The authors' aim was to determine the extent to which error detection contributes to the explanation of a cognitive and motivational model of student performance in an assessment test. A total of 151 science students of secondary education participated in the investigation. Two causal models were developed using a structural equation analysis. This allowed the authors to estimate the effects and relationships between the different variables involved. In addition, they conducted correlational and descriptive analyses of the study variables to further explore the data obtained for the sample. The results indicate that error detection is an important factor that influences student performance to a greater extent than other strategic and motivational variables that have traditionally been considered strong predictors of performance. The findings show that students' detection of their own errors is a mediatory variable that could act as an engagement in the self-regulation cycle.