By · Updated

Practical method

Compare indicators on a consistent scope

Two figures may look comparable while measuring different objects. Prepare a definition record before commenting on differences or change.

01

Define the object and unit

Record what is counted or measured, its unit and population. Distinguish counts, proportions, averages and amounts. A chart label may not be sufficient.

Fictional example: ten enquiries from twenty visits is a proportion; ten enquiries with no visit count is a volume. They answer different questions.

02

Align periods and frequencies

Compare consistent periods and document frequency. A monthly value should not be directly compared with an annual total. Check whether it describes a point-in-time state or events during a period.

Keep reference dates separate from download date. A file received today may describe an old period. State differences before calculating change.

03

Examine denominators and exclusions

For a rate, identify numerator and denominator. Record exclusions and missing values. A population change may explain part of an apparent trend.

For a fictional satisfaction measure, explain how many people answered and who was asked. Do not present respondents as representing every visitor.

04

Retain versions and limitations

Record series versions and source corrections. Where possible compare values from a consistent publication state. A revised series may change earlier calculations.

Keep units, periods and limitations alongside the conclusion. An observed difference does not establish causation; separate hypotheses from observations.

A record to keep with the decision

Minimum evidence for a review
ItemEvidence
ObjectDefinition, unit and population.
TimePeriod, frequency and observation date.
RateNumerator, denominator and exclusions.
VersionSource, revision and comparison limitations.

Download the worksheet to fill in (CSV)

Frequently asked questions

Does a change prove an action’s effect?

It establishes a difference. Examine other changes and measurement design before attributing cause.

Can percentages always be added?

Not without understanding populations and definitions. Percentages do not aggregate like independent volumes.

Reference material

Insee · Données ouvertes

Banque de France · Webstat

The practical checklist is an editorial synthesis to adapt to your service. It does not constitute a certification or an audit result.

Continue with a related decision