Tracing IDEA Diagram Use in Technical Debt Prevention and Repayment: An Observational Case Study in Academic Agile Projects

This repository contains the replication package for the paper "Tracing IDEA Diagram Use in Technical Debt Prevention and Repayment: An Observational Case Study in Academic Agile Projects", accepted at the XXIV Brazilian Symposium on Software Quality (SBQS 2026), Maringá, Brazil. The package includes: (i) the deterministic analysis scripts that reproduce the descriptive tables reported in the paper, and an optional script for the exploratory non-parametric tests, with pinned dependencies in requirements.txt; (ii) the IDEA evidence-decision rules and decision tree used to classify card-level evidence, together with the 0–16 card documentation-quality rubric; (iii) the traceability ledgers (Evidence Ledger and Claim Ledger) and the methodological decision log, linking each manuscript claim to its supporting evidence; (iv) the aggregate descriptive outputs (card-quality results, rubric-dimension breakdowns, aggregate Likert summaries, TD macro-activity and practice summaries, effect sizes, and the perception × documentary triangulation matrix); (v) supporting material, including the GQM metric and qualitative codebooks and a high-resolution TD prevention conceptual map. The 0–16 documentation-quality score was computed by a deterministic, rule-based script. Card-level IDEA evidence combines automatic term matching on card fields with a documented researcher-coded assessment. The README section "Reproducibility Scope" states which steps the package reproduces, which depend on human coding, and which were carried out outside the package. Participant-level data are released only in anonymized form to protect participant privacy. Card examples and qualitative artifacts are in Portuguese, the language of data collection. Version 4 corresponds to the camera-ready version of the paper and is the version cited in its Artifact Availability section. Compared with version 3, it updates paper.pdf to the camera-ready version, adds requirements.txt, and extends the README with environment assumptions and a reproducibility scope section. The data and scripts are unchanged.