Efficacy of ordinal alpha reliability method using teacher literacy and practice of school-based assessment data

Abstract

The credibility of research findings depends to a large extent on the quality of the instruments used to measure the variables involved. As important as reliability index of an instrument is, it is disappointing that many researchers use inappropriate methods, for instance, the use of Cronbach reliability method instead of Ordinal Alpha to establish the reliability index of an ordinal scale, which is inappropriate. A study that will establish the efficacy of the ordinal alpha method and also correct the error of misuse of Cronbach alpha method becomes imperative. Hence, the study adopted Survey approach. Teachers in the basic 1-9 classes in Bomadi, Patani and Ughelli local government areas in Delta State were sampled to complete the instrument. Completion and return rate was 84%. The results showed that estimated reliability index using Cronbach alpha and ordinal alpha methods were 0.64 and 0.71 respectively. Eleven items correlated negatively with the construct when Cronbach alpha method was used; whereas only 3 items fell into this category, when Ordinal alpha method was used. The use Qfordinal alpha leads to higher reliability index and retention of more items. Hence, scale developers and researchers should employ ordinal alpha method anytime ordinal scale is being constructed.