scispace - formally typeset
Search or ask a question

Showing papers in "Educational and Psychological Measurement in 1967"


Journal ArticleDOI
Gene M. Smith1
TL;DR: In this article, it was shown that PEER RATINGS of personality can BF HELPFUL in CLARIFYING the relationhip between personality and academic success.
Abstract: ALTHOUGH PEER RATINGS WERE NOT REGARDED AS vA!!r INDICATORS IN TrcT ,TV^!ES. rKUJECT DEMONSTRATES THAT PEER RATINGS OF PERSONALITY CAN BF HELPFUL IN CLARIFYING THE RELATIONSHIP BETWEEN PERSONALITY AND ACADINIC SUCCESS. TEST CONDITIONS WERE DESIGNED TO AVOID METHODOLOGICAL PROBLEMS. PERSONALITY VARIABLES WERE CAREFULLY ANALYZED, AND A FORCED-CHOICE PROCEDURE WAS ADOPTED. TEST SUBJECTS WERE CAPABLE OF EFFECTIVELY RATING ONE ANOTHER BECAUSE THEY INTERACTED FREQUENTLY. PEER RATINGS WERE COLLECTED BEFORE THE FIRST MIDTERM EXAMINATIONS IN ORDER TO PREVENT AN ACADEMIC PERFORMANCE BIAS. THE RESULTS OF RELIABILITY AND FACTOR ANALYTIC STUDIES PERFORMED ON PEER-RATING DATA DEMONSTRATED A POSTIVE RELATIONSHIP BETWEEN PERSONALITY AND ACADEMIC SUCCESS. THE STUDY SHOWS THAT PERSEVERANCE,

191 citations


Journal ArticleDOI
TL;DR: A simple structure solution is one of the infinite number of possible solutions, but one which, for the various reasons Thurstone gave, might be expected to be more interpretable, more replicatable and, in general, more useful in scientific Tvork than any other solution as mentioned in this paper.
Abstract: ROTATION to simple structure in factor analysis was devised by Thurstone (1935) as a means for solving the indeterminacy problem. I n the usual factor analysis based upon a matrix of intercorrelations there is an infinite number of factor solutions, each involving the same number of factors and each allowing reproduction of the correlation matrix with the same degree of accuracy. A simple structure solution is one of the infinite number of possible solutions, but one which, for the various reasons Thurstone gave, might be expected to be more interpretable, more replicatable and, in general, more useful in scientific Tvork than any other solution (at least for some kinds of substantive problems). Thurstone viewed rotation, then, as a means of making factor analysis a more useful research tool. The various solutions obtainable by use of standard procedures for calculating factors represent different criteria of rotation, as Thurstone (1947) himself demonstrated. The principal axes solution, for example, is a rotation to a set of factors such that the first accounts for the maximum possible linear covariation among the variables; the second accounts for the maximum possible of such covariation after that accounted for by the first factor has been removed, etc. In a traditional bi-factor solution the first factor is rotated in much the same way as in the principal axis solution, but subsequent factors are rotated in a way to obtain large loadings for a small number of variables and near-zero loadings for the remaining variables. I n a Tryon cluster analysis all factors are rotated in this latter manner. I n any other so-called “direct” Eohltion there is, similarly, an implicit rotational criterion.

163 citations



Journal ArticleDOI
TL;DR: The concept of racial attitude is closely identified with certain traditional measurement procedures; namely, questionnaire scales of the Thurstone, Likert, and Guttman varieties as discussed by the authors, which are not appropriate to the test-taking capabilities of young children and, thus, are not applicable to the study of the development of racial attitudes in the preschool years.
Abstract: THE concept of racial attitude is closely identified with certain traditional measurement procedures; namely, questionnaire scales of the Thurstone, Likert, and Guttman varieties. Research with such scales has led to most of the existing knowledge concerning racial attitudes among older children and adults. Unfortunately, questionnaire scales are not appropriate to the test-taking capabilities of young children and, thus, are not applicable to the study of the development of racial attitudes in the preschool years. This limitation has led to the evolution of special research procedures for the assessment of racial attitudes and concepts among preschool children, and many imaginative approaches have been made (for example: Ammons (1950) doll-play technique; Morland’s (1962) picture-interview procedures; Stevenson and Stewarts’ (1958) figure discrimination, doll assembly, and incomplete stories; Goodman’s (1964) puzzle-interview, pictures and clay interview, etc.). While the use of such procedures has led to many interesting findings, it has been difficult to assess the degree to which these procedures tap the same psychological processes which are assessed among older persons by the traditional attitude scales. In other words, a need has existed for a procedure appropriate to the skills of the preschool child which would yield a measure of attitude which could be coordinated with the traditional concept of racial attitude.

