Indicator-corruption principle
Campbell's Law
The more a quantitative social indicator is used for consequential decision-making, the more pressure it faces to be manipulated and to distort the activity it was intended to monitor.
decision pressure on indicator -> gaming + selection + activity distortion
Campbell's law is a social-science principle, not a numerical constant. Corruption depends on stakes, measurement validity, observability, discretion, audit design, multiple objectives, and the ability to shift costs outside the metric.
The target system separates true objective, measured proxy, and rewarded behavior. Measurement validity and audit sampling change how pressure becomes genuine improvement, gaming, or displaced harm.
(%)
The plot, diagram, and calculated result share the same state. Animation runs only when it adds explanatory value.
- CHANGE
- Decision pressure on metric
- WATCH
- improvement + corruption
- MEANING
- The target system separates true objective, measured proxy, and rewarded behavior. Measurement validity and audit sampling change how pressure becomes genuine improvement, gaming, or displaced harm.
A measure changes behavior when it becomes the gate to rewards or sanctions.
Activity flows through a measurement aperture; high pressure redirects effort toward visible score production and away from unmeasured outcomes.
What it actually says
Campbell's law emphasizes reactive measurement. Once an indicator controls funding, ranking, promotion, punishment, or legitimacy, participants adapt to its operational definition. Some adaptation is intended improvement; some is gaming, selection, teaching to the test, relabeling, delay, or transfer of costs.
The problem is not quantification itself. It is the combination of high stakes, incomplete representation, predictable rules, and weak independent observation. Measurement systems can improve by using multiple indicators, qualitative review, rotating audits, outcome validation, and explicit gaming analysis.
"A useful law compresses a pattern. It does not erase the conditions that make the pattern true."
How the idea developed
The modern form emerged through observation, argument, and later refinement. The timeline separates the first insight from the version now used in textbooks and practice.[1]
Donald T. Campbell publishes Assessing the Impact of Planned Social Change.
Campbell states the law in work on social experimentation.
Performance-management research documents gaming and target displacement.
Education, healthcare, policing, finance, platforms, and ESG systems study metric reactivity.
How the pattern works
The relation becomes useful only when its mechanism, measurement process, and operating range are visible.
The indicator captures only part of the objective.
Rewards and sanctions concentrate attention on the measured channel.
People learn classifications, thresholds, timing, and audit rules.
Unmeasured quality, difficult cases, and long-term outcomes can lose resources.
Campbell's law is a social-science principle, not a numerical constant. Corruption depends on stakes, measurement validity, observability, discretion, audit design, multiple objectives, and the ability to shift costs outside the metric.
Where it earns its keep
Applications are strongest when the law changes a decision, measurement, model, or experiment rather than merely providing an analogy.
Design balanced score systems
ApplicationMultiple measures can expose trade-offs and reduce single-target dominance.
Too many correlated metrics can reproduce the same pressure.
Look for selection and boundary shifts
ApplicationIndependent samples can reveal excluded cases and displaced costs.
Audits themselves become gameable when fully predictable.
Separate learning metrics from reward gates
ApplicationLow-stakes diagnostic measures support experimentation.
Do not punish transparent reporting of uncertainty or failure.
Where it stops working
Observed score improvement can combine real performance and gaming; separating them requires independent outcomes, design variation, and institutional knowledge.
Qualitative judgment can also be biased or captured. Replacing metrics with discretion does not remove accountability problems.
"Never set targets"
Better: Targets can coordinate work when measures are valid and safeguards exist."Any metric increase is fake"
Better: Participants can genuinely improve the underlying objective."More metrics automatically solve gaming"
Better: A dashboard can multiply incentives and complexity."Campbell and Goodhart are identical"
Better: They overlap, but Campbell emphasizes social decision pressure and corrupting institutional processes.Sources and further reading
Original publications and serious secondary scholarship are prioritized over summaries.
- Campbell - Assessing the Impact of Planned Social ChangePrimary source for the principle.https://www.globalhivmeinfo.org/CapacityBuilding/Occasional%20Papers/08%20Assessing%20the%20Impact%20of%20Planned%20Social%20Change.pdf
- National Academies - Performance MeasurementAuthoritative measurement context.https://nap.nationalacademies.org/catalog/13083/accounting-for-health-and-health-care-approaches-to-measuring-the-sources
- Bevan and Hood - What's Measured Is What MattersEvidence on gaming in public-service targets.https://doi.org/10.1111/j.1467-9299.2006.00600.x
- Stanford Encyclopedia - Scientific ObjectivityBroader context for measurement, values, and objectivity.https://plato.stanford.edu/entries/scientific-objectivity/