INTRODUCTION TO THE THEORY OF SAMPLING .1. HETEROGENEITY OF A POPULATION OF UNCORRELATED UNITS

The reliability of analytical results depends on the control of all components of sampling errors, i.e. suppression of those components that can be suppressed and estimation and minimization of those components that cannot. All components of the overall sampling error result from the existence of one form or another of heterogeneity, There are two main forms: constitution heterogeneity and distribution heterogeneity. These place importance upon sampling in analytical reliability, and upon the theory of heterogeneity in the theory of sampling, Here, the heterogeneity of a population of uncorrelated units is discussed, and in Part II, the heterogeneity of a time-series of (potentially) autocorrelated units will be addressed, These two problems require different theoretical approaches, which is very seldom acknowledged, the trend being to wrongly process time-series as though they were populations,