Since raw data from sample surveys and experiments often contain errors, attempts are usually made to edit data to make them more useful for analytic purposes. Increasing portions of this editing are now being accomplished by high speed computers. The great variety of such procedures reflects the conflicting desires of providing editing procedures which recognize peculiarities of particular data sets and yet are easily programmed and implemented. This paper proposes a scheme which is hoped to be suitable for a large variety of sample and survey data and yet be relatively easy to implement once a general program has been written.