Revista Electrónica de Investigación Educativa


Vol. 6, No. 2, 2004

Psychometric Properties of a Social
Abilities Evaluation Scale: C-scale

María José Rabazo Méndez   (*)
mjrabazo@unex.es

María Isabel Fajardo Caldera   (*)
mifajardo@unex.es

*  Departamento de Psicología y Sociología de la Educación
Facultad de Educación

Universidad de Extremadura

C/ Luis Álvarez Lencero, 12, 6º A
Badajoz, 06011, España

(Received: March 9, 2004; accepted for publishing: May 2, 2004)

 

Abstract

In this paper the psychometric data of a scale of evaluation of the social abilities (social expression and motor skills) are presented. The scale was constructed to investigate social competence and antisocial conduct in adolescence. The process of the scale’s construction is explained; and the data on its internal consistency, its test-retest reliability, and its concurrent validity are presented. The scale was filled out by 325 teachers from different schools in the city of Badajoz (Spain); however, 20 scales were rejected because they were not properly completed. As well, the scale was analyzed, giving as a result 9 factors that explain 59.4% of the total variance. With the results obtained, the position of situational specificity with regard to social abilities is supported.

Key words: Social competencies, psychometrics, behavior rating scales.

 

Introduction

The issue of the evaluation of social skills has always been very controversial; we could say that there are as many assessment tools as approaches to interpersonal behavior. Perhaps the greatest difficulty lies in the complex nature of interpersonal relationships and lack of agreement on:

  • What constitutes a socially-skilled behavior.
  • The types of conduct included the construct of social skills itself.
  • A significant external criterion for validating the assessment procedures (Caballo, 1993).

Despite these drawbacks, the assessment of social skills is a condition necessary for research and intervention, both clinical and psychoeducational, in the field of social inadequacy, since it seems that most researchers defend the multidimensionality and the situational specificity of the construct.

Evaluation by others is a technique for collecting information on the behavior of a particular subject in the real environment, with the observer-evaluators as part of the subject’s social context.

 

I. Construction of the scale

This C-scale of social abilities, behavioral type, was created to develop a study on social competency and antisocial conduct in adolescence. Echoing the recommendations of Galassi and Galassi (1979), there was made an inventory of specific social skills, composed of subscales (relevant to the population of interest), validated against specific behavioral criteria.

For producing this scale, the following sources were used:

  • Scientific literature on the subject (through the principal psychology databases: PsycINFO, PSICODOC, E-PSYKE, CINDOC, Psycho-Search, TESEO and REDINET.
  • Educational proposal for interpersonal relations through the new educational curriculum (Organic Law 1 / 1990).
  • Analysis of socio-scholastic competency of elementary-school students in the province of Badajoz by means of Memoirs of Teaching Practices developed by the students of the Bajadoz School of Education during three years (1999-2003).

One of the sections on the Memoirs of Teaching Practices is the “Description of the student’s behavioral characteristics”. This section, as its name says, explains how each of the school’s student teachers, after having observed of a group of students during their practice, describes in detail each one. This student teacher is a natural observer of the teacher/student and student/student interactions developed both in the classroom and on the playgrounds. Some student teachers use the technique of the sociogram to analyze the relational constellation inside the classroom and on the playground. Farther along there will be described the behaviors characteristic of each sociometric group. For example:

  • Popular: chosen by “X” percentage of her* peers, cooperates with others, yields her material to others when they need it, and so on.
  • Withdrawn: Not chosen by his peers to conduct classroom activities or to play in the yard; goes to the board voluntarily; when asked about some topic in class, lowers his head and stammers, etc.
  • Aggressive: she is rejected by a very high percentage of peers, annoys them, is disrespectful to the teacher, does not respect the rules of the games, etc.

Other student teachers do not use the sociogram technique, and only describe the behavior of some children who are not adapted to the school context. They write, for example, “The child ‘x’ does not bring the material to class. He is often absent without justification, interrupts others when they are speaking, bothers others when they are working...”

Once all the information was collected, all the behaviors from the sociograms and descriptions were written down, both those that facilitated interaction, and those that inhibited it. Later, those behaviors which were very similar were eliminated, as were those whose frequency of occurrence was minimal, and these modes of conduct were encoded in three broad categories.

  • Aggression. In this category were included behaviors that were related to disruption, noncompliance with rules and regulations, physical and verbal aggression, challenging negativism, etc.
  • Isolation. In this category were included those behaviors related to possible avoidance behavior and anxiety (biting fingernails while at the blackboard, wringing of hands while talking, hanging the head, etc.)
  • Assertiveness. This category is made up of those behaviors facilitating social interaction (participation in games and teamwork), conversational skills (asking questions, asking permission to speak, speaking correctly in front of an audience, etc.), friendly relations (making friends, asking forgiveness, asking for favors, paying compliments, etc.)

