Native language’s effect on Java compiler errors

Novice student compilation behaviors are well documented by prior research, but those findings are derived from instructional contexts that are largely in English-dominant locales. Speakers of languages other than English face unique challenges while learning to program like partial localization of applications and programming language syntax based on English-only keywords. This study examines compiler errors of novice programming students from different native language backgrounds to explore potential differences in their error distributions relative to those in English dominant contexts. For example, it is plausible that students from non-English language backgrounds would experience more “unknown identifier” types of errors while programming with English keywords and API methods. Using data from the BlueJ Blackbox database, we analyzed error distributions for users based on country and language group characteristics. Statistical analysis showed a statistically significant difference in error distributions between native language groups; however, effect sizes were very weak indicating that the differences have little practical significance in terms of guiding either language or instructional design. However, these results may support drawing broader inferences from earlier Java compilation behavior studies to global contexts.