Campbell Law Distortion

When people use a single number to judge things, they change how people behave to look better on that number. This makes the number less useful for judging the real quality or performance.

Campbell's law describes how reliance on a single quantitative indicator induces actors to manipulate behavior and data, degrading the indicator's validity. Persistent metric-focused incentives create perverse feedback that decouples measured values from underlying constructs of interest.

A school is judged solely on standardized test scores, so teachers stop spending time on art, physical education, and critical thinking—subjects not on the test. The test scores may rise, but the broader education the scores were meant to represent actually gets worse.

A hospital network ties department bonuses exclusively to a 30-day readmission rate metric. Clinicians respond by selectively discharging lower-risk patients just after the 30-day window, reclassifying borderline readmissions as new admissions under different diagnostic codes, and avoiding high-complexity cases likely to return. The readmission rate indicator improves on paper while unmeasured dimensions—overall patient outcomes, diagnostic coding integrity, and case-mix complexity—degrade. The original construct validity of the readmission proxy as a signal of care quality collapses, producing systematic measurement bias and multidimensional sensitivity loss across the care-quality construct it was designed to represent.

People chase a visible target so they alter actions to score well on it. Those changes make the target a poor guide to actual quality.

Agents selectively optimize behaviors that increase the proxy metric while neglecting unmeasured dimensions, driven by constraint-laden incentive structures and asymmetric valuation of observable outputs. Structural elements like reporting thresholds and reward gates skew priorities, producing systematic measurement bias.

Use several different measures to judge performance. Reward real outcomes and not just the numbers.

Deploy a balanced set of complementary indicators, triangulating across qualitative and quantitative sources to reduce gaming incentives. Implement audit mechanisms and variable weighting to penalize narrow optimization and preserve construct validity.

gaming of measurement rules; neglect of unmeasured tasks; data fabrication or manipulation

An adversarial actor can deliberately promote a narrow, high-stakes metric in an institution they wish to destabilize or co-opt, knowing that agents will optimize for it at the expense of actual performance, eroding the true quality being measured. In competitive or regulatory contexts, a rival can lobby for adoption of a metric that favors their own optimized behaviors while handicapping competitors who maintain broader, harder-to-game standards. State or corporate actors can also embed Campbell's Law dynamics into accountability systems they nominally endorse, creating a veneer of rigor while systematically hollowing out construct validity and reducing genuine accountability.

Deploy composite indicator sets that triangulate across multiple independent proxies, raising the cost of coordinated gaming across all dimensions simultaneously. Rotate or randomly vary the specific metrics used for high-stakes decisions at irregular intervals to prevent agents from anchoring behavioral adaptations to a fixed target. Supplement quantitative metrics with periodic qualitative audits and construct-validity checks that compare measured values against independent external criteria, flagging divergence as a signal of metric capture.