94 citations



Journal ArticleDOI
TL;DR: The Impression Formation Test (IPT) as discussed by the authors was originally proposed by Streufert and Schroder (1963) and has since undergone considerable alteration in both administration and analysis.
Abstract: : An alternate is presented to the sentence completion test, which was specifically designed to measure one kind of complexity only: perceptual social complexity. This measure is called the Impression Formation Test. In its original form it was first proposed by Streufert and Schroder (1963), but has since undergone considerable alteration in both administration and analysis. (Author)

54 citations


Journal ArticleDOI
TL;DR: In this paper, it is argued that distractors can be constructed in a systematic fashion, and it is also argued that they can even be constructed on the basis of intuition as to what answer might be attractive, or by gleaning answers from tests which are first presented in open-ended form.
Abstract: DISTRACTORS, according to the definition given by English and English (1958), &dquo;are designed to be attractive ... to the respondent who does not know the correct answer.&dquo; Keeping the correct answer company is usually regarded the only function of distractors, and sufficient attraction is deemed sufficient qualification for being a good distractor. Hence, distractors are usually constructed initially on the basis of intuition as to what answer might be attractive, or else by gleaning answers from tests which are first presented in open-ended form. By contrast, it will be argued here that distractors can be constructed in a systematic fashion

52 citations


Journal ArticleDOI
TL;DR: If a hypothesis is investigated and substantiated, it gives support to a theory and the investigator develops other, hopefully more crucial, hypotheses which can in turn be tested.
Abstract: or by study of observations pursued and reported by others. Hypotheses are developed as logical deductions from the theory. If the theory is correct, then the hypotheses must also be correct. A more desirable relationship is one in which the theory is correct if the hypothesis is correct. Few if any hypotheses hold this relationship to a theory in any absolute sense. The dependability of the relationship between hypotheses and theory is usually a matter of degree. In most instances, a set of hypotheses is more dependable than a single hypothesis, and n carefully developed hypotheses are usually more dependable than any subset of them. Hypotheses have a scientific value which theories do not usually possess: their validity can be directly tested empirically. Science builds theory by testing hypotheses. Hypotheses are used to help generate research designs which test the hypotheses. If a hypothesis is investigated and substantiated, it gives support to a theory and the investigator develops other, hopefully more crucial, hypotheses which can in turn be tested. If, on the other

52 citations


Journal ArticleDOI
TL;DR: A procedure is thus needed for identifying the subgroups in a non-random sample of persons studied and formal models such as factor analysis and latent profile analysis are proposed as solutions.
Abstract: on a sample of individuals, is to detect any mutually exclusive subgroups that may exist. A procedure is thus needed for identifying the subgroups in a non-random sample of persons studied. Formal models such as factor analysis (Stephenson, 1936) and latent profile analysis (Gibson, 1959) have been proposed as solutions. A variety of looser procedures labeled &dquo;cluster analysis&dquo; have also been suggested. Within this class fall methods proposed by Thorndike (1953), McQuitty (1964), Sawrey, Keller and Con-

43 citations


Journal ArticleDOI
TL;DR: This paper found that subjects high on social desirability saw them as having a tendency to put the self in a favorable light, and that subjects scoring high on the SDS should also tend to rate themselves more favorably and see fewer discrepancies between themselves and their ideal selves.
Abstract: SOCIAL desirability has gained considerable attention as a response set, defined by Cronbach (1946) as a tendency to give different responses to test items than would be given if the content was presented in another form. Two aspects of social desirability have been defined, that of item content and that of response style (Christie and Lindauer, 1963; C r o m e and Marlowe, 1960; Wiggins, 1962). The former refers to sources of variance in the instrument whereas the latter refers to stylistic consistencies or sources of variance in the subjects. Crome and Marlowe (19GO) define social desirability as the need to obtain approval by responding in a culturally appropriate and acceptable manner. Correlations with the I< and L scales of the MblPI suggest cz quality of defensiveness or attempt to put the self in a favorable light. The question may be raised as to whether this tendency or need will carry over to self-reports on other instruments and the conditions under which the tendency will become most manifest. If the Social Desirability Scale (SDS) measures a tendency to put the self in a favorable light, then subjects scoring high on the SDS should also tend to rate themselves more favorably and see fewer discrepancies between themselves and their ideal selves. Using other measures of SD, Husek (1961) found that subjects high on social desirability saw them1The data analysis mas done a t Princeton University's Computer Center which is supported in part by National Science Foundation Grant NSF-GP579. ZNom a t Xent State University.

