..-~ I . .n. r- ...J V. J. L. .... .V r... L. n: :w ._. I. L. «a n. .. § u c,- ,§¢~ ‘ ~ ~_ ABSTRACT A STUDY OF THE GENERALITY OF DIMENSIONS OF SOURCE EVALUATION by Albert D. Talbott This study investigates the generality of the dimensions of source evaluation developed by Berlo, Lemert, and Mertz (1966) across people and for different kinds of sources. Berlo et al. adapted the semantic differential developed by Osgood, Suci, and Tannenbaum (1957) for evaluation of communi- cation sources. The concepts used are sources or potential sources of messages. The bi—polar adjective scales used are ones deemed relevant to evaluating information sources. The criteria for scale selection grew out of the work of Hovland, Janis, and Kelly (1953). Respondents are asked to use the scales to evaluate the sources in the usual semantic differential fashion. The results are then factor analyzed with each factor representing a dimension of source credibility or source evaluation. Primarily three factorsruunaemerged from their work: safety, qualification, and dynamism. Berlo €2,223: in doing the factor analysis of the adjective scales, summed across both subjects and sources.. This summing implies using all source ratings on a particular scale for each subject as individual observations in computing the correlations . \ n P ~ 0 t Y. Y. . . .— . y do . I . I. o :u o u o u a: . .ru .. . .u u . v u. . .r Fifi P—u FAA §- .- .s J N. o n: u: . nu» F . . .. . 2. . . . 4 L: F . o . p: r . a On .—v 'u 7 u . ‘x o . .n a 4. . pafi . van-.QNAU— do . :~ a: 4 a at ‘ .v auv 1. — A. v r u .r u .~v n. s 6 xv n. .. tau .p ~ . s hr s . a u s . v .tn . u .. ~ u v r. u . . ‘ 1- .: fl. s v Baa a: a. . V . a - v v a. a c C Q . .. .t 2.. .. r . ‘ C Wig ;~ AT. :— 4. -.. .. . g v .- u v .D a E. Albert David Talbott among scales. The number of observations for each correlation would be the number of subjects times the number of sources. The factor structure under such conditions could have been derived from variance attributable to subjects, sources, or an interaction of both. The present study still uses factor analysis, but in such a way that the sources of variance which contribute to the factor structure are not so completely con- founded. In this study, subjects are asked to evaluate two sets of sources on the bi-polar adjective scales develOped by Berlo §£.§ln One set of sources is judged to have beliefs and ideas similar to the subjects' own beliefs and ideas and the other set is judged to have beliefs and ideas different from the subjects' own. The two sets of source evaluation data thus obtained from each subject are factor analyzed individually. For this study, 99 subjects used 34 semantic differential scales to rate 48 communication sources (2 groups of 24 each). In all, 198 factor analyses (3H variable, 24 observation) were computed (2 for each subject). The individual dimensions of source eValuation thus obtained are examined both for differences t>€3tween the two kinds of sources and for their relationship tC> dogmatism. Also, the degree of correSpondence is examined 1Nitween the individual three factor solutions and the three factors isolated by Berlo g3 31; Rokeach, in delineating the defining characteristics D‘ “‘7‘ , Iv ‘5 ' D .'.. ..... 5,- . up - nu "Ifi' ‘1..- l a . ”3-;nr-ga,‘ _- u nn~mnouadl o,- v.- .....,~. A.“ . ‘ — ‘ "iuv~~ v. t "H a a- \— ud .-_‘_ . _ l 'P"y~. w ""'-“I‘A‘M. of. v... \ v - Albert David Talbott between Open and closed systems, focuses on the differences in the way the belief systems of the two types are structured or organized. Closed systems (high dogmatics) exhibit a greater tendency than do Open systems (low dogmatics) for the sub- systems of the disbelief system to decrease in differentiation the closer each is to the disbelief end of the belief-disbelief continuum. If information sources can be viewed as elements of the belief-disbelief systems, then the evaluations of these sources should be organized in a way that is similar to that part of the belief-disbelief systems with which the sources are associated. The two main hypotheses for this study were derived directly from Rokeach's belief system theory. Stated in general form, the hypotheses are: I. The number and/or differentiation of the dimensions of source evaluation will be greater for sources associated with the belief end of the belief- disbelief continuum than for sources associated with the disbelief end. II. The difference between the number and/or differentiation of the dimensions of source evaluation for sources associated with the belief end of the belief-disbelief continuum and for sources associated with the disbelief end will be greater for closed systems than for Open systems. Comparisons between the individual dimensions of source evaluation and the dimensions of source evaluation isolated by Berlo EE.E£° are made. The following two informal general hYpotheses are offered to facilitate these comparisons: --" - P. - “r .oiyA. A .- v.-.A ‘V- .uv- . a .. - ." fi‘wnrc : "-.NV-n~- ... — ‘ . . ' f‘ .‘ “\ .. y- | -... " A..u k .'.“'J ~u- .'.. -.., -,. h“ willv '.. .. .- _A-~ ’y f_‘H\ ..Iu.~.‘."~'~ .- C 'J---‘~ ‘ - _- \ '."'U~ -.. k ' ' ..~A‘fA-‘~ A "\ \ ‘-'~"."~‘~ '- . I fp- ..“-‘ n ‘ \ ‘ ‘.-“‘y~‘~. A A "...‘r vL'- . ~ 0 ‘~2.r r ‘ 3|- ‘U‘. v.‘v -r A ”L A. ~UI‘V :U a \ .. ‘ . . .- . “1:" n A, .\I..." Q \~ ~~ 4.“ ‘ a h ‘4'. ‘ ..; K :F‘ ‘- ‘h‘n‘v‘ \— ‘- ._.' t t M ‘~‘~ fi.‘ '~ ‘4‘ "v 5-. ‘ . ‘O‘ ..-‘,‘J~'V Ur ”.4 Albert David Talbott III. The nature or kind of dimensions of source evaluation used for the sources associated with the belief end of the belief-disbelief continuum will differ from those used for sources associated with the disbelief end. IV. Closed systems will differ from open systems in the nature or kind of dimensions of source evaluation used. Operationalizations of "the number and/or differentiation of the dimensions of source evaluation" used to test Hypotheses I and II included number of factors rotated using the Kiel-Wrigley criterion, proportion of variance accounted for by the one, two, three and four factor solutions, proportion of variance accounted for by the largest factor in the two, three and four factor rotated solutions, and a simple structure index computed for the two, three and four factor rotated solutions. Both hypcmheses for each Operationalization were tested by different effects in the same mixed analysis of variance design. Hypothesis I was generally supported by the results of the analysis. Hypothesis II was not supported. Operationalizations of the degree of correSpondence be— tween the individual three factor rotated solutions and the Berlo gt_al, results used to test Hypotheses III and IV in- cluded a series of dichotomous characteristics. These were designed to make the comparisons both on a factor to factor basis as well as on an overall basis. The statistical analysis included three nonparametric tests all using the chi—square a ‘ a-“ r .-o¢* .-\..«... .--~ I n . 'fQQAr - "r y t.-~-M-AS cs... “P. A. ‘4 y..- - v- . - . . _._ Fr -_,..— '4 ' VO‘uAb-.. .....v a . - ~ ‘ yea 0“.':" ~ 1 a... .4! ...A...- ._ - ‘ -'. -.,_ V: - _ ‘ --....,._-' . . q ‘.¢.. .- y ...‘ --:'~- ‘- - . . ‘Q .‘- C“- 19; ,— ‘.v -vd4-- - . l .._r 2 -Y‘ ""- -~' -. .- . .. " no -, 1 - _ \ h"". V'J‘. . '. .. I ~ 2 -. ~_ -. a...l "—.... ‘ I‘ Q ..“_ d “ 2,. .n...«. V'v-‘ . - ... 1- ;Y‘;r’:“ -‘ F" ~\ v- v..-v~ ‘ I ‘L‘n - ‘ ‘ -v. I ‘~. .- . ‘-V _" ‘I‘ V‘ “-‘ ‘ a T‘ a...“ g- .- l c ‘. V »;':Tr h."'~:. --.'- _ , u. an" ." §*._d’ l U 2V“. “V “n_‘ >- \ . . x- ‘ . l‘ "L 5 fl ‘1 y -d. I ’ o 1', ‘2? 'l .._“- ‘ x .. ~:-‘ 4 -_‘ . I ’- -'~~—-." " 'C —“ ' 3 --. ~n-4 ' r ~- 1‘ v_“-‘ Albert David Talbott distribution. Hypothesis III was generally supported by the results of the analysis. Hypothesis IV was not supported. Conclusions regarding the generality of the Berlo et_al, dimensions of source evaluation are as follows: 1) the generality of the safety dimensions is highest and is acceptable for most conceivable purposes; 2) the generality of the qualification and dynamism dimensions approaches that of the safety dimensions for the belief sources only; 3) for these two dimensions, generality is lower for disbelief sources; 4) generality is lowest if representatives of all three Berlo EE.2$° factors are required among the individual factors; and 5) conclusion four is true for both kinds of sources, except that for dis- belief sources, it is even less likely that all three Berlo 23 al. dimensions will be represented in the individual three factor solutions. Beferences: Berlo, D. K., Lemert, J. B., and Mertz, R. J., Dimensions for ALEvaluation the Acceptability of Message Sources. Report of research‘supported under ContractvOCD-PS-6471, Office of Civil Defense, Department of Defense, as part of Michigan State University's continuing research program on the public acceptance of civil defense messages, 1966. Iknflland, C. I., Janis, I. J., and Kelly, H. H., Communication and Persuasion, New Haven: Yale University Press, I953. OSgood, C. E., Suci, G. J., and Tannenbaum, P. H. The Measurement of Meaning, Urbana: University of Illinois Press,l957. Rokeach, M. J., The Open and Closed Mind. New York: Basic Books, 1960. (/J B“ '. 3 V‘" U \.-.l A t "v V gut ‘0 c"“3 4 H-,, Accepted by the faculty of the Department of Communication, College of Communication Arts, Michigan State University, in partial fulfillment of the require- ments for the Doctor of Philosophy degree. Guidance Committee: 8M®nga Chairman ' Lam/z ‘ "' Z2221; me; 77/ flZém DirectOr of Thesisr A STUDY OF THE GENERALITY OF DIMENSIONS OF SOURCE EVALUATION By Albert David Talbott A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Communication 1968 C Copyright by Albert David Talbott 1968 o “' ‘rPC-i... Va :Rnfig-“ « > n a .,;~r~:r "u... m». _ - P . .‘Y‘,r , qua-a.‘ (I) .. '.A'Y"’, (l- - ‘ ~4-..n__ . I "5‘ I‘V'Py- ~..u-. -.. ~40. ’. ‘, PA ‘ ~ "V- 't 5" “ V .'v H .. 3.";- .' ~ “ ~ v-a v \- ~ LL ~ \ " H - A s.‘ .- ..“ »'”. ..~" ‘n.,' A“‘ ' .._ s ,- Iv '07“. U ‘~..:; I no g ‘ ‘ \ “V V“ -.‘ . I. “‘ ‘ F‘ A ‘ 9- ~‘.‘; A‘ ‘18.... ~‘ .h» o _— ‘- y-“ ACKNOWLEDGMENTS In particular, I would like to acknowledge my debt of gratitude to the following people: Dr. Erwin P. Bettinghaus, chairman of my dissertation committee; Drs. Hideya Kumata, Terrance M. Allen and William A. Faunce, the members of my guidance and dissertation committee; Dr. David K. Berlo, chairman of the Department of Communication, for his generous support of both me and the project; Jeffrey Katzer, for writing the specialized programs needed for the computer analysis; Mrs. Anita Immele, Miss Diane Estfan, Mrs. Teresa Shen, Mrs. Betty Darlington, Mrs. Jean Loh, and Donald Clements, all members of the staff of Research Services of the Department of Communication who got my data through the computer with a Special thanks to Mrs. Immele, who supervised the whole process; the unknown members of the MSU Computer Center staff who made the extensive use of the computer possible; members of the College of Communication Arts Computer Allocation Committee whose generous computer time grants made the analysis possible; Dr. Malcolm S. MacLean, Jr., my former guidance committee chairman and present colleague, director of the School of Journalism at the University of Iowa, for ensuring that I had a Supportive environment after I moved to Iowa in which I could finiSh writing this dissertation; Norman VanTubergen, chief programmer for the Mass Communication Research Bureau of the ii an. I - ‘.r‘n ~o ‘ svotib- v. . A ‘. V (‘- \ J., d-nd ,. ~ g \-r«- A. s...--.- 'v. n- ' \\K "H-Q 04-5.3. r H ‘\ I g V: \f‘y. “on. .A-A...“- “V - ' r “l r" ‘ f"....~. \ '0.-. School of Journalism at the University of Iowa, for help in winding up the computer analysis after my move to Iowa; Mrs. Carol Oukrop, School of Journalism, for her much needed help in editing the text; Miss Cindy DeCounter, for typing the preliminary manuscripts, Mrs. Shirley Sherman, for the final typing of the manuscript and handling that myraid of details that are required in winding up a Ph.D. program; and finally to the large number of scholars who were kind enough to give me time to discuss various aSpects of this research, too numerous to name. Last but not least, I would like to dedicate this endeavor to my wife, Rose, my son, David, and to the memory of my late mother, Mrs. Maude E. Talbott. - "P-‘n - .. . ,— --.-._.","‘~ 1 4 TABLE OF CONTENTS CHAPTER I INTRODUCTION . . . . . . . . . . II STUDY DESIGN AND METHODOLOGY III RESULTS . . . . . . . . . . . . IV DISCUSSION AND CONCLUSIONS . . . BIBLIOGRAPHY . . . . iii Page 25 73 131 150 A) LIST OF TABLES Table Title Page 1 Differential Differentiation Within Belief and Disbelief Systems for Open and Closed SyStemS O O 0 I O O I Q 0 O O O O O O O O O 13 2 Mean and Range of Dogmatism Scores Associated with Three Dogmatism Groups . . 2O 3 Simple Randomized Analysis of Variance Results for the Proportion of Variance Accounted for by the Largest Factor in the One (First Principal Axis Factor), Two, Three, and Four Factor Rotated Solutions . 2O 4 The Number of Times each of the Evaluative Scales had its Highest Loading on the Largest Factor for the Three Dogmatism Groups . . . 23 5 Source Belief-Disbelief Rating Scale Positions Not Used When Not All Nine Were Marked . . 29 6 Frequency Distribution of Loadings on the First Principal Axis Factor . . . . . . . . 31 7 Factor Analysis of the Fifty-three Subjects' Ratings of the Sources (53 variables, 123 Observations): Factor Loading for the Two Factor varimax Rotated Solution and the First Principal Axis Factor . . . . . . . . . . . 32 8 Correlations among Factor Scores for First Principal Axis (lst PA) Factor and the Two Factor varimax Rotated Solution (2- -FS). . . 38 9 Correlation Table among the Factor Scores for the First Principal Axis Factor Computed by Four Methods . . . . . . . . . . . . . . 42 10 Communication Courses Participating in Main Study . . . . . . . . . . . . . . . . 43 iv AA! ‘1 v 4/ ‘1‘ ‘v A. ‘s IList of Tables -- continued Tab 1e ll 12 13 l4 15 l6 17 118 319 2C) 21 22 Title Page Background Information . . . . . . . . . . . 44 Correlations Between Source Belief-Disbelief Mean Ratings of the Regular School Year Students and the Summer Students . . . . . . 47 Quartile Crossed-Classification of the Mean Ratings Given the 80 Sources by Regular School Year and Summer Students . . . . . . . . . . 47 Frequency Distribution of Dogmatism Scores . 60 Mean Scores for Three Dogmatism Groups . . . 62 Study Design . . . . . . . . . . . . . . . . 63 Analysis of variance Source of variation Tab le 0 O O O I O Q 0 O O O O O 0 O O O O O 64 Fourfold Frequency Table for use in Testing the Differences Between the Factor Solutions for the Belief and Disbelief Sources in the Degree to which each Exhibit a Given Characteristic . . . . . . . . . . . . . . . 71 Fourfold Frequency Table for use in Testing the Differences Between Two of the Three Dogmatism Groups on the Degree to which Individual Factor Solutions Exhibit a Given Characteristic . . . . . . . . . . . . . . . 71 DEPENDENT VARIABLE: Number of Factors Rotated in the Factor Solution Meeting the Kiel- Wrigley Criterion Equal to Three . . . . . . 76 DEPENDENT VARIABLE: Proportion of variance Accounted for by the First Principal Axis Factors (One Factor Solutions) . . . . . . . 77 DEPENDENT VARIABLE: Proportion of variance Accepted for by the Two Factor Solutions . . 78 '1... a ..Z , g. -.- S .4 .i~ v —— fit- [a \- u ‘1 ’I- A‘- I! - 1 ,1 ,V‘ ‘8 “W ’1 ~ A ,1 ‘a A- I .4 l u‘ .List of Tables--continued {Table Title Page 23 DEPENDENT VARIABLE: Proportion of variance Accounted for by the Three Factor Solution 79 24 DEPENDENT VARIABLE: Proportion of Variance Accounted for by the Four Factor Solutions 80 25 DEPENDENT VARIABLE: Proportion of variance Accounted for by Largest Factor in Two Factor Rotated Solutions. . . . . . . . . . 83 26 DEPENDENT VARIABLE: Proportion of variance Accounted for by the Largest Factor in the Three Factor Rotated Solutions . . . . . . 84 27 DEPENDENT VARIABLE: Proportion of Variance Accounted for by Largest Factor in the Four Factor Totated Solutions . . . . . . . . . 86 28 DEPENDENT VARIABLE: Simple Structure Index for the Two Factor Rotated Solutions (Proportion of the Total variance Accounted for by the Highest Loadings for Each variable). 87 29 DEPENDENT VARIABLE: Simple Structure Index for the Three Factor Rotated Solutions (PrOportion of the Total variance Accounted for by the Highest Loadings for Each variable . . . . . . . . . . . . . . . . . 88 30 DEPENDENT VARIABLE: Simple Structure Index for the Four Factor Rotated Solutions (Proportion of the Total Variance Accounted for by the Highest Loadings for Each variable) . . . . . . . . . . . . . . . . . 89 31 Number of Factors in the Rotated Solutions in the Individual Factor Analyses Which Met the Kiel-Wrigley Criterion When Set at Three: Frequency Distribution . . . . . . . . . . 92 32 Mean Proportion of variance Accounted for by Each Factor in the Three Factor Solutions of the Individual Factor Analyses and the Berlo Factor Analysis . . . . . . . . . . . . . . 95 vi ... rn - ; .4 .“—. .— AA ‘1 " List of Tables-~continued Table 33 34 35 36 37 38 Title __ Page Mean Index of Factorial Similarity Coefficient Between Each of the Factors in the Three Factor Rotated Solutions of the Individual Factor Analyses and the Three Factors (Safety, Qualification, and Dynamism) Isolated by Berlo 96 Mean PrOportion of variance Accounted for by Each of the Three Groups of Scales (Berlo's Three Factors of Safety, Qualification and Dynamism) on Each of the Factors in the Three Factor Rotated Solutions of the Individual Factor Analyses . . . . . . . . . lOO Comparisons of Berlo‘s Three Factor Solution with the Individual Three Factor Solutions Using the Index of Factorial Similarity: Summary of the Frequency and Percentage of the Number of Different Factors with Which the Three Berlo Factors had their Highest Indices of Factorial Similarity in the Individual Analyses . . . . . . . . . . . . lOl Comparisons of Berlo's Three Factor Solution with the Individual Three Factor Solutions Using the Index of Factorial Similarity: Summary of Frequency and Percentage of Individual Solutions for Which All Three of the Berlo Factors were Judged to have Adequate Matches . . . . . . . . . . . . . . lO5 Comparisons of Berlo's Three Factor Solution with the Individual Three Factor Solutions Using the Index of Factorial Similarity: Summary of the Frequency and Percentage of each Kind of Individual Factors in Terms of Size (Large, Medium, or Small) Which have the - Highest Index of Factorial Similarity with Berlo's Safety Factor . . . . . . . . . . . lO6 Comparisons of Berlo's Three Factor Solution with the Individual Three Factor Solutions Using the Index of Factorial Similarity: The Proportion of Solutions for Which the Safety Factor has the Highest Index of Factorial Similarity with the Largest Factor in the Individual Solutions for the Belief Sources versus that for the Disbelief Sources . . . . . . . . . . . . . 107 vii IList of Tables-~continued Table 39 40 1+1 422 “3 Lu: Title Page Comparisons of Berlo's Three Factor Solution with the Individual Three Factor Solutions Using the Index of Factorial Similarity: the Proportion of Solutions for Which the Safety Factor has the Highest Index of Factorial Similarity with the Largest Factor in the Individual Solutions for both Belief and Dis- belief Sources for the Low Dogmatism Groups versus that for the High Dogmatism Group . . llO Comparisons of Berlo's Three Factor Solution with the Individual Three Factor Solutions Using the Index of Factorial Similarity: Summary of the Frequency and Percentage of each Kind of Individual Factors in Terms of Size (Large, Medium, or Small) Which have the Highest Index of Factorial Similarity with Berlo' s . . . . . . . . . . . . . . . . . 111 Comparisons of Berlo's Three Factor Solution with the Individual Three Factor Solutions Using the Index of Factorial Similarity: Summary of the Frequency and Percentage of each Kind of Individual Factors in Terms of Size (Large, Medium, or Small) Which have the Highest Index of Factorial Similarity with Berlo's . . . . . . . . . . . . . . . . . . 112 Comparisons of Berlo's Three Factor Solution with the Individual Three Factor Solutions Using the Index of Factorial Similarity: The Proportion of Solutions for which the Dynamism Factor has the Highest Index of Factorial Similarity with the Smallest Factor in the Individual Solutions for the Belief Sources versus that for the Disbelief Sources. . . . 113 Classification of the Factors in the Three Factor Individual Solutions According to the Factor Types found by Berlo et a1. (Safety, Qualification, and Dynamism). . . . . . . 115 Classification of Pattern of the Factors in the Three Factor Individual Solutions According to the Factor Types found by Berlo et a1. (Safety—- S, Qualification-—Q, Dynamism-~D, —OEHer--OTH. ) 116 Viii V3 .05- nu. u... hfla . . ffi v ‘4 1' A: ~.. Idst of Tables-—continued 'Table 45 46 47 48 49 50 51 52 53 Title Page Number of Factors in the Three Factor Rotated Solutions of the Individual Analyses Which Could be Classified as One of the Three Factors Found by Berlo gt 31. . . . . . . . 117 Extent to Which the Factors in the Three Factor Rotated Solutions of the Individual Analyses Could be Classified with the Largest as the Safety Factor, the Middle as the Qualification Factor, and the Smallest as the Dynamism Factor . . . . . . . . . . . . . . 118 Extent to Which One of the Factors in the Three Factor Rotated Solutions of the Individual Analyses Could be Classified as the Safety Factor . . . . . . . . . . . . . . . . . 121 Extent to Which the Largest Factor in the Three Factor Rotated Solutions of the Individual Analyses Could be Classified as the Safety Factor . . . . . . . . . . . . . 