Showing papers in "Educational and Psychological Measurement in 1966"

Weak Measurements vs. Strong Statistics: An Empirical Critique of S. S. Stevens' Proscriptions nn Statistics

Abstract: those who belong to what might be called the school of &dquo;strong statistics&dquo; has persisted for a number of years with little apparent change of attitude on either side. Stevens, as the leading spokesman for the weak measurement school, has asserted (1951) and reasserted (1959, 1960) the view that measurement scales are models of object relationships and, for the most part, rather poor models which can lead one far astray from the truth if the scores they yield are added when they should only be counted. At least two current statistics texts intended for psychologists (Senders, 1958; Siegel, 1956) present this view as gospel. Opposing this view, an assortment of statistically minded psycholo gists-e.g., Lord (1953), Burke (1953), Anderson (1961),

Similarity Analysis by Reciprocal Pairs for Discrete and Continuous Data

Abstract: discrete and continuous data. However, in its original form, it has the disadvantages of being complicated and laborious, sometimes requires iteration, and can lead to inconsistencies (McQuitty, 1955). By application of a particular theory of types (McQuitty, 1966) all of these problems can be solved, and a very simple method of hierarchical analysis can be developed, called Similarity Analysis by Reciprocal Pairs, applicable to both discrete and continuous data.

Personality Patterns in Pre-Adolescent Delinquent Boys12:

Abstract: IN a series of prior studies Tve have attempted to determine the principal dimensions of personality associated with juvenile delinquency in samples of institutionalized adolescents. The factorial analysis of data obtained from case histories (Quay, 1964b), questionnaire responses (Peterson, Quay and Cameron, 1959; Peterson, Quay and Tiffany, 1961; Quay and Peterson, 1964), and ratings of problem behavior (Quay, 1964a) has revealed a remarkable conceptual similarity in factor structure. Generally, the majority of variance is accounted for by three factors which we have labeled psychopathic-unsocialized, neurotic-disturbed, and subculturalsocialized. A fourth factor, usually accounting for less than 10 percent of the variance, has also emerged on occasion and we have called this vector inadequacy-immaturity. This structure also bears considerable similarity to that suggested by other studies of problem behavior in nondelinquent as well as delinquent children. (Hewitt and Jenkins, 1946; Jenkins and Glickmnn, 1947; Peterson, 1961 ; Quay and Quay, 1965). Since the development of a system of statistically covarying and

Integrative Complexity: Its Measurement and Relation to Creativity

Abstract: HARVEY, Hunt, and Schroder’s (1961) classification of individuals according to the level of integrative complexity of their personality structure has led to the accurate prediction of the individual’s diversity of perceptions (Driver, 1962), reaction to attitude refutation (Streufert, 1962), and attitude change in a sensory deprived and isolated environment (Suedfeld, 1964). The classification system, when used to compare homogeneous groups, has made possible the prediction of sensitivity to feedback (Lawrence, 1962) and group structure and performance on a variety of measures in a simulated environment (Tuckman, 1964). This evidence provides validity for the classification scheme and its underlying theory, as well as the measure used for identifying the individual’s level of integrative complexity. The measure used for this purpose has been the Sentence Completion Test (SC) developed by Schroder and Streufert (1962), a projective instrument, which can be considered to have been validated by the above studies. The purpose of this study was to evaluate an objectively-scored device to measure integrative complexity by determining the extent to which it was useful in predicting creativity, a characteristic

The Test-Retest Reliability of Children's Ratings on the Semantic Differential:

Abstract: THE present study of the reliability of the semantic differential technique was made to supplement the data from recent investigations (Di Vesta, 1964) providing positive evidence regarding its use with children. Several kinds of reliability might be adequately studied. Among these are item, factor-score, and concept-meaning stability. Osgood and his associates (1957) report highly acceptable levels of these various forms of reliability when the technique is used with adults. Jenkins, Russell, and Suci (1958) also report coefficients of stability of .90 or higher for mean scale values and

Note on Rank Biserial Correlation

Abstract: variables, n consecutive ranks, a dichotomy, a trichotomy, etc. In this paper we shall be concerned with developing a measure of the correlation between one variable comprising n consecutive untied ranks and a second variable comprising a dichotomy. We shall regard the dichotomy (scored 0, 1) as having a ranking variable underlying it. In this regard it is analogous to the assumption made in biserial correlation that a normally distributed variable has been forced into a dichotomy. Of major interest is a surprising equivalence which results between the coefficient found in this paper and a coefficient due to Cureton (1956). It is to present and prove the equivalence of these two coefficients, which though quite different in origin and derivation are algebraically equivalent, that this paper is being written.