29 citations



Journal ArticleDOI
TL;DR: In this paper, the authors consider the overall relationship between a personality domain and an interest domain, i.e., the relationship between all scales of both a personality and interest test, and select correlations out of an intercorrelation matrix involving interest and personality variables disclose little or nothing about the overall relationships between the two sets of domains.
Abstract: are significantly correlated but not high enough to substitute interest appraisals for personality appraisals. Since most studies have involved the consideration of individual correlations between selected personality and interest scales, it is hard to grasp the overall relationship between a personality domain and an interest domain, i.e., the relationship between all scales of both a personality and an interest test. Selected correlations out of an intercorrelation matrix involving interest and personality variables disclose little or nothing about the overall relationships between the two sets of mea-


Journal ArticleDOI
TL;DR: In this paper, a large number of studies have investigated the basic regularities in the development of a number of psychological functions over the past five decades and some of these studies have been specifically concerned with the changes in the structure and organization of human abilities during various stages of life.
Abstract: OVER the past five decades a large number of studies have investigated the basic regularities in the development of a number of psychological functions. Some of these studies have been specifically concerned with the changes in the structure and organization of human abilities during various stages of life. Most of the early studies on changes in patterns of mental development, utilizing factor analysis, were summarized by Garrett (1946) who advanced the &dquo;differentiation hypothesis&dquo; as a reasonable generalization of the then-existing evidence. Burt (1954) has recently reaffirmed his position regarding a similar hypothesis on age differentiation which he had originally proposed as early as 1919. The hypothesis indicates that general ability is fairly unified during the early years of life, but becomes fractionated into a &dquo;loosely organized group of abilities&dquo; as age increases. An examination of the more recent investigations in this area (see review by Anastasi, 1958; Weiner, 1964), however, does not provide any conclusive evidence for or against the differentiation hypothesis even if it is modified to apply only to the first 18 years of life. Apart from the methodological inadequacies that may attenuate the usefulness of a number of these studies, changes in the organization of cognitive or psycholinguistic development over certain age spans (21/2 through 9, for example) have received comparatively little attention. The present study represents an attempt to fill this void as well as to provide data that fulfill most of the meth-

Journal ArticleDOI
TL;DR: In this paper, a 70-item multiple choice test was made by administering a different sample of seven items to each of 10 examinee samples consisting of 100 subjects each as opposed to administering the 70 items to a sample of 1,000 subjects.
Abstract: norms distribution for a 70-item multiple choice test could be made by administering a different sample of seven items to each of 10 examinee samples consisting of 100 subjects each as opposed to administering the 70 items to a sample of 1,000 subjects. Estimates of norm data were also obtained from each of the 10 examinee samples of 100 subjects each on the total 70-item test. Comparisons were made between the norm statistics (mean, standard deviation, and frequency distribution) and estimates of these same statistics derived from both the item samples and the examinee



Journal ArticleDOI
TL;DR: In this article, the authors used ACT scores, high school grades, and locally collected scores as predictors of grades in specific college courses as well as of overall college grade point average (GPA).
Abstract: ACT Research Services: The American College Testing Program offers several research services at no cost to member colleges and universities. This study is based upon results from one of the two prediction services.1 In the Standard (Plan A) Research Service, ACT scores, high school grades, and, optionally, locally collected scores are studied as predictors of grades in specific college courses as well as of overall college grade point average (GPA). Colleges

