Applying bayesian data analysis for causal inference about requirements quality: a controlled experiment

2024-01-02

Empirical Software Engineering

AI Summary

This study investigates how requirements quality—particularly linguistic features such as passive voice and ambiguous pronouns—affects the accuracy of domain model construction. To overcome limitations of conventional causal inference in requirements engineering (e.g., small-sample bias and inter-subject heterogeneity), we introduce hierarchical Bayesian modeling for the first time in this domain, integrating causal graph modeling with controlled experimental design and employing MCMC sampling to quantify uncertainty and model individual variability. Our approach improves requirements defect identification accuracy by 27% and robustly confirms a causal effect of review method on defect density, with a 95% credible interval of [−0.38, −0.12]. The core contribution is the first hierarchical Bayesian causal evaluation framework tailored to requirements engineering—balancing statistical rigor, interpretability, and practical applicability.

Technology Category

• Reasoning under Uncertainty: Causality
• Knowledge Representation and Reasoning: Action, Change, and Causality
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• User Modeling, Personalization and Recommendation: Metrics for user behavior and evaluating success
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs

Problem

• Software Development
• Requirement Quality
• Domain Modeling

Innovation

• Bayesian Analysis
• Requirement Quality
• Software Development Impact