A composite metric, also called an index, combines two or more measures into one standardized score. It can give stakeholders a quick view of a complex process when separate metrics are hard to interpret together. The simplicity is useful for reporting, but it can make the score harder to explain or act on. section 7.3.1
Before combining measures:
- Identify indicators expected to measure the same underlying concept. Examine correlations between them; an unrelated indicator may belong to a separate process. section 6.3.1
- Assess internal consistency with an appropriate method, such as Cronbach’s alpha, rather than assuming the indicators form a reliable scale. section 6.3.1
- Keep the component metrics visible. A combined score can obscure competing trends and reduce differences between entities or over time. section 7.3.1 section 7.4
Combining questionnaire responses
A depression questionnaire can combine answers about different symptoms into one score. First check whether its questions reliably measure the same underlying process; questions unrelated to the others may not belong in the composite. section 6.3.1
Use a composite only when a single view answers a real reporting need and its ingredients can be justified. Otherwise, report the individual metrics: they may show stakeholders what changed and where to act. section 7.3.1 section 7.4