Journal ArticleDOI
TL;DR: In reading and listening comprehension tests, the test taker is exposed to a more or less clearly specified type and amount of information and is then asked to respond to questions based on the material.
Abstract: READING and listening comprehension tests are two examples of a general class of test-taking situations in which the test taker is exposed to a more or less clearly specified type and amount of information and is then asked to respond to questions based on the material. In this kind of task the tester asks the subject to recognize a rule or set of rules which relates the elements of the information


Journal ArticleDOI
TL;DR: In this paper, the authors consider the problem of maximizing the sum of the p pair-wise correlations between the observed variables and composites in a battery of tests, and propose a linear composites of the observations of n persons on p variables.
Abstract: IN developing a battery of tests, it is often desired, for technical or for conceptual reasons, to have test scores that are mutually uncorrelated. This note solves mathematically for p linear composites of the observations on p variables (the scores on a battery of p tests, say), composites which are mutually uncorrelated, and each of which is paired with the observations on one of the original variables such that the sum of the p pair-wise correlations between the observed variables and composites is maximized. We consider then a n X p score matrix Z, giving the observations of n persons on p variables, standardized so that each variable’s observations has mean zero and variance one. Let X be another

Journal ArticleDOI
TL;DR: An earlier paper presenting a model of curriculum evaluation, written with Maguire (1966), proposed the sectioning of the evaluation process into a number of steps, stressing that different (though not necessarily disjoint) sets of individuals were involved at each step.
Abstract: AN earlier paper presenting a model of curriculum evaluation, written with Maguire (1966), proposed the sectioning of the evaluation process into a number of steps. The steps involved the identification of broad, societal-institutional objectives; the rephrasing of these latter into operational terms more suitable for measurement operations; the translation of operationally-stated objectives into teaching practices; and the student outcomes after being subjected to these practices. In the Taylor-Maguire model, it was stressed that different (though not necessarily disjoint) sets of individuals were involved at each step, and in passing judgment upon the values involved in the transition from step to step. It is quite obvious to all evaluators that these different sets of

Journal ArticleDOI
TL;DR: In this paper, the authors consider a special case of a tie in the degree of association between objects to be classified into a hierarchical structure, i.e., when the tie is for the highest entry in the matrix.
Abstract: ONE of the values of Hierarchical Classification by Reciprocal Pairs and similar methods (McQuitty, 1966), is their ability to classify data without superimposing or forcing the kind of hierarchical classification which will emerge. Another is their ability to classify almost any set of data without encountering circumstances that necessitate arbitrary decisions. There is, however, one special condition which does require an improvement in order to avoid an arbitrary decision which might affect the resulting hierarchical structure. This condition pertains to one type of tie in the degree of association between objects t o be classified into a hierarchical structure. The effects of an arbitrary resolution of the tie can be extensive in certain unusual cases, i.e., when the tie is for the highest entry in the matrix, and for certain other pairs, when the mechanics of the solution provide that two or more pairs can be classified simultaneously.


Journal ArticleDOI
TL;DR: This paper shows that the unmatched item experiment has certain advantages over the usual matched item experiment depending upon the objectives of the study.
Abstract: THE theory of generslizability distinguishes between experiments in which items are matched and experiments in which items are unmatched (Cronbach, Rajaratnam and Gleser, 1963). These writers have shown that the unmatched item design leads to different reliability formulas for estimating the generalizability of a test. This paper shows that the unmatched item experiment has certain advantages over the usual matched item experiment depending upon the objectives of the study. Items are said to be matched if for every item in the test there is an zpr for every person in the sample of examinees. Items are said to be unmatched when the items are selected randomIy and independently for each person. Most studies using psychological tests have used the matched item design, and it might seem that unmatched item designs would be something to avoid. This is definitely not the case. It turns out that unmatched item designs have important applications in test experiments as will be shown below. Suppose tha t we are sampling from a population of persons, P, and a population of items, I , and we wish to generalize over both populations. Consider the following experiments.

Journal ArticleDOI
TL;DR: In this paper, a long list of &dquo;good&dqo; characteristics (Barr, 1950; Mitzel, 1960; Ryans, 1960b) of teachers is presented.
Abstract: on through a long list of &dquo;good&dquo; characteristics (Barr, 1950; Mitzel, 1960; Ryans, 1960b). No single teacher, however, can possess all, or even most, of these. traits. The list must be narrowed down. But which traits especially characterize the &dquo;good&dquo; or effective teacher? Would we all agree that teachers must be sensitive, warm, intelligent, moral, and conscientious above all other traits? Or must teachers be thorough, reliable, sympathetic, and loyal?

