In the last few decades there has been renewed interest in the study of self-concept and self-esteem following the contributions of Cognitive Psychology (Markus and Wurf, 1987). Self-esteem, unlike self-concept, refers to evaluative aspects and expresses one’s self-concept in subjective and evaluative terms (Musitu, Román and Gracia, 1988; Lila, 1995). Nevertheless, in practice it is difficult to differentiate the two concepts, and they are often defined in a similar way (Shavelson, Hubner and Stanton, 1976). The term self-esteem is, however, more common in the fields of education and health (Lila, 1991).

Also, within the theoretical models of psychosocial intervention, focused on the empowerment of resources that can improve the person’s psychosocial adjustment and well-being, the term self-esteem is employed. These models emphasise the empowerment of resources such as self-esteem or social support as a more viable means of psychosocial intervention than concentrating simply on repairing deficits. In this line, Rappaport’s (1981, 1987) "empowerment" model highlights the importance of acquiring new resources and abilities that allow the subject increased control or power in his/her context as a means of improving well-being. Similarly, Lin and Ensel’s (1989) model proposes that there are resources and stressors in three areas: social, psychological and physiological. These resources and stressors affect the individual’s well-being, which will be greater when resources outweigh stressors. As far as these resources are concerned, self-esteem is repeatedly identified as one of the most significant, and in fact, the link between self-esteem and psychosocial adjustment is one of the principal reasons for the interest it generates (Herrero, 1994; Gracia, Herrero and Musitu, 1995; Cava, 1995).

Furthermore, any consideration of the empowerment of self-esteem necessarily involves examination of its origin, which can be found in social interaction processes. This approach was first employed by the Symbolic Interactionists (Cooley, 1902; Mead, 1934), and still retains total validity. "Significant others", that is, those that are important for the subject, through the image of him/her that they reflect and through direct interaction, are those that influence the shaping of one’s self-concept. According to Banaji and Prentice (1994), the origin of these self-representations can be found in self-perception processes, social comparisons and direct interaction. Thus, it would appear crucial to consider the social origin of self-concept. Nevertheless, and as previously suggested by the Symbolic Interactionists (Cooley, 1902), not all the people with whom we come into contact are equally relevant to the shaping of our self-concept: only those that are especially important to us. Among the most significant people are the closest family members, peers and teachers.

As far as the family is concerned, this usually constitutes the first socialisation context for the child and the first environment in which s/he begins to develop his/her identity. Indeed, many studies establish relationships between the way a child feels s/he is perceived by his/her parents and his/her self-perceptions (Cook and Douglas, 1998). Moreover, the family also influences the child’s subsequent social relationships and his/her adjustment to school (Larose and Boivin, 1998; Shumow, Vandell and Posner, 1998). However, the importance of the family should not lead us to overlook the existence of other contexts of equal significance in the processes of socialisation and individuation of the child. In particular, school represents an especially relevant context (Cubero and Moreno, 1990), in which children continue the process of development of their self-concept and self-esteem, at the same time as extending their sphere of social relationships. On the one hand, a significant aspect of this context is the relationship with the teacher, since for the majority of pupils s/he is a "significant other" and his/her feedback influences the child’s self-esteem, performance and, possibly, his/her social relationships with peers (Hargreaves, 1978; Machargo, 1991).

On the other hand, the child’s relationship with his/her classmates is another essential element in classroom interaction processes. Classmates, or peers, have similar levels of maturity and development, and their influence on both socialisation and the development of identity is becoming increasingly recognised (Hartup, 1985). Thus, for example, a significant relationship has been established between peer rejection in childhood and low self-esteem (Musitu et al., 1980; Cava, 1998). Children rejected by their peers perceive this rejection as a powerful stressor, interiorise their social status as part of their identity and modify their behaviours in relation to their peer group (Coie, 1990). Moreover, added to these characteristics are the high temporal stability of the "reject" status (Musitu et al., 1984) and the negative consequences of rejection for the child’s psychosocial adjustment, not only in the short term but also in the long term (Kupersmidt, Coie and Dodge, 1990). In fact, all of these circumstances have led to a consideration of children rejected by their peers as "risk group" and, consequently, to the development of different intervention programmes for facilitating their social integration (Parker and Asher, 1987). Nevertheless, such interventions frequently fail to take into account the importance of the peer group in maintaining the status of the rejected child, which may explain the fact that their success has been no more than partial (Hymel, Wagner and Butler, 1990).

