A General Consistency Check Procedure for Machine-Readable Data

Abstract
Several levels of data "cleaning" and editing are discussed, and a standard terminology for data-consistency checking and editing is developed. Consistency checking can occur at the level of data structure, syntax, or semantics. A consistency check either may be unconditional, depending only on the value of a single variable; or it may depend on the joint values of the members of a set of variables. The result of a consistency check may be expressed as the value of a vector whose elements indicate the consistency status of each variable in the set. The vector of return codes from a check can be managed and displayed by adding them to the set of variables, and by using them as input to a report generator or to an editing procedure.