122 Extent to Which One of the Factors in the Three Factor Rotated Solutions of the Individual Analyses Could be Classified as the Qualification Factor . . . . . . . . . . . . . . . . . . . 123 Extent to Which the Middle Sized Factor in the Three Factor Rotated Solutions of the Individual Analyses Could be Classified as the Qualification Factor . . . . . . . . . . . . . . . . . . . 124 Extent to Which One of the Factors in the Three Factor Rotated Solutions of the Individual Analyses Could be Classified as the Dynamism Factor . . . . . . . . . . . . . . . . 125 Extent to Which the Smallest Factor in the Three Factor Rotated Solutions of the Individual Analyses Could be Classified as the Dynamism Factor . . . . . . . . . . . . . . . . . . . 126 Extent to Which One of the Factors in the Three Factor Rotated Solutions of the Individual Analyses Could be Characterized as a Merging of the Qualification and Dynamism Factors . . . 129 ix Last of Tables--continued Tab 1e 54 55 Title ' Page Summary of Findings Related to Hypotheses I and II 0 O O D O O O I O O O O O O O O O O O 132 Summary of Findings Related to Hypotheses III and Iv . . . . . . . . . . . . . . . . . . . 136 .-:..A. Figure 1. LIST OF FIGURES Page Schematic of the data generated by the source evaluation questionnaire cast in the form of a three dimensional matrix or CUbeI I I I I I I I I I I I I I I I I I I I 52 J xi ‘-.n‘-u.v-- uu' ' p LIST OF APPENDICES APPENDIX Title Page A B Pretest Questionnaire Used . . . . . . . . . 154 Mean ratings of the person descriptions (sources) on the scale ranging from persons whose beliefs and ideas are very different from my own (9 position) to persons whose beliefs and ideas are very much like my own (1 position) for both the pretest (Spring 1967) and the Main Study (Summer 1967) . . . . . . 170 Factor scores for the principal axis factor and the two factor rotated solution for the person factor analysis of the belief-disbelief ratings of the 123 person descriptions (sources obtained for the pretest . . . . . . . . . 1 3 List of the sources selected to be evaluated in the main study using the Berlo et al. source credibility semantic differential scales. Descriptions 1—24 are defined as the sources associated with the belief and of the belief- disbelief continuum. Descriptions 25-48 are defined as the sources associated with the dis- belief end of the belief-disbelief continuum 191 Factor loading matrix for the three factor solution for the re-factor analysis of the Berlo et a1. Lansing area data deleting the scale "objective- subjective." . . . . . . . . . . . . . . . 195 Questionnaires administered during Summer, 1967 for the Main Study used to obtain the source sample validation data, the background information, and the dogmatism scale data . 197 The Source Evaluation Questionnaire used during the Main Study: includes the instructions with a sample page used to obtain the semantic differential evaluation of the sample of source descriptions. . . . . . . . . . . . . 210 Overall factor analyses of the source evaluation data . . . . . . . . . . . . . 217 Additional comments on the generality of the Berlo dimensions across subjects . . . . . . 224 x11 .'r'\ «f. “fl\ \ -d"~ -~ g. _’ ‘ \ .- 1 ~..«.v' . -- I. y. -— '\ ~~-L._' _ I . _ ‘ a... - x u... ~H4 I ‘ - A ~ I h.‘ hv4‘-"‘ ‘ I .A . ‘ 7'2 -. hul_‘-" E n. . . 7‘_\~'_‘ .. 4-~“ . ‘— ;~‘;Y- .v-vg“. ‘- ‘-._ _ ~ _y.~ d‘.".‘; ., .‘ l‘ ‘- ~f‘y—- ~ . CHAPTER I INTRODUCTION Considerable research evidence amassed over the past forty years has shown that a person's acceptance of information and ideas is partly based on "who said it." What is it about the source of a message that can make him more effective than another source communicating essentially the same message? This variable, "the source's role in communication effectiveness," has been studied under various names: ethos, prestige, status, per- sonality, charisma, image, character, and more recently, source credibility. Readers interested in reviews or summaries of relevant research can consult Berlo, 1967; Anderson and Clevenger, 1963; Klapper, 1960; Cohen, 1964; Hovland, Janis and Kelly, 1953. The primary concern here is with the Operationalization of source credibility. Anderson and Clevenger (1963) summarized some of the ways in which source credibility (ethos in their terminology) has been measured. They comment: In a few instances the development of a measure of ethos has been the main goal of a research project, but more often the measurement of prestige, credibility, or some other ethical component has been ancillary to the study of such presumed results of ethos as preferences, attitude change, and in— formation gain. The methods of measurement in both types of investigations are the same: (1) rankings, w— a. \ u ‘:°-"‘T"ro ".‘Od. v4. .~..., ;V’\ \v_“ ‘_y I..‘ . v- :A ‘..-.. .. \yJ‘.-v: .s .. ‘F ~‘ \ -‘ ”~41. ," ‘ In“. I‘ ‘. u...“ . “‘A‘I. "~«-'_ ‘-~4-\ -:. A q'- .- ; .. - r- -‘ .A— ‘5— '- \r- 5‘- 7 Ah;— ‘. .‘ 4 ‘n— g._ A" ~ - (2) sociograms, (3) "prestige indexes" obtained from attitude change data, (4) linear rating scales, (5) Thurstone-type attitude scales, and (6) devices similar to Likert scaling techniques, including the semantic differential. To Likert scaling techniques, including the semantic differential. Two more recent research efforts have focused on the Operationalization of the concept of source credibility. McCroskey (1966) conducted a series of studies aimed at Ineasuring "ethos" in oral or Speech communication situations using both standard Likert type scale items and semantic differential bi—polar scales. Factor analysis was used to isolate "dimensions of ethos." Preceding the McCroskey (effort was the work of Berlo and his associates (Berlo, Lemert, 61nd Mertz, 1966; Lemert, 1963).* Their work has two primary Sources or roots. As a theoretic explication of the concept <3f source credibility, they drew on the work of Hovland (Hovland, Janis, and Kelly, 1953). To Operationalize source (aredibility along these theoretic lines, they drew on the work (If Osgood (Osgood, Suci, and Tannenbaum, 1957) in the use of the semantic differential in measurement of meaning. *Throughout the remainder of the text as a means of simplifying reference to the work of Berlo and his associates, it will be Eeferred to by using only the name of the principal author, erlo. . -.-.‘.. r , rut '— unua.._.g._ »-.~- 0,. . .. r ~44. ,_ k . .,. c J . ilb "rn- d-." o “'"fil‘-: , - .4 ‘ .-\,_ reo- :.. .— ”“" -1 ";""r .,g‘ ..‘ . 1.: .- ‘ 'hwvh '- ~....._: . - -. r "'_,3 A .- 0- —'-v F: _ __r‘. “~V Berlo Source Evaluation Studies Berlo and his colleagues note that besides a brief treatment by Hovland et_a1,, little or no effort has been made to place the notions of credibility within some theoretic framework for either conceptual or operational development. Without such effort, particularly Operationalization, they (Berlo §t_a1,, 1966, p. 1) note it "precludes consideration of the variable as a dependent outcome of communication, and makes it difficult to compare and correlate the results of different studies." Berlo (Berlo, et_al., 1966, p. 2) summarized the ideas of Hovland et_a1, in the following way: Although they did not devote a great deal of attention to the question, Hovland, Janis and Kelly did attempt to explicate the concept, both theoretically and operationally. In their review of credibility research, they suggest a two-dimensional conception; perceived expertness and perceived trustworthiness. In some of their studies, they also utilized a single item rating scale for each dimension; however, these ratings were omitted in most experiments. Even when used, their function was simply as a check on the validity of the a .priori high or low credibility values assumed— as attributes of sources such as Oppenheimer, Fortune, Pravda, etc. In their theoretic discussion of preceived expertness and trustworthiness, Hovland et. a1. make a distinction between credibility afia'offier source— related variables such as affection, admiration, power, fear, and awe--but suggest the relevance to credibility variables like intelligence and sincerity. With respect to the function of trustworthiness and expertise, they conclude that credibility varies positively l s’.,.A.- . ~r nth- '-! n. uz- - A'. . s d~-o~-v a: a: with persuasion, although "from the results, it is not possible to disentangle the effects of the two main components of credibility-—trustworthiness and experties——but it appears that both are im- portant variables." Berlo notes a lack of empirical research designed to investigate how receivers evaluate information sources. He suggests that there is a similarity between the source evaluation problem and the general measurement of connotative meaning as developed by Osgood et_a1, (1957) employing the semantic differential. Typically, they assemble a set of adjective pairs judged to be opposites. Each pair of bi—polar adjectives bounds a seven—point rating scale. Respondents are asked to rate a number of concepts on each scale. The data thus obtained are submitted to factor analysis of the scale relationships with the resulting factors termed 11 "dimensions of connotative meaning for the concepts. Thus, the semantic differential is used to tap the "meaning' peOple have for concepts. Osgood §t_al, suggest that people evaluate concepts by applying underlying judgmental dimensions which enable them to differentiate among concepts. Each of these underlying judgmental dimensions is associated with a factor in the factor analysis of semantic differential bi-polar adjective scales. These judgmental dimensions define what they call semantic space. They suggest that one of the basic (Way8_in which people could differ cognitively would be in the . .- -fl....-. n.-. . w ..... number and Kind of Judsmentalnéiaseiionio(soliestively A semantic SPFZRILPhRX use to evaluate concepts. Osgood et_a1, Chavemisolated three major dimensions that they feel are highly general and are used by most people. These are the evaluation, potency, and activity dimensions. Berlo extrapolates the Osgood approach to the study of the "dimensions for evaluating the acceptability of message ‘— .... —-" bu- -— sources." The concepts become sources or potential sources of messages. The bi—polar adjective scales used are ones deemed relevant to evaluating information sources. The criteria for scale selection grew out of the Hovland work. Respondents are asked to use the scales to evaluate the sources in the usual semantic differential fashion. The results are then factor analyzed with each factor isolated representing a dimension of source credibility or source evaluation. Primarily three factors have emerged from their work: safety, qualification, and dynamism. One goal implicit in the work of both Osgood §t_a1. and Berlo is the development of measuring instruments which have a high degree of generality across a wide variety of Contexts, concepts or sources, and people. Both groups of researchers suggest that the problem of generality needs tO be further explored. This is implicit in Osgood's and Darnell's suggestions of concept-scale interaction (Osgood, r u A .. / . .—O --' I 4 guy w o a... - I I o .. .N' . ~ cu on»..- u . ‘- '- - . ‘ Y‘ r .6 5--L._ ‘ a >4“ :'.... 2.4- .'I.-‘--. ..‘ .V‘fio. " "~ \' .A. ‘P\ -‘\‘- . ‘v-a. .v. n. ‘r. a. ‘s 19623 Darnell, 1966). In typical use and construction of instruments based on either Osgood's or Berlo's work, there is implied a high degree of generality. In this project it is proposed to investigate some aspects of the problem of the generality of the dimensions of source credibility. Rokeach's Belief System Theory and Source Evaluation Rokeach (1960) makes a distinction between content and structure of belief systems. The belief systems of two M. _.-—-—--———-.¢-a--w __.._—o—-. people may have widely differing content, but may have highly similar kIndS Of organizations or structures. He lists at least three structural characteristics on which belief systems may vary. These include organization along a central-peripheral dimension, organization along a time- perspective dimension, and organization along a belief- disbelief dimension. The latter dimension has been singled out for focus in this study. Rokeach assumes "that all of a person's beliefs are organized into two interdependent parts; a belief system and a disbelief system." Further, the disbelief system is thought to be composed of a number of sub-systems which can be arranged along a continuum of similarity to the Person's belief system. This dimension is assumed to have Certain properties on which people may vary. Among these m w “r -v— .. on. 9".- v vv..‘ ‘1... _ _r_r ““'h vs. a . ,. ,. ‘xn-I 'Ovbu iv I n o. the. a u“ v.1. "',.. - . , . F; . “” ~v-- .2»: " v. -P \ 'V». :« \ u. v are the degree of differentiation of parts and the degree of isolation or lack of communication within and among parts. The general hypothesis suggested by Rokeach regarding differentiation of parts is: disbelief sub-systems relatively close to the belief system will be relatively highly differen- tiated, while those farther away will be poorly differentiated. The belief system is the most highly differentiated of all. Differentiation of parts is one of the defining prOperties or characteristics of Open and closed systems. A system is closed to the extent that the degree of differen- tiation of the belief system is greater than the degree of differentiation of the disbelief system. Open systems have greater differentiation within the disbelief .system relative to the belief system than do closed. Open and closed systems vary on the degree of isolation within and between parts of the belief and sisbelief systems. Open systems have communi- cation within and among parts; closed systems have isolation within and among parts. It is possible to relate these aSpects of Rokeach's work to the problem of generality of semantic space or credibility dimensions. Degree of differentiation within the belief system or disbelief sub-systems can be interpreted in.the case of semantic Space and/Or credibility dimensions as referring to the number or kinds of judgmental dimensions ~.. v--- a;au H..- p L-‘»-..., o v..- s-.... ‘ . C ‘1~ . ~~ .. ...,..“ . --: ~-',‘-‘. ~ 0‘. :- _‘P{ fl. ‘ .— a. ‘ 4 ‘ -. . a». ,_ > .g :- ..‘-" -A v-_,.‘_ : L..~-d . " P‘i‘r- . ‘- \-"r., ‘ V F "~:~.A ..l P. used which enable one to differentiate between concepts or sources. When Rokeach talks about differences in degree of differentiation within the belief system when contrasted to the disbelief sub-systems, this could be interpreted as placing some limitations on the generality of these dimensions across concepts or sources. Further, Rokeach talks about differences in amount of differentiation as one proceeds toward the disbelief end of the belief-disbelief continuum as a defining characteristic of open and closed systems. This can be interpreted as suggesting differences among peOple in the way they might use bi-polar scales to describe concepts or sources. This suggests limitations on the generality of these dimensions across people. Critique of the Wozniak Study Wozniak (1963) attempted to relate differences in semantic space to Rokeach's theory of open-closed Systems. Wozniak predicted that persons with open systems would employ more judgmental dimensions than would persons with closed systems. He had 241 college students evaluate 19 concepts With 40 semantic differential scales selected from Osgood's work (the Thesaurus study). The 19 concepts could con- ceivably be elements of both the belief and disbelief Systems for most people. They included "my mother," P‘ \- .u. v - -: -‘. ' v--. :4 I '1“‘su-\. ~A fi‘ \—... .' “Ab- ‘. ’P.‘ A a - V‘ .‘q_ ‘ - “~. : \ I -A s-.. ‘,-- .‘ -- .--- -5“ 4‘-." *- . "r-. ,- ‘. ‘.‘ u .’ .fi_ ‘ ~_. ~‘a~" ‘ v , ‘7';- 'v —. I T *‘ ”F; 0 fl fin- -.\a_“ .‘ .- x m. sg. \_~ ‘I ~- '\~ “"a‘ . . ’K - 1‘ "V. r, . 4. - "4 ”me," "John F. Kennedy," "Nikita Khruschev," "engine," ”snow," "boulder," "modern art," "sin," "leadership," "socialism,” "symphony," "birth," "death," "war with Russia," ”hospital," "prison," "United States," and "Communist China." The dogmatism scale was administered also. Three groups were defined: (1) Open-—the 79 lowest scores, (2) Closed——the 81 highest scorers, and (3) Medium--the remaining 81 middle scorers. He did three Osgood type factor analyses of the 40 semantic differential scales, one for each of the three dogmatism groups. Based on the Kiel-Wrigley (1960) criterion (see Chapter II for explanation of criterion), the number of factors rotated for each of the three dogmatism groups varied, six for closed, five for open, and four for medium. Open systems in the Wozniak study could not be differentiated from closed systems in the predicted way. If anything, the suggestion is in the reverse direction: closed systems have greater complexity than open systems due to the larger number of factors extracted. When compared by the index of factorial similarity, the factor structure matrices of both the open and closed groups were quite Similar to each other but both were different from the medium group. The medium group's factor structure resembled Closely Osgood's typical results with evaluative, activity, and potency factors and a fourth factor labeled tautness. . o,- u .,‘ vonv .,.. . .— ... . ‘..'l‘.'. ~ c x-pprw! - u ' ..‘-..-~ . .. . ... . ." *“ 7" r .-.~.‘ N . _ . ‘ ‘1'.“r . ‘17-“-..4 . ... _ Q ”Av.— ,. . v- 0“. -c-‘. ~~J~ - ~ .1. . ‘~'- A_~”,‘1n v‘.— in K . 5,. dAfi 10 For the open and closed groups, the evaluative dimension showed up but the activity and potency did not. The three other factors which did emerge for these groups were labeled dynamism, predictability, and sensory—ennui. There are two major difficulties with the Wozniak study. First, by doing factor analyses of scales for groups and summing across both subjects and concepts, he missed de— termining what sources of variance contributed to the factor structure. Summing across both subjects and concepts implies using all concept ratings on a particular scale for each subject as individual observations in computing the corre- lations. Then the number of observations would be the number of subjects times the number of concepts for purposes of correlation computation. These same procedures were employed in the work of both Osgood and Berlo. The factor structure under such conditions could hwe been derived from variance attributable to subjects, concepts, or an inter— action of both. Rokeach's belief system theory deals with how a person internally organizes or structures the beliefs, values, attitudes, ideas, etc., attached to various aspects Of his environment. The important aSpect from a theoretic POint of view is how the individual structures his own System internally and not some sort of composite structure averaged across both concepts and subjects which is tied to deviations from measures of group central tendency. -....... Q. . x .. \ H.,..." u, , ‘1 a. «7“. I .— 4. ,‘,__ ~'A.".r"r\ ,- “““o-Ax-n. ‘I‘ .-§.— I " \I— -1“ '1' ~'.d. .6. ‘Ar.- h" 'V...u . h -Y‘ri “4..-“; H . -' r . ‘—v“fi:~ I 5'~‘-“ . ..‘ m v I ' y‘: ‘ .Ivi'.‘ — - .,‘. ‘ -_ y: '1 W .4... ~4~ '- 5“ “\ y;: a... A.v~ _. . o‘ n‘fl‘fir- " . Q-“..- 5“. ‘ ~" ‘ ‘1 'v ‘ u -. \.__ 5 . _‘ ‘ ‘ ~~ v“ ‘ '§"‘ . ~ _ ’ Vvh ~V . » "« ’- A~- . ., v __ s v “'... \_ _ _ “P “‘u , ‘. Ix ~ '. ' _ ~'~‘ so ‘- . ._ ‘_ 11 In this study, factor analysis will again be used, but along different lines. If people differ in belief-disbelief system structure or cognitive structure, these differences should be reflected in the ways they use semantic differential scales to order concepts or sources. Factor analysis of a single individual's semantic differential data matrix should reflect the several different ways he uses the scales to evaluate the concepts (or sources). Each predominant, different ordering of the concepts (or sources) would be reflected in a separate factor. Comparisons among number and kind of factors thus obtained from each person could then be made. In this kind of analysis, the structure derived for each subject would be tied to deviations from his own measures of central tendency and not from the group's. The factor structure for a person would be due only to the variance among concepts (or sources) for that individual. The second difficulty with the Wozniak study is re- lated to his interpretation of Rokeach's theory. The differing degrees of differentiation of parts in open and closed systems led him to predict a greater number of factors to explain the variance for the open group than for the closed group. In other words, when he puts all concepts in the factor analysis regardless of whether they were from either the belief or disbelief system, he assumes that the theory predicts open and closed systems differ per EB .F‘ 1.... V‘- . - - - O V'- --.. Lan- - - .._ 'I - ‘ a‘ - . ' U -- . .-_ -‘ .1 A“‘ \- _ _.. s P. -\‘ . 'i. -‘. I \1 ’ , - - 12 in the degree of differentiation of parts. However, it seems reasonable to make a somewhat different interpretation of what the theory predicts, con- sistent with that previously outlined (Rokeach, 1960). In all belief-disbelief systems, there is the tendency for the sub-systems of the disbelief system to decrease in differentiation the closer each is located to the disbelief end. This tendency will be termed here as "de—differentiation" of the belief-disbelief continuum. One way that open and closed systems differ is in the degree of de-differentiation relative to the belief system which occurs within the dis- belief system. Diagrammatically, the present interpretation is outlined in Table 1. The differences generated by the theory are of an interactional nature. The method of examining the differences in differentiation used by Wozniak, in effect, examines it only as a simple effect of openness or closedness of the entire belief-disbelief system. It should be examined as an interaction effect. Although there is a suggestion of a main effect, it would be small and probably not statistically significant. A more appropriate way to test this proposition would be to separate the concepts representing elements of the belief system from the concepts representing elements of the disbelief system. Then two matrices for each individual would be separately factor w. ‘ _ "~ ‘5“-- \. ““ IL: .._v _, r- ‘ u. N- u... 1 “us o‘_ Q - A .v. V" a, , A .