The relevance of teacher-pupil and pupil-pupil relationships for the development of the child’s identity underpins to a large extent the "Galatea" intervention programme, and it is the assessment of this programme’s effectiveness that constitutes the aim of this study. Specifically, the objectives of the Galatea programme are the enhancement of self-esteem and the social integration of children with socioaffective difficulties, within the school context. The two objectives are closely related, since, as we have just pointed out, both the social integration of rejected children –in relation to their peer group– and the improvement of pupil-pupil and teacher-pupil relationships have been proposed as fundamental for the enhancement of self-esteem in this context. In designing this intervention programme, in addition to the theoretical approaches previously outlined, the multidimensional nature of self-esteem (Shavelson et al., 1976; Lila, 1995) has also been considered, together with the possibility of its modification, especially in those areas in which the level of certainty of the self-concept we hold is low (Swann, 1987).

Finally, it should also be pointed out that before the definitive design of the Galatea programme a review was made of some of the intervention programmes in our cultural context with similar aims. As a result of this review, some elements considered to be relevant were included in the programme, such as co-operative learning tasks (García, 1989; García, 1991; Díaz-Aguado, 1994), the implementation of the programme by the subjects’ teachers, the incorporation of the programme’s activities in the school curriculum and the prior creation of a favourable climate (Hernández and Aciego de Mendoza, 1990; Hernández and Garcia, 1992; Martínez, 1994; see Cava, 1998, for a more extensive discussion). The review also revealed some shortcomings: many of the works reviewed did not constitute clearly-structured programmes, but rather series of suggestions for possible activities; the multidimensionality of self-esteem is not always taken into account in the design of the programme; and a rigorous assessment of the programme’s effectiveness is not always included.

It is the intention of the Galatea programme to remedy many of these shortcomings, given its solid theoretical base, its consideration of self-esteem as a multidimensional construct, its design geared to integration in the school curriculum and the fact that it has been revised by several teachers in order to ensure its intelligibility and its adaptability to different sociocultural levels. With regard to the assessment of its effectiveness, we have considered variables related to self-esteem, social integration of children rejected by their peers and teacher’s perception of pupils.

The sample was made up of 537 pupils from three state schools in the Valencia Autonomous Region. 52.8% were males and 47.2% females. Mean age ranged from 10 to 16 years, though the majority (86.6%) were aged between 11 and 14. The children were in grades 5 (n=101), 6 (n=156), 7 (n=82) and 8 (n=198) of the basic compulsory education system in force in Spain at the time (E.G.B.), which has since been replaced by a new system.

Of the three schools contributing to the sample, two of them implemented the Galatea programme (n=441), subjects from the third school constituting a control group (n=96). This control group included one class for each educational level. During the development of the programme we worked with the teachers at these schools, of whom there were a total of 21. As far as pupils’ socio-economic level is concerned, this is similar in the three schools, the majority of subjects being medium-low.

For the measurement of self-esteem we used the questionnaire A.F.5 (García and Musitu, 1998). This questionnaire includes 30 items with 100 response possibilities (0-99). The wide range of possibilities allows more precise measurement of self-esteem and of its variations. For each item, subjects indicate the degree to which they agree with the statement expressed in the item (e.g., "I make friends easily" – Item 1). A response of 99 would show a very high degree of agreement with the item, whilst a response of 0 would express total disagreement. Nevertheless, subjects are reminded of the large number of response possibilities available.