Power of the Analysis of Variance of Independent Groups on Non-Normal and Normally Transformed Data

Abstract: THE science of data analysis, as described by Tukey (1962), (in contrast to statistics), has a goal closely tied to the real world: to derive certain conclusions about the &dquo;state of the world&dquo; from a sample of evidence. Statistics, on the other hand, is a mathematical discipline and can’t be evaluated by the same empirical considerations. Consequently, although data analysis finds statistics very useful, it is helpful to maintain a distinction when approaching the practical problems considered here. Two aspects of data analysis

Some Hypotheses Concerning test Adequacy1

Abstract: AN earlier paper (Fiske, 1963) presented the rationale for using the cumulative homogeneity model in developing psychological tests, especially in the personality domain. It also outlined in nontechnical form a proposed technique for assessing the adequacy of a test in terms of such a model. This paper describes the analytical technique more specifically. It also summarizes the findings from application of the technique to various scales from a wide variety of existing tests. From these findings are derived generalizations in the form of hypotheses. These state the characteristics of a test which are believed to contribute to high degrees of adequacy. The concern underlying this paper is measurement in the service of theory. The goal is to improve personality measurement so that personality concepts can be better studied and refined. The technical approach utilized for this purpose is unsophisticated. It involves simply a breakdown of the total observed variance in a set of test responses to determine t h e relative contribution of the means for persons and of the means for items, and also the remaining variance from all othey, sources, including especially the interaction between persons and items. A simple descriptive approach, rather than one using variance estimates, has been employed to

Social Desirability Estimation and "Faking Good" Well

Abstract: THE apparently widespread tendency of subjects to put their best foot forward when responding to items from personality and adjustment inventories has remained largely unanalyzed in the wake of recent efforts to detect, discourage or bemoan such behavior (Hanley, 1957 ; Voas, 1958 ; Edwards, 1957). Variations in the apparent strength of this tendency under different circumstances of assessment suggest that the phenomenon is in part a situational one. Even within a single assessment context, however, considerable individual differences in this tendency may be observed, suggesting, among other things, a person-situation interaction which reflects motivational characteristics of the subjects. Moreover, there is some evidence which suggests that when subjects are explicitly told to make a good impression, those who do dbestd are likely to be more psychologically sound than those who do less well at this task (Grayson and Olinger, 1957; Canter, 1963). The recent revival of interest in drole-playingd studies has produced a considerable amount of data which supports the observation that even under instructions designed to induce a uniformly high degree of good impression motivation, subjects will vary extensively in their performance (Wiggins, 1959; Walker, 1961; 1962). The present study was designed to test a hypothesis concerning one possible source of these observed individual differences. Whether dpsyching outd a potential employer or cooperating with the explicit instructions of a role-playing experiment, making a

Single and Multiple Hierarchical Classification by Reciprocal Pairs and Rank Order Types

Abstract: method for classifying people (institutions or other objects) into a hierarchical system based on indices of association between them (McQuitty, 1964). The method selects initial pairs of objects which are indicative of types in terms of common characteristics and generally expands and improves the types by incorporating additional objects and by eliminating characteristics which are irrelevant to the descriptions of the types. The method realizes both types of any size and successive levels of classification while at the same time limiting itself to reciprocal pairs. A pair of objects i and j are reciprocal if i is most like j and j is also most like i (i ~ j) ; but not if only i is most like j, or if only j is most like i; i does not need to be most like j just because j is most like i and vice versa.

