Iowa State University Capstones, Teses and Retrospective Teses and Dissertations Dissertations 1971 A theory-based scale for measurement of afective responses to personality and atitude inventories An-Yen Liu Iowa State University Follow this and additional works at: htps://lib.dr.iastate.edu/rtd Part of the Experimental Analysis of Behavior Commons, and the Psychiatry and Psychology Commons Recommended Citation Liu, An-Yen, "A theory-based scale for measurement of afective responses to personality and atitude inventories " (1971). Retrospective Teses and Dissertations. 4557. htps://lib.dr.iastate.edu/rtd/4557 Tis Dissertation is brought to you for free and open access by the Iowa State University Capstones, Teses and Dissertations at Iowa State University Digital Repository. It has been accepted for inclusion in Retrospective Teses and Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information, please contact I I 72-12,568 LIU, An-Yen, 1941- A THEORY-BASED SCALE fOR MEASUREMENT OF AFFECTIVE RESPONSES TO PERSONALITY AND ATTITUDE INVENTORIES. Iowa State University, Ph.D., 1971 Psychology, experimental University Microfilms, A XEROX Company, Ann Arbor, Michigan THIS DISSERTATION HAS BEEN MICROFILMED EXACTLY AS RECEIVED. A theory-based scale for measurement of affective responses to personality and attitude inventories by An-Yen Liu A Dissertation Submitted to the Graduate: Faculty ir. Partial Fulfillment of The Requirements for the Degree of DOCTOR OF PHILOSOPHY Major Subject: Psychology Approved: Signature was redacted for privacy. Ir: Char^^of Major Work Signature was redacted for privacy. For the Major Department Signature was redacted for privacy. For the Graduate College Iowa State University Ames, Iowa 1971 PLEASE NOTE: Some pages have indistinct print. Filmed as received. UNIVERSITY MICROFILMS. ii TABLE OF CONTENTS Page INTRODUCTION 1 REVIEW OF LITERATURE 5 DEVELOPMENT OF THE SCALE 15 The Normative Model 15 The Conceptual Framework 23 The General Procedure 27 The Hypotheses 29 METHOD 31 Subjects 31 Measurement Scales 32 Procedure 34 RESULTS 40 DISCUSSION 63 SUMMARY 72 LITERATURE CITED 74 ACKNOWLEDGEMENTS 80 APPENDIX A: CPI FACTORS SCALES 81 APPENDIX B: PERSONAL TRAITS RATING SCALE 90 APPENDIX C: MANAGEMENT STYLE QUESTIONNAIRE 91 APPENDIX D 96 APPENDIX E 97 1 INTF.ODOCT ION Low internal consistency, considerable instability over time, very limited predictive value, and limited degree of convergent validity are the typical deficiencies of personal- ity scales. Evidence of these deficiencies is the large amount of remainder variance, the variance which is not asso- ciated with either people or item characteristics. The largest proportion of this variance is attriôntable to person- item interaction, the so called idiosyncratic response pat- terns (Fiske, 1963, 1966, 1968; Turner & Fiske, 1968). Due to the functioning of this source of variance, the equivalency and comparability of different persons with the same score are doubtful. A fairly recent summary of validity studies (Guion and Cottier, 1965) indicates little validity in personality measurements. This situation occurs to a lesser degree in achievement tests, but even here the consid- eration of response patterns has led to somewhat better meas- urement (true score estimation). For more than two decades, investigators (e.g. Cronbach, 1946, 1950; Edwards, 1957; Edwards and Walsh, 1964a, 1964b; Jackson, 1967; Messick, 1967) have indicated that variance from response sets, variables identified as social desirabil- ity and acquiescence, intrudes as a source of systematic error into the score obtained from personality inventories. 2 The responses of a person to a personality measure seem to involve response sets as well as responses to the content of the items. Recently, Edwards and Walsh (1964b) also pointed out that a high proportion of variance in personality inventories is attributable to individual difference in the tendency to use the center portion of a scale rather than using the extremes. Specifically, when subjects are offered three alternatives in responding to a question: "true", "?", and "false", they found that the tendency to use a "?" re- sponse is a reliable characteristic of an individual and is independent of personality as measured by the usual scoring procedures. Within the connotations of acquiescence, the ex- planation has been offered that a tendency to respond "true" to a positive statement or "false" to a negative statement contributes to error variance in personality tests. Other evidence indicates that acquiescence occurs more readily when items are weak in content (Trott and Jackson, 1967), or when subjects are uncertain (Peabody, 1964). Generally, there are two approaches to control response sets: (1) reduction cf the extent to which they are permitted to operate by controlling test items and response format and (2) assessment of the degree of their operation after they have occurred and then adjusting for them. Techniques of control and assessment of response sets involve partialing out response sets statistically, using balanced scoring keys. 3 using forced-choice response format, and differentially scoring for separate set and content components (Edwards, 1957; Helmstadter, 1957; Messick, 1961; Webster, 1958). Despite more than a decade of intensive investigation in this area, neither the explanation nor the control of this phenomenon is considered satisfactory. In review articles (Sorer, 1965; Block, 1965) evidence is cited and questions raised concerning the explanations offered by the aforementioned investigations. For example, Scott (1963) strongly argues and supports with data that there are wide differences among individual's conception of the desirable. Figgins (1966) supporting Scott's position, found at least six viewpoints in social desirability judgments of MMPI items. She concluded that forced-choice format is inadequate as a control for social desirability at the individual level. Item reversal, the p-opular technique used for the control of acquiescence, has been severely criticized on methodological grounds by Liberty (1965). This study proposes a normative model which takes ac- count of the individual response sets in scaling affective responses. Both theoretical assumptions and psychometric properties of this model will be explored. The purpose of this study is to reduce the error variance, mainly the person- item interaction, in attitude and personality measurement. The implication of a positive result of this research is a a procedure for obtaining and analyzing affective responses to statements contained in inventories such that the scale has meaning independent of content. Also the measure of a trait or attitude may be relatively free of certain kinds of re- sponse biases. The first part of this thesis presents the normative model and its derivation. The rationale and conceptual framework behind the model will then be discussed, â general procedure for obtaining responses and data processing follows. The second part explores the psychometric proper- ties in context of the model. Data will be analyzed and used to study the psychometric properties of the proposed model. 5 REVIEW OF LITERATURE As mentioned in the introduction section, response sets form a particular class of variables which have long been of concern to the psychometrician. Conventional studies and methods of control have been intensively reviewed by Edwards (1957) and recently in Berg's book (1967). More recent de- velopments in this area can be found in Fiske and Pearson's paper (1970). Here we shall review only the previous re- search which has direct implications for the development of the proposed normative model, and some of the more recent re- search. Dissatisfied with Fechner's law and the fact that the intervals derived by Thurstonian indirect scaling methods are not equal, Stevens (1946, 1957, 1958), based on the psychophysical scaling results from his so-called "direct methods", asserted that subjects make judgments on category scales on the basis of how certain they are that there is a difference between a standard stimulus and another stimulus on the same continuum. In other words, subjects make proba- bility judgments as to differences between stimuli. Support for this hypothesis came from the results of an investigation by Stevens and Galanter (1957). They studied the relation- ship between scaling results from direct psychophysical methods and the conventional scaling procedures such as equal