The design of the questionnaire was made on the basis of a multidimensional consideration of self-esteem (Shavelson et al., 1976), and therefore included items that allowed the assessment of the subject’s concept and evaluation of him/herself in five aspects of the person and relationships: social, academic, emotional, family and physical. The factorial analysis for checking empirically the theoretical validity of the 5 dimensions was carried out with the computer program SPSS 7.5, extracting the factors using the method of analysis of principal components and applying Kaiser’s oblimin rotation, since the dimensions are related. In this factorial analysis (Cava, 1998), 5 factors were obtained, with 6 items in each, explaining 49% of total variance: academic self-esteem –explaining 19.3% of total variance–, emotional self-esteem –explaining 9.4%–, social self-esteem –explaining 8.5%, family self-esteem –explaining 6.8%–, and physical self-esteem –explaining 5%. The semantic content of the items assigned to each factor coincide with the rational factors, defined through the so-called technique of rational assignment. This instrument was used previously by Lila (1995), though on that occasion the response possibilities for each item were limited to three –"always", "sometimes" and "never". Lila’s study obtained a factorial structure made up of these same 5 factors.

This scale allows us to obtain information about the teacher’s perception of each one of his/her pupils. The scale includes 11 items, one of which represents an estimation of the pupil’s success in compulsory education, with 5 response possibilities, from "totally certain" to "definitely not". The remaining 10 items are to be given scores on a Likert-type scale with 10 response alternatives, from 1 –equivalent to the most unfavourable perception– to 10 –equivalent to the most favourable perception. The items refer to the teacher’s perception of each of his/her pupils in terms of classroom behaviour, preparedness to co-operate, level of effort, performance, and relationship with peers.

The sociometric questionnaire is an instrument that permits us to discover the basic structure of relationships of a group, through the responses of its members about their own preferences and dislikes (Arruga, 1974). This questionnaire is based on the consideration that the best way of finding out about the internal structure of a group is to ask the subjects making up the group directly who they like and who they dislike. The information contributed by all the group members provides an overall image of the group in terms of interpersonal relationships. This kind of questionnaire is not a standardised one; rather, it is prepared by the researcher, following specific criteria for the group whose structure of relationships s/he wishes to discover. In our research we drew up a sociometric questionnaire for pupils aged 10 to 16, following the steps described by Arruga (1974).

The questionnaire was comprised of 8 items, grouped in accordance with two sociometric criteria: "teamwork" and "likeableness". In terms of these two criteria, the pupils were asked to give positive selections –"With whom would you prefer to…?", negative selections –"With whom would you prefer not to...?", perceptions of positive selection –"Who do you think has chosen you?", and perceptions of negative selection –"Who do you think hasn’t chosen you?". The method was nominative, with unlimited possibility of choice and weighting in the order of preference. For the data analysis we used a computer program developed by González (1990) –"SOCIO". Using this programme we obtained the sociometric types of the children –popular, rejected, ignored and average–, the sociometric status index of each pupil and a large number of sociometric values. However, in this article we present only the data relating to two sociometric values: rejection status (Sn) –number of rejections received by group members– and rejection impression (In) –number of rejections group members expect to receive from their classmates. These two values permit us to know the possible variations in the number of rejections and in the perception of rejections of the group. Finally, it should be pointed out that, since the questionnaire is based on two sociometric criteria –"teamwork" and "likeableness"–, the sociometric values referred to above (Sn and In) are obtained according to these two criteria.

This programme was developed on solid theoretical foundations, outlined in the introduction to this article, and on the basis of a review of intervention programmes with similar objectives, mainly from our cultural context, though also from others (Cava, 1998). The programme is comprised of 53 activities grouped in 7 modules. Each module includes a set of activities with a common objective, and it is recommended that the modules be implemented in the sequence indicated.

