Abstract
This study compares the statistical methods employed for detecting underachievement, specifically the simple difference method, the regression method and the Rasch method. A sample of 1182 first- and second-year secondary students from 8 high schools in the province of Alicante participated in the study. The results showed a percentage of underachieving students that varies from 14.55% to 30.37%, depending on the statistical method employed. The Rasch method identified the highest number of underachieving students. Statistically significant differences were found between gender and type of student-underachieving and non-underachieving; however, no significant differences were detected between the course and type of student. This study confirms the importance of knowing the measurement properties of the statistical methods, how they affect the detection of underachieving students, and the main educational implications.