Back Cover Blurb
- In experiments and investigations in the human sciences – in psychiatry, psychology, social medicine, criminology, sociology, etc. – the data obtained frequently are 'counts' of one kind or another. Numerous indispensable statistical techniques for the analysis of such data now exist – the most time-honoured being the χ2 test of association. In this book an attempt – the first of its kind – is made to collect the more useful of those techniques together and present them in some sort of logical order.
- In addition to dealing in detail with χ2 tests of association and 'goodness of fit', many other important topics are discussed. For instance it is shown how significant trends in contingency tables may be demonstrated, and how overall χ2 values may be partitioned to provide tests of specific hypotheses about the data under analysis. Ranking methods too are dealt with, and some new tests which are of value in the analysis of serial data – in particular data from learning experiments – are illustrated. There is also a chapter on taxonomy and classification, and another indicating the type of analysis required when constructing 'attitude scales'.
- The book is rather antiquated, and no-doubt the style is a bit stuffy.
- It also antedates the general availability of computer-assistance for the statistical calculations, which probably influences the options and general approach.
- But it’s brief, and the ratio of text to mathematics looks encouragingly high!
Science Paperbacks, Chapman & Hall, London, 1971 Reprint of the original 1961 Edition.
Text Colour Conventions (see disclaimer)
- Blue: Text by me; © Theo Todman, 2020
- Mauve: Text by correspondent(s) or other author(s); © the author(s)