Norman Frederiksen

Measures for the study of creativity in scientific problem-solving

A set of Tests of Scientific Thinking were de veloped for possible use as criterion measures in re search on creativity. Scores on the tests describe both quality and quantity of ideas produced in for mulating hypotheses, evaluating proposals, solving methodological problems, and devising methods for measuring constructs. The tests were administered to 3,500 candidates for admission to graduate school in psychology, using an item-sampling proce dure. Reliabilities based on 45-minute tests were adequate for research purposes. Correlations with GRE scores were low, especially for scores based on number of ideas. Follow-up questionnaires were sent to students asking for information about grad uate school attendance, grades, accomplishments during the first year of graduate study, and self-ap praisals of professional skills. Scores based on quantity (number of responses, number of unusual responses, and number of unusual responses that were also of high quality) were significantly related to self-appraisals and to reports of such profes sional accomplishments as collaborating in re search, publishing scientific papers, and designing and maintaining research apparatus. The quantity scores also were related to indices reflecting the quality of the department attended and to conven tional evaluations of student performance. GRE scores were better at predicting these indices of quality but poorer as predictors of accomplishments and self-appraisals.

Differential Predictability in the Use of Test Scores

THE purpose of this report is to provide an illustration of &dquo;differential predictability&dquo; in the use of test scores. It describes an attempt to improve the usefulness of a test by identifying subgroups of individuals for whom the test is especially appropriate as a predictor. To the extent that such an attempt is successful, one can predict more accurately for the members of that subgroup; for those not in the subgroup the errors of prediction will be greater, and for this subgroup one can avoid

RESPONSE SET AS A MEASURE OF PERSONALITY

THE reliability of various specific response sets has been demonstrated on a number of personality, attitude, and ability measures [12, 13, 26, 31]. Some of the more frequently appearing response styles include: (a) acquiescence, the tendency to agree, like, or respond &dquo;true&dquo; [cf. 12, 16, 22, 24, 26, 33]; (b) evasiveness, the tendency to respond &dquo;indifferent&dquo; or &dquo;?&dquo; [24, 28] ; (c) the tendency to use extreme judgment categories [4, 29, 30] ; (d) inclusiveness, the tendency to give many responses when the number required is unspecified [34, 35, 36] ; and (e) the tendency to respond in a socially desirable or acceptable manner [14, 17, 20]. There are many other response sets not included in this list, such as the set to fake personality items, to skip difficult items, to take a test overly fast or slowly, to be consistent, etc. Some of these tendencies, such as the &dquo;criticalness&dquo; set discussed in the present paper, may be cognitive in nature. While the emphasis in measurement is usually to eliminate response sets because they introduce errors of interpretation in the logical validity of tests [12], the possibility has also been suggested that sets have stable, valid components reflecting consistent individual styles or personality traits [13, 20]. If characteristics of respondents or if interrelationships among variables are interpreted solely in terms of specific item content, then response sets represent errors to be avoided [8, 23, 26] ; but as possible personality variables these stylistic tendencies should be

Replication of a Study of Differential Predictability

THE concept &dquo;differential predictability&dquo; refers to the idea that people may vary in the extent to which their behavior is predictable on the basis of some predictor measure. It may be possible to identify individuals for whom a particular test is especially appropriate as a predictor and others for whom the test is not appropriate. If such knowledge were available, one could use the test for prediction with appropriate individuals, but would refrain from using it for others. This concept was the basis of a study by Frederiksen and Melville (1954) which showed that interest test scores were more closely related to engineering school grades for &dquo;noncompulsive&dquo; students than for &dquo;compulsive&dquo; students. The present study is a replication of the 1954 experiment. &dquo;Compulsiveness&dquo; was thought of as a tendency to be thorough, meticulous, and perfectionistic in one’s work, without regard to the amount of intrinsic enjoyment in the work itself. The hypothesis tested was that certain Strong Vocational Interest Blank scores would be more highly correlated with freshman grades in engineering for noncompulsive students than for compulsive students. The hypothesis was based on the expectation that compulsive students would tend to expend an amount of effort which is unrelated to inter-