- .JVQU 'v v ;-'n» ‘ - v‘v Fri ~*s'_ \~_'— ‘V- In N~ v._ . "_ v. 5.. -.o- A 'v u ‘H (x) C ) («“2 13 Table 1. Differential Differentiation Within Belief and Disbelief Systems for Open and Closed Systems Belief-Disbelief Dimension Belief System Disbelief System Open Relatively High Medium Systems Differentiation Differentiation Closed Relatively High Relatively Low Systems Differentiation Differentiation analyzed. The prediction would be that for open subjects, the difference between the number of factors required to explain the belief matrix and thanumber required to explain the disbelief matrix would be less than the similar difference for closed subjects (using the Kiel-Wrigley criterion to stop rotating factors). Brief Review of Related Research The traditional uses of factor analysis in psychology were designed to allow the investigator to isolate fundamental mental abilities, psychological dimensions, and personality dimensions. The technique emerged from a search for an all- encompassing single common factor, Spearman's G, to multiple factor explanations of psychological and personality dimensions (Guilford, 19543 Thurstone, 19573 Harmon, 1960; 14 Cattell, 1952; Horst, 1955; Thomson, 1951). The factor analyses were of tests. It was assumed that the factors derived from clusters of tests were representative of psychological dimensions which cut across people on which each person could be located. Osgood's use of the factor analytic model grew out of this tradition. Stephenson (1953) and McQuitty (1964, 1966) took a different approach in characterizing the psychological make- up of people in the way they used factor analytic and clustering models. They argue that it is not a single test score that is the most important in explaining psychological processes of individuals but that what is most important is the profile or pattern of scores that a person has on a number of "tests" or items. Their approach is to cluster people on the basis of the similarity of profile or pattern of reSponse to a number of items. Stephenson continued to use the factor analytic model but used people as variables instead of tests as in the more traditional uses. McQuitty developed a whole new set of clustering devices collectively called pattern analytic methods. He prefers non-linear indices of association between persons to the linear corre- lational measures. If items are developed for use with these methods which represent elements of the belief-disbelief system and persons are asked for their degree of endorsement ' I “‘rr‘ A" Uv.-:~--. I . :‘1‘"‘."7‘.: v.. v.’J-l. w I n. v;-,_‘ A A "" ..,_‘ at ca (I) 15 of them, peOple can be grouped by the similarity of their hierarchies of acceptance of these elements. These techniques can help to investigate certain kinds of structural and/or content characteristics of belief-disbelief systems related to source credibility (MacLean, Danbury, Talbott, 1963). Kelly (1955) comes very close to using a clustering method as is prOposed here. His repertory test generates a data matrix for each subject. He developed a non-metric factor analytic technique for explaining the matrix. The factors thus developed are considered by him to be repre— sentative of the psychological space of the individual. In the use of factor analysis as proposed here, one could retain Osgood's notion of semantic Space (similar to Kelly's psychological Space), but it would be the Space of only a single individual and not some kind of group average. Ware (1958), a colleague of Osgood's, did do factor analyses on an individual basis of a typical set of semantic differential scales (20) and concepts (31, e.g., sex, mother, baby, snow, etc.). He had hoped to find sex and IQ differences in several measures of diversity of factor structure he had deveIOped. He found none. Osgood (1963) comments: "But one should never give up a good hypothesis without a fight, and I still think there ought to be a re- .u- a .1)- .. ‘r‘ us.” v. A .4 v u .- v\ ~ ~. A. .. ., - s ~.- ~ u,. . \ u.-. . 1|' (1 16 lation between intelligence and completeness of utiliza- tion of the semantic Space." One would not expect sex or IQ differences for all concepts. Ware used high school students and it is just possible that the sample of concepts he selected were ones which were generally ambiguous for this age group regardless of either sex or IQ, Todd and Rappoport (1964), working in the area of person perception, suggest that factor analysis of data matrices derived from single persons may have utility in the examining the way people perceive other people. Hays (1958) developed a procedure similar to factor analysis in which groups perceived characteristics of persons which are mutually implicative. His procedure is like an oblique factor analysis for ordinal data. Todd and Rappoport con- trasted the use of factor analysis with the Hays procedure as means of characterizing the way people perceive other persons on a very small number of persons in a highly ex— ploratory fashion. They were sufficiently encouraged with this approach that they expected to use it in more theo- retically based studies. McQuitty (1953) has also used factor analysis in a way similar to that proposed here in a study of personality integration. In it, McQuitty factor analyzed data matrices obtained from a single subject as a means of assessing aspects of personality integration. ~w 1? Re-Analysis of Some of Wozniak Study Data Some of Wozniak's data were re-analyzed along the lines suggested here (Talbott, 1964). To apply the notions directly, two groups of concepts would be needed. One group would be associated with the belief end of the belief-dis- belief continuum and the other with the disbelief end. Wozniak used only 19 concepts. This is too few concepts to separate them into two such groupings. Also, Wozniak had no measure to sort out which end of the continuum to associate with each concept. Hence, individual factor analysis can not be used to investigate the differentiation differences be- tween belief and disbelief sub-systems which Rokeach suggests are among the defining differences between open and closed individuals. However, another hypothesis was suggested by Rokeach's belief system theory. The largest factor (in terms of variance accounted for) isolated by Osgood was typically the evaluative factor. It usually contained scales such as good-bad, kind- cruel, clean-dirty, and fair—unfair. Rokeach's theory states that the magnitude of rejection of the disbelief sub-systems along the disbelief continuum is relatively greater for closed system individuals than for open system individuals. The scales typically associated with the evaluative factor imply the notion of rejection. As a consequence, closed 18 individuals are apt to have more scales which have a significant evaluative component than are open individuals. This would suggest that the evaluative factor for closed systems would be larger than for open systems. Since the evaluative factor most typically is found to be the largest, the largest factor in the two, three, and four factor rotated solutions should be larger for closed systems than for open systems. The same prediction could be made for the one factor solution or the first principal axis factor. The scheme in the re—analysis of the Wozniak data involved the separate factor analysis of the data matrices of 60 individuals from the Wozniak study: 1) the 20 indivi— duals with the most Open systems (the 20 lowest dogmatism scorers), 2) the 20 individuals with the mat closed systems (the 20 highest dogmatism scorers), and 3) the 20 individuals nearest the median dogmatism score among all 241 subjects he used. Table 2 contains the dogmatism score average and range for each of the three groups. Scales were used for variables (40) and the concepts (19) for observa- tions in computing the correlation matrix. Principal axis factors were extracted using unities as communalities followed by varimax rotation. 19 It was predicted that the average amount of variance accounted for by the first principal axis factors in the individual factor analyses would be greater for closed system individuals than for open system individuals. In addition, the same basic prediction was made for the largest factor (most variance accounted for) in the two, three, and four factor varimax rotations. The results are contained in Table 3. The average amount of variance accounted for by the first principal axis factors in the individual factor analyses was reliably larger for closed system than for open system individuals. Also, support was found for the hypotheses regarding the largest factor in the two and three rotated factor solutions but not for the fourth. However, although not reliable at the .05 level, the differences diSplayed in the four factor solution were in the direction of the pre- diction. In selecting the scales to be used, Wozniak selected those which were representative of the factors which Osgood §£_§1, had isolated in their factor analytic studies. Five scales were chosen to represent the evaluative factor. The number of times these scales had their highest loadings on the largest (accounted for the most variance) factors for 20 Table 2. Mean and Range of Dogmatism Scores Associated With Three Dogmatism Groups. Mean Group Dogmatism Score Range Open 115.75 91—128 Medium 156.05 155-157 Closed 191.95 183-202 Note--Wozniak used the 40—item Form E scale. Table 3. Simple Randomized Analysis of Variance Results for the PrOportion of variance Accounted for by the Largest Factor in the One (First Principal Axis Factor), Two, Three and Four Factor Rotated Solutions Mean Proportion of variance Accounted For: Open Medium Closed P Proportion of Variance Mean s.d. Mean s.d. Mean s.d. value First Principal Axis ‘ Factor .256 .058 .302 .063 .304 .056 .05 Largest Factor: Two Factor Varimax Rotation .237 .056 .275 .065 276 .056 .05 Largest Factor: Three Factor varimax Rotation .215 .047 .249 .055 .249 .044 .05 Largest Factor: Four Factor Varimax .213 .055 .238 .067 .250 .062 .10 are}; A.” .a | VI.“ v. . l‘ ’n; Orr; dun In. ‘u s . k a)”- L4 or 'rro ‘.‘ A P P‘- '““- an". a. V." A v ”3. Univ .J" .‘ “'v ‘ AA Unb Q Mr N . ~T‘h,.',~ IF“ "f'T' '- b'» - 4 ,4 M .. q \’ \ ‘9‘ YfAA A“v~ .“ v ‘ l‘ “I ‘I ~ 4‘35!- “¥: Cr». &. bv' 21 each of the twenty people in the three dogmatism groups on the three factor rotated solutions was counted. The results are tabulated in Table 4. This informal analysis suggests that the largest factors isolated in the three factor solution of the individual factor analyses for the medium and closed system individuals were more likely to be similar to the evaluative factor isolated by Osgood than they were for the open system individuals. This result supports the explanation given as to why the first principal axis factor and the largest factor in the rotated solutions should be larger for closed systems than for open systems. Present Study The present study investigates some of the possible differences in the dimensions of source evaluation across persons and for different kinds of sources. Subjects are asked to evaluate two sets of sources on the source evalua— tion biapolar adjective scales developed by Berlo. One set of sources is associated with the belief end of the belief- disbelief continuum and the other set with the disbelief end. The two sets of source evaluation data obtained from each subject are factor analyzed individually. The dimensions of source evaluation thus obtained are examined both for differences between the two kinds of sources and for their n 1‘ ‘4 I-§ _q. . A 5*..av‘y- w rcn‘ .‘.ha¢'- N '\ 'wr‘ vb" ., (i Jvo..\- L I r" f‘ !- D . P“ ‘I‘ i . e.gnoa a \- ~..~-~ \ \ i‘k s , .p-ad- 1| ft. urn .m a. i: a. n . .. . . . vi ~:L n. . .flu .7“ -. o 2.. «C .n . Me. 4 i . c .1 I. ~tl‘ nflu and .14 .C : . ”C A}. . 7.. .h)» A; . v 2. . A a. O .. ‘ AL M. e A: «C h . 2‘ .r.. urn Cu 2. ... u «C . . a: .14 a x » II. 22 relationship to dogmatism. Rokeach, in delineating the defining characteristics between open and closed systems, focuses on the differences in the way the belief systems of the two types are structured or organized. Closed systems (high dogmatics) exhibit a greater tendency than do Open systems (low dogmatics) for the sub-systems of the disbelief system to decrease in differentiation the closer each is to the disbelief end of the beliefedisbelief continuum. If in- formation sources can be viewed as elements of the belief- disbelief systems, then the evaluations of these sources Should be organized in a way that is similar to that part of the belief-disbelief systems with which the sources are associated. The two main hypotheses for this study were derived directly from Rokeach's belief system theory. Stated in general form, the hypotheses are: I. The number and/or differentiation of the dimensions of source evaluation will be greater for sources associated with the belief end of the belief-disbelief con- tinuum than for sources associated with the disbelief end. II. The difference between the number and/or differentiation of the dimensions of source evaluation for sources associated with the belief end of the belief-disbelief con- tinuum and for sources associated with the disbelief end will be greater for closed systems than for open systems. a. . .r . .w a rd r“ L v .3; s V x. .i . r. r. .., l .2 a .7. AA“ L.- X. .1 .1 a r... L V EH“ 0 23 {Fable 4. The Number of Times each of the Evaluative Scales Had its Highest Loading on the Largest Factor for the Three Dogmatism Groups Closed Medium 9222. Good-bad l7 17 11 Heavenly-hellish 18 16 9 Kind-cruel l9 l9 l4 Clean-dirty 17 16 14 Fair-unfair l9 l7 15 Average number of times the five scales had highest loadings 18.0 17.0 12.6 on the largest factor . ,, ”5*.” '3': an .F Y vii—JAVQVa- .- e ‘ ~.:-T' "a ars "'NH ado—L .' ‘u A“ l "3";335.‘ arc. ,- bv-onvuy~ . V k a T’T r fi--0 . I u l “r-v I . n.‘ . Fv . " \ IU .Q ‘n. ‘L— ‘9 "A 1: ' . 5-3 ‘ ‘- VQ 1 . v-gv "‘“‘*‘U . I n. f5 .4 H- . a I I y. 1‘ nu u‘“... \ P 24 Comparisons between the individual dimensions of source evaluation and the dimensions of source evaluation isolated by Berlo are made. The following two informal general hy- ‘potheses are offered to facilitate these comparisons: III. The nature or kind of dimensions of source evaluation used for the sources associated with the belief end of the belief-disbelief continuum will differ from those used for sources associated with the disbelief end. IV. Closed systems will differ from open systems in the nature or kind of dimensions of source evaluation used. It is hoped that this study will indicate what some <>f the limits may be on the generality of the construct of 'the dimensions of credibility or dimensions of source evaluation as explicated by Berlo and his colleagues. mm» .o.‘ . ~ .Hv VK. Av]. PHJ . fl. ‘ n91. QM CV | . u ..\ q. . . . .C C Hr .C n; E .C ( F; (\ C E .. :1. Y .3 c \1 ) , 1-. \.l QN. A ll} t. d 3 u. x) ”4 - Z J rap A.» u: u... CHAPTER II STUDY DESIGN AND METHODOLOGY Introduction Briefly summarized, the major tasks involved in the execution of this study were as follows: 1) Selection of a set of 48 sources which can be split into two groups on the basis of with which end of the belief-disbelief continuum each is associated. Evaluation by a group of subjects of the 48 sources using the source credibility semantic differential scales (34) used by Berlo et al. (1966) in his public sample. __'_—' Formation of two data matrices for each subject (24 sources-- rows, 34 scales-—columns) with one containing the sources associated with the belief end and the other containing sources associated with the disbelief end of the continuum; and then, factor analysis of each separately. Construction of the indices which purport to measure the dependent variables. Classification of the subjects in terms of dogmatism. Statistical analysis of the relationship among the dependent variables and the two independent variables. The data for this study were gathered in two phases. The first phase or pretest study consisted of selecting the 48 sources to be used in construction of themain instrument. 25 , on» F"’\" “ \— u v yqu‘ -~ (\Y‘ .-\.u a "7‘“ fl .5 __......A..-‘v < ~va . . “HA... ' y ~ 5 ‘ i‘v vuu — . uvL.- .— 0 I“ Erlfcy‘r‘a P. A ‘4 26 This was done Spring term, 1967, using college students from two sections of an introductory course in the Department of Communication at Michigan State University. The data gathering for the second phase or main study consisted of two parts con- ducted during Summer term, 1967, using student volunteers from five communication courses at MSU. The first part of the main study represented a validation of the sample of sources selected in the pretest using different students, administering the dogmatism scale and obtaining some back- ground information on each student. The data gathering for the second part consisted of obtaining the semantic differential evaluations of the selected sources. Pretest: Selection of Sources Brief descriptions of 123 possible sources were assembled. Care was taken to include a wide variety of persons. Some of the categories of sources deliberately represented included family, personal friends and acquaintances, political figures, religious leaders, heads of government, newsmen, social protesters, military figures, law enforcement officials, communists, and others. Sources were selected to represent various degrees of remoteness from the usual college student's range of experience. Also, . LC .hu_ o - . u o. . . ‘4 v“ L. r: ~ \ .. . 7““ .U ‘4 H IV fill ~¥ aHd 1* ‘ru WA.- sde ‘h‘ «V11. C r\ ( 1‘ .a D? r so &v z: .. . .2” ,6 it .. . at. +t a . 3., it to 3 . . . . . . . . r. .nm a” N»; 1 a S S a: no 4 .a. I. a. r a S C r. 3 .1.» .3“ at A u/ and 7| f0 :1. 1%. 3‘4 2 1; 4; “fl Se “u. 83 h. a: . .u‘ .«u 3» at L . A. . A v . W: Z . .. «I. : .. w a . . .Mv. . . .L n. L m L. V . . n 2. .r “ Mr“. has. r ... m u "rpm we. Q «\~ Riv 3. Ali .0‘ .21... 3. ~ 27 care was taken to represent a variety of cultures. Some of the descriptions included the name of the person described, while others did not. For some, a specific name was not appropriate, e,g,, my best friend, my mother, etc. Fifty-three (53) students from two sections of the Spring, 1967, class, Communication 100, Department of Communication, Michigan State University, volunteered to evaluate the 123 descriptions of persons on the following 9-point scale: 9 - has beliefs and ideas extremely different from my own 8 — very much different 7 - has beliefs and ideas quite a bit different from my own 6 - only slightly different 5 - can't decide or has about the same amount of beliefs and ideas which are like my own as are different from my own 4 - only slightly like 3 - has beliefs and ideas quite a bit like my own 2 - very much like 1 - has beliefs and ideas extremely like my own. For a list of the items (descriptions of persons), and the 'way reSpondents were asked to evaluate them, see the questionnaire in Appendix A. 28 The means and standard deviations for each of the 123 items averaged across the 53 subjects will be found in Appendix B, ordered from those having beliefs and ideas judged "most different from my own" to "those most like my own." There was considerable variation in the means. There would be no particular difficulty encountered if one wished to order the person descriptions by the mean rating on the scale. Individual scale marking tendencies were examined also. Twenty-seven of the fifty-three subjects used all nine marking positions of the scale. Twenty-one more used eight marking positions. About ninety percent of the subjects used either all or all but one of the nine positions on the scale. Only four of the fifty-three subjects used seven of the nine positions and only one subject used as few as six positions. No one used fewer than six positions. Table 5 shows which marking positions on the scale were not used when not all nine positions were utilized. From examination of the table, it is apparent that scale position 1 or "has beliefs and ideas extremely like my own" was by far the least often used. Thtsanalysis shows that almost all of the subjects used all or nearly all of the possible range of the rating scale. .. . T. a»¢u ,. A r 29 Table 5. Source Belief-Disbelief Rating Scale Positions Not Used When Not All Nine Were Marked Scale Positions No. of cases Percent of Not Used Involved Total Cases One scale position 1 19 73.08 not used 2 1 3.85 6 1 3.85 Two scale positions 1 and 3 l 3.85 not used 1 and 4 l 3 85 l and 9 2 7 69 Three scale positions 1, 6 and 7 l 3.85 not used Total cases 26 Ly .f u - .I . v y x q r e . . b E . \ S .3 ,. .4 E T. .- .. _ J S -. t . .. i E a; . s E -T. C a... O a S E .u x . a. f :3 r... a v -. a; .. u v. 3: -t : _ .. .. L rt C S E S E S C: ., a r.“ A v ...m w . 7. .h .nu : . ~ it Pl; ., .ng : g r... .0. .n. 3: a . L. :1; » .~ . . x . A: a . .2 TC . v .i . 0 3 .1. . v -r t . _ w.“ T. A a. . C. .l .3 E T L... T F L . . ml. ,MW 2. S; W». . A 2. PM A}. v . a. Cu O .. . r C r1. .. u .5 P. a I u ab :4 At r.“ . . -. . to :o r . No .L a... my. .wu cu n3 Lu my .fi. to :1. 3.. a y A: a... A .. . an.“ . NV Uh. Cy & u a y a... 3L Hr. nu. .fi; 9:. n. L a.» F . A...» In. L V MI} . . ~\~ Q» A!‘ . . 4 4 . n ~ My. «7. C. .: .14 Cu I: mm . Mm «To 2. r..