This study explores the use of the Multidimensional Support Model for Robotics Learning Activities (MuSRA) to help novice programmers develop debugging skills. MuSRA combines instructional strategies with digital tools to support metacognition. Central to the model is student-recorded digital data—screenshots, recordings, video captures—collected during robotics tasks. It positions students as reflective "scientists" analyzing their programming processes. The study involved 48 lower secondary school students with minimal programming experience in MuSRA-guided activities. The research examined (a) programming errors and (b) students' ability to correct them. Results showed common novice mistakes and improvements in debugging. Crucially, digital data reflection supported error detection. These findings suggest MuSRA is an effective approach to teaching debugging in educational robotics.