Comparison of Two Approaches—Jan and Prof—for Capturing Rater Strategies

Abstract: IN an earlier paper the present authors (Naylor and Wherry, 1965) have described the use of the Bottenberg-Christal JAN technique (Bottenberg and Christal, 1961) for the isolation of rater policies. In that paper we reported in detail the JAN outcome for one of four air force specialties analyzed. Actually four specialties were studied and the present paper is based upon the complete study. The four specialties consisted of two supervisory levels of a mechanical specialty and two versions of an administrative (housekeeping) type of specialty, both versions being at the same supervisory level. The JAN technique is based upon defining the capturing of rater policy as the extent to which one can predict the actions of a rater

The Relationship of Self Report to Inferred Self Concept

Abstract: BOTH the self report and the self concept are currently among the most popular subjects for psychological research. Wylie (1961) for example, has reviewed 463 such articles. Many of the studies reported by Wylie as well as many more reported more recently have utilized the self report as a direct measure of the self concept. This use of the two terms as though they were synonymous has been vigorously challenged by Combs and his associates (Combs and Snygg 1959; Combs and Soper 1957; Combs, Soper, and Courson, 1963). They point out that the self concept is an internal organization of the individual’s perceptions about himself, whereas, the self report is a behavior representing what the individual is willing and able to say about himself when he is asked to declare his position. They insist that the self concept must be approached through some form of inference based on the individual’s behavior. This research was designed to explore the relationship of self report to inferred self concept in the hope of shedding some light upon this confusion.

ρ m As an "Error-Free" Index of Rater Agreement

Abstract: are all instances of the application of regression analysis to the evaluation of performance strategies in probability learning tasks. Similarly, Naylor and Wherry (1965), Wherry and Naylor (1966) and Madden (1963) have used the multiple regression model to capture the strategies or policies of judges in attitude research. There is an important distinction, of course, between these two kinds of research investigations. The major distinguishing feature lies in the presence or absence of a criterion of &dquo;correct&dquo; behavior.

Simplex Structure in the Grading of Essay Tests1

Abstract: IN December 1963, the 80,842 candidates electing the English Composition Test (ECT) in the College Board test administration were required to write a 20-minute essay. The test mas the first ECT to include free writing since December 1956 and the first to include a 20-minute essay since April 1947 (CEEB, 1947). I n a period of five days during the subsequent week, 145 readers gave each of the essays two independent readings. This study was designed to determine whether under these conditions the reading reliability would approach that obtained previously under somewhat less demanding circumstances. It cdminates ti 15-year period of research and development during which the College Board EngIish Composition Test has moved from an all-essay examination through various stages of objective and semi-objective format to its present specifications. The decision to reintroduce a 20-minute essay had not been made arbitrarily. Using data collected in 1961, SNineford (Noyes, 1963) had demonstrated that 20-minute essay scores based on five independent readings contributed uniquely to the validity of a onehour English Composition Test for predicting ft reliable essay criterion. The reading had been carried out during a five-day ex-

The Differential Prediction of College Grades from Biographic Information

Abstract: A present-day concern of researchers into higher education is prediction of academic accomplishment from nonacademic variables. Holland (1961) was interested in establishing a new basis for awarding scholarships to high aptitude students through the use of nonintellectual indices of creative performance. Other investigators such as Astin (1964) and Heilbrun (1965) have focused upon the dropout problem in higher education and the personal and environmental factors associated with failure. French (1963) wanted to compare aptitude, interest, and personality measures as differential predictors of success in major-field areas as a step in developing

Evidence on Problems in Estimating Common Factor Scores

Abstract: WITH the advent of the computer in psychological research, problems in estimating factor scores have come increasingly to the fore. In a recent article Horn (1965) compared various such estimation methods in terms of their theoretical-mathematical properties and by intercorrelating the scores estimated by the different procedures. His findings showed that the scores estimated by three complete methods were highly similar, as were those estimated by three so-called incomplete methods, but that these two broad classes of techniques yielded scores which correlated only .80 on the average. He found, also, that the intercorrelations among different orthogonal factors estimated by each of the three complete methods averaged .08 (zero being the desired value), whereas these intercorrelations for the incomplete methods averaged .14. This kind of information no doubt has relevance for questions concerning the usefulness of these various techniques in research and applied work. But an essential item of information is lacking. It is important to know not only how the factors estimated by the different procedures are interrelated among themselves, but also how each relates to the variables upon which a factor analysis is based. One purpose of the present investigation is to supply this kind of information.

