Determining assessment performance in Applied Statistics with ROC analysis

Abstract

Difficulties in learning and teaching statistics have been a subject of interest for researchers and many studies have attempted to describe the factors influencing academic performance in the subject of statistics. The aim of this research is twofold: a) to test for the effects of attendance to statistics sessions, interest in the subject, collective feedback, satisfaction, and the existence of previous experience with applied statistics, in the students’ assessment results in applied statistics; and b) to build an “instrument” to predict which students will find problems to pass the subject. Sample consisted of 166 students of statistics in psychology. Logistic regression and ROC curve analysis were used, and class attendance, collective feedback utility and previous experience with statistics emerged as good predictors of statistics achievement.