The first module, called "Preparing the ground: a positive climate", includes 10 activities, whose aim is to foment an atmosphere of confidence and support in the classroom. The second module is called "Personal resources", and its 7 activities have the common objective of focusing on the potential positive aspects of all pupils: resources, qualities, abilities and skills of their own and of their classmates. This module also takes into account pupils’ successes and achievements. The aim of the third module, "My identity", comprising 21 activities, is to improve self-knowledge, assuming the global nature of the module’s defining features (both positive and negative), and to increase acceptance of one’s own body. The fourth module, "My projects", with 3 activities, focuses on pupils’ reflections with regard to their goals and aspirations, and includes some trials with easy-to-achieve goals. The fifth module, also with 3 activities, is called "It’s not always easy", and its objective is the analysis of the impediments, sometimes internal, that make it difficult for us to achieve our objectives. The sixth module, "My friends", made up of 6 activities, refers to peer relationships and friendship. Finally, the seventh module, "My family", with 3 activities, deals with pupils’ relationships with their respective families.

The programme, as presented to the teachers, contains an introduction specifying the conditions for its application and an approximate schedule for the duration of each module (Box I).

The duration of the entire programme may vary from three to five months, depending on the number of activities carried out and the intervals between them. Box I also shows the activities making up each module and the minimum number of activities considered necessary.

The teacher has some degree of freedom to choose within each module those activities s/he considers most adaptable to his/her pupils’ characteristics and his/her own. Nevertheless, it is also possible to carry out all the activities that make up a module, or even to repeat an activity, where considered appropriate. For each activity details are given of its specific objectives, its approximate duration, the materials required by participants, the resources available to the teacher, the group dynamic to employ, the procedure for its development, its adaptability, possible variations and some suggestions to aid the achievement of the objectives.

At the beginning of the school year, teachers from the two schools that were to implement the Galatea programme attended a seminar on self-esteem and interaction processes in the school context. They were also given information on the programme. Fundamental to achieving the objectives of the programme is the involvement of the teachers themselves (rather than external agents) in the implementation of the programme and their active participation. Also of essential importance is prior training of teachers.

Over a period of two months, those teachers that volunteered to implement the programme attended the seminar, which was made up of two parts. The first part (10 hours) involved the analysis of the following topics: self-esteem, teacher-pupil relationships and relationships between pupils. In the second part (6 hours), the content of the intervention programme was described in detail and special emphasis was placed on the importance of the way the activities were developed (even more than their specific content) and on the relevance of the teacher’s attitudes and expectations. This second part made use of some of the activities involved in the programme.

On completion of the seminar, those teachers that were interested in the programme began its application. Prior to starting the activities, pupils filled out the self-esteem and sociometric questionnaires, and teachers the EA-P Scale. These assessments were carried out both in the two schools that constituted the experimental group and in the control group school. The same assessments were carried out a second time at the end of the programme, shortly before the end of the school year. A Nonequivalent Comparison Group Design (Cook and Campbell, 1979) was used. It is considered a quasi-experimental design, since subjects were not assigned randomly to the different groups (Pascual, García and Frías, 1996).

For evaluating the effectiveness of the intervention we compared pre-test and post-test means in the experimental and control groups. Comparison of means was made using the Student t test. Below are presented, first, the results obtained in the experimental group, and second, those for the control group. Table 1 shows the results obtained on comparing pre- and post-test means (using Student t) in the five self-esteem dimensions in the experimental group.

From this table it can be seen that the means of those pupils in whose classrooms the Galatea programme was implemented are superior in the post-test compared to the pre-test. Nevertheless, the difference between means is only significant in two of the five dimensions. Specifically, the increase is statistically significant in family self-esteem (t=-2.13; p=.033) and physical self-esteem (t=-2.72; p=.007). Table II shows the results of comparing means before and after the intervention, in the experimental group, in the variable teacher’s perception of pupils.

