An international comparison using a diagnostic testing model: Turkish students' profile of mathematical skills on TIMSS-R

摘要

This study illustrates how a diagnostic testing model can be used to make detailedcomparisons between student populations participating in international assessments. Theperformance of Turkish students on the TIMSS-R mathematics test was reanalyzed with adiagnostic testing model called the Rule Space Model. First, mathematical and cognitiveskills (' attributes ') measured by the test were determined. One hundred sixty-two items werecoded in terms of their attribute involvement, creating an incidence matrix—the Q-matrix.Using the Q-matrix and the student response data, each student's attribute mastery profile wasdetermined. Mean attribute mastery levels of Turkish students were computed and comparedto those of their American peers. It was shown that Turkish students were weak in algebra andprobability/statistics. They also demonstrated poor profiles in skills such as applying rules inalgebra, approximation/estimation, solving open-ended problems, recognizing patterns andrelationships, and quantitative reading.