~. «r. »U fin navy mum. .5.» 1A.. AH” .M\H “l". ”.lw uurdlv Ht?» Wt w! ” IFIW. had "V. p N NA— .NA u. a v My. An. «I u. . ‘ ~ NII 30 To get further information on the degree to which subjects agreed in the way they assigned scale ranks to the persons, a 53x53 matrix of correlations was formed by correlating each subject's set of ratings on the 123 descriptions with every other subject's. This matrix was then submitted to a principal axis factor analysis with unities in the diagonals, followed by a Varimax rotation. The analysis yields essentially a one-factor solution. Approximately forty-six per cent of the variance was accounted for by the first principal axis factor. Loadings on the first principal axis factor are summarized in Table 6. This suggests that there was considerable agreement among most of the subjects on the ratings assigned the 123 person descriptions. The varimax rotations were examined to determine how unambiguously the 53 subjects could be classified into groups on the basis of rotated factor loadings. The two factor rotated solution yielded the best results in terms of this criterion (see Table 7). In the two factor solution, there is one substantial factor (accounting for 41% of the total variance and 76% of the common variance) which rather clearly clusters forty-one of the fifty—three subjects. The loading for three subjects are Split between both factors (both .loadings in the .40's or .50's). Only nine of the fifty-three 31 Table 6. Frequency Distribution of Loadings on the First Principal Axis Factor Magnitude of No. of Per cent of Cumulative the loading; Subjects all Subjects Percentage .80's 11 20.8 20.8 .70's 16 30.2 51.0 .60's 9 17.0 68.0 .50's 7 13.2 81.2 .40's 2 3 8 85.0 -30'8 5 9.4 94.4 .20's 2 3.8 98.2 .lO's 1 1.9 100.1 MI... -N flaw uh”..an "I “VEHWRV‘ m. 32.. L C O~.~p wvuhm in- flavflwv~ «an; .—-Q «DHXJmu at“ \MV pUHi‘ “u N.— ~ ~n...q «omwdfiafiuvm , 7 L400 hie»); u Anzava‘Qawx/L Humangflv M. PAH U 3&1; I >5 L T; XQEfiL aw> . zlwafieb :5 m .L 7 ; - T. _.r..>.,~.\~..:< .880 L3»; ONGQ Aw “V SIN. a. .4 \ UH~S~ fun.» m. Hx< HEQH 02 DH. w .NCQLCQ; 51.3.? 323. MC; 5... 3 1 CL 23?. \ t 3 fl. L 3.. 3 32 mama. omms. mmas. mmmm. mozm. mmom. pa aims. omsw. 03mm. Hmom. mmmm. wmbs. H r—IHt—ir—It—‘Ir—ir—{r‘il—ll—Ir—i Q: U) oocmflmm< HOQomm gonoom me< Haaaoeasm pagan mmms.o comm.o mozm.o wamm.o mm:s.o osm:.o mamm.o mom:.o mwom.o mmmm.o wms:.o :mam.o hpfiamcssaoo MHN0.0 mmma.o mmmo.o mNHN.O mmmo.o NmSm.o mNO0.0 bdwm.o mmma.o mmsa.0n mmmH.O mmdd.o N howomm mm:m.o :mos.o mmms.o mmom.o mmmm.o sw::.o mwms.o swam.o ommm.o swam.o smom.o 0355.0 H sopomm CUM-:I'LflKOb-CDON H gonadz pommnsm moaomm mflx< Hmmfiocfimm pmham msp cam soapsaom cmpwpom xweflum> honowm 039 039 How mmqflcmoq honomm ”Amcoapm>ammno mma «moanwamm> mmv mmogsom was no mwafipam .mpommnsm mmg:p-»pmam map mo mamsfimc< nopomm .N mHQmB m... a. : (NP .kKL ~VNL ON by x .. PV 0 O ) Ins 0 Hum .fi .. «umuk . O \OCQ Au IPVH s I‘J} vN.U\--- ll‘ldllil" u ... . ~.u..¢ figs 5 «.20..» 2.3:; ~.. .\ .~ :, V‘s .vfl.‘.o n u I. l \ o .‘I\.\. ~ A .y .. c ~ ~ - ~l. |.5.‘-‘-. . . . I ~ \H. t. . O 3:03 0 3. _ ~.::w~fi..nv ONO: . axmwNLKL 0 1.1.3.93. n.» —.NH .N 5de m4d~3fl .5-—v~«H . 4. d3 Alb a“ 33 moam. mmmw. mass. ammo. :mwz. :me. mmas. came. pomp. moaw. mamm. momm. omom. jaws. Hams. hopomm me< Hmmaosanm umsam r-lr-lr—ir-ir-ir-lu-ir—ir—lr—ir-it—lr—IN N omcmfimm< honomm msam.o :mam.o mmmm.o mm::.o om:m.o ommm.o smom.o Hmm:.o mom:.o omo».o mmam.o msas.o sawm.o Hmam.o :zmm.o mpHHmssaaoo amm:.o smma.o smmm.o ommm.o smmm.o mmom.o :mmm.o mzmfi.o mea.o s:mm.o Hmao.o omma.o mmmo.o msmH.o mmms.o N sonomm smao.o m:mm.o mmHs.o mmmm.o mmo:.o m::m.o masm.o mmmm.o mamm.o moms.o mmmm.o momw.o Homw.o msms.o somm.o H gonowm SN mm mm am mm mm Hm om mH ma NH ma mH :H MH gonasz somnpsm cossfipcoonnb mHowH 0:38. H mem.o smam.o- ssws.o m: mssm. m s:wm.o mwma.o o:mm.o m: mmmm. H mmsm.o mmmm.o mms:.o H: mmms. H Hmmm.o mmmo.o mamm.o o: mHms. H momm.o mmsH.o mmms.o mm mmmm. m somm.o mHHs.o HHmH.o mm mama. H :smm.o ommH.o oaHs.o am my moms. H Hmm©.o mamm.o moms.o mm moHs. H ssom.o :HmH.o momm.o mm mamm. pHHam kmmz.o Ham:.o momm.o Hm swam. H mmmm.o mmHm.o smmm.o mm msmo. H st:.o osom.o mmHm.o mm ammm. H mmmm.o mmom.o mmss.o Hm Hmmm. m momm.o Hmzs.o HHmH.o om mmHH. m smom.o Hmos.o mmOH.o- mm mamm. m maHm.o mmss.o mmmH.o mm nepomm me< .mmmmmmmm ‘HHHHmcsssoo m H ampssz meHoQHpm pmsHm souowm gonowm QOpowm pommpsm omschQOOIIN mHQwB 33me . 04”wa . CmHyb . camo.o moH:.o H:m&.o rt no:H.o mamk.o ya I". . . I . .. .. , :Hmkw . 0 \.~ M ~.€..th~_—~,M——MA we ofi al.-s“ . A. . 35% I ‘.\‘1 .-|. ~ .. \ . .. . s . ‘ s...\ 35 50.00H 50.00H 50.00 00.05 0000H00> 000000 Rm.m: RN.:m RO.MH RN.H: mocmth> H0005 HHOm. 0 0000.0 0000.0 0:0H.0 00 0005. H m5H0.0 :HHm.0 5005.0 00 5000. 0 0000.0 5H55.0 0000.0- Hm 0Hz5. H 0000.0 0000.0 0000.0 00 0000. pHHgm 0000.0 0000.0 00H:.0 0: 0005. H 0000.0 005H.0 0005.0 0: 0000. H mH00.0 0:Hm.0 0505.0 5: 0000. H 00H5.0 H0mH.0 H000.0 0: 0Hm0. H 0000.0 00H:.0 H:05.0 m: 0005. H :Hm0.0 mm:H.0 0H05.0 :0 Eflmmfifl 300%“ 3:22.... Home Ems mwmfim U®5QH9200115 mHQmB =""=vs are ud~,.'urv U V V!" "I-“ . ' r- ”r” “"1 «Iv Jgnqu Sr” -;.'h/~¢'-¥ ‘ud ‘*"U\4VV 0'. V .‘fi - . l A < J l I ‘FJ- q .0 Q .. . L ..._, ‘a‘. 5" '1 ‘5‘9, I 9..” o. I- Q ‘."‘~— v I ‘. (‘A ‘5. ‘Q “M... Q} h :1 ~~k~ ‘5.- c‘, v \rh,.1 ‘ VA“- 3 .F‘A V: 0 O .5; ~1 . li"‘n‘_~ . D“\ID 1‘ 36 subjects are associated clearly with the second factor. Factor scores were computed for both the one factor (first principal axis factor) and the two factor solution. The method of computation was as follows (MacLean, Danbury, and Talbott, 1964; Creaser, 1955): l) 2) 3) 4) Each subject is assigned to a factor only if it can be done unambiguously (loading must be high on the factor with no secondary loadings of consequence on the other factor). The loading of only the assigned subjects are transformed to factor weights by the following formula: Wi = Where: W- is the weight of 2 t e ith variable; Fi ) ii the loading of i h variable on the factor being estimated. The factor scores for each factor then are computed by applying the factor weights to the standard scores of the raw data for only the subjects assigned to it. The weighted factor scores for each factor then are converted to standard scores. The factor scores for a given factor, then, are nothing more than the distribution of the weighted average standardized rankings given each of the 123 person descriptions by only the subjects unambiguously associated with the factor. It should be noted that this method does not use the factor weights which are directly derivable from the general factor analytic model. If the clusters of variables do in fact fall very close to the reference vectors constructed 37 by application of the factor analytic model, the factor scores obtained by this approximate method and the formally derived ones will be very nearly the same. If the clusters do fall some distance from the reference vectors, then the approximate method employed here will tend to produce corre- lated factor scores instead of orthogonal ones. This is because the approximate method uses only information about central tendencies of the group clustered and ignores completely its relationship to the other clusters. The formally derived factor weights take into account both kinds of information and hence the factor scores are uncorrelated or orthogonal. The primary interest here was to represent each group central tendency as closely as possible, hence the approximate method was selected. As can be noted in Table 8, the first principal axis factor cannot be distinguished from Factor 1 (the largest). in the two factor rotated solution. Factor 2 is substantially correlated with each of them. Examination of the factor score differences between Factor 2 and each of the other two factors shows that there are only 22 items out of the 123 for which the difference between standard scores exceeds one standard deviation. There are no items for which such differences exceed two standard deviations. The standard so ores for all three factors are reported in Appendix C. Arc Table 8. 38 Correlations among Factor Scores for First Principal Axis (lst PA) Factor and the Two Factor varimax Rotated Solution (2-FS) 1) 1st PA. --_ 2) 2-FS: l .998 -—- 3) 2-FS: 2 .764 .725 -__ "1 ,. T),- ...A-‘ V‘ 6X3 .7“""" A‘“V~‘ .- EEC F331 '6' . r." we“ an: alb war .0vuo an! into :Lrector c ten ' 0 -00 Q» Q» 2 .. 1. r0 nu: 39 From examination of the small differences that do exist between Factor 2 and the other two factors, one could easily conclude that students of factor 2 were of a more liberal bent than those on the other factors. Factor 2 students perceive social protesters and some left wing types of persons as having beliefs and ideas which are more like their own than do students on the other factors. The kinds of persons Factor 2 students perceive as having beliefs less like their own than students on the other factors are some prominent republicans, military leaders, director of the FBI, and other persons usually character- ized as right of center politically. Since the refinements that would be introduced by going to a two factor solution would be small and the complexity of analysis very great, it was decided that the ratings of the person descriptions in terms of similarity of their beliefs with subjects' own beliefs could be essentially and adequately represented by a single factor solution (first principal axis factor). This means that in selecting the source descriptions, only one rank order in terms of belief similarity needs to be taken into account for the students participating in the Pretest study. Factor scores were computed for the first principal axis factor in three additional ways. First, the simple 40 means were computed across the 53 students for each of the 123 items using the raw data. Second, the simple means were computed across the 53 students on each item using the raw data transformed to standard scores. The third other method used the factor weights derived from the model. In this particular situation, factor scores computed by any one of the four methods were not interpretably different from each other (see Table 9). Since the simple means of the raw data were the easiest and least time-consuming to compute and interpret, it was decided to use them to select the two sets of sources, each to be associated with different ends of the belief-disbelief continuum. The means of the belief ratings of the 123 person descriptions were rank ordered (see Appendix B). The top twenty-four person descriptions (largest means) were defined as sources associated with the disbelief end of the belief-disbelief continuum. And conversely, the bottom twenty-four person descriptions were defined as the sources associated with the belief end of the belief-disbelief continuum. See Appendix D for a list of the source descriptions thus selected. 41 Selection of Source Credibility Semantic Differential Scales Since one of the main purposes of this study was to investigate the generality of the source credibility dimensions of the Berlo results, it was decided to use his semantic differential scales in order to facilitate com- parisons. His results showed that one scale had a very low communality (objective-subjective). This scale was there- fore dropped. Hence, thirty-four of Berlo's original thirty-five scales were selected for use in this study. The Berlo data were re-factor analyzed deleting the objective—subjective scale. The resulting factor matrix will be found in Appendix E. This minor refinement re- sulted in no changes in the previous interpretation of the factor analysis results. Main Study: Participants Participants for the main study were volunteers from five courses of the Department of Communication, Michigan State University, taught during the 1967 summer term. The names and enrollments of the five courses are in Table 10. Students were not required to participate, but were en- couraged to do so by the instructors. Special credit was given to students who did participate in the project. The background information obtained from the volunteers from the five classes is contained in Table 11. 42 Table 9. Correlation Table among the Factor Scores for the First Principal Axis Factor Computed by Four Methods 1) derived factor weights ...... 2) approximate factor weights .99963 ...... 3) simple means of standardized data .98735 .98429 ______ 4) simple means of raw data ,99113 ,98897 .99932 ____ 1 2 3 u 43 Table 10. Communication Courses Participating in Main Study Course Summer Term Enrollment Communication 100 First Five Week 39 Communication 100 Ten Week 54 Communication 300 First Five Week 28 Communication 300 Ten Week 19 Communication 470 First Five Week 12 TOTAL 152 (I) (I) 44 Table 11. Background Information N = 133 Item "Response _Percentage Sex male 60.9% female 37.6 no answer 1.5 Year in School freshman 15.0% sophomore 12.0 junior 22.6 senior 42.9 graduate 6.8 no answer 0.8 Age 17 2.3% 18 12.0 19 8.0 20 1 .O 21 18.0 22 14.3 23-24 8.3 25-27 7-5 28—30 4.5 31-49 5-3 no answer 0.8 Marital Status married 24 8% single 73 7 no answer 1. Estimated Parent's Total Income under $3,000 0 % Last Year 3,000 to 4,499 4 4:500 to 5,999 5- 6,000 to 7,999 12. 8,000 to 9,999 15. 10,000 to 14,999 26. 15,000 to 19,999 12. 20,000 to 29,999 30,000 and over no answer How do you consider yourself conservative politically? neither liberal Have you traveled in a foreign yes <3ountry outside the United States no and Canada? no answer & [.4 O w-P‘kfi ONUUUO OUTUTOUUOOUOUTCI) Ul- - 0 a... 45 Main Study: Validation of Sample of Sources The sample of sources was based on ratings from a different set of subjects than would be used for the main study. The pretest used student volunteers from the regular school year. The main study was toxuuastudents attending summer sessions. The sources selected from the one set of subjects should be validated for the second set. The forty sources rated with "beliefs and ideas most different from my own" and the forty sources rated as "most similar to my own" by the regular school year students were again rated by the summer school students using the same rating scale. Tables 12 and 13 very readily show the high degree of similarity between the relative orders of the belief-dis- belief mean ratings for the two groups or students. The orders of the means for two groups did not differ greatly, but what about the magnitude of the means themselves? In- dependent groups t-tests were run comparing the means for the two groups of students on each of the 80 source mean ratings. There were differences for only 8 of the 80 pairs of means. From such results, it seemed reasonable to use the sample of sources selected using the Spring group for the summer group. The background data reported earlier was collected at 46 A the same time as the above validation data. Also admin- istered at the same time was the dogmatism scale. Appendix F contains the questionnaires used to obtain the validation data, the background information, and the dog- matism score. Construction of Source Evaluation Instrument The basic instrument for obtaining data for assessing the individual dimensions of source evaluation (credibility) basically consists of the semantic differential evaluation on a 9-point scale of the forty-eight sources (twenty-four associated with each end of the belief-disbelief continuum) using the thirty-four Berlo credibility adjective scales. The task required of a subject in reSponding to such an instrument would be substantial. Hence, certain procedures were followed with an aim toward both alleviating and dis- tributing the fatigue effects involved in such a task. The questionnaire was set up in typical semantic differential fashion, the source description at the t0p of the page (legal size) with the thirty-four scales double- Spaced down the page. A different independent random order of scales was assigned to each of the forty-eight source descriptions. For each order, the polarity of the scales was independently randomized. Each questionnaire contained 47 Table 12. Correlations between Source Belief-disbelief Mean Ratings of the Regular School Year Students and the Summer Students I". "‘1 .. _.. _.‘ . Rank Order Pearson P-M Sample of Sources on which Correlation Correlation Correlation is Based of Means of Actual Means All 80 sources .955 .983 40 sources having beliefs similar to my own .774 .885 40 sources having beliefs different from my own .870 .844 Selected sample of 48 sources .945 .989 24 sources having beliefs similar to my own .713 .861 24 sources having beliefs different from my own .843 .771 Table 13. Quartile Crossed-Classification of the Mean Ratings Given the 80 Sources by Regular School Year and Summer Students (Number of Means = 80) Regular Summer Students School Year Quartile l 2 3 4 Students (similar) (different) 1 (similar) 18 2 0 0 2 2 18 O O 3 0 0 19 1 4 (different) 0 0 l 19 48 forty-eight pages of semantic differential scales with a different source description on each plus the instructions. The semantic differential portion of each questionnaire was assembled using a different independent random order. It was hOped that whatever serial effects of fatigue that might be present in completing the questionnaire would be randomly distributed across the scales, polarity of the scales, and the concepts. A computer program with a random number generator was used to generate the three independent randomi- zations needed in constructing the physical questionnaire. The questionnaires were to be coded as the scales appeared on each page from left to right (l-left, 9-right). The randomization program also output the parameters for another program, the purpose of which was to unscramble the random orders that would be involved in the coded raw data. For a sample page from the questionnaire along with the instructions, see Appendix G. Subjects were instructed when they evaluated the jperson described with the scales to think of some communi- caticmlsituation in which the person could have something ‘to say to them. The instructions further stated: This could be a face-to-face conversation, a Speech, something the person has written in a letter, neWSpaper, magazine, or other written document, news film on TV, or some other kind of 49 Situation in which the person could communi— cate something to you. Also, the thing that the person would be communicating would be a reasonable thing for that person to be doing. Obviously, the attempt here was to induce a set so that the subjects would use the scales to describe each person as a source in some kind of communication situation. In order to decrease the probability that subjects would skip around in the questionnaire, and thereby defeat to some degree the purpose for which the randomizations were employed, subjects were Specifically instructed to complete each page before going on to the next in serial order. Also, a complete list of the person descriptions was included in the questionnaire at the close of the in- structions with Specific instructions to look it over. Some skipping around would probably be prompted by wanting to know what kinds of persons are to be evaluated. Pro- viding the list would help in this reSpect. Also, it was felt that subjects should have some idea of‘the range of the types of individuals before starting the task so that they could develop anchoring points for the ends points of the scales relative to the selected sources (person descriptions) very early in the evaluation process. One other procedure was employed to reduce fatigue effects of the lengthy questionnaire. The questionnaire was SO self administered in the residence of the participating student. The students were asked Specifically not to complete the questionnaire in a single sitting. Rather, they were encouraged strongly in separate verbal in- structions at distribution time to do them in several sittings, at least four, and more if possible. They were given two weeks in which to complete the task. Also, in- structors in the participating classes gave student volunteers extra credit for participating in research projects. Main Study: Source Evaluation Data Approximately two weeks after the validation data for the selected sample of sources was obtained, the source evaluation questionnaire was distributed to student volunteers in the hast 10 minutes of the participating classes. The completed questionnaires were to be turned in to the in- structor. The turn-in date was set for approximately two weeks after distribution. It was decided that if a subject failed to reSpond to more than ten per cent of the sources or five pages, his questionnaire would not be included in the analysis. In all, 105 completed questionnaires were turned in from the five participating classes. Only two incomplete questionnaires had to be discarded. There were 51 only six additional questionnaires which had whole pages where the data were missing (3 with 1 page missing, 1 with two pages missing and 2 with 4 pages missing). Throughout the rest of the questionnaires, there was little problem of missing data for individual scales. There were no more than a dozen such instances. On the whole, the subjects seemed to have reSponded to the task in a serious and deligent manner. A missing response on usable questionnaires was coded as 5, the midpoint of the 9-point scale used. In addition, four more questionnaires had to be discarded because the dogmatism score was mtsing for them. In all cases, they were from students attending only the first five week summer session who could not be reached to obtain the missing score before they left the campus. After editing, there were 99 usable source evaluation questionnaires. The data thus obtained by administration of this questionnaire forms a three dimensional data matrix or cube. It has 161,568 cells (99 x 48 x 34) with each containing the semantic differential evaluation score (ranging from l.'to 9) for one of the 99 subjects for one of the 48 souices on one of the 34 scales. A schematic of this cube (N=66> Total Berlo Factor ‘ Safety' .789 .565 .224 .605 .443 .332 .697 .504 .278 Qualifi- catitui .618 .676 .360 .554 .543 .380 .586 .610 .370 Dynamism .268 .408 .584 .289 .353 .337 .279 .381 .461 (N=99) 97 safety factor tends to hold across all sub—group solutions. For the qualification factor, the general pattern still holds up for the sub-groups except for the factor which is slightly favored. For belief sources, the tendency for slightly favoring the middle factor is present, as in the overall averages. However, the slight variation comes with the disbelief sources. There, instead of slightly favoring the middle factor, there is a tendency for the largest factor to be slightly favored. The most variation is found in the pattern for the way the dynamism factor maps into the three individual factors. Indices for the dynamism factor are weaker than those for the other two factors. The overa11 pattern holds up for the belief sources with a slightly greater tendency for the dynamism factor to map into the smaller factor. The variation comes with the disbelief sources. Across all solutions for disbelief sources, there is no tendency for the dynamism .factor to map into any particular factor in the individual solutions. However, when the dogmatism sub-groups of the disbelief sources are examined, one notes differing tendencies. For the high dogmatism group, there is a slight tendency to favor the middle factor. For the middle group, the snallest factor is favored. For the low dogmatism group, thermais no particular tendency to favor any Specific in- d ividual factor. 