Hove etc.: Lawrence Erlbaum Associates. xv, 567 p. (1988).
This is a handbook for behavioral or biosocial scientists who use statistical inference. It is assumed only that the reader knows how to proceed to perform a test of statistical significance. The first chapter presents a general description of statistical hypotheses testing. There are described four parameters of statistical inference: power, significance criterion , sample size and effect size .
Chapters 2 to 10 present different statistical tests. They are organized in the following way: 1. The test is introduced; 2. the index is described and discussed; 3. the power tables are presented together with the method of their use; 4. the sample size tables are presented and the method of their use is described. Chapters 2, 3, 4, 5, 6 and 8 are equipped with numerous illustrative examples. One can find the following tests: The test for means (ch. 2); the Pearson product-moment correlation coefficient (ch. 3); testing of hypotheses concerning differences between population correlation coefficients (ch. 4); the sign test and the test that some defined characteristic is one- half (ch. 5); testing of hypotheses concerning differences between independent population proportions (ch. 6); chi-square tests (ch. 7); the analysis of variance (ch. 8); multiple regression and correlation analysis (ch. 9); set correlation and multivariate methods (ch. 10).
The major changes in comparison with the first edition are: there is a new chapter 10 and a new chapter 11 treating the power analysis in more integrated form: effect size, psychometric reliability and the efficacy of “qualifying” dependent variables. Also, the references have been updated.