How do researchers respond to this question? Do they have a clear notion of instruments measuring social competence and their validity in terms of the feasibility to assess relevant criterion behavior? What is relevant criterion behavior? And are we capable of developing valuable intervention studies to influence social competence or, stated differently, are we capable of identifying valid “conditions – mechanisms – outcome chains?”

 This symposium is meant to present empirical research that is seeking answers to these questions. Given the multicultural nature of our modern societies the presentations will all explicitly describe and justify the way the studies dealt with cultural diversity issues and how these affected the outcomes.

How do pupils in multicultural elementary schools perceive pupils with different ethnic backgrounds? And how is their perception influenced by ethnically heterogeneous grouping? In this study we investigated popularity and perceived cooperativeness in multiethnic pupil teams in elementary schools. First of all, we hypothesized that pupils’ popularity and perceived cooperativeness would increase as a function of SCL time. Secondly, we hypothesized that SCL time increases the popularity and level off the difference in perceived cooperativeness between national and immigrant pupils. Finally, many studies have found that interethnic contact reduces interethnic bias because it facilitates the formation of interethnic friendships. We propose in this paper that SCL time not only increases interethnic friendships, but also augments the popularity and perceived cooperativeness within ethnically heterogeneous teams, but not of ethnically homogeneous teams.

94 Pupils (26 teams) from five multiethnic elementary schools participated in a structured cooperative learning (SCL) curriculum. Both teachers and pupils had no prior SCL experience. The curriculum entailed two SCL training lessons, followed by a SCL math curriculum of nine lessons. The results show that SCL time increased both popularity and cooperativeness across ethnic background. In addition, experience with SCL increased the popularity of immigrant pupils and decreased the perceived cooperativeness difference between immigrant pupils and national pupils. Lastly, SCL time only augmented the popularity and perceived cooperativeness within ethnically heterogeneous teams. This last result shows that enduring interethnic contact facilitates not only interethnic friendships, but also interethnic popularity and perceived cooperativeness.

Introduction

How do pupils in multicultural elementary schools perceive pupils with different ethnic backgrounds? And how is their perception influenced by ethnically heterogeneous grouping? The intergroup contact hypothesis (Allport, 1954) tells us that interethnic barriers will only decrease under specific conditions (i.e., cooperation instead of competition between members of different ethnic groups, equal status and common goals for participating pupils, and support for inter-ethnic relationships and collaboration of authorities and institutions). Pettigrew and Tropp (2006) showed that these conditions may be important for a reduction in interethnic bias, but that it is the exposure time to ethnically distinct groups that is essential for a decrease in bias. That is, the more people from different ethnic groups get to know each other, the more they are inclined to become friends. Whereas Pettigrew and Tropp defined interethnic liking as friendship, we define it here as popularity and perceived cooperativeness. We hypothesized that pupils’ popularity and perceived cooperativeness would increase as a function of structured cooperative learning time. Secondly, we hypothesized that structured cooperative learning time increases the popularity and level off the difference in perceived cooperativeness between national and immigrant pupils. Finally, we propose that popularity and perceived cooperativeness increase as a function of structured cooperative learning time. This holds in the ethnically heterogeneous teams only.

Method

Ninety-four pupils from 26 teams in five multi-ethnic elementary schools in the Netherlands participated in a structured cooperative learning curriculum. Neither teachers not pupils had prior structured cooperative learning experience. We used the following measures:

Popularity.         A list was filled in twice by all pupils: at the start of the structured cooperative learning curriculum (T1) and at the end (T2). Pupils were required to rate their team members as perceived by the whole class on the behavioural characteristic: ‘is well liked by everyone’. Scores were averaged per pupil, excluding their own scores.

Perceived cooperativeness.       Pupils were required to nominate team members at the end of every lesson who did not implement the structured cooperative learning rules. They wrote down the name(s) of the non compliant team members.