In Table II it can be observed that the teacher’s perception of the pupil’s level of co-operation in class (t=-5.68; p=.001), academic performance (t=-2.88; p=.004) and level of acceptance by peers (t=-3.48; p=.001) shows statistically significant differences in the measurements carried out before and after the intervention. In these three dimensions the mean was found to be statistically superior in the second measurement. Thus, there was seen to be a more favourable perception of pupils by the teacher after the implementation of the Galatea programme. Table III shows the results of comparing the means obtained in the two measurements, considering the sociometric values of rejection status (Sn) and rejection impression (In), themselves obtained in accordance with the criteria "teamwork" and "likeableness".

It can be seen from Table III that there are statistically significant differences in the sociometric value of rejection status for both criteria. That is, the number of rejections received by subjects decreases significantly in the post-test (t=3.04; p=.003 for the "teamwork" criterion and t=3.46; p=.001 for the "likeableness" criterion). The rejection impression value, that is, the number of rejections the subjects of the group expect to receive from their peers, also decreases significantly for the "teamwork" criterion (t=2.51; p=.012). No significant decrease was observed, however, in rejection impression for the "likeableness" criterion.

The next table shows the results relating to the comparison of pre- and post-test means in the control group, beginning with the five dimensions of the self-esteem variable (Table IV).

In Table IV it can be seen that none of the differences between the means was statistically significant. Thus, in the control group there is no evidence of (significant) variations in any of the five self-esteem dimensions considered.

As for the differences between means in the four dimensions of teacher’s perception of the pupil, in Table V it can be seen that there are no statistically significant differences. Thus, in the control group there was no significant variation in teachers’ perceptions of their pupils. Finally, the results of comparing the means in the sociometric values of rejection status and rejection impression for the "teamwork" and "likeableness" criteria are shown in Table VI.

From these results it can be appreciated that the control group children present no significant variations in the sociometric values of rejection status and rejection impression. There are no significant differences in the means of these values for either the "teamwork" or the "likeableness" criterion.

Various suggestions have been made on the basis of theoretical models relevant to psychosocial intervention, such as Rappaport’s "empowerment" model (1981, 1987), in relation to the enhancement of intra- and interpersonal resources, and there has been a demand within the school context for instruments that allow teachers to achieve socioaffective goals in formal education. However, up to now there has been a lack of intervention programmes capable of responding to these suggestions and needs. In this article we present the main results with regard to the assessment of the effectiveness of an intervention programme in the school context, the objectives of which are the enhancement of self-esteem and improvement in the social integration of children with socioaffective difficulties.

As regards the results relating to the self-esteem variable, these demonstrate the empowerment of two of its dimensions –family self-esteem and physical self-esteem– in the pupils that took part in the intervention programme. In this line, even though not all the self-esteem dimensions were enhanced by the intervention, it should be pointed out that comparisons of the means pre- and post-intervention were made considering all participating pupils, that is, including those that did not initially present low self-esteem. This circumstance should be taken into account on analysing the dimensions that were enhanced by the programme.

Moreover, this result confirms once more the need to consider self-esteem as a multidimensional construct, as previous researchers have proposed (Shavelson et al., 1976). With regard to the self-esteem dimensions that were enhanced, it would seem to be those dimensions that are in a phase of restructuring for all pupils that are easiest to modify. Pre-adolescence and adolescence are considered as periods in which the bodily changes experienced by subjects lead them to a restructuring of their physical self-concept (Block and Robins, 1993). Likewise, these periods are characterised by a reconsideration of family relationships, since the adolescent is required to take a more active part in family decisions and processes (Fuligni and Eccles, 1993). Thus, it would appear that both self-esteem dimensions –family and physical– are in a phase of reformulation for all the pupils in the experimental group. As far as the possibility of modifying the self-concept is concerned, Swann (1987) postulates in his theory of self-verification that we (subjects) tend to look for feedback that confirms our own view of ourselves, and that we actively resist disconfirmatory feedback. Nevertheless, according to Swann (1987), the degree of certainty and importance of the self-concept constitute decisive elements with regard to its modifiability. Thus, a self-concept that is given less importance or of which the subject is less certain will be more easily modified. The result obtained in the present research may lend some support to this theory, given that those dimensions that are in a period of reformulation were those that were able to be modified.

