Reasonableness is a property of a measurement that can be used to determine whether it may be erroneous without knowing what the correct measurement should be (for example by correlation). Reasonableness test typically take one of two common forms:
- Ratio – where two independent measurements are compared to one another to see if the result is realistic. For example, dividing the Area taken up by workstation spaces by the number of workstations (Resource Count) gives the area per workstation. A reasonableness test of this ratio might be that it is 3sqm (30sqft). If data is measured – say from a CAFM system – that is less than 3sqm then that suggests that either the Area is being under-reported or the number of workstations is being over-reported.
- Trend – this compares a measurement to the other recent measurements of the same item. For example, global Headcount might be measured monthly and a movement in one month of more than 1% might suggest that some misreporting has taken place. In this example, it is easy to see that reasonableness testing doesn’t necessary identify an error, it only shines a light on data that is suspect: the Headcount might have changed significantly from one month to the next if there had been a corporate acquisition or significant redundancies.
See Error Detection and Correction for more details.« Back to Glossary Index