Definition

A framework distinguishing whether a measure's level is correct (validity) from whether its methodology stays consistent so changes can be compared over time (reliability), with a further cross-sectional problem that even reliable data fails when different units use different methodologies. Illustrated by the bathroom scale: being off a pound consistently is fine for tracking your own change, but a scale recalibrated in the night ruins time-series comparison, and comparing yourself to a neighbor's different scale ruins cross-comparison.