Use of Item Response Theory to Validate Cyberbullying Sensibility Scale for University Students

Abstract

A thirteen-item cyberbullying sensibility scale (CSS), developed by Tannkulu, Kmay, and Ancak (2013) and extensively used by researchers, was used to measure the cyberbullying sensibility levels of high school students. Unlike other similar concepts, such as cyberbullying and cyber victimization, there are no scales developed to measure the cyberbullying sensibility among university students. In this study, the data obtained from 727 university students were analyzed based on item response theory (IRT) techniques, and psychometric evidences were obtained to evaluate whether it is appropriate to use the scale on the university students. Accordingly, a parameterization of CSS items was performed by using the graded response model. Using the discrimination parameters and item fit statistics, some items were removed from the original scale and a seven-item CSS version was developed since preliminary exploratory and confirmatory factor analyses provide inadequate evidence for the validity of a one-dimensional structure of cyberbullying sensibility. However, an IRT-based item removal process yielded an acceptable improvement. In this way, despite the six items being removed from the original CSS form, the scale retained 64% of the information it provided. The reliability values computed based on the classical approach and IRT were above .8 after the item elimination process with only a minor drop. With the validation process, the CSS will be a valuable measurement tool to determine the level of cvberbullying sensibility among university students and allow academicians to conduct research with this population.