Supporting novice programmers through meaningful feedback from a pedagogical programming environment implementing a case-based reasoning approach

Giving meaningful feedback that provides more information on underlying programming misconceptions as opposed to cryptic error messages can help students learn programming for the first time. A Case-Based Reasoning (CBR) approach to programming misconception diagnosis and remediation can provide this feedback in real time. However, CBR approaches generally need large sets of cases to accurately provide feedback. CBR-C is a Pedagogical Programming tool developed to give meaningful feedback following a CBR approach to students learning the C Programming Language for the first time that is able to give feedback despite having insufficient cases. Experiments for evaluating the Feedback Generation of CBR-C conducted with Manila Science High School students learning the C language for the first time suggested that receiving feedback from CBR-C is better than not receiving any feedback. The improvement in code quality of the students who used CBR-C was mildly statistically significant than those who did not, with a p-value of 0.0548 given the chosen significance level of (0.05) using the Wilcoxon Rank Test. Focus group discussions and content analysis on their code suggested that the tool helped them by giving them a starting point for addressing their programming misconceptions. However, CBR-C did not have any effect on the overall academic performance of the students