ARC Laboratory Sharing

The Rasch rating scale (Andrich, 1978) has been widely applied in different testing situations for decades when more and more groups get to learn and use it. However, its depth of understanding has not gone in the same great pace of its width of understanding. While more and more researchers adopt using questionnaire and survey items and then analyzing in their research projects, sometimes they did not follow correctly what is put in the literature, in some cases, it is even worse that they have wrongly assumed what was not said in the literature as truth and regard it as a myth, convention or principle which everybody should adhere to. One of such myths is about the disordered thresholds. There have been controversies that reversed delta is taboo so that any disordering should be avoided and eliminated when discovered. This article examines a variety of derivations of the model illuminate the situation. The authors showed that the so-called parameter disorder and order of the response categories are separate phenomenon: reversed deltas are not necessarily evidence of a problem. One situation is that even though an item shows a disordering of deltas, it still fits the particle credit model. However, it should be noted that disordered parameter estimates may indicate that the item is not functioning as well as it was intended so it is worth reviewing these items regarding its functioning.