Thus we arrived at the first version of the scale on the observable dimension of the child’s behavior in class; the scale consisted of 90 items.

In a second stage, bearing in mind a bibliographic review of the issues of social competence, social skills and the educational proposal of competency for interpersonal relationships through the new educational curriculum, we grouped our behaviors as they contained implicitly a verbal component, a non-verbal one, or a paralinguistic one.

Finally, guided by the contents that Monjas (1993) includes in his program of social skills aimed at children and young people, we made the C-scale presented in Annex I.
This scale includes, at the end, a section where teachers are asked to make a general assessment of the student she has been evaluating, according to the following criteria:

As can be observed, this type of scale falls within the category of rating scales, in which the tutor must rate certain behaviors of his students observed over time. It is a strategy of indirect evaluation. This type of scale has certain advantages, such as: the rapidity with which it can be answered, and which allows one to obtain data from very diverse situations in which the student manifests himself spontaneously. In spite of the advantage of its ecological validity, we are conscious of the limitation of an informational bias which supposes the evaluation of an observer, and because of this, each student was evaluated by two different teachers: the tutor and a teacher/specialist (in English, physical education, or music).

This instrument (C-scale) consists of five Likert scales with three response alternatives: 3 = Always, 2 = Sometimes, 1 = Never, each of which we believe to measure a dimension of social competence.

 

II. Psychometric analysis of the C-scale

Given the multidimensionality of the social skills and the independence of these dimensions between themselves, we believed it would be useful to have psychometric data regarding them. While an individual may be relatively skilled in one class of behaviors, he may have major problems in another. In these cases, data concerning each of these different dimensions would be most useful.

2.1. Procedure

Teachers filled out the scales for observation of students in November, when they had a more or less reliable knowledge of the group.1 In the month of March they completed the scale again.

2.2. Results

The mean and standard deviation of the C-scale with N = 295 subjects were 74.37 and 17.31, respectively. Cronbach’s for internal consistency was 0.9484. The test-retest reliability (with an interval of four months between the first and the second time that the scale was done) for N = 70, was 0.90. Concurrent validity (with the Peer Questionnaire taken as a criterion) with N = 50 was 0.86.

We carried out the factor analysis of the scale with N = 295 subjects. We used SPSS for Windows, version 6.0, and used the varimax rotation of principal components with orthogonal dimensions.

In the first stage there were obtained nine factors, with an eigenvalue 1.0 that explained 59.4% of the total variance. Table I shows the analytic-factorial structure of the C-scale. It was considered that an item saturated significantly in a factor if its weight were at least in that factor, which is substantially correct for purposes of interpretation (Stevens, 1986).

Table I. Factorial analysis of the C-scale

The nine factors obtained in the factor analysis of the C-scale approach the five dimensions the C-scale hypothetically evaluated. The factor Self-control explains 34.1% of the variance, and is centered on the expression of feelings and opinions in interpersonal conflictive situations. Factor II is conversational skills, and explains 5.3% of the variance. Factor III, Skills of cooperation and assistance during teamwork explains 3.7%. Factor IV, Defending one’s rights is an expression of discomfort and displeasure, and explains 3.2%. Factors V, VI, IX have been annulled on the grounds that their items saturated in more than one factor. With regard to factors VII and VIII, we have considered it advisable to regroup them, because they are both related with basic skills for socio-scholastic interaction. The items to which we refer below did not participate in any empirical factor, and did participate in the rational factors. We therefore decided to leave them included in the factor called basic skills for socio-scholastic interaction:

21. He presents himself as a nice guy with an easy smile.
34. The child respects the center’s facilities.
29. When she has to ask for something, she says “please”, and then says, “Thank you”.

The definitive C-scale consists of five factors, which are:

  • Factor I. Self-control skills (expression of feelings and opinions in conflictive interpersonal situations).
  • Factor II: Conversational skills.
  • Factor III: Skills of cooperation and help during work.
  • Factor IV: Skills related to the defense of personal rights.
  • Factor V: Basic skills for socio-scholastic interaction.

Then we obtained the mean and standard deviation for each of the factors (see Table II).

Table II. Means and typical deviations of the C-scale

Finally, to determine and analyze the possible differences between adaptive, aggressive and withdrawn groups (according to the teacher’s overall assessment) on the implementation of each of the dimensions of the social skills of a behavioral nature, we obtained a series of ANOVAS, where the independent variable was always the teacher’s assessment on three levels: adaptive, withdrawn, aggressive; and the dependent variable was the mean obtained in each dimension of social skills.

