If repeatability asks whether you get the same result when you repeat everything identically, reproducibility asks whether you get it when something changes. That change may be the operator, the instrument, the laboratory, the time or the environment. Because more sources of variation add up, the dispersion is equal to or greater than that of repeatability.
Which conditions are varied
- Different operators applying the same procedure.
- Different instruments or equipment of the same type.
- Different laboratories or sites.
- Different days or seasons, with variation in the environment and in the condition of the instrument.
When reporting reproducibility, you must state which conditions changed, because the concept is relative to them. There is also an intermediate case, intermediate precision, which varies only some factors, typically within a single laboratory over time.
What the engineer sees in practice
Two shifts measure the same part and systematically get different results. If each operator is consistent with himself or herself but there is a difference between them, the problem is one of reproducibility: it may be due to placement technique, contact pressure, interpretation of the procedure or differences between instruments. A repeatability study does not provide that information.
How it is reduced
- Write the procedure precisely, including the measurement conditions.
- Train operators to the same criteria.
- Control temperature and other relevant environmental factors.
- Calibrate instruments against common references.
When it does not apply
If there is only one operator, one instrument and one location, reproducibility in the strict sense cannot be evaluated; what is observed in that case is repeatability or, over time, intermediate precision. One must not be presented as the other.