Validity of Some Objective Scales of Motivation for Predicting Academic Achievement

Abstract: THE main purpose of this study was to test the hypothesis that a self-rating scale, designed to yield a fairly direct measure of academic achievement-motivation, will show higher correlations with achievement than will a measure of a person’s generalized need for achievement. This hypothesis was suggested by Myers (1965) but his study did not present a test of it with data on the same individuals. Other purposes were to check on the comparative predictive validity of (1) similar scales such as those derived from the Stern Activities Index (SAI), (2) a simple preference between academic and nonacademic subjects (Mayhew, 1965), (3) conventional verbal and nonverbal measures of general ability, and (4) selected combinations of these variables.

Three Equivalent Forms of a Semantic Differential Inventory

Abstract: scores. We used the forms to trace the momentary attitude changes towards one’s own voice at three successive but closely spaced times. Our experimental purpose, that of obtaining independent yet valid assessments of momentary attitude towards one’s own voice, directed that the items on the three forms be equivalent but not identical to each other. The instruments finally constructed proved to be most satisfactory and we present the forms here for the use of any colleagues who might wish to employ them in their own work. The forms, of course, were constructed to yield information about the subject’s own voice, and the concept to be rated by the polar adjectives is &dquo;my voice.&dquo; Nevertheless, there may be other concepts for which the items could be appropriate. Some students may wish to test whether these forms maintain their equivalence when they are used to describe other concepts. We are also suggesting our method of inventory construction as one that can assure equivalence. Too often investigators cull items from the Osgood, Suci, and Tannenbaum book (1957) without pretesting their relevance or their factor structure for the concept they wish to study. We therefore believe that whether or not the forms are appropriate for as-

Judging Complex Value Stimuli: An Examination and Revision of Morris's Paths of Life:

Abstract: a conception of &dquo;the good life.&dquo; Morris (1956) has epitomized these paragraphs as follows: Way 1 Preserve the best that man has attained. Way 2: Cultivate independence of persons and things. Way 3: Show sympathetic concern for others. Way 4: Experience festivity and solitude in alternation. Way 5: Act and enjoy life through group participation. Way 6: Constantly master changing conditions. Way 7 : Integrate action, enjoyment, and contemplation. Way 8: Live with wholesome, carefree enj oyment. Way 9: Wait in quiet receptivity. Way 10: Control the self stoically. Way 11: Meditate on the inner life.

Inhibitory Effects of a Pretest on Opinion Change

Abstract: IN recently reported studies (Lana and Rosnom, 1963; Lana, 1964b) it has been shown that the use of a pretest and its manner of presentation can affect the direction of opinion change when two opposed arguments on the same topic are utilized. Primacy refers to the success in changing opinion of an initial argument of two opposed communications. Recency refers to a similar success of an argument presented second. When the pretest is hidden from the subject (Lana and Rosnom, 1963) or missing altogether (Lana, 1964b), in the pretesbtreatment-posttest opinion change research design (Solomon, 1949), change occurs usually in the direction of the first presented of two opposed arguments (primacy). Recency, however, does not necessarily occur when the pretest is exposed. The principal concern of this study is to examine the effect of the pretest on opinion change regardless of direction (primacy or recency). The initial focus of the studies mentioned above (Lana and Rosnom, 1963; Lana, 1964b) mas on the effect of the present on the resulting directional effect (primacy-recency) , but a more important consideration may be whether or not the pretest generally in-

Need-press and expectation-press indices as predictors of college achievement.

Abstract: WHEN a college applicant chooses a particular educational institution it may be assumed he does so largely because he expects it will most adequately satisfy his personal needs. If he is accepted he may or may not find the directive influences or &dquo;press&dquo; of the college in keeping with his needs or his expectations of the institution. It has been hypothesized that the extent of agreement (congruence) between one’s internal forces (needs) and the external environmental forces (press) he encounters is positively related to his adaptation to that environment (Stern, 1960; Pace, 1961). This may be referred to as the need-press hypothesis. It may be, however, that it is not necessarily so much the congruence of his needs and the press as it is the congruence of what he expects (expectations) and the press he subsequently encounters that more strongly influences his adaptation. This may be referred to as an expectation-press hypothesis. These hypotheses involve abstract concepts, i.e., press, needs, and expectations, and the defining and measurement of their referents pose certain difficulties; Getzels and Guba (1957), for example, have effectively stressed the need for establishing distinctions, classifying such terms, and stating their interrelationships with behavior. Of the three terms, the press of collegiate environments has perhaps been most successfully defined and measured, by such instruments as the College Characteristics Index (CCI) (Pace

