Abstract:
This study aims to address reverse scored item and the number of response categories problems in Likert-type scales. The Fear of Negative Evaluation Scale (FNES) and the Oxford Happiness Questionnaire (OHQ) were used as data collection tools. The data of the study were analyzed according to the Rasch model. It was found that the observed and expected test characteristic curves were largely overlapped, each of the three rating scales worked effectively, and the differences between response categories could be distinguished successfully by the participants in straightforward items. On the other hand, it was determined that there were significant differences between the observed and expected test characteristic curves in reverse scored items. According to the results the participants could not distinguish the response categories of the reverse scored items at three, five and seven-point rating versions of both scales. Hence, the reverse scored items were removed from the data file, and the analysis was repeated. The results revealed that item discrimination, reliability coefficients for person facet, separation ratios and Chi square values calculated for the facets of person and items were higher in five-pointed rating compared to three and seven pointed rating. Based on these results it can be said that the scale categories in reverse scored items could not be discriminated by responders at all type of rating, and that reverse scored items did not measure the same latent structure as straightforward items did.
