The Principles of Correlation

Abstract

In the early chapters we were concerned with the computation of “descriptive ” statistics. We sought simple numerical descriptions of distributions, such as central tendency or variability. Later we went beyond the mere description of the distribution within the sample. By considering the general theory of the reliability or significance of statistics, we learnt something of the conditions under which it is permissible to generalise from the sample to the parent population. Thus, we have been concerned so far mainly with statistics that describe quantitatively certain characteristics or attributes of either a sample or a total population.