Verification of six personality factors

Abstract: ous items, rather than to analyze single items (Comrey, 1961, 1962, 1964, 1965, 1966; Comrey and Schlesinger, 1962). Items have been developed to fit a certain homogeneous item pool, but have been required to exhibit a statistical as well as a logical belonging before being actually assigned to that item group. On the basis of these previous findings, six major personality factors were selected for further study. Six factored homogeneous item dimensions (FHIDs) were chosen to represent each hypothesized factor. Each FHID, in turn, consisted of six items intended to be relatively homogeneous in content. Most of the FHIDs selected for a given factor had been good marker variables for that factor in previous analyses; some were newly developed for this study. The major purpose of the present investigation was to determine whether the postulated six-factor personality structure would

Nonintellective Predictors of Achievement in College

Abstract: DESCRIPTORS*PREDICTIVE ABILITY (TESTING), COLLEGE STUDENTS, *ACADEMIC ACHIEVEMENT, *COCURRICULAR ACTIVITIES, COMPETITIVE SELECTION, HIGH SCHOOLS, BIBLIOGRAPHIES, ACHIEVEMENT RATING, *COLLEGE ENTRANCE, EXAMINATIONS, OEL BEHAVIOR INVENTORY, NATIONAL MERIT SCHOLARSHIP QUALIFICATIONS TEST, VOCATIONAL PREFERENCE INVENTORY, ADJECTIVE CHECK LIST, CALIFORNIA PSYCHOLOGICAL INVENTORY, NATIONAL SCIENCE FOUNDATION, CARNEGIE CORPORATION, FORD FOUNDATION, EVANSTON

Test-Retest Reliability of the Epps

Abstract: IN his manual for the use of the Edwards Personal Preference Schedule, (EPPS), Edwards (1959) reported test-retest relibility coefficients for each of the 15 need scores plus a consistency score for a group of 89 University of Washington students who were retested after a one week interval. Subsequent studies have reported test-retest reliabilities of need scores after intervals of three weeks (Mann, 1958) and one week (Horst and Wright, 1959). Coefficients from these studies are reported in Table 1. Borislow (1958) has also reported test-retest reliabilities ranging from .65 to .91 for the profile of need scores for Ss after a two week interval. All these studies have demonstrated satisfactory short-term reliability for the test and consequently call into question the argument of Levonian, Comrey, Levy, and Procter (1959) who stated that the basic form of the EPPS item is one that encourages low reliability of response since the S must choose which of two statements seems more descriptive of himself while at the same time this

Improved Hierarchical Syndrome Analysis of Discrete and Continuous Data

Abstract: IN the isolation of statistically defined types, Hierarchical Syndrome Analysis has the advantage of being simple, brief, and applicable to both discrete and continuous data. In its original development, it does, however, have shortcomings; once Individual i has classified with Individual j then he cannot by himself classify with any individual other than j, and this is true even though j and k each have i most like them. In order to correct for this limi-

Comparison of Personality and Attitude Variables

Abstract: area of political and social attitudes resulted in the definition of several substantially correlated primary level factors which in turn produced a second-order general factor of Radicalism-Conservatism (Comrey and Newmeyer, 1965). The present investigation was undertaken for the purpose of determining the interrelationship, if any, of these two domains and in particular to ascertain whether any of the attitude measures might prove useful in measuring the main personality factors. A 216 item inventory was prepared which consisted of 51 factored homogeneous item dimensions, or FHIDs (Comrey, 1961), a social desirability scale of eight items, and a four item validity scale designed to detect random marking. Each FHID consisted of four items which were written to measure a specifically conceived variable and hence to correlate substantially with each other. Each FHID, except one, was included in the analysis because it had a substantial loading on one of the personality or attitude factors in at least one previous analysis, or it was newly conceived for this investigation with the hope that it would be highly related to one of these factors.