معامل ألفا الرتبي : تقدير لمعامل ثبات درجات الاختبار باستخدام البيانات الرتبية

Abstract EN

Test score reliability is an important source of evidence about the validity of the inferences, thus the researchers aspire obtaining the highest reliability coefficients values to increase confidence in the accuracy of the results obtained.

Alpha coefficient is the most widely used method to calculate the reliability coefficient, despite this, there is a misunderstanding of this coefficient and using it without respecting their important assumptions required.

There is ongoing debate about the use of alpha for ordinal data based on the Pearson correlation matrix because alpha assumes that the item responses are continuous, it was found that it underestimates the real reliability compared with the ordinal alpha coefficient which offers greater values because it is appropriate with the ordinal data calculating based on the polychoric correlation matrix.

For this reason, this study aims at showing the preference of ordinal alpha coefficient for ordinal data compared with alpha coefficient, and give a practical example using R program