98 Interpretation Limits on Index of Factorial Similarity-- No one has developed a sampling distribution for the index of factorial similarity, but Wrigley has what he terms a lower tnund of good fit for the index: .1. lower bound of = l Hik— where k is the number of 800d fit 2 factors in each of the two solutions being compared. In this situation, k would equal three with the lower bound of good fit being .789. With this standard, the only indices in the table one coulitalk about across most sub-groups (all sub-groups which include belief sources) would be some of the largest ones for the safety factor. Only one of the nine sub- groups has indices for the qualification factor which exceed this lower bound value. None does for dynamism. Another approach to this problem would be to generate some coefficients from random data. Factor analyses of random data of the size of the Berlo factor analysis and ten of the size of the individual analyses were computed and compared on the same basis as in this study using the index of factorial similarity. A distribution of 250 indices was generated. The absolute range was from -.37 to +.47, the ninety-fifth per- czentdle range was from -.31 to +.42, the ninetieth percentile range was -.28 to +.39, and the median value was between +uCfl+ and +.O5. Using this standard, one couLi make much wider 99 interpretation involving the magnitude of the indices. All factors would have a number of interpretable indices. Ivariance Distribution Pattern Across the Three Scale Types in the Individual Solutions-- There is an alternative way of getting some idea of the mapping of the individual factors into the Berlo three source evaluation factors. The scales are classified into three groups, each representing a Berlo factor. Then, the way the variance associated with each group of scales is distributed across each of the three in- dividual factors (largest, midele sized, smallest) is examined. Table 35 presents this tabulation. At the top of the table, the variance associated with each group of scales is broken down for the Berlo factor analysis. This breakdown is useful for interpreting the similar breakdowns for the individual factor analyses. As can be seen, the variance associated with each scale group is distributed essentially on a single factor. The secondary loadings of the scales on other than the principal factor were quite low in the Berlo analysis. When the similar tabulations in the remainder of Table 34 are examined for the various sub-groups of the in- dividual solutions, the distribution pattern for the three scale groups is nowhere as clean, clearcut, and as striking 100 Table 34. Mean Proportion of Variance Accounted for by Each of the Three Groups of Scales (Berlo's Three Factors of Safety, Qualification and Dynamism) on Each of the Factors in the Three Factor Rotated Solutions of the Individual Factor Analyses FaEtor Berlo ---------- Factor Analysis Large Medium Small Safety Scales .313 .025 .003 Qualification Scales .027 .126 .009 Dynamism Scales .006 .009 .094 Individual Factor Analyses (Number equals 198) Belief Sources Disbelief Sources Total Factor Factor Factor Low Large Medium Small Large Medium Small Large Medium Small Dogmatism (Open) Safety .224 .088 .046 .194 .099 .066 .209 .094 .056 Qualifi- cation .057 .079 .033 .083 .056 .037 .070 .068 .035 Dynamism .034 .049 .059 .059 .063 .038 .046 .056 .049 (N=33> (N=33) (N=66> Medium Dogmatism Safety .229 .082 .044 .189 .109 .064 .209 .096 .054 Qualifi- cation .065 .085 .017 .072 .067 .031 .068 .076 .024 Dynamism .028 .038 .076 .064 .052 .044 .046 .04 .060 (N=33> High Dogmatism (Closed) Safety .199 .107 .041 .177 .094 .073 .188 .101 .057 Qualifi- cation .064 .073 .026 .063 .079 .020 .063 .076 .023 Dynamism .034 .037 .066 .064 .051 .033 .049 .044 .050 Total Safety .217 .092 .044 .187 .101 .068 .202 .097 .056 Qualifi- cation . 062.079 . 025 . 073 . 068 . 029 . 067 .073 . 027 Dynamism .032 .041 . 067 . 062 . 055 . 039 . 047 .048 .053 (N=99) (N=99) (N= 198) 101 Twible 35. Comparisons of Berlo's Three Factor Solutions with the Individual Three Factor Solutions Using the Index of Factorial Similarity: Summary of the Frequency and Percentage of the Number of Different Factors With Which the Three Berlo Factors had Their Highest Indices of Factorial Similarity in the Individual Analyses —' No. of'Factors f6r Disbelief Sources No. of Factors --------- for Belief Two or Ikogmatism Sources Less Three Total Iwa Two or Less freq. 3 6 9 (Open) pct. 9.09 18.18 27.27 Three freq. 11 13 24 Pct. 33-33 39-39 72-72 Totala freq. l4 19 33 pct. 42.42 57.57 99.99 Idedium Two or Less freq. 4 5 9 pct. 12.12 15.15 27.27 Three freq. ll 13 24 Pct- 33.33 39.39 72-72 Total freq. 15 18 33 pct. 45.45 54.54 99.99 Itigh Two or Less freq. 6 3 9 (Closed) pct. 18.18 9.09 27.27 Three freq. l8 6 24 pct. 54.54 18.18 72.72 Total freq. 24 9 33 pct. 72.72 27.27 99.99 Two or Less freq. 14 13 27 Total pct . 14.14 13 13 27.27 (Across all Three Three freq . 32 4O 72 Ckroups) pct. 32.32 40.40 72.72 Total freq. 46 53 99 pct. 46.46 53.53 99.99 aThe chi square for comparison between the totals for the disbelief sources only for the Low Dogmatism Group versus the High Dogmatism group = 5.061 (df = l), p value less than .05. 102 as in the factor analysis which produced the scale groupings. By this measure, there is a marked tendency for the safety :fictor to map into the largest individual factor. However, the variance accounted for by the safety scales on the middle and smallest factors cannot be ignored. Often the variance accounted for by these scales is larger than the accounted for by either of the other two scale groups. variance associated with the qualification is mostly dis- tributed on the largest and middle individual factors. For the belief sources, the tendency is to favor the middle factor and for disbelief sources, the tendency is to favor the largest factor. The dynamism scales' variance is dis- tributed differently for belief sources than it is for dis- belief sources in a more pronounced fashion than for quali- fication. For belief sources, the smallest factor has the :largest amount of the variance associated with the dynamism sscales. For belief sources, the tendency is for the largest and huddle factors to have most of the variance, with the Ilargest favored slightly. The most interesting aspect of Table 34 is that when averaged across all solutions, the safety scales accounted for more variance on each of the three individual factors than did either of the other two types of scales. This same pattern holds generally for dis- 103 belief sources. The variation that does come for the belief sources is with the dynamism scales on the smallest factor. There the variance accounted for by them exceeds that of the safety scales. Mapping of all Berlo Factors into Individual Three Factor Solutions-— To examine these data more systematically, the number of different individual factots into which the Berlo factors map are cross-tabulated with the dogmatism groups and the source types (see Table 35). Three factors are used in a higher proportion of the solutions in mapping the Berlo factors into the individual factors using the index of factorial similarity for belief sources (73%) than for disbelief sources (54%). This proportion is constant across all three of the dogmatism groups for belief sources (73%). It does vary for the disbelief sources. It is highest for 'the low dogmatism group (58%) and lowest for the high dog- matism group (28%). Just by examining the largest indices of factorial shmilarity and finding that each Berlo factor maps into a sseparate individual factor, one cannot conclude that the match is good. Quite often the difference between indices is quite small, making clearcut assignment nearly impossible. However, it is a necessary but not sufficient condition for an adequate matching. For the solutions which map into 104 three factors, one could examine the way the variance of the scales is associated with each factor to define an adequate match. For present purposes, two conditions had to “be met for the match to be defined as adequate. More than half of the variance associated with each scale category had 'to be distributed on a single separate factor. Further, the scale category associated with each factor had to account for more than half of thevariance accounted for by that factor. Table 36 contains the tabulation of this classification of the individual factor solutions. A higher prOportion (36%) of the solutions for belief sources were judged to be adequate than that (19%) for the disbelief sources. These overall proportions for both kinds of sources differed very little among the three dogmatism groups. The proportion of solu- tions for low dogmatics which were judged to match inadequately the Berlo results for both kinds of sources (36%) was lower than that (61%) for high dogmatics. Mapping of Safetquactor into Largest Individual Factor-— Another interesting aSpect of this matching procedure is to exandne more systematically for each Berlo factor separately 'the degree to which it matches with the individual factors (large, medium, or small). The safety factor is summarized ixl'Table 37. The most obvious difference between the solutions ijszfor the belief and disbelief sources. Berlo's safety 105 Table 36. Comparisons of Berlo's Three Factor Solution With the Individual Three Factor Solutions Using %te Index of Factorial Similarity: Summary of Frequency and Percentage of Individual Solutions for which All Three of the Berlo Factors were Judged to have Adequate Matches ‘Factor MatChing for Disbelief Sources Factor Matching ----------- Dogmatism for Belief Sources Adequate Inadequate Total Low Adequate freq. 3 12 15 (Open) pct. 9.09 26.36 45.45 Inadequate jfreq. 6 l2 18 pct. 18.18 36.36a 54.54 Total freq. 9 24 33 pct. 27.27 72.72 99.99 Medium Adequate freq. 4 7 11 pct. 12.12 21.21 33.33 Inadequate freq. 2 2O 22 pct. 6.06 60.60 66.66 Total freq. 6 27 3 pct 18.18 81.81 99.99 High Adequate freq. 1 9 10 (Close) pct. 3.03 27.27 30.30 Inadequqte freq. 3 2O 23 pct. 9.09 60.60a 69.69 Total freq. 4 29 33 pct. 12.12 87.87 99.99 Total. Adequate freq. 8 28 36 (Across a1 pct. 8.08 28.28 36.36 Three ' Inadequate freq. ll 52 63 Groups) pct. 11.11 52.52 63.63 Total freq. 19 8O 99 pct. 19.19 80.80 99.99 aTTua chi square for comparison between the proportion of the ImvaDogmatism Group for which the match between Berlo's results amui'the individual analyses is judged to be inadequate for bcflfli kinds of sources and that for the High Dogmatism Groups = 1+L127'(df = l), p-value less than .05. 106 Table 37. Comparisons of Berlo's Three Factor Solution with the Individual Three Factor Solutions Using the Index of Factorial Similarity: Summary of the Frequency and Percentage of each Kind of Individual Factors in Terms of Size (Large, Medium, or Small) Which have the Highest Index of Factorial Simili- larity with Berlo's __ Belief Safety Factdr ‘__ Sources Disbelief Sources Factor Dogmatism Factor Large Medium Small Total Low (Open) Large freq 16 6 3 25 pct. 48.48 18.18 9.09 75.75 Medium freq 4 O 3 7 pct. 12.12 0.00 9.09 21.21 Small freq. 0 O l 1 pct. 0.00 0.00 3.03 3.03 Total freq. 20 6 7 33 pct. 60.60 18.18 21.21 99.99 Medium Large freq 13 10 3 26 pct. 39.39 30.30 9.09 78.78 Medium freq. 2 2 l 5 pct. 6.06 6.06 3.03 15.15 Small freq. 1 O 1 2 pct. 3.03 0.00 3.03 6.06 Total freq. 16 12 5 33 pct. 48.48 36.36 15.15 99.99 High. Large freq 9 7 4 20 (Closed) pct 27.27 21.21 12.12 60.60 Medium freq 6 3 2 11 pct. 18.18 9.09 6.06 33.33 Small freq 2 O 0 2 pct. 6.06 0.00 0.00 6.06 Total freq 17 10 6 33 pct. 51.51 30.30 18.18 99.99 TotaJ. large freq 38 23 10 71 (Across pct. 38.38 23.23 10.10 71.71 all Three freq . 12 .12 5 .05 6. O6 23 .23 Groups) Small freq . 3 O 2 5 pct. 3.03 0.00 2.02 5.05 Total freq. 53 28 18 99 pct. 53.53 28.28 18.18 99.99 107 Table 38. Comparisons of Berlo's Three Factor Solution with the Individual Three Factor Solutions Using the Index of Factorial Similarity: The Proportion of Solutions for which the Safety Factor has the Highest Index of Factorial Similarity with the Largest Factor in the Individual Solutions for the Belief Sources versus that for the Disbelief Sources Disbelief SOurces Factor which has Highest Index of Factorial Similarity Belief Sources Factor Medium or which has Highest Index Largest Smallest of Factorial Similarity Factor Factor Total Largest Factor freq. 38 33 71 pct. 38.38 33.33 71.71 Medium or freq. l5 13 28 Smallest Factor pct. 15.15 13.13 28.28 Total freq. 53 46 99 * pct. 53.53 46.46 99.99 chi square = 6.021 (df = l), p-value less than 105 108 factor is more likely to match most closely with the largest factor in the individual solutions for belief sources (72%) than for disbelief sources (53%) (see Table 38). In terms of the differences between dogmatism groups, the only difference worth mentioning is in the degree to which the safety factor maps into the largest factor in the disbelief sources' solutions for the extreme groups. Among low dogmatics (see Table 39), the safety factor is more likely to match the largest factor (48%) than among high dogmatics (27%). Mapping of Qualification Factor into Middle-Sized Factor-- Table 40 summarizes the results for the qualification factor mapping. There are no particular differences between sub-groups of solutions in the way this mapping proceeds. For about half the solutions, the match is with the middle factor; for another third, the map is into the largest factor; and for the remainder, the smallest factor. Mapping of Safety_Factor into Smallest Individual Factor5-cThe matching for the dynamism factor is contained in Table 41. Again, as with the safety factor, the most Obvious differences are between kinds of sources. ,The match for belief sources is more likely to be with the smallest factxxr (67%) than it is for disbelief sources (26%) (see Table 42). All differences among the dogmatism groups are small. 109 Classification of Factors in Individual Three Factor Solutions-- Another approach to the question of the degree to which the individual solutions correspond to the Berlo results is to examine each of the individual three factor rotated solutions and classify each factor in terms of the results found by Berlo. Each scale in the individual solutions is assigned to the factor on which it had the highest loading. The highest loadings for each factor is examined. If more than half of the scales in one of the Berlo factor-scale categories is loaded on the same individual factor and these loadings in that scale category comprise more than half the highest loadings for the individual factor in question, then that factor is classified as representing that Berlo factor. Alternatively, if more than half of the scales from two or three Berlo scale categories have their highest loadings on the same factor and this comprises more than half of all the scales having their highest loading on this factor, then that individual factor is classified as a combination of the Berlo factors involved. Finally, all remaining unclassified factors are designated as "other." These are the individual factors assumed to be not classifiable in terms of the Berlo results. Tluazresults of classification of the 198 individual three- factxxr rotated solutions are found in Tables 43 and 44. One tablja summarizes the individual classification elements for 110 Table 39. Comparisons of Berlo's Three Factor Solution with the Individual Three Factor Solutions Using the Index of Factorial Similarity: The Proportion of Solutions for which the Safety Factor has the Highest Index of Factorial Similarity with the Largest Factor in the Individual Solutions for both Belief and Disbelief Sources for the Low Dogmatism Groups versus that for the High Dog- matism Group *Factor for BEth Belief and Disbelief Sources which has Highest Index of Factorial Similarity Medium or Largest Smallest Doggmatism Group Factor Factor Total Low (Open) freq. l6 17 33 pct. 48.48 51.51 99.99 High (Closed) freq. 9 24 33 pct. 27.27 72.72 99.99 Total freq. 25 41 66 pct. 37.87 62.12 99.99 chi square = 3.863 (df = l), p-value less than .05 111 Table 40. Comparisons of Berlo's Three Factor Solution with the Individual Three Factor Colutions Using the Index of Factorial Similarity: Summary of the Frequency and Percentage of each Kind of Individual Factors in Terms of Size (Large, Medium, or Small) Which have the Highest Index of Factorial Similar- ity with Berlo's QualificatI0n*Factor Belief Disbelief Sources Factor Sources Dogmatism Factor Large Medium Small Total Low Large freq 1 5 4 10 (Open) pct. 3.03 15.15 12.12 30.30 Medium freq 5 5 6 16 pct 15.15 15.15 18.18 48.48 Small freq 5 O 2 7 pct. 15.15 .0.00 6.06 21.21 Total freq ll 10 12 33 Pct 33-33 30.30 36.36 99-99 Medium Large freq 5 5 2 12 pct 15.15 15.15 6.06 36.36 Medium freq 7 8 4 19 pct. 21.21 24.24 12.12 57.57 Small freq 2 O O 2 pct 6.06 0.00 0.00 6.06 Total freq l4 l3 6 33 pct. 42.42 39.39 18.18 99.99 High Large freq 2 9 l 12 (Closed) pct 6.06 27.27 3.03 36.36 Medium freq 8 6 2 16 pct. 24.24 18.18 6.06 48.48 Small freq 1 4 O 5 pct. 3.03 12.12 0.00 15.15 Total freq ll 19 3 33 pct 33.33 57-57 9-09 99-99 Total. large freq 8 19 7 34 (Across all pct. 8.08 19.19 7.07 34.34 Three Medium freq 2O 19 12 51 Gimnups) pct. 20.20 19.19 12.12 51.51 Small freq 8 4 2 14 pct 8.08 4.04 2.02 14.14 Total freq 36 42 21 99 pct 36.36 42.42 21.21 99.99 112 Table 41. Comparisons of Berlo's Three Factor Solution with the Individual Three Factor Solutions Using the Index of Factorial Similarity: Summary of the Fre- quency and Percentage of each Kind of Individual Factors in Terms of Size (Large, Medium, or Small) Which have the Highest Index of Factorial Similarity with Berlo's DynamISm Factor Belief Disbelief Sources Factor Sources Dogmatism Factor Large Medium Small Total Low (Open) Large freq. 4 l O 5 pct. 12.12 3.03 0.00 15.15 Medium freq. 3 4 l 8 pct. 9.09 12.12 3.03 24.24 Small freq. 6 8 6 20 pct. 18.18 24.24 18.18 60.60 Total freq. l3 l3 7 33 Pet. 39 39 39-39 21 21 99 99 Medium Large freq 0 l l 2 pct. O 00 3.03 3.03 6.06 Medium freq 2 2 pct. 6.06 6.06 6.06 18.18 Small freq. l2 5 25 pct. 36.36 15.15 24.24 75.75 Total freq. 1 8 ll 33 pct. 42.42 24 24 33.33 99.99 High, Large freq. 1 O 2 3 (Closed) pct. 3.03 0.00 6.06 9.09 Medium freq 4 5 O 9 pct. l2 12 15.15 0 OO 27 27 Small freq 9 6 21 pct 18 18 27.27 l8 18 63 63 Total freq ll 14 3 pct. 33.33 42.42 24.24 99 99 Total. Large freq. 5 2 3 10 (Across pct. 5.05 2.02 3.03 10.10 all. Medium freq. 9 ll 3 23 Three pct. 9.09 11.11 3.03 23.23 Groups) Small freq . 