli, F., Cords, M., & van Schaik, C. P. (2002). Conflict resolution following aggression in gregarious animals: A predictive framework. Animal Behaviour, 64, 325–343. Aureli, F. & de Waal, F. B. M. (2000). Natural conflict resolution. London: University of California Press. Aureli, F. & van Schaik, C. P. (1991). Postconflict behaviour in long-tailed macaques (Macaca fascicularis): II. Coping with the uncertainty. Ethology, 89, 101–114. Aureli, F., van Schaik, C. P., & van Hooff, J. A. R. A. M. (1989). Functional aspects of reconciliation among captive long-tailed macaques (Macaca fascicularis). American Journal of Primatology, 19, 39 –51. Bar-Tal, D., Raviv, A., & Goldberg, M. (1982). Helping behavior among preschool children: An observational study. Child Development, 53, 396 – 402. Butovskaya, M. L. (2001). Reconciliation after conflicts: Ethological analysis of post-conflict interactions in Kalmyk children. In J. M. Ramirez, & D. S. Richardson (Eds.), Cross-cultural approaches to aggression and reconciliation (pp. 167–190). Huntington, NY: Nova Science. Butovskaya, M. L. & Kozintsev, A. (1999). Aggression, friendship, and reconciliation in Russian primary schoolchildren. Aggressive Behavior, 25, 125–139. Butovskaya, M. L., Verbeek, P., Ljungberg, T., & Lunardini, A. (2000). A multicultural view of peacemaking among young children, In F. Aureli, & F. B. M. de Waal (Eds.), Natural conflict resolution (pp. 243–258). London: University of California Press. Call, J., Aureli, F., & de Waal, F. B. M. (2002). Postconflict third-party affiliation in stump-tailed macaques. Animal Behaviour, 63, 209 –216. Call, J., Aureli, F., & de Waal, F. B. M. (1999). Reconciliation patterns among stumptailed macaques: A multivariate approach. Animal Behaviour, 58, 165–172. Caplan, M. Z. & Hay, D. F. (1989). Preschoolers’ responses to peers’ distress and beliefs about bystander intervention. Journal of Child Psychology and Psychiatry, 30, 231–242. Castles, D. L. & Whiten, A. (1998). Post-conflict behaviour of wild olive baboons: II. Stress and self-directed behaviour. Ethology, 104, 148 –160. Cords, M. & Aureli, F. (1993). Pattern of reconciliation among juvenile long-tailed macaques. In M. E. Pereira & L. A. Fairbanks (Eds.) Juvenile primates: Life history, development, and behaviour (pp. 271–284). Oxford, United Kingdom: Oxford University Press. Cords, M. & Aureli, F. (2000). Reconciliation and relationship quality. In F. Aureli & F. B. M. de Waal (Eds.) Natural conflict resolution (pp. 177–198). Berkeley: University of California Press. Cords, M. & Thurnheer, S. (1993). Reconciliation with valuable partners by long-tailed macaques. Ethology, 93, 315–325.

PEACEMAKING AND CONSOLATION IN PRESCHOOLERS Das, M., Penke, Z., & van Hooff, J. A. R. A. M. (1998). Postconflict affiliation and stress-related behaviour of long-tailed macaque aggressors. International Journal of Primatology, 19, 53–71. Denham, S. A. (1986). Social cognition, prosocial behavior, and emotion in preschoolers: Contextual validation. Child Development, 57, 194 –201. de Waal, F. B. M. & Aureli, F. (1996). Consolation, reconciliation, and a possible cognitive difference between macaques and chimpanzees. In A. E. Russon, K. A. Bard, & S. T. Parker (Eds.), Reaching into thought: The minds of the great apes (pp. 80 –110). Cambridge, United Kingdom: Cambridge University Press. de Waal, F. B. M. & van Roosmalen, A. (1979). Reconciliation and consolation among chimpanzees. Behavioral Ecology & Sociobiology, 5, 55– 66. de Waal, F. B. M. & Yoshihara, D. (1983). Reconciliation and redirected affection in rhesus monkeys. Behaviour, 85, 224 –241. 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, United Kingdom: Cambridge University Press. 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. Fujisawa, K., Kutsukake, N., & Hasegawa, T. (2005). Reconciliation pattern after aggression among Japanese preschool children. Aggressive Behavior, 31, 138 –152. Hartup, W. W., Laursen, B., Stewart, M. I., & Eastenson, A. (1988). Conflict and the friendship relations of young children. Child Development, 59, 1590 –1600. Hemelrijk, C. K. (1990). Models of and tests for reciprocity, unidirectionality and other social interaction patterns at a group level. Animal Behaviour, 39, 1013–1029. Hinde, R. A. (1972). Non-verbal communication. Cambridge, United Kingdom: Cambridge University Press. Hinde, R. A. (1974). Biological bases of human social behavior. New York: McGraw-Hill. Hoffman, M. L. (1982). Development of prosocial motivation: Empathy and guilt. In N. Eisenberg-Berg, (Ed.), Development of prosocial behavior (pp. 281–313). New York: Academic Press. Ihaka, R., & Gentleman, R. (1996). R: A language for data analysis and graphics. Journal of Computational and Graphical Statistics, 5, 299 – 314. Iskandar, N., Laursen, B., Finkelstein, B., & Fredrickson, L. (1995). Conflict resolution among preschool children: The appeal of negotiation in hypothetical disputes. Early Education and Development, 6, 359 –376. Killen, M., & Turiel, E. (1991). Conflict resolution in preschool social interactions. Early Education and Development, 2, 240 –255. Kutsukake, N. & Castles, D. L. (2001). Reconciliation and variation in post-conflict stress in Japanese macaques (Maraca fuscata fuscata): testing the integrated hypothesis. Animal Cognition, 4, 259 –268. Kutsukake, N. & Castles, D. L. (2004). Reconciliation and post-conflict third-party affiliation among wild chimpanzees in the Mahale Mountains, Tanzania. Primates, 45, 157–165. Laursen, B. & Hartup, W. W. (1989). The dynamics of preschool children’s conflicts. Merrill-Palmer Quarterly, 35, 281–297. Ljungberg, T., Westlund, K., & Forsberg, A. J. (1999). Conflict resolution

