Empirical Research Design: A Security-by-Design Evaluation Framework for Ontology-Grounded Code Generation

Pre-registration of the empirical research design for Paper 4 of the AppSec Core series. This study evaluates whether deterministic, ontology-grounded policy delivery (via MCP server) produces higher security-by-design execution fidelity than probabilistic RAG and ungrounded generation, across code, IaC, and pipeline artifacts. The design document specifies the evaluation framework (M_SbD metric, multi-tier oracle), experimental conditions (G0/G1/G2), task set (12 tasks, 4 artifact types), statistical analysis plan, and infrastructure checklist — registered prior to experiment execution. Companion papers: AppSec Core Normalized Ontology (Paper 1), Coverage-Preserving Knowledge Compilation (Paper 2), Ontology-Grounded Retrieval via MCP (Paper 3).