Multiple-choice and constructed-response item formats are widely used in Philippine classrooms where tests are primarily summative. Conducting item analysis provides teachers the professional possibility to independently improve their test items while also allowing them to make informed decisions about test improvement and ensure that their assessments provide meaningful scores. Comparisons of item analysis approaches have been extensively used in standardized tests but not so much in small-scale testing such as classroom assessments. Hence, this research reports an empirical study where a summative test for classroom use is analyzed using psychometric techniques from the Classical Test Theory (CTT) and the Item Response Theory (IRT). This study employs an action research design where the primary author being a teaching practitioner-researcher, reports how an item analysis of his teacher-made summative test was conducted to improve the test items. Results from this study show that interpretations of the item parameter estimates from both psychometric approaches, CTT and IRT agree on most of the recommendations for the items. The items where the two approaches provide contrasting recommendations call for a further qualitative evaluation of the flagged items putting the teacher in a judicious position to discern which recommendations to follow regarding the concerned items. In the context of this study, an example of a guiding insight in deciding would be reflecting on the learning objectives measured by the test and the context in which the test was created.
Read full abstract