57

in 5-year-old boys: Does postconflict affiliative behaviour have a reconciliatory role? Animal Behaviour, 58, 1007–1016. Martin, P., & Bateson, P. (1985). Measuring behavior. Cambridge, United Kingdom: Cambridge University Press. Palagi, E., Paoli, T., & Tarli, S. B. (2004). Reconciliation and consolation in captive bonobos (Pan paniscus). American Journal of Primatology, 62, 15–30. 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. Sackin, S. & Thelen, E. (1984). An ethological study of peaceful associative outcomes to conflict in preschool children. Child Development, 55, 1098 –1102. Sigel, S. & Castellan, N. J. (1988). Nonparametric statistics for the behavioral sciences (2nd ed.). New York; McGraw-Hill International Editions. Smith, P. K. (1984). Play in animals and humans. (2nd ed.). Oxford, United Kingdom: Basil Blackwell. Strayer, F. F. & Strayer, J. (1976). An ethological analysis of social agonism and dominance relations among preschool children. Child Development, 47, 980 –989. Strayer, J. (1980). A naturalistic study of empathic behaviors and their relation to affective states and perspective-taking skills in preschool children. Child Development, 51, 815– 822. Underwood, B. & Moore, B. (1982). Perspective-taking and altruism. Psychological Bulletin, 91, 143–173. Veenema, H. C., Das, M., & Aureli, F. (1994). Methodological improvements for the study of reconciliation. Behavioral Processes, 31, 29 –38. Verbeek, P. & de Waal, F. B. M. (2001). Peacemaking among preschool children. Journal of Peace Psychology, 7, 5–28. Vespo, J. E. & Caplan, M. Z. (1993). Preschoolers’ differential conflict behavior with friends and acquaintances. Early Education and Development, 4, 45–53. Watts, D. P., Colmenares, F., & Arnold, K. (2000). Redirection, consolation, and male policing: How targets of aggression interact with bystanders. In: F. Aureli & F. B. M. de Waal (Eds.), Natural conflict resolution (pp. 281–301). Berkeley: University of California Press. Wellman, H. M., Cross, D., & Watson, J. (2001). Meta-analysis of theoryof-mind development: The truth about false belief. Child Development, 72, 655– 684. Wittig, R. M. & Boesch, C. (2003). The choice of post-conflict interactions in wild chimpanzees (Pan troglodytes). Behaviour, 140, 1527–1559. Zahn-Waxler, C., Radke-Yarrow, M., Wagner, E., & Chapman, M. (1992). Development of concern for others. Developmental Psychology, 28, 126 –136. Zahn-Waxler, C., Robinson, J., & Emde, R. N. (1992). The development of empathy in twins. Developmental Psychology, 28, 1038 –1047. Zahn-Waxler, C., & Smith, K. D. (1992). The development of prosocial behavior. In V. B. von Hasselt & M. Hersen (Eds.), Handbook of social development (pp. 229 –256). New York: Plenum.

Received June 1, 2005 Revision received September 16, 2005 Accepted October 7, 2005 䡲