The curriculum entailed two structured cooperative learning training lessons, in which pupils were instructed in the use of general structured cooperative learning rules (e.g., ‘do not interrupt each other’, ‘do not make fun of each other’), and rules of helping behaviour (e.g., ‘ask precise questions, ‘give help when needed’). These rules were practiced in a cooperative math assignment. The structured cooperative learning training was followed by a structured cooperative learning math curriculum of nine lessons (two lessons per week). The pupils completed cooperative math assignments in fixed teams, under supervision of the teacher. In each lesson two authentic math assignments with a common theme (e.g., the zoo) had to be solved by the pupils. Authentic math assignments are mathematical tasks with a strong narrative structure that are embedded in contexts familiar to the children and to which multiple solutions are possible.

Results

The results show that structured cooperative learning time increased both popularity and cooperativeness across ethnic background. In addition, experience with structured cooperative learning increased the popularity of immigrant pupils and decreased the perceived cooperativeness difference between immigrant pupils and national pupils. Lastly, structured cooperative learning time only augmented the popularity and perceived cooperativeness within ethnically heterogeneous teams.

The findings show that Pettigrew’s (1998) interpretation of the contact hypothesis, focusing on the actual exposure time instead of characteristics of the contact that should be satisfied provides a better explanation for the interethnic bias in ethnically diverse classrooms than does the original contact hypothesis . In addition, our findings indicate that structured cooperative learning may provide a solution to interethnic problems in multiethnic elementary schools in the Netherlands.

A Multisource Assessment of Social Competence Scale was developed and examined to test the consistency of the self, peers’, teachers’, and parents´ ratings of Finnish school-aged children's (n = 985) social competence. The correlations between the four social agents were statistically significant, albeit quite low, indicating that the different sources tend to provide divergent pictures of a child’s social competence. To study the possible concomitants for these divergences, we compared linguistically and mathematically high-, average- and low-achieving students in regard to self, peer, teacher, and parent ratings. The differences between high- and low-achieving children were all statistically significant indicating, that the low-achievers had less co-operating skills and empathy behaviour and more impulsive and disruptive behaviour than high-achievers. Outstandingly, the effect sizes concerning the differences for teachers´ ratings were especially high.

So what makes the differences in teacher ratings so high? School culture and the dominant norms and values defining the correct behaviours are important in the understanding of the social competence in the classrooms. Earlier studies suggest that both rater culture and rated culture specific differences may play a role (see Mpofu, Thomas, & Chan, 2004). The teacher ratings of low-achieving students were particularly low. Could it be that teacher use prejudicial expectations while rating students’ social competence? This and other possible cultural and rater-based explanations for the divergences and their effects on students’ socio-emotional well-being, learning opportunities and intervention designs are discussed in the presentation.

Introduction

It has been argued that social competence is best studied using multiple sources of data (Renk & Phares, 2004). Different raters with different cultural norms, values and attitudes may give different ratings of a child’s competence; nevertheless these may all be valid perspectives on a child’s social behavior. In the school context self-ratings, peer ratings, teacher ratings, and parent ratings are especially relevant. Most studies on children’s social competence have adopted the perspective of one or two of these types of raters. Studies on the consistency of these different perspectives on children’s social competence, however, have found conflicting results, which suggests that manifestations of a child’s social competence are ambiguous (see Renk & Phares, 2004).

Different cultural backgrounds and related expectations may be playing a role. According to Mpofu, Thomas and Chan (2004) ethnic minority culture students are rated significantly lower on social competence, especially by their teachers. As another “minority group” also children with learning disabilities are at greater risk of social rejection and social problems than average- to high-achieving children. A plausible explanation, according to Nowicki (2003), is that children who have difficulties processing academic information also face substantial challenges in untangling a constantly changing and norm-related environment of social interactions from many sources, nonverbal cues, situational / cultural contexts, and so on.

