The Compatibility between the Two-Parameter Logistic Model and Classical Theory in Constructing an Achievement Test in General Science for 6th Grade Students

Authors

Husni Salem, Yarmouk University - JordanFollow
Nidal Al-Shraifin, Yarmouk University - JordanFollow

DOI

https://doi.org/10.35192/jjoas-h.v22i1.9

Abstract

This study aimed to assess the compatibility between the two-parameter logistic model and classical theory in constructing an achievement test in science for sixth grade students. To achieve this purpose, a science achievement test consisting of 50 multiple-choice items was constructed and applied to a sample of 567 sixth graders from eight randomly selected public schools in Jarash. The responses were analyzed using SPSS, BILOG-MG 3, NOHARM, and LDID programs to obtain statistical indicators for both examinees and items. The results indicate that the items of the sixth-grade science test can be classified as either matching or non-matching according to item response theory using the two-parameter logistic model and classical theory. There is a notable similarity between the two theories in classifying the items, with an agreement rate of 92% between the conventional indicators and the two-parameter logistic model indicators. This means that 92% of the 50 test items were judged similarly by both theories, while the remaining 8% showed discrepancies. Additionally, the results revealed a significant statistical difference (α = 0.05) in reliability between the two theories. Using the modified Feldt equation for non-independent samples, the analysis showed F = 1.99 in favor of the item response theory using the two-parameter logistic model. Furthermore, there was a significant statistical difference (α = 0.05) in criterion validity between the two theories, with a t-test for non-independent samples yielding t = 2.16 in favor of criterion validity according to item response theory. Finally, the results indicated a strong correlation between the capabilities of test examinees and their raw scores or standardized scores, which were positively correlated and statistically significant (α = 0.05), classified as a very high correlation.

Recommended Citation

Salem, Husni and Al-Shraifin, Nidal (2020) "The Compatibility between the Two-Parameter Logistic Model and Classical Theory in Constructing an Achievement Test in General Science for 6th Grade Students," Jordan Journal of Applied Science-Humanities Series: Vol. 22: Iss. 1, Article 9.
DOI: https://doi.org/10.35192/jjoas-h.v22i1.9
Available at: https://digitalcommons.aaru.edu.jo/jjoas-h/vol22/iss1/9