24 22 2O 66 pct. 24.24 22.22 20.20 66.66 {VWC fi. Total freq. g8 35 26 99 Pct. 3 38 35.35 26.26 99-99 113 Table 42. Comparisons of Berlo's Three Factor Solution with the Individual Three Factor Solutions Using the Index of Factorial Similarity: The Proportion of Solutions for which the Dynamism Factor has the Highest Index of Factorial Similarity with ~the Smallest Factor in the Individual Solutions for the Belief Sources versus that for the Dis- belief Sources Belief Sources Factor Medium or which has Highest Index Smallest Largest of Factorial Similarity Factor Factor Total Smallest Factor freq. 2O 46 66 pct. 20.20 46.46 66.66 Medium or freq. 6 27 33 Largest Factor pct. 6.06 27.27 33.33 Total freq. 26 73 99 pct. 26.26 73.73 99.99 chi square = 29.25 (df = l), p—value less than 105 114 each individual factor separately and the other summarizes according to the pattern of classification elements found across the three individual factors from largest to smallest. Extent of all Berlo Factors Represent in Individual Solutions-- The first question that might be asked is whether there are differences among sub-group solutions in the number of Berlo factors found to be represented in the individual three factor solutions. See Table 45 for such a tabulation. The solutions for belief sources differ from those for dis- belief sources, both in the number of Berlo factors repre- sented and in the number of times that all three were repre- sented. Factor solutions for belief sources tend to have more factors representative of the Berlo results (mean 2.10) than do the solutions for the disbelief sources (mean 1.63). The proportion of times that all three Berlo dimensions are present is higher for belief sources (49%) than it is for disbelief sources (30%). In the Berlo results, the three factors in order by amount of variance accounted for are safety, qualification, and dynamism. Does this same order hold up in the individual solutions? As might be expected (see Table 46), the number of solutions for belief sources which represent the three Berlo factors in the same order of magnitude of variance extracted is krger (35%) than that for 115 Table 43. Classification of the Factors in the Three Factor Individual Solutions According to the Factor Types found by Berlo et a1. (Safety, Qualification, and Dynamism) Individual Dis- Analysis Factor Over— Be—.be-~ Low Medlfigh Low Medfmh Low Medffigh Factor Size T e all lief lief Do .Do .Do .Do .Do .Do .Dog.Do .Dog largest Saggety I55 57 E9 355 3‘55 3 T95 205 1 1'6— 155-18—— Qual. 14 9 5 3 6 5 2 3 4 1 3 1 Dynamism 4 2 2 2 2 O 2 O O O 2 O SQ Comb. 13 9 4 6 5 2 3 4 2 3 l 0 SD Comb. 2 O 2 O O 2 O O O O O 2 OD Comb. 31 4 27 12 9 10 2 1 l 10 8 9 SQD Comb. 2 2 O l 1 O 1 1 O O O O Other 26 16 10 7 8 ll 4 4 8 3 4 3 "" -‘ " Total 198 99 99 65 56 65 33 33 33 33 33 33 Medium Safety 49 21 7128 l5 18 16 8 4 9 7 Qual. 73 45 28 23 28 22 25 l8 l2 8 10 10 Dynamism 33 16 17 16 10 7 6 6 4 10 4 3 SQ Comb. O O O O O O O O O O O 0 SD Comb. O O O O O O O O O O O 0 QD Comb. 12 4 8 2 2 8 1 l 2 l 1 6 SQD Comb. O O O O O O O O O O O O Other 31 l3 18 10 8 l3 3 4 6 7 4 7 Total 198 99 99 55 66 56 33 33 33 33 33 33 SmalleSt Safety 0 O O O O O O O O 0 O O Qual. 2O 8 12 ll 4 5 4 1 3 7 3 2 Dynamism 70 5O 2O 21 28 21 15 19 19 6 9 5 SQ Comb. 0 O O O O O O 0 O O O 0 SD Comb. O O O O O O O O O O O O QD Comb. l O l O O 1 O O O O O l SQD Comb. O O O O O O O O O O O O Other 107 41 66 34 34 39 l4 l3 14 2O 21 25 __ Total 198 99 99 65 66 66 33 33 33 33 33 33 Total Safety 155 778 .77 5O 5’ 52 27 24 3 9 25 (Across Qual. 107 62 445 37 3 32 21 22 19 l6 16 13 All Three Dynamism 107 68 339 ’39 4O 28 23 25 2O 16 15 8 SQ Comb. l3 9 4 6 5 2 3 4 2 3 1 0 SD Comb. 2 2 0 0 2 0 0 0 0 0 2 QD Comb. 44 8 36 14 11 19 3 2 3 11 9 16 SQD Comb. 2 2 O l 1 O l l O O O O Other 164 70 94 51 5O 63 21 21 28 30 29 35 Total 594 297 297 198 198 198 99 99 99 99 99 99 Abbreviations: Qual.--Qualification3 S-—Safety; Q--Qualification3 D--Dynamism; Comb.--Combination3 Dog.--Dogmatism 116 Table 44. Classification of Pattern of the Factors in the Three Factor Individual Solutions According to the Factor Types found by Berlo et a1. (Safet —- S, Qualification--Qi,Dynamism--D, Other-~OTH. Individual Analysis Belief Disbelief Factor Dis- Med- Over- Be- be- Low Med .High Low MedHigh Low Med High Large ium Small all lief lief Dog.Dog.Dog.Dog.Dog.Dog.Dog.Dog.Dog. D 53 35 18 17 23 13 11 16 Q 15 10 }_l 40 U'IIDOl-J l—"\'|OO Ski—'00 OI—‘l—‘i—‘U‘ll-‘HO Ol—‘HO WNW mm upmm Opt/J03 fart: U 00 camera COMUD CDC/300 O *3 F 0TH. 0TH. a) O a :t: . 0TH. 0TH. S 0TH. 0TH. Q 0TH. 0TH. D 0TH. 0TH. 0TH. D QD S 0TH. S QD 0TH. |-'|'\) '\]|\)l\)f\) i—‘OO\l—’ NUTIUON I'DUOUUU'lUll-‘HH l-‘UONCM UOCI) S 0TH. QD QD 0TH. 0TH. SQ D 0TH. 1 SQ 0TH. D SQ 0TH. 0TH. SD Q 0TH. SQD 0TH. 0TH. 0TH. 0TH. 0TH. NONH OLJOO\l-’ (DI-Jl—‘(DO NNm-P'CDOOH l-'l\)|—’O\ F—‘l—' [DI-JON O-P'UUO IUKOOI'U F—‘Ol—‘I—J NOOO OOl—JU'I [UNKO [UP—'00 OUIOO I'OKOI—‘l-J I-'l\)|\)l\) HOOO I-‘i-‘l—‘JI' l—‘UUUO (MONO I-'l-’U0l—‘ CIDNH—‘UU Ol—‘ON [UPI-4H ONON OWL» l—‘i—‘OH OM00 HMOI—J 0000 [0000 OOH-l: l\)l\)|\) [DI—’00 O-F—‘OO l—Jl—JOH OOHO HOOO OOON ONH NOOO l—‘I—‘OO {UPI—’H Oi-‘OI—J Ml-‘I-‘O Ol—‘O-P’ OUOIUCI) |--’OO|--J OIUUOO l--’\]OI-’ |-'O|—'|-' OOOO OOO|—' OONQ OOOO OI—‘OO HOOP-'0 l—‘NI—‘N OOOO |-'|-'I-'|'\) l—‘l—‘NNI l—‘ONO OOUOI—J O\O\Ol\) OOOI—J OOOI—’ Ol—‘OUO OI—‘l—‘U‘I \O \O \0 KO 0\ ON C‘\ O\ O\ 0\ DU 00 U0 U0 U.) U) DO UL) U) LU U.) W Total 198 117 Table 45. Number of Factors in the Three Factor Rotated Solutions of the Individual Analyses Which Could be Classified as One of the Three Factors Found by Berlo et al. *Number of7Factors in‘Solution Which Could be Classified as One Found by Berlo et al. O—' 17 2 ’3 Total Overall freq. 17 72 30 79 198 (All solutions) pct. 8.59 36.36 15.15 39.90 Belief Sourcesa freq. 8 23 19 49 b 99 pct. 8.08 23.23 19.19 49.49 Disbelief Sourcesa freq. 9 49 11 30 b 99 pct. 9.09 49.49 11.11 30.30 Low Dogmatism freq. 8 2O 8 3O 66 pct. 12.12 30.30 12.12 45.45 Medium Dogmatism freq. 3 25 8 3O 66 pct. 4.54 37.88 12.12 45.45 High Dogmatism freq. 6 27 l4 19 66 (closed) pct. 9.09 40.91 21.21 28.79 Low Dogmatism freq. 3 6 7 17 33 Belief Sources pct. 9.09 18.18 21.21 51.51 Low Dogmatism, freq. 5 14 1 13 33 Disbelief Sources pct. 15.15 42.42 3.03 39.39 Medium Dogmatism, freq. 3 8 3 19 33 Belief Sources pct. 9.09 24.24 9.09 57.57 Medium Dogmatism, freq. 0 l7 5 ll 33 Disbelief Sources pct. 0.00 51.51 15.15 33.33 High Dogmatism, freq. 2 9 9 13 33 Belief Sources pct. 6.06 27.27 27.27 39.39 High Dogmatism, freq. 4 l8 5 6 33 Disbelief Sources pct. 12.12 54.54 15.15 18.18 aThe Sign test comparison between the distributions for Belief Sources versus Disbelief Sources: chi square = 7.347 (df = l), p-value less than .05. The chi square for comparison of the proportion of Belief Sources having All Three Factors each Classified as one of the Berlo et al. factors and that for Disbelief Sources = 6.353 (df = l), p-value less than .05. 118 Table 46. Extent to Which the Factors in the Three Factor Rotated Solutions of the Individual Analyses Could be Classified with the Largest as the Safety Factor, the Middle as the Qualification Factor, and the Smallest as the Dynamism Factor _f Belief Sources Disbelief Sources Have Have Above Above Classification? Dogmatism Classification? Yes “No 7Total Low Yes freq. 1 10 11 (Open) pct. 3.03 30.30 33.33 No freq. 5 17 22 pct. 15.15 51.51 66.66 Total freq. 6 27 33 __> pct. 18.18 81.81 99.99 Medium Yes freq. 4 12 16 pct. 12.12 36.36 48.48 No freq. 3 l4 17 pct. 9.09 42.42 51.51 Total freq. 7 26 33 pct. 21.21 78.78 _99.99 High Yes freq. 1 7 8 (Closed) pct. 3.03 21.21 24.24 No freq. 4 21 25 pct. 12.12 63.63 75.75 Total freq. 5 28 33 g_ pct. 15.15 84.84 99,99 Total Yes freq. 6 29 35 (Across pct. 6.06 29.29 35.35 All Three Groups) No freq. 12 52 64 pct. 12.12 52.52 64.64 Total freq. 18 81 99 pct. 18.18 81.81 99.99 chi square for Total = 6.244 (df = l), p-value less than 105 119 disbelief sources (18%). The observed differences for these characteristics for the three dogmatism groups seemed to be all well within chance expectations. At least two more kinds of comparisonseae suggested by this classification of the individual factors. First, to what extent is each of the Berlo factors represented in the indivi- dual solutions? Second, to what extent does each factor in the individual solutions represent each Berlo factor in the same order of magnitude of variance accounted for (i.e., the largest represent safety, the middle-sized represent quali— fication, the smallest represent dynamism)? The results are summarized in Table 46 through 52. The differences found follow the same sort of pattern thus far found. What re- liable differences there are can be accounted for by the source categories, with none accounted for by the dogmatism groups. Extent Safety Factor is Represented in Individual Solutions-- Across all 198 factor solutions, the Berlo safety factor was represented in 155 (78%). This pattern was fairly general throughout all sub-groups of solutions (see Table 47). For slightly more than half of all solutions (54%), the largest factor was classifiable as safety. Again, the overall pattern was general for all sub- groups of solutions (see Table 48). No differential 120 tendencies were found for the way the safety factor was represented in the sub—groups of individual solutions. As will be clear from the following descriptions, the safety factor will have the greatest representation of the three Berlo factors among the individual factors. Extent Qualification Factor is Represented in Indivi- dual Solutions-- On an overall basis, the qualification factor is represented in the individual solutions in a little more than half of the 198 solutions (54%) irreSpective of the amount of variance accounted for. When size of the individual solution factor is considered, qualification is represented by the middle-sized factor in slightly more than a third of all solutions (37%). The solutions for belief and disbelief sources differ on both characteristics. Belief source solutions are more likely to have a factor representative of the Berlo qualification dimension (63%) than are the dis- belief source solutions (45%) (see Table 49). Similarly, the factor representative of qualification is more likely to be the middle-sized one for belief sources (54%) than it is for disbelief sources (28%) (see Table 50). Extent Dynamism Factor is Represented in Individual Solutions-- In terms of representation in all solutions, the dynamism factor is tied to qualification (54%). Dynamism is representative of the smallest individual factor in about a 121 Table 47. Extent to Which One of the Factors in the Three ‘ Factor Rotated Solutions of the Individual Analyses Could be Classified as the Safety Factor Disbelief Sources Have Belief Sources Safety Factor? Have Safety Dogmatism Factor? Yes No Total Low Yes freq. l9 8 27 (Open) pct. 57.57 24.24 81.81 No freq. ’4 ‘ '2 " “ ‘6'" .I: pct. 12.12 6.06 18.18 Total freq. 23 10 33 pct. 69.69 30.30 99.99 Medium Yes freq. 21 3 24 pct. 63.63 9.09 72.72 No freq. 8 l 9 pct. 24.24 3.03 27.27 Total freq. 29 4 33 pct. 87.87 12.12 99.99 High Yes freq. 2O 7 27 (Closed) pct. 60.60 21.21 81.81 No freq. 5 1 6 v pct. 15.15 3.03 18.18 Total freq. 25 8 33 pct. 75.75 24.24 _99.99 Total Yes freq. 6O 18 £8 pct. 60.60 18.18 7 .78 No freq. l7 4 21 pct. 17.17 4.04 21.21 Total freq. 77 22 99 pct. 77.77 22.22 99-99 Table 48. 122 Extent to Which the Largest Factor in the Three Factor Rotated Solutions of the Individual Analyses Could be Classified as the Safety Factor Largest Factor for Disbelief Sources Class- ified as Safety? Largest Factor for Belief Sources Classified as Dogmatism Safety? Yes No Total Low (Open) Yes freq. 9 10 19 pct. 27.27 30.30 57.57 No freq. 7 7 14 pct. 21.21 21.21 42.42 Total freq. 16 17 33 pct. 48.48 51.51 99.99 Medium Yes freq. 9 11 20 pct. 27.27 33.33 60.60 No freq. 6 7 13 pct. 18.18 21.21 39.39 Total freq. l5 18 33 __ pct. 45,45 54.54 99.99 High Yes freq. 8 10 18 (Closed) pct. 24.24 30.30 54.54 No freq. lO 5 15 pct. 30.30 15.15 45.45 Total freq. 18 15 33 pct. 54.54 _45.45 99399. Total Yes freq. 26 31 57 (Across) pct. 26.26 31.31 57.57 No freq. 23 19 42 pct. 23.23 19.19 42.42 Total freq. 49 5O 99 pct. 49.49 50.50 99.99 123 Table 49. Extent to Which One of the Factors in the Three Factor Rotated Solutions of the Individual Analyses Could be Classified as the Qualification Factor ' Disbelief Sources Have Belief Sources Qualification Factor Have Qualification Dogmatism Factor? Yes No Total Low Yes freq. 7 14 21 (Open) pct. 21.21 42.42 63.63 No freq. 9 3 12 pct. 27.27 9.09 36.36 Total freq. l6 17 33 pct. 48.48 51.51 99.99 Medium Yes freq. 10 12 22 pct. 30.30 36.36 66.66 No freq. 6 5 11 pct. 18.18 15.15 33.33 Total freq. l6 17 33 ,pct. 48.48 51.51 99.99_ High Yes freq. 6 l3 19 (Closed) pct. 18.18 39.39 57.57 No freq. 7 7 14 pct. 21.21 21.21 42.42 Total freq. 13 2O 33 ___ PCt- 39-39 60-50 99-99 Total . Yes freq. 23 39 62 (Across pct. 23.23 39.39 62.62 all,Three Groups) No freq. 22 15 37 pct. 22.22 15.15 37.37 Total freq. 45 54 99 pct. 45.45 54.54 99.99 Chi square for Total = 3.879 = l), p-value less than .05 Table 124 50. Extent to Which the Middle Sized Factor in the Three Factor Rotated Solutions of the Individual Analyses Could be Classified as the Qualification Factor Middle Factor for Belief Sources Middle Factor for Disbelief Sources Class- ified as Qualification? Classified as Yes No Total Dogmatism Qualification? . Low Yes freq. 2. ~13 * 15 (Open) pct. 6.06 39.39 45.45 No freq. 6 l2 18 pct. 18.18 36.36 54.54 Total freq. 8 25 33 _pct. 24.24 75.75 99199 Medium Yes freq. 6 l2 18 pct. 18.18 36.36 54.54 No freq. 4 ll 15 pct. 12.12 33.33 45.45 Total freq. 10 23 33 PCt- 30-30 59-69 99-991_ High Yes freq. 2 10 12 (Closed) pct. 6.06 30.30 36.36 No freq. 8 13 21 pct. 24.24 39.39 63.63 Total freq. 10 23 33 ,pct. 30.30 69.69 99.99 Total Yes freq. 10 35 45 (Across pct. 10.10 35.35 45.45 All Three Groups) No freq. 18 36 54 pct. 18.18 36.36 54.54 Total freq. 28 71 99 pct. 28.28 71.71 99.99 chi square for Total = 4.830 (df 1), p-value less than .05 125 Table 51. Extent to Which One of the Factors in the Three Factor Rotated Solutions of the Individual Analyses Could be Classified as the Dynamism Factor DisEelief Sources Have Belief Sources Dynamism Factor Have Dynamism chi square for Total = 15.373 (df = l), p-value less than .05 Dogmatism Factor? Yes No Total :Low Yes freq. ll 12 23% (Open) pct. 33.33 36.36 69.69 No freq. 5 5 10 pct. 15.15 15.15 30.30 Total freq. l6 17 33 ___ pct. 48.48 51.51 399.99 Medium _Yes freq. 11 147 25 pct. 33.33 42.42 75.75 No freq. 4 4 8 pct. 12.12 12.12 24.24 Total freq. 15 18 33 1.. pct. 45.45 54.54 9939 High Yes freq. 6 14 20 (Closed) pct. 18.18 42.42 60.60 No freq. 2 ll 13 PCt- 6-06 33-33 39.39 Total freq. 8 25 33 fi_ ppct. 24.24 75.75 99.99 Total Yes freq. 28 4O 68 pct. 28.28 40.40 68.68 No freq. 11 2O 31 pct. 11.11 20.20 31.31 Total freq. 39 6O 99 pct. 39.39 60.60 99.99 Table 52. Extent to Which the Smallest Factor in the Three Factor Rotated Solutions of the Individual Analyses Could be Classified as the Dynamism Factor Smallest Factor for Smallest Factor for Disbelief Sources Class- Belief Sources ified as Dynamism Classified as '— Dogmatism Dynamism Yes No Total Low Yes freq. 3 12 15 (Open) pct. 9.09 36.36 45.45 No freq. 3 15 18 pct. 9.09 45.45 51.51 Total freq. 6 27 33 pct. 18.18 81.81 99.99 Medium Yes freq. 4 l5 19 pct. 12.12 45.45 57.57 No freq. 5 9 14 pct. 15.15 27.27 42.42 Total freq. 9 24 33 pct- 27.27 72.72 99-991 High Yes freq. 4 l2 16 (Closed) pct. 12.12 36.36 48.48 No freq. 1 16 17 pct. 3.03 48.48 51.51 _.Tota.l ..fres-_., ....5. .928 33 __ g __ pct. 15.15 84.84 99.99, Total Yes freq. 11 39 5O (Across pct. 11.11 39.39 50.50 All Three Groups) No freq. 9 4O 49 pct. 9.09 40.40 49.49 Total freq. 2O 79 99 pct. 20.20 79.79 99.99 chi square for Total 17.571 (df = l), p-value less than .05 127 third of the solutions (35%). The general pattern of the sub-group differences for dynamism is about the same as it was for qualification. The solutions for belief sources are more likely than the solutions for disbelief sources not only to have one of the factors representative of the Berlo dynamism dimension (69% vs. 39%) but also for it to be the smallest one (50% vs. 20%) (see Tables 51 and 52). Magging of Qualification and Dynamism in Individual Solutions-- These findings show that on an overall basis, safety was more likely to be represented than either quali— fication or dynamism. When this tendency is examined for the two kinds of sources separately, the tendency still holds but is much more pronounced for the disbelief sources than for the belief sources. Can these differences be accounted for in terms of Berlo dimensions? Until now, the discussion has been only in terms of when an individual factor was representative of a single Berlo dimension. However, in Table 43, it can be noted that in 61 of the 198 individual solutions (31%) there was one factor which was representa- tive by our definition of two or more Berlo factors. Nearly three-fourths of such cases (44) involved a combination of the qualification and dynamism dimensions. Table 53 examines the degree to which the various sub-groups of solutions exhibit 128 F the tendency to merge Berlo's qualification and dynamism factors. As is obvious from the table, the overwhelming tendency is for such merging to occur for the disbelief sources (36%) rather than for the belief sources (8%). If the four pairs of solutions for which the merging is ex- hibited for both the belief and disbelief sources are ignored, only four solutions merged the two Berlo factors for the belief and not for the disbelief sources, whereas for thirty- two solutions, the tendency was in the reverse direction. Merging the Berlo factors for the disbelief but not for the belief sources was eight times more likely with this sample than was the opposite. The merging of the two dimensions occurs in the solution sub-groups inversely with the degree to which the two dimensions are represented separately by individual factors. Brief Summary of Findings Four hypotheses were tested in this study. Two hy- potheses concerned the differences in the dimensions of source evaluation which would be employed for two kinds of isources. The two kinds of sources included those judged to have beliefs and ideas similar to the subjects' own beliefs and ideas and those judged to have beliefs and ideas different from their own. The Specific Operationalizations used in the Table 53. Extent to Which One of the Factors in the Three Factor Rotated Solutions of the Individual Analyses Could be Characterized as a Merging of the Qualification and Dynamism Factors Merging—Present for Disbelief Sources? Merging Present for Dogmatism Belief Sources? Yes No Total Low Yes freq. 2 l 3 (Open) pct. 6.06 3.03 9.09 No freq. 9 21 30 pct. 27.27 63.63 90.90 Total freq. ll 22 33 pct. g33.33 66.66 99.99 Medium Yes freq. 1 l 2 pct. 3.03 3.03 6.06 No freq. 8 23 31 pct. 24.24 69.69 93.93 Total freq. 9 24 33 pct. 27.27 72.72 99.99 Low Yes freq. 1 2 3 (Closed) pct. 3.03 6.06 9.09 No freq. 15 15 30 pct. 45.45 45.45 90.90 Total freq. l6 17 33 ___ pct. 48.48 <_51.51 99.99 Total Yes freq. 4 4 8 (Across pct. 4.04 4.04 8.08 All Three Groups) No freq. 32 59 91 pct. 32.32 59.59 91.91 Total freq. 36 63 99 pct. 36.36 63.63 99.99 chi square for Total = 20.250 (df = l), p-value less than .05 130 tests were designed to index the number of dimensions of individual source evaluation, the degree of differentiation of them, and the degree to which they corresponded to the Berlo results. The predictions were that the number and differentiation of the individual dimensions of source evaluation would be greater for belief sources than for dis- belief sources. It was also predicted that there would be differences between the two kinds of sources in the degree to which the individual dimensions of source evaluation would correspond to the Berlo dimensions of source evaluation. These hypotheses were generally supported (Hypotheses I and III). Two other hypotheses were concerned with the relationship between dogmatism and the individual dimensions of source evaluation. These were concerned with the same characteristics of the dimensions of source evaluation for the same two kinds of sources as were the other two hypotheses. These hypotheses were generally not supported (Hypothesis II and IV). Little can.be inferred from the results of this study about the re- lationship between dogmatism and the individual dimensions of source evaluation. CHAPTER IV DISCUSSION AND CONCLUSIONS Summary of Findings Hypothesis I-- Hypothesis I predicted that the in- dividual factor solutions for the belief sources would have a greater number of dimensions of source evaluation and would exhibit greater differentiation than would the factor solutions for the disbelief sources. Table 54 summarizes the findings for Hypothesis I for the eleven operationalizations of the dependent variable, "number and/or differentiation of dimensions of source evaluation or credibility." Results for eight of the eleven indices were statistically significant and in the direction predicted by Hypothesis I. The conclusion is that the results of this stidy do generally support Hypothesis I. Individual subjects' dimensions of source evaluation (credibility) for sources judged as having beliefs similar to the subjects' own beliefs are more numerous and more greatly differentiated than are the correSponding dimensions for sources judged not to have 'belief'similar to the subjects' own. Hypothesis II-- Hypothesis II predicted that the difference between the number of dimensions and their degree 131 132 .