Psychometric evaluation of the Bangla-Translated Rotter’s Internal-External Scale through classical test theory and item response theory

Abstract

There is no psychometric tool to assess locus of control for Bangla-speaking people. Hence, we attempted to translate the 23-item Rotter’s Internal-External scale into Bangla and validate it on Bangladeshi adult participants. In Study 1 (N = 300), we translated the items into Bangla and conducted an exploratory factor analysis, which gave a one-factor solution with 12 items. In Study 2, we conducted a validation study (N = 178) to accumulate evidence on the structural and concurrent validity of the 12-item scale. Structural validity assessed by confirmatory factor analysis yielded the best fit for the one-factor model with 11 items (CFI = 0.98, TLI = 0.97, RMSEA = 0.00). The scale’s significant correlations with Internal Control Index, which is also a measure of locus of control (r = −0.22, p < 0.01), Neuroticism (r = 0.21, p < 0.01) and Openness to Experience (r = −0.22, p < 0.01) demonstrated its satisfactory concurrent validity. Reliability coefficient of this 11-item scale was satisfactory (McDonald’s Omega total = 0.72). The item quality was assessed on the combined samples of Study 1 & 2 (N = 478) using the item response theory (IRT), which showed that the scale covered a sizable range of the underlying locus of control with items varying in difficulty (−1.09–2.79). Item discrimination analysis indicated sufficient discriminating power of the items (0.49–2.21). The test information curve showed the scale’s adequate ability to discriminate between external and internal locus of control. IRT analysis also indicated satisfactory marginal reliability for the scale (0.72). These psychometric properties suggest the usability of the Bangla version of Rotter’s Internal-External scale.