Error-Grid Framework: A Pedagogical Approach for Diagnosing Programming Mistakes and Enhancing Learning Outcomes in Secondary Education

Abstract

Programming in secondary schools is still a difficulty because learners are transitioning to text-based languages; this is the case of programming. The dropping pass rates in Mauritius are indicative of endemic conceptual and procedural discontinuities. This study constructs and tests the Error-Grid Framework, a low-cost classroom intervention that uses errors as a kind of diagnostic message to inform specific instruction. Content analysis of 30 Grade 11 students (n=90 handwritten scripts; 2,987 lines) was conducted, providing a code of recurring errors in 11 concepts of core programming. Reliability was studied among four educators using Krippendorff’s alpha, and effectiveness was studied with the help of a pre/post design with two groups of students (n=11; n=13). Classroom utility was measured using survey data from 39 teachers across 20 schools. The findings indicated moderate inter-rater reliability (1=0.67; 2=0.64), substantial error decreases (p<0.001), and continued difficulties in loops, arrays, and functions. Teachers showed dense usability and diagnostic worthiness in the study. The framework facilitates differentiated teaching and timely feedback. In the future, this work should be extended to other languages, where automation is introduced to provide scalability and investigate long-term effects.