‘J‘ r -’ Table 54. Summary of Findings Related to Hypotheses I and II F" Hypothesis I “If Belief (B) vs. Dis- Hypothesis II Dependent Variable belief(D)_Sources Low (L) vs. NUMBER OF IMENSION High (H) Dogmatism 1 Number of factors in B greater than D not significant rotated solution meeting Kiel-Wrigley criterion set at three 2 Proportion of variance not significant not significant accounted for by one factor solution 3 Proportion of var— B less than D i not significant iance accounted for by two factor solution 4 Proportion of variance B less than D not significant accounted for by three factor.solution 5 Proportion of B less than D not significant variance accounted for by four factor solution DIFFERENTIATION 0F DIMENSIONS 6 PrOportion of total ndzsignificant not significant variance accounted for by largest factor in two rotated solution 7 Proportion of total not significant not significant variance accounted for by largest factor in three factor rotated solution 8 Proportion of total B less than D not significant variance accounted for ‘ by largest factor in four factor rotated solution 9 Simple structure index B less than D L less than H or proportion of common variance accounted for by largest loadings in two factor rotated solution 13 Table 54--continued 10 Simple structure in- dex or proportion of common variance accounted for by largest loadings i 3 B less than D n three factor rotated solution 11 Simple structure index or proportion of common variance accounted for by largest loadings in four factor rotated solution B less than D not significant not significant v 134 of differentiation for belief sources and for disbelief sources would be greater for low dogmatics (open system) than for high dogmatics (closed systems). Table 54 summarizes the findings related to this hypothesis fbr each of the eleven separate operationalizations. The results support the hypothesis in only one of the eleven tests. The con- clusion is that the results of this study fail to support Hypothesis II. Consequently, nothing can be inferred from this study about the differences between high and low dog- matics regarding the individual dimensions of source evalua- tion for sources either judged to have beliefs similar to the subjects' own or judged to have beliefs which are different. Hypothesis III-- The primary focus of the analysis for Hypothesis III was on the degree to which individual dimen- sions or source evaluation correSponded to the results ob- tained by Berlo for the two kinds of sources (belief vs. disbelief). The hypothesis stated only that the nature or kind of dimensions of source evaluation for belief source inculd differ from those of disbelief sources. The hypothesis was offered to facilitate the comparisons of the three factor rotated individual solutions with the overall Berlo three factor analysis solution. Sixteen more or less systematic dichotomous variables were developed in hopes of 135 getting at aspects of the degree of correspondence. Table 55 summarizes the findings related to this hypothesis for these variables. Results for thirteen of the sixteen indices Show statistically significant differences between the two kinds of sources. The conclusion is that Hypothesis III is supported by the results of this study. The dimensions of source evaluation and their degree of correspondence with the I3erlo dimensions for sources judged to have beliefs similar 130 the subjects' own beliefs do differ from the dimensions fiax'sourcesjudgedix>hmye beliefs not similar to the subjects' cywn. On an overall basis, the dimensions for the belief sources correspond more to the Berlo dimensions than do the Climensions for the disbelief sources. On a factor to factor (dimension to dimension) basis generally, no differences were :found for the Berlo safety factor. The differences come into Ifilay for the qualification and dynamism factors. Belief amlrces exhibit a greater tendency for these factors to be pzwasent than do the disbelief sources. Finally, there was cflnserved what can be termed a merging of the Berlo qualifi- catixn1ahd dynamism factors into a single dimension. The teruiency to do this was much greater for the disbelief sources than for the belief sources. Hypothesis IV-- Hypothesis Iv mimored Hypothesis III. Instead of exploring the differences between source categories, 136 Table 55. Summary of Findings Related to Hypothesis III and Iv Hypothesis III Hypothesis IV Dichotomous Characteristic Belief (B) vs. Low (L) vs. of Individual Three Factor Disbelief (D) High (H) Dogmatism Rotated Solution Spurces 1 Degree Kiel-Wrigley B less than D not significant criterion solution is three factor solution Number of different B greater than D L greater than H factors into which the (for disbelief three Berlo factors map sources only) using index of factorial ‘ similarity Degree three factor B greater than D L less than H solutions are judged an (both B & D adequate match with Berlo judged inadequate) factors using index of fac. Sim. Degree Berlo safety factor B greater than D L greater than H maps into largest factor (both B & D map using index of factorial into largest) similarity Degree Berlo qualifi- not significant not significant cation factor maps into middle-sized factor using index of factorial similarity Degree Berlo dynamism B greater than D not significant factor maps into smallest factor using index of factorial similarity Number of factors repre- B greater than D not significant senting Berlo dimensions or factors Degree all three Berlo B greater than D not Significant factors are represented in three factor solution Degree all three Berlo B greater than D not Signigicant factors are represented in three factor sulution with: safety, largest; qualification, next; dynamism, smallest 137 Table 55--continued 10 11 12 13 l4 15 16 Degree safety factor not significant is represented by any one of the factors Degree safety factor is not significant represented by largest factor Degree qualification B greater factor is represented by any one of the factors Degree qualification B greater factor is represented by mildle-sized factor Degree dynamism factor B greater is represented by any one of the factors Degree dynamism factor B greater is represented by smallest factor than D than D than D than D Degree one of factors B less than D represents merging of qualification and dynamism dimensions not not not not not not not significant significant significant significant significant significant significant 138 however, IV explored the differences between groups of subjects, Specifically, Hypothesis Iv predicted that high and low dogmatics would differ in the nature or kind of dimensions of source evaluation used and in the degree of correspondence of these dimensions to the Berlo dimensions. The same set of variables was used for both hypotheses (III and IV). The summary of findings is contained in Table 55. Statistically significant differences were found for only three of the sixteen variables. Also, these differences are difficult to interpret. The conclusion is that Hypothesis IV is not generally supported by the results of this study. Consequently, nothing can be inferred from this study about the differences between high and low dogmatics regarding the nature or kind of individual dimensions of source evaluation and their degree of correSpondence to the dimensions of source evaluation found by Berlo. Rokeach's Belief System Theory Two hypotheses for this study (I and II) were derived directly from notions expressed in Rokeach's belief system theory. These hypotheses were concerned primarily with structural aspects of the dimensions of source evaluation as they relate to the structure of the overall belief- disbelief systems. One of the basic notions in Rokeach's “—— W _ - . PM’S El 139 theory is that there is variation in the structure of all pe0ples' belief—disbelief systems. There is greater differentiation of the structure of sub-systems of the dis- belief system the more closely the sub-system correspond to the individual's own belief system. Rokeach talks about this aSpect in terms of the location of the various sub- systems along the belief-disbelief continuum. In this study, two groups of communication sources were characterized by their location along the belief-disbelief continuum. Hypothesis I predicted that the individual dimensions of snurce evaluation for the sources located near the belief end of the belief-disbelief continuum would be structurally more differentiated than would the dimensions for the sources located near the disbelief end. This hypothesis, as in- dicated, was generally supported. Hence, one of the more laasic notions which Rokeach draws upon in building his higher level theory was supported by the results of this study. In developing the defining Characteristics of the open and closed mind, or dogmatism, Rokeach draws upon the general characteristic that structural differentiation decreases in the sub-systems of the belief—disbelief systems the closer they are located to the disbelief end of the belief-disbelief continuum. For the closed mind or high dogmatic, this decrease in differentiation will be much more pronounced than it will 140 be for the open mind or low dogmatic. Hypothesis 11 made this prediction about the relationship between the individual dimensions of source evaluation and dogmatism. The results of this study failed to support this hypothesis and hence Rokeach's more general notions. Speculation as why this happened is contained in the next section. Generality of the Berlo Dimensions of Source Evaluation Across Subjects Hypotheses II and IV were concerned primarily with the generality of the dimensions of source evaluation across subjects. The results failed to support both hypotheses. Hence, little can reasonably be inferred. From a statistical standpoint, failure to detect differences could have been primarily due to one of two situations. Either the variance of the dependent measures over all observations was so small that virtually any sub-grouping of subjects would preclude the finding of reliable differences or the observed variance was large enough but it was clearly not distributed to produce reliable differences for the categories employed in the study. The latter case obtained in this study. The dogmatism (categories failed to account for the variance. The between subjects variance for the dependent measures included in this stmdy will have to await future research to be explained. 141 Speculation as to why the expected differences did not emerge can be offered from at least two bases. It is possible that the revised scale as used in this study failed to operationalize dogmatism in a way consistent with Rokeach's theory for some reason with this particular set of subjects. The part of the theory which was confirmed in this study was not dependent on the measure of dogmatism employed. This would suggest that this study's application of the theory had :nme degree of validity. The other basis relates to the degree to which the stimuli the subjects were asked to, in effect, classify can be considered to have high syndrome relevance. White, Alter, and Radin (1965) did observe differentiation differences in the conceptual categories employed by high and low dogmatics in the classification of highly syndrome-relevant stimuli. With stimuli of low syndrome relevance, no differences were observed. It is possible that the evaluations of sources required in this study were sufficiently Specific and common across people that they cannot be considered to have high syndrome relevance to dogmatism. The same may be said of the situation implicit in each source description in which any meaningful evaluation could occur. Source evaluation can be considered as a highly Specific aSpect of the general notion of person perception (Tagiuri and Petrullo, 1958; Brunner and Tagiuri, 1954). Possibly Rokeach's dogmatism 142 theory may be more useful in explaining differences in the general way people perceive other people rather than in some highly detailed Specific aSpect such as source credibility. Attention had been given to syndrome relevance of the stimuli and it was concluded that it was high. It is possible this assumption was in error. In the secondary analysis of the Wozniak data, differences were found which related to dogmatism. None were found for this study. Why? Another possible explanation might lie in the difference between the extreme dogmatism groups. In this study, the mean dogmatism score difference between the high and low groups was about 20 points on a 20 item scale. In the analysis of the Wozniak data, the difference between the high and low groups was nearly 80 points on a 40 item scale. The two groups were nearly twice as extreme in their mean dogmatism for the Wozniak analysis as they were for this study. This could be another reason WHY no differences were found. In summary, nothing can be inferred from the results Of this study about the generality of the Berlo dimensions of Source evaluation or credibility across people. ——.,_ -4 A.) -_. . 143 Generality of the Berlo Dimensions of Source Evaluations Across Sources Hypotheses I and III were concerned primarily with the generality of the dimensions of source evaluation across sources. The results generally support both hypotheses. Differences were observed in the individual dimensions of source evaluation between the two kinds of sources used for this study. Belief sources tended to exhibit dimensions of source evaluation which more closely corresponded to the Berlo dimensions than did disbelief sources. When specific factors are examined individually, this difference between the two kinds of sources is exhibited only for the qualification and dynamism dimensions. Source differences for the safety dimension were minimal. Only about half of the belief sources' solutions could lie characterized as having factors representative of all three Berlo dimensions. For the disbelief sources, this dropped to less than a third. On this kind of overall basis, the gen- erality of the Berlo dimensions to the individual dimensions fbr the belief sources leaves something to be desired. General- ity is even less for the disbelief sources. When concerned with some kind of multi-dimensional evaluation<1f sources, one must take into account the kinds of sources being evaluated. One cannot assume that the same pattern of dimensions will cutacross all sources. .13“? 144 If one is willing to settle for something less and is content with evaluations along a Specific dimensions, the generality outlook can be improved. Slightly more than ninety per cent of all individual three factor solutions had at least one factor representative of one of the Berlo dimen- sions. This was true for both kinds of sources. Confined Specifically to the safety factor, nearly eighty per cent of all solutions had a factor representative of this dimension. This held for both kinds of sources. The Berlo safety factor, then, is quite general across the individual dimensions of source evaluation across both kinds of sources. The generality of the qualification and dynamism factors approaches this generality for belief sources only. About two-thirds of the solutions for belief sources have a factor representative of “the Berlo qualification factor. The same is true for the dynamism dimension. The correSponding figures for the dis- t>elief sources for the two dimensions are Slightly less than half of the solutions for qualification and even less (about 40%) for dynamism. The generality here leaves much to be (tsired. The Berlo results Showed the safety factor to be the largest, qualification to be the next in size, with dynamism being the smallest. If this further restriction is taken into account when assessing the generality, it will be 145 lowered for all the cases discussed above. However, one 'probably would not want to get this restrictive for most purposes. In final summary, the generality of the safety dimension is the one which approaches an acceptable level for most conceivable purposes. At least, it does for the kinds of sources included in this study. The same is probably true for the qualification and dynamism dimensions but for belief sources only. For the latter two factors on the disbelief sources, the generality is lower. Generality is low if there must be representatives of all three Berlo factors among the individual dimensions of source evaluation. This is true for both kinds of sources, except that for disbelief sources, it is even less likely that all three Berlo dimensions will be represented.* *For additional comments on the generality of the Berlo dimensions across subjects see Appendix I. 146 Suggestions for Future Research One obvious area for future research is in the matter of the correlates of the differences in the individual dimensions of source evaluation. Dogmatism was offered as one of the correlates in this study. However, no reliable relation- Ehip was found between dogmatism and variation in the individual dimensions of source evaluation. Further research is needed to isolate some of the correlates. Linear correlation measures and factor analysis were used to impose structure on the subjects' evaluations of the snurces using the semantic differential bi-polar adjective scales. Other techniques are available for imposing structure on source evaluation data. McQuitty (1964, 1966) offers a whole series of clustering techniques using non-linear indices of association. A number of these techniques provide a means of characterizing the structure in a hierarchical manner. Hays (1958) and Kelly (1955) offer techniques which can briefly be described as non-metric factor analyses. Other techniques for imposing structure on source evaluation data sshould be explored. It is quite possible that one or more of these alternative techniques for characterizing structure may prove to be highly useful or theoretically meaningful in the study of the dimensions of source evaluation. Characterizing the dimensions of source evaluation in a hierarchical manner has a certain appeal. P: tnfiflnn‘ 147 The type of factor analysis used in this study and by Berlo yielded orthogonal or uncorrelated factors. Does this imply that the dimensions thus isolated are psychologically independent? It is not difficult to argue that evaluations of at least some kinds of sources along a safety dimension should be curvilinearly related to evaluations along a dynamism dimension. For example, consider two groups of important national or international political leaders, one rated high in safety (low threat) and the other low in safety (high threat). It would seem quite reasonable that both groups would be rated high in dynamism (highly active). Factor analysis need not necessarily produce uncorrelated factors. There are a number of oblique rotational models available which yield correlated factors. This would be useful if the dependency among factors was linear. McQuitty's non-linear techniques might be useful for exploring dependencies of a non-linear nature. Further work needs to be done in exploring the "psychological de- pendence or independence" of the dimensions of source evalua- tion. To what extent does information about the ratings of various kinds of sources on one dimension imply information about the rating they have on one or more of the other dimen- sions? FIN." 91. 148 Suppose that people do differ reliably in the nature or kind of dimensions of source evaluation. Can typologies of the dimensions of source evaluation be developed and characterized? Is the variation of the dimensions of source evaluation Spread evenly over a wide range? Or if typologies are to be developed, is the variation Spread over a wide range but with a few areas of high concentration with most other areas of rather low concentration? If typologies are possible, can they be indexed economically? In other words, what are the correlates of the typologies? If typologies can be developed, there is the obvious implication that source influence in the communication process would differ among them. Would this mean that different communication strategies would have to be developed for the various typologies for maxi- mum communicative effectiveness? Further research is needed to explore such questions. There is at least one important area which seems to 138 begging for further research. This is consideration of 'the dimensions of source evaluation as the dependent outcome of manipulation of elements in the message or communication situation. If source evaluation or credibility is to become a viable concqpt, it must be more than something a source statistically carries with him through a communication situation. Theory and research must be developed which will 149 map changes in manipulatable elements of the message or communication situation into changes in the value or structure of the dimensions of source evaluation. Such research is con- tingent on the development of useful and reliable means of characterizing the dimensions of source evaluation. It is hoped this study has laid some of the groundwork for achieving this objective. However, much more needs to be done. BIBLIOGRAPHY Anderson, K., and Clevenger, T., Jr., A Summary of Ex- perimental Research in Ethos. Speech Monographs, 1963, 39. 59-78. Berlo, D. K., The Process of Communication. New York: Holt, 1960. Berlo, D. K., Attitude Change Abstracts. Unpublished research, Department of Communication, Michigan State University, 1967. Berlo, D. K., Lemert, J. B., and Mertz, R. J., Dimensions for Evaluating_the Acceptability of Message Sources. Report of research supported under Contract #OCD-PS-647l, Office of Civil Defense, Department of Defense, as part of Michigan State University's continuing research program on the public acceptance of civil defense messages, 1966. Brown, R. W., Is Boulder Sweet or Sour?, Contemprapy Psychology, 1958, 3, 113-115. Brunner, J. S., and Tagirur, R., The Perception of People. In G. Lindzey (ed.), Handbook of Social Psychology, Vol. II. Reading, Mass.: Addison-Wesley, 1954, 634-654. Carroll, J. B., Review of Measurement of Meaning. Language, 19593 :2: 58'77 . Cattell, R. B., Factor Analysis. New York: Harper, 1952. Cohen, A. R., Attitude Change and Social Influence. New York: Basic Books, 1964. Creaser, J. W., An Aid in Calculating Q-Sort Factor-arrays. ‘Jgurnal of Clinical Psychology, 1955, 11, 195-196. Darnell, D. K., Concept Scale Interaction in the Semantic Differential, Journal of Communication, 1966, 16, 104-115. 