DEVELOPMENT AND VALIDATION OF A SELF-ESTEEM SCALE

Abstract

The construct of Self-esteem has been explicated within the indigenous social context through development and validation of a Self-esteem Scale. Firstly, the dimensionality of the Self-esteem construct and internal consistency/reliability of the selfesteem scale were ascertained For that, an item pool was developed from qualitative data obtained from two pilot studies and the translations of foul' existing Self-esteem Scales. After an extensive scrutiny and evaluation of the items, 72 items, most relevant to the construct and expressing evaluations of a global self and its various aspects, were phrased in self-reported statements with a five-point scale. This scale was given to a sample of 300 participants (150 boys and 150 girls). The Principal Componellt Factor Analysis revealed that most of the items of Self-esteem Scale were positively loaded on first four factors that explained 22.5% of the total variance. The eigenvalues for these factors were 7.4, 4.0, 2.5 and 2.2, respectively. The factor solution was rotated to get clear and intelpretable dimensions of the self-esteem. The contents of the items with >.30 factor loadings on the four factors in rotated solution were examined in detail. The rotated factor solution was found to be more meaningful in terms of the theoretical interpretation of its factors. Following the criteria of Kline (1986), only those items were selectedfor further examination which had >. 30 factor loading. All examination of the contents of these items yielded four factors which were labelled as Self-Acceptance, SelfCompetence, SOCial, and Physical Self-Acceptance and, Academic oS'ef[-Competence. There were 11 items with >.30 factor loading on first factor, i.e., Self-Acceptance. On secondfactor of Self-Competence 6 items were having >.30 factor loadings and on third factor, Social and Physical Self- Acceptance, 7 items were found to be haVing >.30 factor loading, whereas on fourth factor of Academic Self-Competence, there were five items which were having >.30 factor loadings. The Self-Esteem Scale was reduced to only those 29 items which received high factor loadings on four dimensions of self-esteem. These constituted the four subscales of the Self-Esteem Scale. These selected 29 items were positively correlated with the total score with an average correlation of. 42. The Self-Esteem Scale (29 items) was found to be internally consistent and reliable as indicated by the alpha coefficient value .83 (p<.00). The split-half reliability wasfound IV to be . 72 (p<.00) with Spearman Brown correction. Boys scored higher on the SelfEsteem Scale as compared to girls supporting the hypotheses formulated in this regard. The difference of scores between boys and girls was found to be nonsignificant on the dimension of Academic Self-Competence. In the second phase of the research, five validation studies were carried 01lt to test the validity of Self-Esteem Scale. Study I conducted 011 a sample oj 60 participants tested the concurrent!convergent validity of Self-Esteem Scale hy finding its correlation with Rosenherg (1965) Self-Esteem Scale (r=. 62 p<. 00). The scores of the four suhscales were also positively related with scores oj Rosenberg Scale. Study II was conducted on sample oj 60 participants to test the convergent validity of the suhscale oj Academic SelfCompetence by finding its correlation with scores obtained through Academic SelfConcept Scale developed by Ahmed (1986) and achievement scores obtained ill school examination. The results showed that the scores on the Academic Self-Concept Scale were positively related with scores of the subscale Academic Self-Competence (r=.46, p<.OO) , whereas the positive correlations with the other three subscales were less in magnitude and non-significant. The highly positive correlation coefficient provided the evidence of convergent validity of Academic Self-Competence scale and, non-significant and less positive correlation oj Academic Self-Concept Scale with other suhsea/es indicated the discriminant validity oj these subscales. The correlation hetween the Academic Self-Competence and achievement scores indicated the concurrent validity oj this subscale (r = .29, p<.05). The other three studies were carried out jor construct validation of Self-esteem hy examining its relationship with Anxiety, Delinquency and Depressioll. Study III was conducted on a sample oj 150 participants to explore the relationship between selfesteem and anxiety. High self-esteem and anxiety were found to be negatively related to each other (r=-. 48, p<. 00) and the hypotheses that low self-esteem individuals score high on Anxiety scale (t-value =4.55, dj90, p<.OO) was supported. Study IV was conducted on a sample oj 100 participants to explore the relationship between self-esteem and self reported delinquency. The results showed that there was significant negative correlation hetween high self-esteem and delinquency (I' =-.23,p<.01) and the participants with high v self-esteem scored low all delinquency scale (t-value=2.53 p<. 01), thus indicating that the self-esteem and delinquency are negatively related to each other. The relationship of delinquency with subscales of Self-Competence and Academic Self-Competence was found to be negligible and non-significant. Study V, conducted on a sample of 145 participants, examined the relationship between depression and self-esteem. The analysis of data revealed the negative relationship between high self-esteem and depression (r=.53, p<.OO). The low self-esteem individuals scored high on SSDS and significantly differed from individuals scoring high on Self-esteem Scale (t-value = 7. 50, df=86,p<.00). The findings of the present research have revealed a theoretically intefpretable multidimesiollal structure of self-esteem within an indigenous context. The Self-Esteem Scale, was found to be a valid and reliable measure. The implicationsforfuture research have been discussed with reference to further validation and improvement ill methodology.