Category Discipline: Define It Before You Count It
A number can be accurate and still invite the wrong comparison. Before deciding that something rose, fell or broke a record, ask what qualified for the count and whether that rule stayed the same.
Category discipline means making those inclusion rules visible. It starts with a concrete question: what must be true for one person, application or event to enter this total?
Start With The Definition
Labels such as completed applications, active members and reported incidents sound self-explanatory. They are not complete definitions. An application might count when submitted, when all documents arrive, when staff verify it or when a decision is issued. Each measure can be useful, but each answers a different question.
Write the rule in a sentence before interpreting the trend. Include what is excluded and whether the unit is a person, a case, an event or a transaction. One person can submit multiple applications; a count of applications is not automatically a count of people served.
A Hypothetical Example
Imagine a local program reports 80 completed applications in one month and 120 in the next. These are invented numbers for illustration, not a report about an actual program.
In the first month, completed means every required document has been checked. In the second, completed means a form has been submitted, even if documents are still missing.
The published total rose by 50 percent: (120 − 80) ÷ 80 × 100. The arithmetic is correct. It does not establish that staff processed 50 percent more fully checked applications, because the two totals use different entry rules.
Suppose a review finds that only 75 of the second month’s 120 submissions met the original document-check standard. On that consistent basis, the count fell from 80 to 75, a decrease of 6.25 percent. That additional figure is also hypothetical. Its purpose is to show why a comparable count can change the interpretation.
Different Categories Can Both Be Useful
The Bureau of Labor Statistics provides a real example of precise classification. Its household survey distinguishes employed people, unemployed people and people outside the labor force. In general, unemployment requires having no job, being available for work and actively seeking work in the preceding four weeks. People on temporary layoff awaiting recall have a job-search exception.
Someone without a job therefore does not automatically fall into the official unemployed category. That distinction does not erase the person’s circumstances. It tells us which question a particular statistic answers. Broader labor-underutilization measures answer related questions using different inclusion rules.
A Definition Change Is A Question, Not A Verdict
A changed rule does not by itself establish deception. An agency might improve a questionnaire, expand coverage or replace an outdated classification. The useful question is how the change affects the comparison.
The Census Bureau’s American Community Survey guidance cautions that changes in variables and geography can affect comparability. Differences in the population covered, question wording, residence rules, reference periods and tabulation can also matter. Reading the methodology is part of reading the result.
Ask whether the earlier total has been recalculated under the new rule, whether both definitions were measured during an overlap period, or whether the source explicitly warns against a direct comparison. When a consistent series is unavailable, describe the limitation rather than manufacturing a precise trend.
Use This Comparison Checklist
1. Define the unit. Are we counting people, jobs, applications, incidents or dollars?
2. State the entry rule. What conditions must an item meet, and what is excluded?
3. Match the scope. Use comparable populations, geography and reference periods.
4. Check for changes. Read methodology notes, revisions and breaks in the series.
5. Recalculate consistently where possible. If it is not possible, narrow the conclusion.
6. Separate measurement from explanation. A comparable increase still does not, by itself, establish what caused it.
Category And Denominator Work Together
The category tells you what qualifies for a count. The denominator tells you what that count is compared against. Both need to match the claim. A correctly calculated percentage can describe a real change while answering a different question from the headline.
Before sharing the next impressive total, try this: write its inclusion rule in one sentence. If you cannot find that rule, you have identified the next piece of evidence to look for.
Read next
After defining who or what is counted, check the total used to calculate the comparison.
Comments