Coding agents in the social sciences

## Executive Summary This report presents findings from a February–March 2026 survey of 1,260 quantitative social scientists on AI and coding agent use: 81% have used generative AI for research, but only 20% regularly use coding agents (e.g., Claude Code). Adoption is concentrated among early-career researchers, men (by name classification), and researchers at high-status universities; common uses are generating code and editing prose. Coding agent users report starting more projects, posting more working papers, and submitting more grant proposals, though there is no evidence yet of increased journal submissions; respondents are optimistic AI will raise paper-writing productivity but are less confident about broader field-level benefits.