Until recently, two happy vices stole a good share of my time. Of those burglars, the first was the reading of science-fiction stories, a source of entertainment that I discovered at the age of eight, but kept manfully in check because I never stumbled upon a continuous supply. Of late, however, real life has mocked this branch of fiction so closely that I now miss in it the elements of fantasy and escape that used to delight. You might say that, being crowded out by the real thing, we armchair spacemen face technological obsolescence.

My second time-stealer has been “Murder for Pleasure,” as Howard Haycraft calls the detective story. Fortunately, I did not discover its existence until my twenties, so I had a little time for study before then. The great value of the detective story, quite apart from its educational aspects– after all, where else can one gather so much data about useful everyday questions such as tasteless poisons, homicidal law, and criminal psychology–is the fine training one gets in applied psychology and in the logic of everyday life.

Guttman's principal components for the weighting system are the item scoring weights that maximize the generalized Kuder-Richardson reliability coefficient. The principal component for any item is effectively the same as the factor loading of the item divided by the item standard deviation, the factor loadings being obtained from an ordinary factor analysis of the item intercorrelation matrix.

The question is raised as to whether the null hypothesis concerning the number of common factors underlying a given set of correlations should be that this number is small. Psychological and algebraic evidence indicate that a more appropriate null hypothesis is that the number is relatively large, and that smallness should be but an alternative hypothesis. The question is also raised as to why approximation procedures should be aimed primarily at the observed correlation matrix R and not at, say, R−1. What may be best for R may be worst for R−1, and conversely, yet R−1 is directly involved in problems of multiple and partial regressions. It is shown that a widely accepted inequality for the possible rank to which R can be reduced, when modified by communalities, is indeed false.

A theory for discrimination learning which incorporates the concept of an observing response is presented. The theory is developed in detail for experimental procedures in which two stimuli are employed and two responses are available to the subject. Applications of the model to cases involving probabilistic and nonprobabilistic schedules of reinforcement are considered; some predictions are derived and compared with experimental results.

A three-component model for comparative judgment which allows for individual differences in preference is proposed. An implication of the model is that errors in the observed proportions due to sampling individuals in paired comparisons experiments are correlated. By neglecting this correlation, Mosteller's test for the method of paired comparisons tends to accept falsely the goodness of fit of the Case V solution. It is shown that bounds may be set for the correlation effect which make a valid test possible in some cases and provide useful standard errors for the estimated affective values.

Three techniques are commonly employed to capitalize on a concomitant variate and improve the precision of treatment comparisons: (1) stratification of the experimental samples and use of a factorial design, (2) analysis of covariance, and (3) analysis of variance of difference scores. The purpose of this paper is to compare the effectiveness of these alternatives in improving experimental precision, to identify the most precise design and the conditions under which its advantage holds, and to derive, in the case of the factorial approach, recommendations as to the optimal numbers of levels.

A formal set of axioms is presented for the method of successive intervals, and directly testable consequences of the scaling assumptions are derived. Then by a systematic modification of basic axioms the scaling model is generalized to non-normal stimulus distributions of both specified and unspecified form.

The expected value of mean square concept is used to determine the effects of the presence of interactions in the single Latin square design on F tests. The results indicate that as the number of random effects included in the experiment increase, more F tests are unbiased, and that some of these are valid F tests. However, when F test bias does occur it is almost always of a negative nature so that the conclusions stated are conservative ones. Positive F test bias may occur when the triple interaction is extant and when zero or one random variate is included in the experiment.

A coefficient of association τ′ is described for a contingency table containing data classified into two sets of ordered categories. Within each of the two sets the number of categories or the number of cases in each category need not be the same. τ′=+1 for perfect positive association and has an expectation of 0 for chance association. In many cases τ′ also has a -1 as a lower limit. The limitations of Kendall&s τ′a and τ′b and Stuart&s τ′c are discussed, as is the identity of these coefficients to τ′ under certain conditions. Computational procedure for τ′ is given.