The place of social intelligence in a taxonomy of cognitive abilities

Abstract Studies have yielded inconsistent results regarding the place of social intelligence in a taxonomy of cognitive abilities. An investigation is described in which measures of social behaviors obtained from interviews were related to a wide variety of cognitive variables. Significant relations were rare. It is suggested that the question to be posed is not the simple one of the place of social intelligence in the cognitive domain but rather one of the relationships of the elements of two large taxonomic systems. A Taxonomy of social intelligence would provide a set of definitions procedures that would improve efforts to replicate and build on previous research. More important would be the heuristic value of a structural representation of “social intelligence” based on theoretical relationships involving environmental, physiological, and genetic as well as psychological factors. Development of a taxonomy of situations would facilitate the study of interactions involving performance and environmental conditions.

Public And Private School Graduates In College

Intercorrelations among the predictors and Freshman Average Grade for the public and private school graduates who completed the freshman year are given in Table 1, along with means and standard deviations. The average of the freshman average grades for public school graduates was found to be 2.8 and for the private school graduates 3.1-the boys from public school were found to earn better freshman grades,

The Construction and Validation of an Arithmetical Computation Test

THIS article is a condensation of two reports (2, 5) prepared by National Defense Research Committee Project N-io6 during World War II under the direction of Harold Gulliksen, Project Director. The summary is presented with no thought that the particular test under consideration is of unusual interest, but rather to illustrate several methods of test construction and validation which are unique or rarely employed and which deserve to be brought to the attention of those interested

The Influence of Timing and Instructions on Cooperative Reading Test Scores

lections carefully first,&dquo; then answer the questions. The timing of the test is one of the two problems under consideration in this study. The administration of the test is somewhat simplified by permitting students to begin Part II whenever they finish Part I, in that less careful proctoring of the examination is necessary than would be required with separate timing. The decision not to time Parts I and II separately was based on the assumption that separate timing would not make enough difference in scores to warrant the complications of test administration which would be required by separate timing. The validity of this assumption has not been checked, however, and it seems reasonable to suppose that the variability of the comprehension scores and their correlations with the vocabulary

MEASURES FOR THE STUDY OF CREATIVITY IN SCIENTIFIC PROBLEM‐SOLVING

A set of Tests of Scientific Thinking were developed for possible use as criterion measures in research on creativity. Scores on the tests describe both quality and quantity of ideas produced in formulating hypotheses, evaluating proposals, solving methodological problems, and devising methods for measuring constructs. The tests were administered to 3,500 candidates for admission to graduate school in psychology, using an item-sampling procedure. Reliabilities based on 45-minute tests were adequate for research purposes. Correlations with GRE scores were low, especially for scores based on number of ideas. Follow-up questionnaires were sent to students asking for information about graduate school attendance, grades, accomplishments during the first year of graduate study, and self-appraisals of professional skills. Scores based on quantity (number of responses, number of unusual responses, and number of unusual responses that were also of high quality) were significantly related to self-appraisals and to reports of such professional accomplishments as collaborating in research, publishing scientific papers, and designing and maintaining research apparatus. The quantity scores also were related to indices reflecting the quality of the department attended and to conventional evaluations of student performance. GRE scores were better at predicting these indices of quality but poorer as predictors on accomplishments and self-appraisals.

Predicting Mathematics Grades of Veteran and Nonveteran Students

THE need has often been felt in the Counseling Service at Princeton University of a sound basis for predicting success in the introductory courses in mathematics. Since mathematics is a key subject for students who concentrate in the physical sciences and in engineering, the prediction of success in beginning mathematics courses may be quite helpful in counseling students regarding the advisability of entering such fields. An adequate method of predicting achievement in mathematics is also of value to faculty advisors, who, for example, frequently