Talk to us
Hampton Wilson Institute · Research Brief 01

Fit for the Decision.

Data qualityHampton Wilson InstituteEVOLVRS ResearchOctober 2026

Public organizations usually judge their data against the system that produced it. This brief argues they should judge it against the decision it must support. The same record can be fit for one decision and unfit for the next, and AI raises the stakes: a model will summarize whatever it is given, fluently, whether or not the records can carry the conclusion.

Key findings
  1. 01Accuracy is necessary, and it is not enough.

    The people who use data judge its quality by relevance, timeliness, completeness and access, not correctness alone.

  2. 02Most failures are failures of fit.

    The information was never collected, is fragmented, answers a different question, lacks context, or arrives too late.

  3. 03AI makes fit a precondition.

    Federal policy already asks agencies to document data quality before high-impact AI is used.

Figure 1 · Five tests of fit
  1. DecisionWhich decision must this information support?
  2. ObservationDoes it observe what that decision is about?
  3. CoverageWho and what does it cover, and who is missing?
  4. ComparisonWhat is it being compared with?
  5. ProtectionWho could be exposed by it?

Each test is asked against the decision, not the record.

The question is not “is this data good?” It is “is it good enough for this decision?”

For leaders

Questions to ask.

Before the next decision reaches the table.

  1. Which decision is this information meant to support?
  2. Does it observe what that decision is actually about?
  3. Who is missing from it?
  4. What would we need to see before relying on it?
  5. Is it ready to reach a model, or a public report?
References and limits

This brief does not claim

  • The five tests are a design framework, not a validated instrument.
  • It makes no claim about the data quality of any specific organization.
  • Federal AI requirements cited apply to federal agencies.
  1. Wang, R. Y., & Strong, D. M. (1996). Beyond accuracy: What data quality means to data consumers. Journal of Management Information Systems, 12(4), 5–33. doi:10.1080/07421222.1996.11518099
  2. Ackoff, R. L. (1967). Management misinformation systems. Management Science, 14(4), B147–B156. doi:10.1287/mnsc.14.4.B147
  3. Office of Management and Budget. (2025). Accelerating federal use of AI through innovation, governance, and public trust (M-25-21). Executive Office of the President.
  4. Federal Committee on Statistical Methodology. (2005). Report on statistical disclosure limitation methodology (Statistical Policy Working Paper 22, 2nd version). doi:10.21949/1529877

Cite this briefHampton Wilson Institute. (2026). Fit for the decision: Data quality in public organizations (Research Brief 01). EVOLVRS. https://evolvrshq.com/research/fit-for-the-decision/

Brief 02 · Decision qualityPrepared, Not Decided.Read Brief 02Brief 03 · Outcome qualityDid It Work?Read Brief 03

See clearly. Decide with evidence.

Bring the decision. The Institute will help you test what the evidence can carry.

Talk to the InstituteAll research