n.a portion, within a body of records, selected for permanent retention through an appraisal process that identifies those records or types of records within that series that are expected to be more important to saveKepley 1984, 239When one employs statistical sampling, one must choose a series that is as homogeneous as possible and in which the individual file is not important. The statistical sample will target typical cases. If the spectacular cases are to be culled, then a subjective sample should be employed. There are two main types of statistical samples: systematic and random.Sly 1987, 60Indexing and cross referencing are an important subsidiary to arrangement. If the records were created with cross-referencing linkages, a probability sample would destroy that continuity, whereas a subjective sample could preserve them.
Subjective samples may be created via a set of thoroughly vetted criteria designed to identify those records of the highest importance or by a relatively unsystematic process. In either case, subjective samples are never statistically valid, so they are useful only in cases where the body of records is heterogeneous and there is a clear distinction to be made between records that are strong candidates for retention and those that are not.