Differences in institutional or contextual social expectations for different groups of children could lead to differences in perceived social competence, especially if we chose to use multiple raters with different backgrounds. Based on our previous research (Junttila, Voeten, Kaukiainen, & Vauras, 2006; Junttila, Vauras, & Laakkonen, in press) we argue that there exist rater-related differences between genders and between learning disability and other students. Therefore our aim in this presentation is to deepen our understanding of the possible factors explaining these differences in the perceived social competence of school-aged children. 

Method

The 15 elementary schools participating in this project were from urban, small town, and rural communities in southern Finland. The sample (n=985) consisted of two consecutive cohorts of fourth-grade (mean age 10 years and 3 months) students from these mainstream and special education schools, including their teachers and parents.

Children’s social competence was rated using Multisource Assessment of Children’s Social Competence Scale (Junttila et al., 2006). The scale includes four factors of social competence: Co-operating skills (e.g. “effectively participates to group activities”) and Empathy (e.g. “is sensitive to the feelings of others”) to assess the prosocial dimension, and Impulsivity (e.g. “has a short fuse”) and Disruptiveness (e.g. “argues and quarrels with peers”) to assess the antisocial dimension of social competence.  The dimensions were evaluated by the children themselves, their peers, teachers, and parents.

Reading ability was assessed using a Finnish Standardized Reading Test for elementary school children (Lindeman, 1998). In the decoding test, the children’s task was to identify as many words in long continuous word chains as they could within a period of 3 minutes and 30 seconds. In the reading comprehension test, the children were given two narrative texts and twelve multiple-choice questions about the texts they had read.

The mathematical skills were assessed on two time-limited tests. One (Räsänen, 2004) assessed how well children understood number relationships, and the other assessed how well the children knew the basic arithmetic operations (Räsänen & Koponen, 2005).

Results

In our study the social competence of low-achieving children was rated lower than the social competence of average- or high-achieving children, which has also been reported in the meta-analytic studies by Swanson and Malone (1992), Kavale and Forness (1996), and Nowicki (2003) . The differences between high- and low-achieving children were all statistically significant, indicating that the low-achievers showed less co-operating skills and empathy behaviour and more impulsive and disruptive behaviour than high-achievers. As can be seen in Figure 1 the differences were especially large concerning the ratings of teachers. For example, the difference in low- and high-achieving children's co-operating skills was more than five times larger concerning the teacher- than the self- and parent ratings.

Figure 1: Differences between self, peer, parent,  and teacher in ratings of four aspects of social competence  (about here)

The ratings of children’s social competence seem to be strongly related to the rater. For a comprehensive view of a child’s social competence, we have to take all four rater perspectives into account. When studying social competence, a careful choice must be made as to which of the perspectives are relevant for the given social context and the purpose of the study.

But what makes the differences in teacher ratings of differently achieving students so high? School culture and the dominant norms and values defining the correct behaviours are important for understanding social competence in the classrooms. Earlier studies suggest that both rater culture and rated culture specific differences may play a role (see Mpofu, Thomas, & Chan, 2004). The teacher ratings of low-achieving students were particularly low. Could it be that teacher use prejudicial expectations while rating students’ social competence? This and other possible cultural and rater-based explanations for the divergences and their effects on students’ socio-emotional well-being, learning opportunities and intervention designs are discussed in the presentation.

Next to the relationship between classroom characteristics and achievement, there always has been an interest in classroom characteristics and behavior, as the outcome of schooling and as an intermediate variable. Class size is one of these characteristics. Finn e.a. (2003) conclude that students in small classes display less disruptive and more prosocial behavior than do students in large classes.

The research also suggests that teachers get to know each student more intimately in small classes and their tolerance for a broader range of student behaviors is increased. They discuss two theoretical perspectives to support why small classes work: ‘visibility of the individual’ and ‘sense of belonging’. Several questions remain on differential effects and mechanisms to explain why small classes affect students’ academic and social behavior. To adresses these issues a large scale study has been started in primary education in the Netherlands. In this study three cohorts of students are followed for a period of 3 years (age 4-8, 200 schools). Their behavior was measured once a year, their teachers were questioned and students and teachers were observed during the lessons.

