How Do Code Smells Affect Skill Growth in Scratch Novice Programmers?

TL;DR

This study analyzes how code smells in Scratch projects relate to computational-thinking skills, using large-scale data and machine learning to inform education and tooling for novice programmers.

Contribution

It provides the first large-scale, detailed mapping of the relationship between CT skills and design flaws in novice Scratch projects, with implications for education and software maintenance.

Findings

Identifies specific CT skills associated with certain code smells.

Develops predictive models linking CT proficiency to code quality.

Provides open dataset and analysis pipeline for further research.

Abstract

Context. Code smells, which are recurring anomalies in design or style, have been extensively researched in professional code. However, their significance in block-based projects created by novices is still largely unknown. Block-based environments such as Scratch offer a unique, data-rich setting to examine how emergent design problems intersect with the cultivation of computational-thinking (CT) skills. Objective. This research explores the connection between CT proficiency and design-level code smells--issues that may hinder software maintenance and evolution--in programs created by Scratch developers. We seek to identify which CT dimensions align most strongly with which code smells and whether task context moderates those associations. Method. A random sample of aprox. 2 million public Scratch projects is mined. Using open-source linters, we extract nine CT scores and 40 code smell indicators from these projects. After rigorous pre-processing, we apply descriptive analytics, robust correlation tests, stratified cross-validation, and exploratory machine-learning models; qualitative spot-checks contextualize quantitative patterns. Impact. The study will deliver the first large-scale, fine-grained map linking specific CT competencies to concrete design flaws and antipatterns. Results are poised to (i) inform evidence-based curricula and automated feedback systems, (ii) provide effect-size benchmarks for future educational interventions, and (iii) supply an open, pseudonymized dataset and reproducible analysis pipeline for the research community. By clarifying how programming habits influence early skill acquisition, the work advances both computing-education theory and practical tooling for sustainable software maintenance and evolution.