The development and evaluation of a web based adaptive testing system: The case of probability unit

The main purpose of this study is to develop and evaluate a computerized adaptive testing system for permutation, combination, binomial expansion and probability which are sub-topics of probability unit in 11th grade mathematics course. The developed computerized adaptive testing system was integrated into UZWEBMAT, which is an adaptive intelligent web based environment to learn mathematics. An exam application including the developed adaptive testing system was implemented at a high school in Trabzon in the 2011-2012 academic year. The capability of the developed adaptive testing system to set the ability levels of students about the unit of probability was evaluated based on the data obtained from the implementation of exam. Also, the relationship between students? ability level estimations obtained from the exam application and their learning styles was investigated. System logs were used for data collection. The capability of the developed adaptive testing system to set the ability levels of students was determined taking into consideration the standard error amount and reliability coefficient. The amounts of standard error obtained from the all exam sessions were transformed into the classical reliability coefficients and evaluated. Kruskal Wallis H-Test was employed for investigating the relationship between the ability level estimations of students that were obtained from exam application and their learning styles.Research findings demonstrate that the developed computerized adaptive testing system can set the ability levels of students about the unit of probability at a very high level of reliability. Additionally, the data obtained from exam application show that there is no statistically significant correlation between the ability level estimations of students and their learning styles.Keywords: Computerized Adaptive Testing System, Ability Level Estimation, Learning Styles