The Effect of Distracters Selection Method in Multiple Choice Tests Items on their Psychometric Properties Based on Classical Measurement and Modern Measurement Theories

(Nayel Odeh Al- Kaabnh and Ahmad Sulaiman Oude)

---

Abstract

The objective of this study was to investigate the effect of distracters selection method in multiple choice test items on their psychometric properties (difficulty, discrimination, and guessing) and ability parameters estimation based on classical and modern measurement theories. To achieve the objective of the study, the researchers prepared an achievement test for 9 th graders consisting in its final format of 41 items. Three forms of items distracters (A, B, C) were used. SPSS and Bilog-MG programmes were used to estimate statistical indicators. Findings revealed no effect for selection method of multiple test item distracters on examinees ranks on continuum ability scaling. Significant differences were found for standard error estimations of ability parameters estimation means. Significant differences were found between examinees abilities estimated using raw score due to the method used in item distractors construction. Significant differences were found between difficulty parameter estimation and discrimination parameter estimation. No significant differences were found in standard errors estimation for difficulty parameter. No significant differences were found in guessing estimation parameter. Significant differences were found between standard error for guessing parameter estimation. Significant differences between difficulty coefficients and discrimination coefficients values were found according to the Classical Theory. Key Words: Classical measurement theory, Distractors' selection strategy, Modern measurement theory.
PDF

References