many-analysts

This paper evaluates LLM-based coding agents (Claude Code and Codex) in social science analysis, finding they match or exceed human methodological diversity while remaining vulnerable to interpretation bias through verdict-layer manipulation.