Preliminary analyses show that extraversion is negatively correlated to agreeableness and emotional stability. In smaller classes children are rated more extravert and less agreeable and emotional stable. This is in line with other research that indicate that children in smaller groups are more on the ‘firing line’, they have to be attentive, children are more willing to approach the teacher and teachers tolerate more noise. On the other hand this could indicate that in larger groups children receive less space and opportunity to act out. Teachers might interpret this as more agreeable and emotionally more stable. Further analyses need to be executed on the development of social behavior from grade 1 to 4 and the differential effects.

Introduction

Next to the relationship between classroom characteristics and achievement, there always has been an interest in classroom characteristics and behavior, as the outcome of schooling and as an intermediate variable. Class size is one of these characteristics. Finn e.a. (2003) conclude that students in small classes display less disruptive and more prosocial behavior than do students in large classes. The research also suggests that teachers get to know each student more intimately in small classes and their tolerance for a broader range of student behaviors is increased. They discuss two theoretical perspectives to support why small classes work: ‘visibility of the individual’ and ‘sense of belonging’. Several questions remain on differential effects and mechanisms to explain why small classes affect students’ academic and social behavior.

Method

To address these issues a large scale study has been started in primary education in the Netherlands. In this study three cohorts of students are followed for a period of three years (age 4-8, 200 schools). The cohorts contain pupils from various socio-economic and ethnic backgrounds and therefore classes considerably differed with respect to class compositions. Students’ behavior was measured once a year, their teachers completed a questionnaire and students and teachers were observed during the lessons. Behavior was assessed by using a valid and reliable teacher rating scale, the School Behavior Checklist – Revised (SCHOBL-R), which was developed by Resing et al. (1999). It consisted of 52 items of a semantic differential type, that the teacher had to score for each of his or her students and results in scores on four factors: extraversion, agreeableness, attitude towards schoolwork and emotional stability. According to Resing et al (1999) the factorial structure of the scales remained stable over age as well as gender and ethnicity.    

Results

Our preliminary analyses show that the low socio-economic pupils and the pupils with a 'non-Dutch' background differ from the pupils with a 'Dutch', high/middle class background with regard to their social behavior. Pupils with a high socio-economic status have a better attitude towards schoolwork, are more extravert and more agreeable. National pupils, with a low socio-economic status have the lowest attitude towards schoolwork, while immigrant pupils (pupils who were born abroad or who have at least one parent who was born abroad) with a low socio-economic status are rated lowest on extraversion and highest on emotional stability. Moreover class size correlates with the pupil population in the school: the more low socio-economic status and ethnic pupils, the smaller the class size. In smaller classes children are – in general - rated as being more extravert and less agreeable and emotionally stable. This is in line with other research that indicates that children in smaller groups are more on the ‘firing line’, they have to be attentive, children are more willing to approach the teacher and teachers tolerate more noise. On the other hand this could indicate that in larger groups children receive less space and opportunity to act out. Teachers might interpret this as more agreeable and emotionally more stable.

Though Resing et al (1999) do not report significant age-effects, preliminary results indicate some development in social behavior during the first four years of schooling. Pupils tend to be rated most extravert in grade 2 (5-year-olds). Their attitude towards schoolwork is the highest in this grade and decreases when they become older. Further analyses need to be executed on the development of social behavior from grade 1 to 4 for pupils from different socio-economic and ethnic background. Moreover, analyses on differential effects of class size on social behavior for pupils with different socio-economic and ethnic background, taking into account the class composition, are being executed.

