The original purpose of this book was to serve as a basis for training personnel of the United States Civil Service Commission and of the United States War Department who were to be directly engaged in the preparation of written or performance tests of achievement for predicting job performance of public personnel. It was the intention to present basic concepts and methods, not details of operational procedures peculiar to one or both of these particular agencies. The original draft of the book has been used as the basis for several training programs in both agencies. It is believed that it may be of interest to other agencies engaged in examining work through U. S. Civil Service Committees of Expert Examiners and. Boards of U. S. Civil Service Examiners, to State and local civil service agencies, and to private industry. Universities also may find it of value in preparing students for examining work, especially in the public service.
This book goes considerably further than merely listing a set of principles for the construction of test items. It attempts to provide an understanding of how a test fits into a total examination, of how to plan test content or how to determine what items should be constructed, of how items behave in combination with each other and with other parts of the examination, of how to measure job success, and of how to tell whether items are, in fact, contributing to the prediction of job success.
The contents of this book do not correspond to any academic course of instruction nor to any program for a graduate degree. It contains things not at present in any university curriculum and excludes a multitude that are. The sole basis for selection has been practicality. Concepts used in or directly pertinent to test development are included. Purely theoretical considerations found no place. The orientation throughout, unlike that of any available book on test development, is public personnel administration; most of the examples and illustrations used were obtained by probing into the experience of public personnel agencies.
Means, standard deviations, standard errors of differences, tetrachoric and point biserial correlation coefficients, and multiple regression equations are discussed. Some of these terms may sound technical; they are. They could have been lightly dismissed with a categorical statement that they are "beyond the scope of this book." The point is, however, that the methods or concepts indicated are . . .