Evidence Brief: Why Engagement Surveys Don’t Measure What Matters

Engagement surveys have become the default tool for measuring workplace culture. Organizations spend over $2 billion annually on them. But the evidence shows they measure sentiment — how employees feel in the moment — not the structural conditions that actually predict turnover, disability claims, absenteeism, and psychological safety.

This evidence brief synthesizes research showing that only 3.6% of published studies measure psychological safety correctly at the team level using intraclass correlation (ICC) validation. The rest use individual-level self-report, which conflates personal perception with group-level climate — a fundamental measurement error.

The brief makes the case for a structural measurement approach: assessing the conditions (leadership behaviour, accommodation processes, disclosure climate) that create psychological safety, rather than asking employees if they feel safe and hoping the aggregate tells you something actionable.