Comparison of International TIMSS 2011 Proficiency Levels and Cut-off Scores Set by Using Cluster Analysis

Abstract

The aim of the study is to compare TIMSS 2011 proficiency levels with the proficiency levels defined by the researchers using cluster analysis for Turkey, Korean, Norway, and Morocco in 4th and 8th grades in the fields of science and mathematics. Therefore, it is tried to be reached that these cut-off scores for each country can serve the evaluation of each country itself. For this research, the data gathered from related countries’ students was taken from TIMSS 2011 database. Statistical analysis was performed with SPSS Version 21.0 statistic software package. The cut-off scores for these four countries selected in this study for each grade level and course type were defined using cluster analysis. Then, proficiency levels according to these cut-off scores were compared to the general TIMSS 2011 proficiency levels, and so the difference between these levels and percentage of agreement have been examined. According to the results, cut-off scores set by using cluster analysis for Korea were higher than TIMSS international benchmarks. Cut-off scores set for Morocco, Norway, and Turkey were lower than TIMSS international benchmarks. the percentage of agreement of the proficiency levels was found to be between 8.1% and 70.0%, and in general, it has been found that the percentage of agreement was low. Consequently, it is suggested that countries should make a standard-setting study for their own samples instead of using TIMSS international benchmarks for their own evaluations.