Statistics Tables for Mathematicians, Engineers, Economists, and the Behavioural and Management Sciences

Statistics Tables for Mathematicians, Engineers, Economists, and the Behavioural and Management Sciences

Statistics Tables for Mathematicians, Engineers, Economists, and the Behavioural and Management Sciences

Statistics Tables for Mathematicians, Engineers, Economists, and the Behavioural and Management Sciences

Excerpt

For several years I have been teaching a first-year undergraduate Statistics course to students from many disciplines including, amongst others, mathematicians, economists and psychologists. It is a broad-based course, covering not only probability, distributions, estimation, hypothesis-testing, regression, correlation and analysis of variance, but also non-parametric methods, quality control and some simple operations research, especially simulation. No suitable book of tables seemed to exist for use with this course, and so I collected together a set of tables covering all the topics I needed, and this, after various improvements and extensions, has developed into the current volume. I hope it will now also aid other teachers to extend the objectives of their own applied Statistics courses and to include topics which cannot otherwise be covered very meaningfully except with practice and use of some such convenient set of tables.

In preparing this book, I have recomputed the majority of the tables, and have thus been able to extend several beyond what has normally been previously available and have also attempted some kind of consistency in such things as the choice of quantiles (or percentage points) given for the various distributions. Complete consistency throughout has unfortunately seemed unattainable because of the differing uses to which the various tables can be put. Two particular conventions should be mentioned. If, in addition to providing critical values, a table can be used for forming confidence intervals (such as tables of normal, t, χ and F distributions), then quantiles q are indicated, i.e. the solutions of F(x) = q, where F() represents the cumulative distribution function of the statistic being tabulated, and x the tabulated values. If a table is normally used only for finding critical values (such as tables for non-parametric tests, correlation coefficients, and the von Neumann and Durbin-Watson statistics), then significance levels are quoted which apply to two-sided or general alternatives. To make clear which of the two conventions is relevant, quantiles are referred to in decimal format (0·025, 0·99, etc.) whereas significance levels are given as percentages (5%, 1%, etc.).

I would like to express my gratitude to a number of people who have helped in the production of this book: to Dr D. S. Houghton, Dr G. J. Janacek, Cliff Litton, Arthur Morley and Peter Worthington, who have ‘vetted’ the work at various stages of its progress, to the staff of the Cripps Computing Centre at Nottingham University who have helped in very many ways, to Tonie-Carol Brown for help with preparing data for some of the computer programs, and to Betty Page for typing the text. Responsibility for any errors is mine alone—there should not be many, as the tables have been subjected to many hours of checking and cross-checking, but I would be grateful to anyone who points out to me any possible mistakes, whether they be substantiated or merely suspected. I would also greatly appreciate any suggestions for improvements which might be incorporated in subsequent editions.

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.