The influence of item format on Turkish eighth graders’ performance on the TIMSS

Abstract

This study examined the influence of item format (multiple choice and constructed response items) on Turkish students’ performance on the TIMSS 2011 eighth grade mathematics assessment, together with cognitive (knowing, applying, and reasoning) and content (number, algebra, geometry, and data and chance) domains. In addition, the ranking position changes of Turkey based on different item types were investigated among other participant countries. The study is a descriptive research study. Using the results of the TIMSS 2011 eighth-grade mathematics assessment, students’ average percent corrects for each item type were calculated, separately. According to the analyses, Turkish students performed significantly better on multiple-choice items than on constructed response items. Moreover, it was found that the performance of the students was statistically different depending on the cognitive level of the items, and the performances were decreasing from low to high cognitive level of items. It was found that students’ performance on data and chance items was significantly higher than their performance on number items as well as on algebra items. However, the interaction effects of item format and cognitive level of items and item format and content domains of items were non-significant. In addition to this, it was observed that, the country rankings based on different item types were significantly similar to the ranking based on overall items, and also the average performance of Turkish students on different item types was generally under international averages, but there was no significant difference between the averages.