It would be of interest in future research to study potential variations in self-esteem, subsequent to the implementation of the Galatea programme, considering specifically those children with problems of social integration –rejected children. In these children, variations in self-esteem may possibly cover a greater number of dimensions. In fact, despite the lack of data as such, we did ask the participating teachers for a qualitative assessment in this regard, and their response was that the intervention programme is particularly useful in the case of children with socioaffective problems (Cava, 1998).

With regard to the teacher’s perception of pupils, the results show a more favourable perception in the post-test in the experimental group and no variation in the control group. Specifically, of the four dimensions of teacher’s perception of pupils, three –co-operation, performance and acceptance– present a statistically significant higher mean after the intervention. The implementation of the programme appears to have increased teachers’ knowledge of their pupils in a wider range of aspects –over and above performance and certain disruptive behaviours, which are commonly the basic elements contributing to the formation of teachers’ expectations (Veiga, 1995). Improvements in the perception of performance may also be related to the inclusion in the intervention programme of co-operative learning tasks. This type of task has been linked to improvements in relationships between pupils, in self-esteem and in academic performance (Ovejero, 1990; Díaz-Aguado, 1994). We should also stress the importance of teacher’s perception of better acceptance of pupils by their peers, since teachers tend to be well acquainted with the social reality of the classroom (Taylor, 1989). However, there was no significant pre-programme/post-programme difference in teachers’ estimation of future success in compulsory education. Thus, although teachers manifest more favourable expectations with regard to their pupils, they appear to exhibit a degree of caution in modifying their perception of pupils’ future academic achievement. Modification in this dimension may require more time to elapse

Finally, in the experimental group there was observed a reduction in the number of rejections received by pupils in the two sociometric criteria considered –"teamwork" and "likeableness"– and a decrease in rejection impression in the criterion "teamwork". In the control group there were no significant variations in these sociometric values The reduction in the number of rejections may confirm the viability of attempts to influence the integration of children with peer relationship difficulties, considering their classmates as a fundamental element for social integration. For many years, the predominant models in intervention with rejected children have been the "Social skills deficit model" (Ladd and Mize, 1983) and the "Individual deficit hypothesis" (Asher and Renshaw, 1981). These models assume that deficiencies in social functioning are the principal cause of social difficulties in rejected children, and that modifying these deficiencies will lead to greater acceptance by peers. This approach attributes the majority of the responsibility for rejection to the child him/herself, largely ignoring the group processes that often perpetuate the status of children in the group. However, the results, sometimes less positive than expected, have led to the proposal of other approaches to intervention that include the peer group itself. This is the perspective underlying the intervention programme presented here.

Furthermore, we observed not only a decrease in the number of rejections received but also a reduction in the number of rejections group members expect to receive from their classmates –rejection impression. This result may be of interest in that the anticipation of rejection by peers has been proposed as a factor contributing to the maintenance of rejection status, since such anticipation would lead rejected children to modify their behaviour, show signs of expected rejection and convert their expectations into self-fulfilling prophecies (MacDonald and Cohen, 1995).

The intervention programme presented here –the Galatea programme– would appear, then, to have facilitated pupils’ social integration and to have contributed to enhancing their self-esteem and improving teachers’ expectations towards them. These three variables appear, moreover, to be closely inter-related. Thus, improvements in relationships between pupils and in teacher-pupil relationships may create a context for the development of new and more positive self-concepts. Another advantage of this programme is that it is useful both for children with socioaffective difficulties and for socially well-adjusted children, lending it both therapeutic and preventive validity. Finally, we consider that the programme not only responds to a demand within the school context, but that it can also be adapted to other contexts and other needs, modifying some activities but maintaining its philosophy. Moreover, it is a programme that can be implemented by different kinds of professionals, such as psychologists, educationalists, teachers or social workers, as long as those involved receive prior training in relation to its theoretical foundations and to the attitude and climate necessary for its correct implementation.