Analysis and planning of experiments by the method of maximum likelihood

the long version

The present volume carries further than any previous publications, consideration of the treatment and analysis of experimental data, particularly the determination of parameters necessary for the description of curves, and of the systematic errors in regression analysis. The methods suggested for calculations are generally more complicated than the underlying statistical ideas, mainly because of the implicit reduction of actual problems to linear ones, without which the solution usually entails progressive approximations with very many arithmetic operations. High speed electronic computors are becoming available for such operations; but even so the concepts and techniques of linear regression analysis are usually necessary for the selection of the proper solutions.