Summative Assessment Development for Classroom Use: Insights from Classical Test Theory and Item Response Theory


This presentation reports an ongoing empirical study where a summative test for classroom use is analyzed for further improvement using psychometric techniques from Classical Test Theory and Item Response Theory (e.g. Rasch analysis and 2PL or 2-parameter logistic model). The research aims to show a comparison of these techniques by applying them on the same set of item responses to a test on polynomial functions which was devised as a summative assessment. The said test consists of fifteen (15) multiple-choice items that were scored dichotomously, and five (5) constructed-response items which were polytomous items. The analyses were conducted using R, which is an open-source statistical software with packages that are widely used in the field of psychometric assessment. Insights that are drawn from the results can be very useful in understanding the process of test development especially that less attention is provided to analysis of test items after summative assessment in the classroom. Given that Philippine teachers are burdened with heavy workload and other administrative tasks, it is understandable that classroom teachers tend to devote more time to scoring and grading tests and may have no time to analyze and improve the test. Additionally, results from this study can help inform the existing body of literature in test development regarding effective psychometric techniques that help improve the individual quality and functioning of test items.

Author Information
Mark Lester Garcia, Ateneo de Manila University, Philippines
Kevin Carl Santos, University of the Philippines, Philippines
Catherine Vistro-Yu, Ateneo de Manila University, Philippines

Paper Information
Conference: ERI2023
Stream: Assessment and Learning Analytics

The full paper is not available for this title

Virtual Presentation

Comments & Feedback

Place a comment using your LinkedIn profile


Share on activity feed

Powered by WP LinkPress

Share this Research

Posted by James Alexander Gordon