GlossaryStatistics and interpretation
Peer comparison
Peer comparison uses agencies or places selected for relevant similarities. Learn why population size and rates alone do not establish comparability.
A peer comparison compares agencies or places selected because they share characteristics relevant to a question. It can help put a measure in context, but the comparison depends on the selection criteria and compatible data. Similar population size alone does not make two organizations equivalent in responsibilities, demand, or reporting.
Choose peers for the question
For staffing, relevant differences may include service responsibilities and geography. For response intervals, call types, travel conditions, and the measured milestones matter. For reported crime, the population base, reporting coverage, offense definitions, and local context all affect interpretation.
The FBI's archived 2010 caution against ranking explicitly warns against comparing reporting units solely by population coverage or enrollment. It is cited for that methodological caution, not for current crime levels or a current list of participating agencies. FBI comparison caution.
Example
Hypothetical example: An analyst compares three agencies serving populations near 50,000. One provides patrol only; another also operates a jail; the third serves a large visitor destination. Their total expenditures per resident differ.
The figures may accurately describe spending under each agency's accounting rules. They do not establish that the highest-spending agency is least efficient. The analyst needs compatible service scope, spending definitions, and information about demand before making that judgment.
Common confusion
A peer group is not necessarily a representative sample of all agencies. Selecting only agencies with available data can leave out organizations whose circumstances differ. The comparison should explain the selection and exclusions.
A rate adjusts for its denominator, not every difference between places. Dividing counts by resident population does not account automatically for commuting, tourism, land area, service obligations, or reporting practices.
Make selection and limits visible
State why the peers were selected and whether the criteria were chosen before looking at the results. Selecting different peers until a preferred outcome appears can make an ordinary observation look exceptional.
Use matching periods and definitions. Separate missing values from zero, identify provisional figures, and avoid combining incompatible reporting systems without a documented method. Report the number of peers and the range or distribution where useful, rather than only an average.
A good peer comparison supports a focused question: how a specified measure differs among defined organizations. It does not convert a single workload, spending, or crime statistic into an overall verdict on safety or agency quality.