Max D. Engelhart

A Note on the Analysis of Gains and Posttest Scores

basis of pretest scores, the matched groups formula yields the same t whether posttest scores or gains are analyzed. Where the preand posttest yield comparable scores and the groups are not initially equivalent, analysis of covariance yields the same F and identical adjusted differences between posttest means and mean gains. All of this is true even when the correlation between pretest scores and gains is negative. In view of the problems confronted in analyzing experimental data involving growth or gains, it seems strange that, so far as this author knows, no text or other publication explains what is stated in the paragraph above. When subjects are paired on a pretest which yields scores comparable to those of the posttest, to prove that the formulas

Book Reviews : Arthur R. Jensen. Educability and Group Differences. New York: Harper and Row, 1973. Pp. viii + 407.

Jensen’s article in the Harvard Educational Review, Winter of 1969, &dquo;How Much Can We Boost IQ and Scholastic Achievement?&dquo; was followed by the criticisms of Jerome S. Kagan, J. McV. Hunt, James F. Crow, Carl Bereiter, David Elkind, Lee J. Cronbach, and William F. Brazziel, and by Jensen’s reply to his critics. All of these became the content of Environment, Heredity, and Intelligence, Reprint Series No. 2. Further criticisms of Jensen’s efforts by Richa.rd J. Light and Paul V. Smith, Arthur J. Stinchcombe, F. S. Fehr, Thomas J. Cottle, and Martin Deutsch became the content of Sci-

10.00

This impressive anthology edited by the author of ~~er~~ tal Design: Procedures for the Behavioral Scieneee is intended for students of statistics at the beginning and intermediate level. Those of us who have taught statistics recognize perhaps three groups of students: those who view it as a required course, those who see it as a useful tool, and finally, those who come to appreciate the inner charm and beauty of the subject. This book should help enlarge the latter group at the expense of the former two. Moreover, the serious investigator may find that his conceptions o~ certain methods and tests are not shared by a sizable fraction of his colleagues. The issues are presented from a variety of viewpoints. The arguments given are not limited to matters of procedure and mathematics; they are just as apt to extend to the philosophy and the interpretation. Presented in this form the book should be most useful to a wide audience. The book is a selection of 45 papers grouped into 10 chapters. Most of the selections are reprinted from such journals as Psychological Bulletin, American Psychologist, American Statistician, Psychological Review, and Science. Several selections were especially written for this anthology, however, to complete the cov-

Book Reviews : Roger E. Kirk (Ed.) Statistical Issues, A Reader for the Behavioral Sciences. Monterey, Cal.: Brooks/Cole 1972. Pp. × + 401.

EDUCATIONAL AND PSYCHOLOGICAL MEASUREMENT in recent years. Its comprehensiveness makes it an excellent source of readings for use in the training of educational researchers and researchers in other behavioral sciences. It contributes the understanding of the philosophy of our science which should undergird training in technical procedures. Philosophy of Educational Research can become a very worthwhile additions to one’s professional library. It is gratifying to have so many well-remembered classic papers immediately available for further study. This book is primarily concerned with empirical research in education. &dquo;The readings in Part I have been selected because they shed light on the conditions that would have to be met if educational research were to make good its claims of being scientific.&dquo; (p. vi) In education and other behavioral sciences the questions of methods are not settled. &dquo;Accordingly, the readings in Part II, which deal with such topics as observation, inference, testability, causation, and models are especially relevant to all scholarly workers in education. For much of the scholarly writing on educational problems, although not empirical, is concerned with findings and methods of empirical research.&dquo; (p. viii) Among the 13 selections contained in the three chapters of Part I

5.95 (paperback:

This comprehensive text in educational measurement gives &dquo;more attention to practical problems and procedures than to theoretical formulations and issues.&dquo; After chapters devoted to the history of educational measurement and its functions in the process of education, Ebel discusses cognitive outcomes as objectives. The following items quoted from Ebel’s summary of chapter 3 are indicative of his educational philosophy: 8. A major goal of education is to develop in the student a command of substantive knowledge. 9. A person’s knowledge consists of everything that he has experienced as a result of his perceptions of external stimuli or internal thought processes. 10. Knowledge is a structure built out of information by processes of thought. 11. Verbal knowledge is a very powerful, uniquely human form of knowledge. 13. Command of knowledge is demonstrated by its use in problem solving, decision making, explanation, argumentation, and predictions.

Book Reviews : Harry S. Broudy, Robert H. Ennis, and Leonard J. Krimerman (Eds.) Philosophy of Educational Research. New York: Wiley, 1973. Pp. xvi + 942.

he is highly critical. In the tenth chapter, &dquo;Published Tests: An Evil or a, Blessing&dquo;? there is a good discussion of intelligence testing on a very general and elementary level, giving some important precautions for interpretation of IQ’s. There is no specific help, however, in interpreting obtained IQ’s. No mention is made of the terms &dquo;gifted,&dquo; and &dquo;mentally retarded.&dquo; It is emphasized that the interpretation of scores should always take into account measurement error, but there is no mention of how it should be taken into account. There is also a brief

14.95

test manuals intelligently. At the end of each chapter is a summary of it and a section called &dquo;Points to Ponder.&dquo; These points, very useful study guides, are among the many pieces of evidence that identify the authors as experienced and effective teachers. The first chapter &dquo;Measurement in Education&dquo; is the foundation for the other four, containing treatment of the basic principles of testing as well as the statistical concepts which are needed to understand reliability and validity. (This reviewer has never seen a better beginner’s explanation of those two concepts.) The three middle chapters treat aptitude tests, achievement tests, and interest, personality and attitude inventories. (The authors make a real effort to give a clear simple differentiation among these types of instruments, and yet they are not at all simplistic.) A final chapter &dquo;Educational Testing: A Broader Viewpoint&dquo; gives consideration to testing first as a school wide program, then as

Book Reviews : Robert L. Ebel. Essentials of Educational Measurement. (2d ed.) Englewood Cliffs, N. J.: Prentice-Hall, 1972. Pp. xiv + 622.

These are two generally impressive but quite different anthologies. Problems in Human Assessment ranges over a great variety of content while Principles of Educational and Psychological Measurement is much more restricted in scope. The former contains 74 selected readings and the latter }j. There are only six selections common to both and all six are classic papers on validity. In reading these two books, each from cover to cover, this reviewer sought to contrast in his mind the status of assessment and measurement at the time of his own graduate training with their status today. His first impression was one of awe with reference to the proliferation of concepts and constructs, methodologies, and instruments. With the expansion of intelligence testing, achievement testing, and educational research in the years immediately following World War I, it is commonly believed th^t both testing and research methodologies were uncritically applied. While this was generally true, the leaders of the testing and research movements were not uncritical. Consider this quotation from one of the founding fathers of both movements, S. A. Courtis:

9.95

teachers. The four major sections are concerned with innovative practices in administration and supervision, in elementary education, in secondary education, and in special adaptations at grade levels from one through 12. The readers of EDUCATIONAL AND PSYCHOLOGICAL MEASUREMENT will find those reports devoted to evaluation and assessment of greatest interest although the number of such studies is small. There are special projects concerned with the academically talented and with specific problems in guidance that may also be of some interest. Irrespective of one’s particular specialization in measurement and evaluation the contents of this volume do afford to keep one abreast of current practices and experimental programs in the public school setting. For this reason its value as a reference volume may be expected to last for a period of five to ten years.