Although much research in the past decade has concentrated on peer relations, this research has mostly focused on peer relations in the form of (best and/or reciprocated) friendships and sociometric status. Research into the broader form of peer relations, social networks, has been relatively sparse, mainly due to the statistical problems surrounding the analysis of social network data. These social networks seem important to study, because they are the prime setting where socialization can occur and social competence can develop. Presently, new techniques are available, both for network identification and analysis of network data, that can be used to study interesting phenomena such as which factors govern selection of peers for social networks.

It is well known that ethnic background is one of the selection factors for friendships during child and adolescent years. The question is whether this selection effect occurs as strongly within social networks. This was explored in a Dutch sample of 8^th graders.

Ethnic identity played an important role in the selection of peers for social networks. However, a lot of the networks were multi-ethnic in make-up, supporting a view that integration does take place in Dutch schools during peer interactions.

Introduction

It is well known that ethnic background is one of the selection factors for friendships during child and adolescent years. The question is whether this selection effect occurs as strongly within social networks.

Social competence and social cohesion are keywords in research on adolescents today, even though there are many different opinions on what those terms exactly stand for. Social cohesion can be thought of as the ‘tightness’ of a group of people: are these people interconnected, is there a feeling that they belong to this group? In school settings, this translates to (enough) social interactions with peers, perhaps combined with a sense of belonging to a group or network of the peers, but most ideally also a sense of school belonging. Even though groupings of students are normal in education, for instance in the practice of classrooms or homerooms, this does not automatically mean that these groups are socially cohesive: we assume that students interact with each other daily, but this does not automatically imply that the peer relations are of such a nature that we can speak of a socially cohesive group.

Although much research in the past decade has concentrated on peer relations, this research has mostly focused on peer relations in the form of (best and/or reciprocated) friendships and sociometric status. Research into the broader form of peer relations, social networks, has been relatively sparse, mainly due to the statistical problems surrounding the analysis of social network data. These social networks seem important to study, because they are the prime setting where socialization can occur and social competence can develop. Presently, new techniques are available, both for network identification and analysis of network data, that can be used to study interesting phenomena such as which factors govern selection of peers for social networks.

It is well known that ethnic background is one of the selection factors for friendships during child and adolescent years. The question is whether this selection effect occurs as strongly within social networks. Dutch governmental educational policy is directed at integrating children at all ages in multicultural schools. One of the aims of multicultural schools is to form opportunities for students to interact with other students from different backgrounds. With regard to best friends, this interaction does not seem to take place: there are still relatively few interethnic pairings of friends. However, it is possible that interethnic contact does take place, but within social networks rather than the more intimate form of peer relations that friendship is. Whether or not classrooms are socially cohesive, is in part decided by the interconnectedness of these students: even though teacher might perceive a class as a (tight) group, if there is no interaction between different groups of students, there is no real social cohesion.

Method

To test which role ethnic background plays in the formation of social networks, we compared the ethnic make-up of naturally occurring peer groups in a selection of 8^th grade classrooms. These peer groups or social networks were identified by using each participant as an observer of the whole classroom. Each participant was asked to describe, within the classroom, “who hangs out with whom”. The responses were transcribed and, using the Social Composite Map procedure of Kindermann and Cairns (Kindermann, 1993) combined to form a Social Composite Map of the classroom. On this map, social networks are visible within the classroom, and these networks were used for the further analyses.

Secondly, both the ‘objective’ and ‘subjective’ethnic identity of the participants was assessed. Objective identity was based on the birthland of both parents and the child itself: if one of them was born in a country outside of the Netherlands, that county of origin was used to assign the identity of the child. Secondly, the child was asked how he or she identified themselves. This was labeled the subjective identity. Both measures were used in the analysis.

Results

Ethnic identity does seem to play an important role in the selection of peers for social networks. However, a lot of the networks are multi-ethnic in make-up, supporting a view that integration does take place in Dutch schools during peer interactions, albeit in more general peer interactions than in (best) friendships.