n.choosing a portion of a body of records for permanent retention through a statistical process that ensures each file within the series has an equal chance of being included in the retained set; statistical samplingHindus, Hammett, and Hobson 1980, 42Random sampling is the key that allows sound inferences to be drawn on the basis of limited samples from large populations. Simple random sampling can be refined, however, in ways that will increase accuracy in some limitations. One of the refinements employed in the current study is stratification; often a population can be considered as being composed of several subpopulations or strata, each of which is relatively homogeneous. For both substantive and statistical reasons, it may be of interest to consider each of them separately. Even if an overall estimate of a population characteristic is desired, it may be more efficient to find separate estimates for each stratum and then combine them into an overall estimate.Hull 1981, 11Any form of selection other than some types of systematic sampling and all random sampling will, in fact, be subjective even though the criteria laid down may be legal, fiscal or administrative in character, but whereas in theory it is perfectly proper for the officer of an agency to determine what will be useful to him and his successors for the purposes of the agency served, if a third party determines selection with research interests in mind then questions of validity and subjectivity inevitably arise.Hull 1981, 13As will be seen later ‘random’ sampling not only involves a haphazard selection but also requires that every piece in a series has an equal chance of being selected, the very an[t]ithesis of ‘systematic’ sampling.Cook 1991b, 38Three common types of sampling are simple random, systematic random and stratified random. Simple random sampling is where the 1,400 chosen numbers are applied randomly across the entire population. It is thus most appropriate for very homogenous, very short-lived series without significant internal geographical, gender, chronological or other bias.Ham 1993, 77Random sampling. This method is designed to assure that each case, file, or other unit is as likely to be selected as any other. This method involves assigning each unit in the series a unique identification number, or utilizing existing unique numbers. A random number table is used to select the units that will be in the sample. Random sampling is usually applied to files where the units are filed by some non-random predetermined order such as place of residence or income level.Lyle 2004, 8Random, or statistical, sampling is, in theory, the most objective of the three main sampling methodologies. It allows for every record within a sample equal opportunity to be chosen.choosing a portion of a body of records for permanent retention through a process that employs a preset pattern of selection to produce a sample of the whole; systematic random samplingKepley 1984, 239–240Random sampling has been described as “purer” statistically because it dramatically reduces the possibility of the sample being biased by the systematic selection process itself.Cook 1991b, 38Perhaps the most appropriate method for archivists, therefore, is systematic random sampling, where only the first number is chosen randomly, and then every nth number thereafter is chosen until the full sample size (say 1,400 cases) is attained.choosing a portion of a body of records for permanent retention through a process that ensures the retention of a statistical sample of each type or category of file within the series in the retained set at the same percentage as it was within the whole; stratified sampling or stratified random samplingCook 1991b, 38Stratified random sampling is where the whole is broken down into logical “strata” (which may be defined as parts or subgroups or geographic areas or file blocks of the whole—like the categories in the United States Justice Department litigation case files mentioned earlier), and then each stratum is randomly or systematically sampled, thus ensuring that no part is overlooked.
