Random sampling of time records is often necessary in wage and hour cases. Some employers maintain handwritten or machine produced records that are not efficiently converted to an electronic format. It is not always possible to convert these types of records in a timely and cost effective manner.
Random sampling is a tool that is frequently called upon in these situations. For instance, it is common to randomly sample a population of employees and then randomly sample a time period from that sample of employees. This type of cluster random sampling is common in these situations.
See for example: Sharon Lohr, Sampling: Design and Analysis, 2nd Edition, Brooks/Cole, 2010.