Applying Rasch analysis in refinement and validation of interpersonal skills measure for gifted children

Abstract

BackgroundInterpersonal characteristics of gifted adolescents is important because of the potential influence on individuals’ psychological health and future professional success. Understanding the interpersonal characteristics requires valid and reliable measures. This study attempted to explicitly describe the application of Rasch analysis in the validation and development process of an existing measure of interpersonal skills among gifted adolescent in Ohio.MethodsWe extensively evaluated the psychometric properties of the 40-item scale measuring the interpersonal competence of adolescents among the gifted population in Ohio (N = 127) using Rasch analysis. Multiple aspects of reliability and validity were tested including dimensionality, rating scale functioning, and fit statistics.ResultsThe internal consistency reliability of the scale was confirmed with an adequate fit to the Rasch model. However, the scale demonstrates relatively poor performance in terms of unidimensionality with our sample. Also, the rating scale categories were confusing given that our sample could not effectively distinguish some adjacent categories. Corresponding refinements have been made and the revised scale formed a meaningful linear progression with improved performance on unidimensionality, rating scale functioning, and fit statistics.ConclusionThe study provided evidence that the construct of interpersonal skills is measurable. Based upon the original Interpersonal Competence Questionnaires consists of 40 items, the author extracted and piloted a refined measure consisting of 31 items that performed a meaningful, theoretically consistent linear progression measure that could be used to measure the level of interpersonal skills of gifted children.