Do sociolinguistic factors influence program writing styles?

When programmers write programs, they may have specific coding preferences, including naming conventions, comments, or decision logic. While making stylistic decisions, programmers have to follow the syntactic rules of the programming language, and these decisions may demonstrate some insights about a programmer’s coding choices or demographic background. Several studies have investigated how the use of natural languages reflects the sociolinguistics background, including gender and region of users. However, few studies have focused on finding the imprints of programmers’ coding styles. For example, do programming languages carry marks that indicate a programmer’s stylistic choices? Do programmers’ gender or region influence how they will shape the code? This work investigates these questions by analyzing programming contest programs using statistical and machine learning techniques. Using concepts from sociolinguistics and software metrics in C++ programs, we identified programming features or components that are significant for classifying programmers based on their gender and region. Our goal was to identify sociolinguistic and software metric factors that might help us identify imprints of a programmer’s program writing choices. Initial efforts have resulted in prediction accuracies of 90.61% (for gender) and 79.73% (for region), based on programmers’ program writing style. This study indicates that, just as natural language, programming language also conveys information about programmers’ choices when they write code. The data sets and code are available at https://github.com/deen-abdullah/Dataset-ACDSA2025.