Introduction

The previous chapters have been concerned with a review of basic descriptive statistics. We now move to the much more important topic of inductive statistics. This chapter and chapter 6 will be devoted to an exploration of some of the implications of working with data from samples. The basic purpose of inductive statistics is to make it possible to say something about selected characteristics of a population on the basis of what can be inferred from one or more samples drawn from that population (see §1.3.4).

Samples can be obtained in various ways. Typically, the historian works with information extracted from surviving documents such as parish registers, manuscript census schedules, household inventories, farm accounts, private diaries, and records of legal proceedings. She will usually be working with a sample of such records, either because that is all that was ever available (only some people keep diaries), or because only some of the original documents have survived, or because the records are so detailed that it would be too expensive and time-consuming to extract all the information they contain.

Whenever sample data are used it is necessary to pose the fundamental question: How good are the results from the sample? How much can we learn about the population as a whole from the data available to us, and with what confidence? The overall answer to these questions involves three separate issues that must be considered briefly before turning to the statistical aspects.