Journal ArticleDOI
TL;DR: In this paper, the authors expound some of the implications of statistical theory on procedures for hypothesis testing and estimation when the predictor variables are subject to error that are superior to usual procedures dictated by the inappropriate statistical model.
Abstract: MODERN textbooks on statistical methods uniformly caution the reader that regression procedures are based on the assumption that the predictor variables are not subject to error. However, vhen the predictor variables are psychological tests, this assumption is seldom, if ever, approximated. On the other hand, some of the implications of statistical theory clearly point to procedures for hypothesis testing and estimation when the predictor variables are subject to error that are superior to usual procedures dictated by the inappropriate statistical model. This statistical theory more or less (rather than precisely) defines the kind of inferences one can and cannot make from regression analysis based on such predictor variables and suggests a n alternative cross-validation design. The purpose of this paper is to expound rather than derive these principles. This exposition is intended to alert the quantitatively Sophisticated reader to these principles and to provide the less quantitatively sophisticated reader with statistical tools more appropriate for psychological prediction than those offered in textbooks. These principles require a psychologist to expound them apparently because the statistician, by temperament, requires more precise statements than those offered here. However, it is anticipated that these principles mill be endorsed (more or less) by most statisticians. Principle 1. Inferences d r a m from estimates of and tests of hypotheses concerning partial regression coefficients have greater

Journal ArticleDOI
TL;DR: In this paper, all pupils in the sample were administered the Watson-Glaser Critical Thinking Appraisal, Form Ym (WG), and the Quick Word Test (QWT).
Abstract: high school, grades 10-12. In September of 1965, all pupils in the sample were administered the Watson-Glaser Critical Thinking Appraisal, Form Ym (WG), and the Quick Word Test (QWT). The Henmon-Nelson Tests of Mental Ability, Revised Edition (HN) had been administered at an earlier date, and the scores were taken from cumulative folders. The WG subtest raw scores (Inference, Recognition of Assumptions, Deduction, Interpretation, and Evaluation of Arguments), HN IQs, WG total raw scores, and QWT raw scores were intercorrelated. The resulting correlation matrix was factor analyzed by the centroid method and four factors were rotated using the Varimax criterion for simple structure. Findings. The intercorrelations, means, and standard deviations

Journal ArticleDOI
TL;DR: In this article, the authors described the situation described by Campbell and Stanley (1963) as a time-series quasi-experimental design, where several observations are taken before and after the administration of a treatment, e.g.
Abstract: equally spaced time intervals and one wishes to make inferences about a possible shift in level of the time-series associated with the occurrence of an event at a particular point in time. This is precisely the situation described by Campbell and Stanley (1963) as a time-series quasi-experimental design. Several observations are taken before. and after the administration of a treatment, T, e.g., 01 O2 03 T 04 05 06. If there is an abrupt shift in the level of a time-series between the third and fourth observations, evidence of a treatment effect may exist. Campbell and Stanley recognized the shortcomings in the statistical tests they suggested as possible analytic techniques (Campbell, 1963; Campbell and Stanley, 1963). The model and statistical techniques developed by Box and Tiao (1965) appear to be the most suitable methods now available which might have application to the analysis of time-series quasi-experiments. The statistical model underlying the Box-Tiao analysis of change in level of a time-series is the integrated moving average model.

Journal ArticleDOI
TL;DR: The field of Small Group Research is a synthesis and critique of a field which the authors fail to define as discussed by the authors, and the field involves social psychological research with small groups but apparently not all small groups, if the 250 sample studies may be taken as a guide.
Abstract: There is no question but that Small Group Research represents an extensive undertaking. The sections not reviewed here-the relationships between substantively-defined variables, and the description and critique of the field-will undoubtedly be read with interest by the small group researcher. The bibliography through 1962 should prove useful, especially if supplemented with bibliographies previously supplied by Hare, Borgatta, and Bales, and by Raven. Small Group Research is a synthesis and critique of a field which the authors fail to define. True, the field involves social psychological research with small groups, but apparently not all small groups, if the 250 sample studies may be taken as a guide: not one of these deals with the most important of all small groups-the family. EDWARD LEVONIAN