Lewis R. Goldberg

In Response to Jackson's Challenge: The Comparative Validity of Personality Scales Constructed by the External (Empirical) Strategy and Scales Developed Intuitively by Experts, Novices, and Laymen.

Abstract Fifteen graduate students in psychology and 15 individuals with no formal psychology training were each paid to construct one 20-item scale to measure Sociability, Achievement, or Dominance. These 30 scales, plus the Personality Research Form and the California Psychological Inventory, were subsequently administered to 168 college females from seven living organizations. Average peer rankings were employed as criteria in order to compare the validity of personality scales constructed by different strategies. While the validity of Intuitive scales constructed by the average nonpsychologist was lower than that of the CPI External scales, the validity of scales constructed by the average psychology student and of the most reliable scales constructed by the nonpsychologists was essentially the same as that of the External scales. Moreover, the most reliable scales constructed by psychology students and the PRF scales were of approximately equal validity, considerably higher than that of any of the CPI scales.

The Reliability of Reliability: The Generality and Correlates of Intra-Individual Consistency in Responses to Structured Personality Inventories

tionally, such intra-individual score variability has been averaged across a sample of individuals to furnish one type of estimate of the reliability of the scale. Analogously, the intra-individual variability in the responses to a single item has been averaged across a sample of individuals in order to provide an estimate of the stability (and, indirectly perhaps, the ambiguity) of the item. Personality scales (and items) have traditionally been compared on the basis of these reliability estimates, unreliability being equated with measurement error.

Student Personality Characteristics and Optimal College Learning Conditions: An Extensive Search for Trait-by-Treatment Interaction Effects.

The goal of this research project was to discover those personality characteristics of college students which predispose them towards learning more effectively from one, rather than some other, particular instructional format. Over 800 students in each of two college courses were taught by one of four different methods, and three broad classes of criterion information were assessed: (a) knowledge of course content, (b) amount of extra-curricular reading, and (c) degree of student satisfaction. Each student completed an extensive battery of personality measures, which yielded over 350 test scores. The ratio of significant interaction effects to the number expected by chance was only 4 to 3. Consequently, new interaction scales were developed empirically in each course, and these were then cross-validated in the other course. In general, these new scales did not produce statistically significant interaction effects upon cross-validation, and scales constructed from items which produced significant interactions in both courses showed low internal consistency and low convergent validity. Factors which could have attenuated the strength of trait-by-treatment interaction effects are discussed.

The validity of rating procedures to index the hierarchical level of categories

Abstract Mean ratings of Concreteness vs Abstractness, Imagability, Categorizability, Meaningfulness, Familiarity, Number of Attributes, and Pleasantness were tested as predictors of the hierarchical level of 440 categories within 22 hierarchies. Neither Concreteness nor Imagability was related to hierarchical level, either monotonically or curvilinearly, whereas Number of Attributes was a near-perfect predictor. A second study investigated the process by which subjects rate the number of attributes in a category, using five instructional conditions, two focused on the number of associated attributes, two on the number of common attributes, and one on the number of instances. Only one kind of process seemed to be involved in all five types of ratings: The broader the category, the larger the numerical value that is naturally associated with it.

The Comparative Validity of Questionnaire Data (16PF Scales) and Objective Test Data (O-A Battery) in Predicting Five Peer-Rating Criteria

Thirty tests from the 1955 edition of Cattells Ob jective-Analytic (O-A) Test Battery, plus Forms A and B of the Sixteen Personality Factor Question naire (16PF), were administered to 82 male under graduates. In addition, each subject was rated by 7 to 11 close associates on each of 20 bipolar rating scales, 4 scales tapping each of 5 peer-rating fac tors. These peer ratings were used as criterion vari ables to be predicted by the 16PF scales and by the O-A Battery. The O-A Battery measures were slightly more highly related to one peer-rating fac tor (Culture); the 16PF scales were slightly more highly related to another (Conscientiousness); and the two sets of test variables were essentially equiv alent in predicting the other three factors (two of which showed no significant relationships with either instrument). The lack of any consistent su periority of the objective test scores over the ques tionnaire scales, coupled with some criticisms of the objective tests on purely logical grounds, should make one cautious in accepting the claims being made for the comparative validity of the O-A Bat tery.

What If We Administered the "Wrong" Inventory? The Prediction of Scores on Personality Research Form Scales from Those on the California Psychological Inventory, and Vice Versa

Equations are presented for estimating the scores on each of 20 PRF scales and each of 19 CPI scales from the other inventory. Estimates of the cross-validity of these inter-inventory predic tion equations ranged from about .40 to approxi mately .80. Scores on the typical CPI scale are predicted reasonably well by one or two PRF con tent scales, plus the Desirability scale. While scores on most PRF scales can be equally well es timated by a few CPI scales, one PRF scale—Sen— tience—appears to be unpredictable from the CPI.

A general scheme for the analytic decomposition of objective test scores: Illustrative demonstrations using the Rod-and-Frame test and the Müller-Lyer illusion☆

Abstract Objective tests of personality typically include a number of items or trials; the total score on the test is the sum of the subjects “correct” responses across all such trials. Normally, the trials are varied systematically across various facets of the test design, so that the total score represents a composite measure of accuracy averaged across these test facets. However, since only one score is computed for each subject, some potentially important kinds of individual differences—namely all those associated with each particular variation in the test design—are treated solely as measurement unreliability. Such a psychometric stance may serve to obscure more differentiated types of individual differences, with the result that composite scores from trials based on one type of experimental design may not be highly related to such scores from trials using a somewhat different design. The present paper presents a general procedure for scoring objective tests more analytically. To illustrate this general rationale, and to demonstrate its potential utility, data have been reanalyzed from two previous studies, one using the Rod-and-Frame test, the other the Muller-Lyer illusion. In both cases, the traditional global accuracy score did not correlate significantly with other theoretically related variables, while a number of component scores were quite highly related.