The hypothesis from which we began was as follows:

There are statistically-significant differences between the three groups: adaptive, withdrawn, aggressive, in all dimensions of social skills, the adaptive group being that which obtained the highest scores, and the aggressive group, the lowest.

Table III reflects a summary of the various analyses of variance of a single factor, which we have done for that effect.

Table III. Means pertaining to the dimensions of the social skills
for the groups adaptive, withdrawn and aggressive

From observations of the coefficients and their statistical significance in Table IV, we deduced that the hypothesis is fully confirmed, as there are significant statistical differences among the three groups of overall assessment by the teacher: adaptive, withdrawn, and aggressive, in all dimensions of social skills, with a level of significance by p < 0.0000*** for all the dimensions. Our hypothesis is therefore fully confirmed.

We next present the analysis of the group-to-group comparison, for which we have used the LSD Bonferroni modified multiple range test with a significance level of 0.05 and a critical range of 3.40.

Table IV. Group-to-group contrasts
for all the dimensions of social skills

A close look at Table IV shows that there are significant differences when contrasting all groups and dimensions, except the comparison between the adaptive group and the withdrawn group in the self-control dimension; these have proved non-significant.

 

III. Conclusions

Psychometric data obtained in this study on the C-scale provide suitable reliability and validity for the use of the scale, both at a research and a psychopedagogical level.

The results of this work also support the concept of the multidimensionality of the construct of social skills. Many of the dimensions we found here have also been obtained in other studies, other social skills inventories and with other populations (e.g., Monjas, 1995; Caballo and Buela, 1988; Caballo, 1993).

We can say that there is a whole set of dimensions that appear as clear components of the construct of social skills and behaviors that are important in our social life, especially in the child/youth population in the school context. The initiation of relationships, expression of feelings, defense of personal rights and confrontation with authority, are some examples.

However, this C-scale fails to explain 30% of the variance. Something similar happens in the factor analysis in other scales such as Caballo’s EMES-M (1993).

 

References

Caballo, V. E. (1982). Los componentes conductuales de la conducta asertiva. Revista de Psicología General y Aplicada, 37 (3), 473-486.

Caballo, V. E. (1983). Asertividad: definiciones y dimensiones. Estudios de Psicología, 13, 52-62.

Caballo, V. & Buela, G. (1988). Factor analyzing the College Self-Expression Scale with a Spanish population. Psychological Reports, 63, 503-507.

Caballo, V. & Buela, G. (1989). Diferencias conductuales, cognoscitivas y emocionales entre sujetos de alta y baja habilidad social. Revista de Análisis del comportamiento, 4, 1-19.

Caballo, V. & Carrobles, A. (1988). Comparación de la efectividad de diferentes programas de entrenamiento en habilidades sociales. Revista Española de Terapia del Comportamiento, 6 (2), 93-114.

Caballo, V. E. (1993). Manual de evaluación y entrenamiento de las habilidades sociales. Madrid: Siglo XXI.

Caballo, V. E., (Comp.). (1998). Manual de técnicas de terapia y modificación de la conducta. Madrid: Siglo XXI.

Universidad de Badajoz. (2003). Memorias sobre las prácticas de enseñanza. Unpublished document, Universidad de Badajoz, Facultad de Educación.

Galassi, J. P. & Galassi, M. D. (1979). A comparision of the factor structure of an assertion scale across sex and population. Behavior Therapy, 10, 117-129.

Ley Orgánica 1/1990, October 3, 1990, de Ordenación General del Sistema Educativo. Boletín Oficial del Estado No. 238 (October 4, 1990).

Lowe, M. R. (1985). Psychometric evaluation of the social performance survey schedule: Reliability and validity of the positive behavior subscale. Behaviour Modification, 9, 193-210.

Monjas, I. (1993). Programa de entrenamiento en habilidades de interacción social. PEHIS. Salamanca: Trilce.

Monjas, I. (1995). La educación del alumnado con necesidades educativas especiales en la educación secundaria obligatoria. Un difícil y complejo reto. Siglo Cero, 26 (4), 5-23.

Stevens, J. (1986). Applied multivariate statistics for the social sciences. Hillsdale, NJ: Lawrence Eerlbaum.

Translator: Lessie Evona York-Weatherman

UABC Mexicali

1In Spain the school year runs from September to June.

Please cite the source as:

Rabazo, M. J. & Fajardo, M. I. (2004). Psychometric properties of a social abilities evaluation scale: C-scale. Revista Electrónica de Investigación Educativa, 6 (2). Retrieved month day, year, from: http://redie.ens.uabc.mx/vol6no2/contents-rabazo.html