Construct validity and reliability in Rosenberg’s self-steem scale for Brazilian older adults who practice physical activities

Abstract

The present research aimed at investigating the construct validity and reliability of the Rosenberg’s Self-esteem Scale (RSES) (1965) for Brazilian older adults. 292 elderly participants (67.54, DP = 6.90 years) who do physical activities in Florianopolis, Santa Catarina, Brazil. The analysis of the data was made in two steps: the analysis of the principal components and analysis of the factor loadings. To examine the construct validity we carried out the factor analysis described in two stages: principal component analysis and factor weights. After, we proceeded with the oblique rotation, using the method Oblimin. To check the reliability of the scale, we used the Cronbach’s alpha. It was observed that there were two factors that together accounted for 51% of the total variance of the results of the scale. The first of these factors solely explained for 38% of the variance, and the second by only 13.4% of the variance. The factor analysis without rotation indicated that the scale can be considered one-dimensional. However, after the rotation, it was found that the negatives are the first factor and the second positive factor. Cronbach’s alpha was established 0,812, considered good. It was observed that the RSES had acceptable psychometric qualities for research on Brazilian elderly who practice physical activities, thus can be considered a valid and reliable instrument for use with this population.