Validation of Persian Translation of the Intercultural Sensitivity Scale Using Rasch Model

Abstract

This study subjected the Intercultural Sensitivity scale consisting of 24 items to Rasch analysis in a sample of 200 participants. The scale was translated into Persian and administered to graduate and undergraduate university students. Analysis of data showed that there were no misfitting items. Furthermore, no item manifested gender differential item functioning (DIF). All the thresholds were ordered and respondents could distinguish well between categories of the scale. The results of principal component analysis (PCA) of standardized residuals revealed that there were two contrasts with eigenvalues above two. Deleting positively and negatively loading items separately did not improve model fit. Thus, the content of items was investigated and it became clear that most of the positively loading items covered those items which have negative content and conversely, negatively loading items encompassed the items which have positive content. This brings about two psychometric dimensions in this scale. 1) Wording of the items revealed that the existence of items with negative wording in the scale results in statistical artifacts, and 2) The secondary dimension here could be interpreted to be an artifact of the wording. It was concluded that the scale is unidimensional and enjoys acceptable psychometric properties.