Campbell's Law
High-stakes use of a social indicator can invite pressure that distorts both the measure and the activity it was meant to represent.
A public indicator can help people see whether an institution is meeting its responsibilities. Trouble can arise when the indicator becomes the basis for consequential rewards, penalties, or reputational judgments.
A hospital is ranked by its average discharge time. The ranking may reveal bottlenecks, but if staff rush discharge planning to improve placement, the statistic could rise while continuity of care suffers. A second measure can make that tradeoff visible.
Campbell's Law describes a tendency: the more a quantitative indicator is used for social decision-making, the more it may be subject to corruption pressures and distort the process it monitors. It is a caution, not an iron law.
Imagine a school funding formula tied closely to one exam score. Educators may narrow lessons toward tested material, administrators may redirect resources to likely test-takers, and families may experience pressure around a single result.
Some adaptations can improve the measured outcome honestly. Focused practice may help students master material. The concern is not that every response is fraudulent, but that an indicator can reshape priorities and conceal unmeasured parts of the mission.
This issue overlaps with Goodhart's Law, yet Campbell emphasizes the institutional and social stakes of using indicators to allocate power or resources. The stronger the consequence, the greater the reason to examine who can influence the number.
Suppose a city allocates grants using a reported neighborhood crime rate. Agencies might change reporting practices or prioritize visible incidents. The rate still matters, but funding decisions should examine reporting access and residents' experience too.
A test score may contain useful information while still being an incomplete proxy for learning. If it drives high-stakes choices, decision-makers should also consider student work, course access, retention, and other context relevant to the specific purpose.
Build safeguards before attaching consequences. Use more than one indicator, review changes in the underlying process, and allow people affected by the rule to describe effects the official dashboard does not capture.
Separate diagnosis from judgment where possible. A metric can identify where support might be needed without immediately declaring a person or institution successful or deficient. This reduces the incentive to hide bad news instead of addressing it.
Review the policy over time. Once people adapt, the indicator may mean something different from what it meant at launch. A transparent process for recalibration can protect its usefulness without pretending measurement is neutral.
An indicator can clarify a public problem and also change it. What additional evidence would make you comfortable using a score to make decisions that affect other people?
Psychological perspective
Campbell's Law is a framework for thinking about institutions and incentives, not a diagnosis of the individuals who work within them. People respond to systems, and different responses can occur even when everyone intends to serve the public.
Measurement can affect attention, identity, and trust. Treating an indicator as provisional evidence leaves room to hear from those measured and to revise policies when the consequences differ from the intention.
Thinking clearly
Before linking a metric to a consequence, write down the outcome it represents and the important parts it leaves out. Ask who can alter the recorded value and who bears the cost of errors.
Use independent evidence alongside the score, including case reviews and feedback from affected people. Compare patterns across groups to see whether a seemingly neutral indicator hides unequal access or reporting.
Set a review date and a way to pause or revise the policy. Keep the metric useful for learning, rather than allowing yesterday's proxy to become an unquestioned verdict.
The concern is not that every response is fraudulent, but that an indicator can reshape priorities and conceal unmeasured parts of the mission.
When a measurement carries consequences, whose experience should count alongside the number?