http://worldcat.org/entity/work/id/1321187

Statistical procedures and their mathematical bases

"This volume is a revision and extension of a book by the same title published privately in lithoprinted form in 1935. The wide demand for the preliminary edition showed that there was a need for the type of presentation of statistics here offered. The characteristic feature of the book is the effort to explain the mathematical origins of the most widely used statistical formulas in terms that persons with comparatively little mathematical training can easily follow. We believe that, if statistical workers do not take their tools as magic but understand them in the light of their origins and assumptions, they will use these tools more intelligently and more safely. In order to make such understanding available to persons of little mathematical training we give the derivations in much detail. It is a well-known fact that the source of difficulty in mathematical reading by relatively untrained persons is largely the omission of steps which are supposed to be obvious. When these steps are supplied and when the use of specialized mathematical terminology is reduced to a minimum, much that would otherwise be closed to the reader is readily understandable. The title of the book is somewhat too pretentious. It might better be called Some Statistical Procedures and a Little Insight into the Mathematical Bases of a Few of Them. It is not, of course, a comprehensive treatment of the mathematical bases of statistics. It is intended to bridge the gap between the elementary courses, in which the formulas are given purely authoritatively, and the original contributions in the monographic press, which are often highly mathematical in character. In this edition we have included many of the statistical techniques advocated by R.A. Fisher and have undertaken to bring them into synthesis with classical statistics. We do not believe that the Fisher techniques will prove to have the importance for research in the psychological and social sciences that they have in the biological sciences, because in the former fields it is unnecessary to work much with small samples and with rough exploratory research. Certainly the Fisher techniques will only supplement and not supplant the classical methods in the psychological and social sciences. Nevertheless, we believe that the workers for whom we are writing in these fields are entitled to know what these techniques are. We have attempted to take the magic out of them, as we did also out of classical statistics, by explaining them in very simple terms and by showing how they fit in with the older methods. In this way we hope to bring it about that the workers in the fields for which we are writing will find some useful elements in them without grasping at them as some 'new magic' and unwarrantedly throwing away the vastly useful techniques of classical statistics as 'antiquated'"--Preface. (PsycINFO Database Record (c) 2010 APA, all rights reserved).

Open All Close All