150 151 Guilford, J. P., Psychometric Methods. New York: McGraw- Hill, 1954. Gulliksen, H., How to Make Meaning More Meaningful, Contemporary Psychology, 1958, 3, 115-119. Harmon, H. H., Modern Factor Analysis. Chicago: University of Chicago Press, 1960. Hays, W. L., An Approach to the Study of Trait Implication and Trait Similarity. In R. Tagiuri and L. Petrullo (Eds.), Person Perception and Intertpersonal Behavior. Stanford:IStanford'UniversityIPress, 1958. pp. 289—299. Horst, P., Factor Analysis of Data Matrices. New York: Holt,Il965. Hovland, C. I., Janis, I. J., and Kelly, H. H., Communi- cation and Persuasion. New Haven: Yale University Press, 1953. Kelly, G. A., The Psychology of Personal Constructs. New York: Norton, 1955, Vol. I. Kiel, D. F., Effects upon the Factorial Solution of Rotating varying Numbers of Factors with Differing Initial Communality Estimates. Unpublished master's thesis, Michigan State University, 1966. Kiel, D. F., and Wrigley, 0., Effects upon the Factorial Solutions of Rotating varying Numbers of Factors, American Psychologist, 1960, i3, 487-488. (Abstract) Klapper, J. T., The Effects of Mass Communication. Glencoe: Free Press, 1960. Lemert, J. B., Dimensions of Source Credibility. Paper presented to the A.E.J. ConventiOn, Lincoln, Nebraska, 1963. Lindquist, E. F., Design and Analysis of Experiments in Psychology and Education. Boston: Houghton Mifflin, 1953- McCroskey, J. C., Scales for Measurement of Ethos. Speech Monographs, 1966, 33, 65-72. m‘a Fimgrfif 152 MacLean, M. 8., Jr., Danbury, T., and Talbott, A. D., Civil Defense Belief Patterns:(ll) Source Credibility. A report prepared for the Office of C1v11 Defense, Department of Defense by the Department of Communi- cation, College of Communication Arts, Michi an State University, East Lansing, Michigan, April 19 3. MacLean, M. 8., Jr., Danbury, T., and Talbott, A. D., Civil Defense Belief Patterns: (VIII) Technical Summa . A report prepared for the Office of C1v11 Defense, Department of Defense by the Department of Communication, College of Communication Arts, Michigan State University, East Lansing, Michigan, March 1964. McQuitty, L. L., A Statistical Method for Studying Per— sonality Integration. In 0. H. Mowrer (Ed.), Psychotherapy Theory and Research. New York: Ronald Press, 1953, pp. 414-462. McQuitty, L. L., Capabilities and Improvements of Linkage Analysis as a Clustering Method. Educational and Psychology Measurement, 1964, 24, 441-456. McQuitty, L. L., Single and Multiple Hierarchical Classifi— cation by Reciprocal Pairs and Rank Order Types. Educational and Psychology Measurement, 1966, 26, 253-265. Osgood, C. E., Studies on the Generality of Affective Meanéng Systems. American Psychologist, 1962, 31, 10-2 . Osgood, C. E., Suci, G. J., and Tannenbaum, P. H., The Measurement of Meaning. Urbana: University of IlIinoIS’Pfess, 1957. Powell, F. A., Open- and Closed-Mindedness and the Ability to Differentiate Source and Message. Journal of Abnormal Social Psychology, 1962, 63, 61—64. Rokeach, M. J., The Open and Closed Mind. New York: Basic Books, 1960. Siegel, S., Nonparametric Statistics. New York: McGraw- Hill, 1956. 153 Stephenson, W. J., The Study of Behavior, Chicago: University of Chicago Press,'1953. Tagiuri, R., and Petrullo, L. (eds.), Person Perception and Inter:personal Behavior. Stanford: Stanford University Press, 1958} Talbott, A. D., Indexing Belief-Disbelief Systems. Un— published research, 1964. Thomson, G., The Factorial Analysis of Human Ability. London: University of Lendon Press, 1939. Thurstone, L. L., Multiple—Factor Analysis. Chicago: University of Chicago—Press, 1947. Todd, F. J., and Rappoport, L. A., Cognitive Structure Approach to Person Perception. Journal of Abnormal Social Psychology, 1964, 66, 469-478. Troldahl, V. C., and Powell, F. A., A Short Form Dogmatism Scale for Use in Field Studies. Social Forces, 1965, ii: 211-214. Ware, E. E., Relationships of Intelligence and Sex to Diversity of Individual Semantic Meaning Space. Unpublished doctoral dissertation, University of Illinois, 1958. ‘White, B. J., Alter, R. D., and Radin, M., Authoritarianism, Dogmatism and Usage of Conceptual Categories. Journal of Personality and Social Psychology, 1965, 3, 293-295. Winer, B. J., Statistical Principles in Experimental Design. New York: McGraw-Hill, 1962. 'Wozniak, D. F., A Factor Analytic Study of Semantic Structures of Closed, Qpen, and Medium Belief-Disbelief Systems. Unpublished doctoral dissertation, Michigan State University, 1963. 154 APPENDIX A Pretest Questionnaire Used 155, Part I of Phase 01 INSTRUCTIONS Number HAS BELIEFS AND YCu are participating in a study of how people evaluate other persons. In this part of the study, you will be asked to look at the brief description of a number of persons and then to rate each on the degree to which each person has beliefs and ideas which are like your own or are different from your own. Below are a couple examples of the kinds of descriptions of persons you will be asked to look at: YOUR FAVORITE LOCAL TV NEWSCASTER MAYOR OF THE CITY OF EAST LANSING, GORDON THOMAS You are to use the ladder scale on the right of each page to rate each person. Please examine this scale. You are to place each person somewhere along the scale according to which position best describes how much that person has beliefs and ideas which are either like your own or are different from your own. After you have decided what position best describes your feelings about the person described, then you would put the number of that position in the blank space to the left of the description. For example, if you thought that "YOUR LOCAL TV NEWSCASTER" was best described by "HAS BELIEFS AND IDEAS QUITE A BIT IDIFFERENT FROM YOUR OWN," you would mark it as follows: 7 YOUR FAVORITE LOCAL TV NEWSCASTER 7 6 4 3 9 IDEAS EXTREMELY DIFFERENT FROM MY OWN 8 very Much Different HAS BELIEFS AND IDEAS QUITE A BIT DIFFERENT FROM MY OWN Only Slightly Different Can't Decide or Has About the Same Amount of Beliefs and Ideas Which Are Like My Own as Are Different From My Own Only Slightly Like HAS BELIEFS AND IDEAS QUITE A BIT LIKE MY OWN 156 Another example, if you thought that "MAYOR OF THE CITY OF EAST LANSING, GORDON THOMAS" was best described as having beliefs and ideas "very Much 2 Like” my own, you would mark it as follows: MAYOR OF THE CITY OF EAST LANSING GORDON THOMAS Very Much Like HAS BELIEFS AND IDEAS EXTREMELY LIKE MY OWN Before you mark the scale for the first description, please glance over all of the descriptions of persons so you will get an idea of the kinds of persons you will be asked to evaluate. Then go back to the beginning of the descriptions and work your way through without looking back and forth through the items. Do not worry over or puzzle about individual descriptions. It is your immediate feelings about each person described that we want. On the other hand, please do not be careless, because we want your true impressions. Also, please write the numbers in the Spaces as clearly and legibly as you can. questions, please begin. If you have no 157 CONVICTED KILLER OF LEE HARVEY OSWALD, JACK RUBY MEMBER OF THE COMMUNIST PARTY IN ENGLAND SECRETARY GENERAL OF THE UNITED NATIONS SCOTLAND YARD AGENT IN ENGLAND HEAVYWEIGHT BOXING CHAMPION CASSIUS CLAY (MUHAMMAD ALI) THE PERSON YOU ARE LEAST LIKELY TO GO TO FOR PERSONAL ADVICE PRESIDENT OF GENERAL MOTORS THE MAYOR OF THE CITY I LIVE IN SPOKESMAN FOR THE GROUP OF MICHIGAN TEACHERS PICKETING THE STATE LEGISLATURE FOR MORE FINANCIAL SUPPORT TO EDUCATION A PERSON WHOSE OPINION YOU NEITHER RESPECT NOR DIS- RESPECT REPUBLICAN SENTATE LEADER AND SENATOR FROM ILLINOIS, EVERETT DIRKSEN PRIME MINISTER OF GREAT BRITAIN GOVERNOR OF A SOUTHERN STATE A UNITED STATES SECRET SERVICE —_II_IAGENT MY MOTHER HAS BELIEFS AND IDEAS EXTREMELY DIFFERENT FROM MY OWN Very Much Different HAS BELIEFS AND IDEAS QUITE A BIT DIFFERENT FROM MY OWN Only Slightly Different Can't Decide or Has About the Same Amount of Beliefs and Ideas Which Are Like My Own as Are Different From My Own Only Slightly Like 158 VICE-PRESIDENT OF THE UNITED STATES SYNDICATED NEWSPAPER ADVICE COLUMNIST, ANN LANDERS THE PERSON I AM MOST LIKELY TO GO TO FOR PERSONAL ADVICE MY NEXT DOOR NEIGHBOR "YANKEE GO HOME" DEMON- STRATORS IN A LATIN AMERICAN COUNTRY CONVICTED GERMAN WAR CRIMINAL AND ONE TIME ADOLF HITLER'S DEPUTY, RUDOLF HESS HAS BELIEFS AND IDEAS QUITE A BIT LIKE MY OWN very Much Like HAS BELIEFS AND IDEAS EXTREMELY LIKE MY OWN low—ll. . 159 CUBAN PRIME MINISTER HAS BELIEFS AND A TEACHER IN THE SCHOOL SYSTEM 9 IDEAS EXTREMELY WHERE I LIVE DIFFERENT FROM MY.OWN THE PERSON I KNOW WHOM I DISLIKE THE MOST NATIONAL COMMANDER OF THE AMERICAN LEGION 8 Very Much Different COMMANDER IN CHIEF OF THE RUSSIAN ARMY MEMBER OF THE COMMUNIST PARTY IN HAS BELIEFS AND 7 IDEAS QUITE A RED CHINA BIT DIFFERENT FROM MY OWN MEMBER OF THE GOVERNMENT OF FRANCE WHO IS A COMMUNIST NATIONAL PRESIDENT OF THE CHAMBER OF COMMERCE DEMOCRATIC SENATOR FROM 6 Only Slightly NEW YORK, ROBERT F. KENNEDY Different BRITISH DEMONSTRATORS WHO ARE AGAINST THE U.S. NUCLEAR Can't Decide or SUBMARINES DOCKING IN ENGLAND Has About the Same Amount of PRESIDENT OF THE UNITED STATES 5 Beliefs and Ideas Which are Like My THE POPE, SPIRITUAL LEADER OF Own as Are Different THE CATHOLIC CHURCH From My Own NUMBER ONE MAN IN NATION-WIDE CRIME SYNDICATE, COSA NOSTRA 4 Only Slightly Like GRAND DRAGON OF THE KLU KLUX KLAN <__ ‘_ COMMUNIST PARTY LEADER OF NORTH HAS BELIEFS AND 'VIETNAM 3 IDEAS QUITE A BIT LIKE MY OWN 160 COMMANDER OF THE TROOPS IN NORTH ATLANTIC TREATY ORGANI- ZATION (NATO, OUR ALLIES IN ___EUROPE) DICTATOR OF A LATIN AMERICAN COUNTRY PRESIDENT OF FRANCE A PERSON I KNOW WHO IS MAKING A CAREER OUT OF SERVICE IN THE MILITARY (ANY BRANCH) NATIONAL PRESIDENT OF THE UNITED AUTOMOBILE WORKERS, WALTER RUETHER Very Much Like HAS BELIEFS AND IDEAS EXTREMELY LIKE MY OWN .IQ‘JJ. ..I 161 LEADER OF THE AMERICAN NAZI PARTY, GEORGE LINCOLN ROCKWELL THE PERSON I MARRIED OR AM GOING TO MARRY A YOUNG MAN WHO DEMONSTRATES AGAINST THE WAR IN SOUTH VIET- NAM AND BURNS HIS DRAFT CARD GOVERNOR OF A NEIGHBORING STATE THE MEDICAL DOCTOR I MOST OFTEN GO TO A CHAPLAIN IN THE MILITARY ARCHBISHOP OF CANTERBURY, HEAD OF THE CHURCH OF ENGLAND EDITORIAL WRITER OF PRAVDA, THE RUSSIAN NATIONAL NEWSPAPER POLICE CHIEF IN A SMALL COUNTRY TOWN IN LATIN AMERICA DEMONSTRATORS AGAINST THE WAR IN SOUTH VIETNAM WHO GREETED PRESIDENT JOHNSON WHEN HE VISITED AUSTRALIA HEAD OF THE RUSSIAN SECRET POLICE FORMER SENATOR AND REPUBLICAN PRESIDENTIAL CANDIDATE, BARRY GOLDWATER EDITOR-IN-CHIEF OF THE NEW YORK HAS BELIEFS AND IDEAS EXTREMELY DIFFERENT FROM MY OWN Very Much Different HAS BELIEFS AND IDEAS QUITE A BIT DIFFERENT FROM MY OWN Only Slightly Different Can't DeCIde or Has About the Same Amount of Beliefs and Ideas Which are Like My Own as Are Different From My Own TIMES FORMER PRESIDENT OF THE UNITED STATES AND GENERAL OF THE ARMY, DWIGHT D. EISENHOWER THE POLICE CHIEF IN THE CITY WHERE I LIVE # Only Slightly Like HAS BELIEFS AND IDEAS QUITE A BIT LIKE MY OWN 162 FORMER GOVERNOR OF MICHIGAN AND DEMOCRATIC SENATORIAL ___CANDIDATE, G. MENNEN (SOAPY) WILLIAMS LOCAL RELIGIOUS LEADER OF MY FAITH MEMBER OF THE STATE LEGISLATURE FROM MY DISTRICT REPORTER ON THE NEWSPAPER I REGULARLY READ very Much Like HAS BELIEFS AND IDEAS EXTREMELY LIKE MY OWN l ""- .a 163 HEAD OF THE NATIONAL EDUCATION ASSOCIATION HAS BELIEFS AND IDEAS EXTREMELY DIFFERENT FROM AMERICAN U2 PILOT WHO WAS CONVICTED BY THE RUSSIANS OF SPYING ON 9 RUSSIA, GARY FRANCIS POWERS MY OWN MY FAVORITE LOCAL TV NEWSCASTER ,_ MY FATHER COMMANDING GENERAL OF ALL 8 very Much BRITISH MILITARY UNITS Different A PROSTITUTE HAS BELIEFS AND IDEAS QUITE A BIT DIFFERENT FROM MY OWN U.S. SECRETARY OF DEFENSE, ROBERT MCNAMARA 7 EVANGELIST, DR. BILLY GRAHAM A PERSON WHOSE OPINION YOU HAVE VERY LITTLE RESPECT FOR 6 Only Slightly MEMBER OF THE COMMUNIST PARTY Different IN RUSSIA EDITORIAL WRITER FOR THE LEADING NEWSPAPER IN FRANCE Can't Decide or ”II.- . .v n 1 I . ." i r A NEGRO CIVIL RIGHTS LEADER AND HEAD OF THE SOUTHERN CHRISTIAN LEADERSHIP CON- FERENCE, DR. MARTIN LUTHER ‘“‘KING, JR. MY BEST FRIEND CHIEF JUSTICE OF THE UNITED STATES SUPREME COURT, EARLE WARREN FORMER REPUBLICAN VICE- PRESIDENT AND PRESIDENTIAL CANDIDATE, RICHARD M. NIXON Has About the Same Amount of Beliefs and Ideas Which are Like My Own as Are Different From My Own Only Slightly Like 164 A POLICEMAN IN MY COMMUNITY REPORTER ON THE LONDON TIMES, A LEADING BRITISH NEWSPAPER U.S. SECRETARY OF STATE, ——DEAN RUSK BUDDHIST MONK IN SOUTH VIETNAM COMMANDER OF THE NORTH VIETNAMESE COMMUNIST TROOPS FIGHTING IN SOUTH VIETNAM HEAD OF A COMMUNIST FRONT ORGANIZATION IN THE UNITED STATES A STRIP-TEASE DANCER HAS BELIEFS AND IDEAS QUITE A BIT LIKE MY OWN Very Much Like HAS BELIEFS AND IDEAS EXTREMELY LIKE MY OWN I I n ‘ 165 DIRECTOR OF THE FEDERAL BUREAU OF INVESTIGATION (FBI) “"' - HAS BELIEFS AND LEADER OF THE AMERICAN 9 IDEAS EXTREMELY COMMUNIST PARTY DIFFERENT FROM MY OWN CHAIRMAN OF THE JOINT CHIEFS OF STAFF OF THE UNITED STATES MILITARY SERVICES COMMANDING GENERAL OF THE UNITED STATES MARINE CORPS 8 Very Much Different MAJORITY LEADER OF THE UNITED STATES SENATE THE POLICE CHIEF OF PARIS, ___FRANCE HAS BELIEFS AND GOVERNOR OF THE STATE OF 7 IDEAS QUITE A BIT ___MICHIGAN, GEORGE ROMNEY DIFFERENT FROM MY OWN MEMBER OF THE CITY COUNCIL ___WHERE I LIVE __ A MEMBER OF THE COMMUNIST PARTY IN MY HOMETOWN 6 Only Slightly _-' Different THE LEADER OF THE COMMUNIST PARTY IN THE CITY WHERE I LIVE Can't Decide or HEAD OF THE AMERICAN MEDICAL Has About the Same ASSOCIATION 5 Amount of Beliefs and Ideas Which are COMMUNIST PARTY CHAIRMAN OF Like My Own as RED CHINA Are Different From My Own RUSSIAN NOVELIST WHO REFUSED THE NOBEL PRIZE FOR LITERATURE, BORIS PASTERNAK SPOKESMAN FOR THE JOHN BIRCH 4 Only Slightly SOCIETY, JOSEPH WELCH Like STATE DIRECTOR OF CIVIL DEFENSE ~:-.* ‘- ll 166 MANY TIMES SOCIALIST CANDIDATE FOR PRESIDENT OF THE UNITED STATES, NORMAN THOMAS LOCAL DIRECTOR OF CIVIL DEFENSE A PERSON I KNOW WHO WAS DRAFTED RED CHINESE STUDENTS DEMONSTRATING IN MOSCOW IN PROTEST TO RUSSIA'S POLICIES TOWARD RED CHINA SOCIAL WORKER WHO WORKS IN THE SLUMS OF NEW YORK 0R CHICAGO A PRINCIPAL IN THE SCHOOL SYSTEM WHERE I LIVE F anti-u 'u..' l m -4 1 HAS BELIEFS AND IDEAS QUITE A BIT LIKE MY OWN Very Much Like HAS BELIEFS AND IDEAS EXTREMELY LIKE MY OWN 167 NATIONAL RELIGIOUS LEADER OF MY FAITH MICHIGAN STATE UNIVERSITY STUDENTS PICKETING THE EAST LANSING CITY HALL FOR ALLEGED ___RACIAL DISCRIMINATION IN HOUSING ACCUSED KILLER OF EIGHT STUDENT NURSES IN CHICAGO, RICHARD SPECK FORMER RUSSIAN PREMIER, KHRUSCHEV NEGRO CIVIL RIGHTS LEADER AND HAS BELIEFS AND IDEAS EXTREMELY DIFFERENT FROM MY OWN Very Much Different HEAD OF THE STUDENT NONVIOLENT COORDINATING COMMITTEE, STOKLEY CARMICHAEL A HINDU HOLYMAN IN INDIA A NEGRO CIVIL RIGHTS DEMON- STRATOR IN THE SOUTH OR A HAS BELIEFS AND IDEAS QUITE A BIT DIFFERENT FROM MY OWN LARGE CITY PREMIER OF SOUTH VIET NAM, NGUYEN CAO KY TEAMSTERS UNION PRESIDENT, JIMMY HOFFA Only Slightly Different MEMBER OF THE BRITISH PARLIAMENT A NETWORK NEWS COMMENTATOR ON TV A PERSON WHOSE OPINION I HIGHLY RESPECT A NEGRO RIOTER IN THE LOS ANGELES Can't Decide or Has About the Same Amount of Beliefs and Ideas Which are Like My Own as Are Different From My Own WATTS RIOT OF LAST YEAR A CONVICTED DOPE PEDDLER HEAD OF SANE, NATIONAL COMMITTEE Only Slightly Like TO ABOLISH NUCLEAR BOMB TESTING A WHITE PERSON WHO DEMONSTRATES FOR NEGRO CIVIL RIGHTS HAS BELIEFS AND IDEAS QUITE A BIT LIKE MY OWN 4"! “1“qu 168 A MINISTER, PRIEST, OR RABBI THAT I KNOW PERSONALLY REPORTER FOR THE OFFICIAL NEWSPAPER OF THE RED GUARD IN COMMUNIST CHINA FEDERAL DIRECTOR OF CIVIL DEFENSE PRESIDENT OF MICHIGAN STATE UNIVERSITY, JOHN HANNAH Very Much Like HAS BELIEFS AND IDEAS EXTREMELY LIKE MY OWN ‘mmmm ,. _~. A50 ._ ¢-_ ~ 6 '..‘ 169 F- Please complete the following: Name: Student Number: Age: Sex: Male Female (please circle) Year in School: Freshman, Sophomore Junior Senior Graduate (circle one) Major: Do your parents live in Michigan? yes no (circle one) Have you traveled in a foreign country? yes No (circle one) How do you consider yourself politically? liberal (check one) conservative neither Marital Status: Married Single (circle one) Please check the income range which would cover your parent's total income last year: ____under $1,500 _ 8,000 to 9 999 ____1,500 to 2,999 ___1O,OOO to 14,999 ___3,000 to 4,499 ___15,000 to 19,999 4,500 to 5,999 ____2O,OOO to 29,999 ___6,000 to 7,999 ___30,000 and over What is your father's occupation? —— 170 APPENDIX B Mean ratings of the person descriptions (sources) on the scale ranging from persons whose beliefs and ideas are very different from my own (9 position) to persons whose beliefs and ideas are very much like my own (1 position) for both the pretest (Spring, 1967) and the main study (Summer, 1967). “1.4511?“ in IL 171 S©SH.H as mmma.s mmam.H so oasm.s omsm.a ms mfiaa.a Hmom.a mo oamo.s mam:.H ms mmms.s demm.m mm mmmm.s mmmm.a as msam.s ooHH.H ms omsm.s mama.a om m:sm.m mmma.a we Hema.w mmmfi.a as Hmoo.w m:ma.a as prN.@ .mwmmm zoom Coo: ASCEEva hodpm SHOE mocm.a mp mmsm.a emphases cease moose m:mm.a ow wzmw.s pmficsssoo m we 053 mucosa mo pCoECoo>ow oep mo menace mamm.a as omsa.s Sousa smassssoo CmOHaoE< one mo socmoq mwmm.a we Smom.w mopmpm popHCD hep CH Coapmm uficmmso pcosm pmflCSESOU osp mo comm :mam.a me nmom.w whom uaocdm .absaoo soaoam uaooa peas coo pas HmCHEHSO ~83 sesame popOH>coo Hmww.a as m:mm.~ ooa>o¢ Homomsom mom 09 03 on haoxfiq pmmoq com 50% acmsom one mmmm.a ms mmmm.w scopOH> spmoz mo socmoq zpsmm pmfizzssoo mmom.a we mmaa.w mafiso pom mo cmEmHmso apsmm pmficsssoo moma.a ea moma.m Rooam oaocoam .owooaso ca momssz scoossm bewam co aoaaam consooa oamw.a we moma.w cCfieo com CH zpsmm pmflcsssoo map Mo moose: ommo.a ms maom.m Haosxoom saooaaq owaooo .Spaoa Hnmz cocasos< one yo socmoq mmoa.o om :Qmw.m node stx Sam oep mo Comwmm comma .>mQ xzmm C662 869H .1 .osopm AWCflsmmv pmmpmhm 172 mwNm.H :m Hmfio.w NS®©.H Sm wwwa.w Com poommom mappflq hCo> o>cm 50% CoHCHQo mmozz Commmg < masm.a mm memo.a mamw.a mm meow.» dammsm Ca hpsmm pmHCCEsoo on mo MODEM: mmmw.a w: Hwam.© mmua.a mm m:wm.s opppm Choepdom 6 mo COCCo>ow mmm:.a ow mH©H.N wmo:.a ow ommm.w mCHCo pmHCCsEoo CH UCCCU pom on mo sogwmmzoz HoaoauCo one hot sooaoaom mmw:.a so Hmam.w mm:m.a Hm omwm.w Coapoom omom copua>Coo < mma:.a mm mmam.» mmm©.H mm omwm.w OOHHom posoom Cmfimmsm on mo comm mmom.H Sm mwmo.w mmam.a mm maom.n Commmmzoz HmCOHpmz Cmflmmsm one .ocsosa Co assess Hoaaobaom moo:.a mm momm.s wam:.a so mmmm.s ossamsm Ca zpsmm pmflCCEEoo on mo RODEO: swam.a mm mmHH.S mamm.a mm assm.s solos nachos .aooaoom Cosflm CCOh on sou Cosmoxomm mmmm.a oa smoa.a ammm.H we :sam.a scam Coon .oaosmo mo>smm cog mo COHHHM oop0fi>Coo mmam.a ma wwmm.w mmwm.a no NHFC.N THpmoz omoo «cocoaoCmm OEHCU opfizuCOAsz CH Cc: oCo gonadz mmam.a Hm wmmm.s mamm.a mm mamm.s Soapofl> Cocom Ca wCHpCmam mmoose pmHCCCsoo omoEMC upma> Cpmoz men yo mochsEoo .>om 2Com Cmoz .>o C 117 (El .Jn .oCmpm .oComm x mm Coo: EmpH Aaosssmv Sodom Cass AmCHanV pmmposm a“ Us. ll‘dfil 173 wjmm.H mm MHmH.N :moo.m m: madm.w opCpHmeHm C mamm.H CC mCHm.m memo.H CC mCmm.m CooCom omsoCICHCCm a mo:©.H mm omwm.@ bmH®.H m: mHCm.m MCHCU com UHCSOB mOHoHHom m.memCm 0p pmopoam CH soomoz CH wCprHmeosom mpCoonm mmoCHCo com mHmm.H N: MOCm.© mmmm.H w: ONHN.m moConm mo pCochoam OHmm.H m: CmHC.m momm.H a: sama.m pmoz oCH oCHH umHQ H EOCB zoCM H Comamm 6C9 momm.H m: MHmm.m womm.H m: bimh.m hapCCoo CCOHCoSC CHpmH m CH mCOHCCHCCosom zosom ow moemez CmHm.H m: Hmom.m swam.H m: mmma.o oHoCH CH CoCaHom CoCHm C mmmb.H mm de®.w Hmmm.H om MHHw.© SmemH> Epsom CH MCoz HmHCUpCm osmm.H ma mmHo.m Hmom.H Hm momm.m aCszoo CCOHSCCC CHHCH C CO COHCHOHQ sHmo.H mo mmoC.a ono.m mm wmmm.m AHHC ooaCoCsz HCHU mCHmmoo nCOHQCHCO mConm prHozh>mom wHHm.H om ssHa.m HHCC.H mm HHmm.m SCCC CoHnmCm on Ho HOHCU CH sochasoo MHmm.H Hm Cmma.m comC.H Cm oooo.s SoComCCCC «COHEOHm CmHmmCm Hossom mHHC.H mm omem.» camo.H mm memo.s o>HH H oCoCs aoHo on CH Hosea pmHCCesoo on Ho Cmcwoq 6C9 Hmmm.H mm oomo.s mmmm.H mm HmmH.s Czoposom as CH mpawm pmHCCssoo OCH mo 90950: < .>om mem Coo: .>om Mcmm C662 EmpH .oCmpm .ocopm AHmEEdmv hodpm SHOE AwCHHmmv pmmpmhm 174 mmmm.m H: mwom.m mmCH.N m: mHCm.m mmmm.H ommw.H mmmm.m Cmmm.H mmmm.H mNH©.H momN.H mmww.H NONH.N momo.m Smom.m Smom.m dm:m.m meo.© mdmo.w Comm.m Comm.© wmmm.w E 396 dmom.m OHHOHHOCC OOHHOH> OC COCS COOCCow HCOO uHmOHm OOHOOHU 0C3 COCHOH> CHCom CH HOB meHOwO OHOHOHmeoCOQ OCOHow CH wCHCoom OOCHHOC loam COOHoCz .m.C OCH meHOOC OHO 0C3 OCOHOHHOCOCOQ COHHHHm HOHOBOHOO zCHOm «OHOOHOCOU HOHHCOOHOOHm COOHHCCCOm UCO HOHOCOm HOCHom OoCOCm CH COCOCOBOZ wCH IOOOH OCH Com HOHHHB HOHHOpHcm OOHHOCC CHHOH CH C308 SCHCCoo HHOSO O CH HOHCO ooHHoa OHHom HCSHH -HHCOOHOOHm COHCD OHOHOCOOB HOOCOHEHOO mOHCoum qOOppHssoo mCHpOCHOHooo pCOH0H>uCoz pCOOCpm OCH Ho OOOm OCO HOOOOH mprHm HH>HU OCsz HOOw HOOH Ho HOHm mppOz OOHOwC< OOH OCH CH COHOHm OHsz < COHCCU OHHOCpOo OCH Ho COUOOH HOCHHHHCm qOmom OCH CHOU CHOCC mHC OCHCm OCO COC -COH> Cpsom CH Caz oCC meHcOC OOHOHmeoCOm 0C3 CO2 wCCow C .>OQ CCOm COOS .CCOHO Asosssmv ROCCO CHOz .>OQ .oCOom CCOm COOS AwCHHQmV mepOHm EOHH 175 ©0HC.H momH.m mHOHoz HOHOCOU mo HCOUHmOHm m:mo.m sme.m mCooCOH CCC .HmHCsCHoo O0H>O< HOCOCOSOZ OOHOOHOCHm me0.H omwm.m HOOCOOH :OHQ HOC HOOCmOm COCHHOz 50% COHCHQO OOOCz CoOHOm C mHsm.H mmmm.m CoCCOHOz Coom bxoz Hz mm:w.m NHNC.m OOHOHm UOHHCD OCH mo HCOUHOOHmIOOH> momm.H wom:.m COHwOH COOHHOCC OCH Ho HOOCOCCOU HOCOHHOZ HmmO.H msam.m mssoCH Ccaaoz .moposm oopHCp oCH Ho oCoonosa Com OHOOHOCOU HOHHOHOom OOCHB hCOz msmO.H mzmm.m am ooo Cosswz nCO2 HOH> CHCom mo HOHCOCm mmoo.H mzmm.m OoCOCm .mHCca Ho HOHCO ooHHoa oCH SHsm.H mmmm.m oCOHow Ho CoCCCo oCH Ho OOOm aFHCCCOHCOU Ho COCOHCCOHC mmoo.m mmmm.m OOHOHO COHHC: oCH Ho HCoonoaa moom.H aCma.m msOHHHHz Aaasomv CoCCoz .o qOHOOHoCOo HOHHOHOCOm OHHOHooCOQ OCO COmHCOHS mo COCCO>00 HOCHom SHCm.m wcwm.m COCOCO HHHHO .CC .pmHHoOCcsm .>OQ oCOm COOS .>OQ xCOm COO: COHH .CCOHm .CCOHm AHOssdmv hCCHm CHOZ AMCHHCmV HOOHOHm nigh“ It, 176 AHOEECWV hCCHm CHOZ mwmm.H mmmm.: OCOzom mHoCOHm HCOO qOHOmCm Co wCHHCm Ho OCOHOOCm OCH an OOH0H> :Coo mOs 0C3 HOHHH m: COOHHOC< MCmC.H HHmm.C CoCHOsm COHHOC «OHOCHoz OHHCoCoPCC OOHHCD OCH Ho HCOOHOOHH HOCOHHOz Hmmw.o oooo.m HCOCOHHHOH COHHHHH OCH Ho HOCCOZ momH.H oooo.m OOOH>HOm KSOHHHHZ mOHOHm OOHHCD OCH Ho HHOHm H0 OHOHCU HCHOh OCH H0 COCCHOCO Hmmw.H oooo.m O>HH H OCOC3 SOHO OCH CH HOHCO ooHHoa oCH NH::.H oooo.m CHOHHHm HOOHO Ho HOHOHCHE OCHCH HHzm.H Nmmo.m OCOCOZOE HHOCom flOmCOHOQ Ho HHOHOCOOm .m.b Cmmo.H mmso.m mHHCC SCOHHHHC CmHHHCm HHO Ho HOHOCOU mCHcCOCCoo mmmm.H mCmo.m ACoCOCC SCOV HCOHHHHC OCH CH O0H>HOm Ho H50 HOOHOO O wCHCOz OH 0C3 SOCC H COOCOH < wmmS.H mmHH.m CH OSHH H SOHO cCH Ho Coho: oCH mmmO.H HmmH.m COCCOHOOC mHCom qOHCHOHOHHH HOH ONHCH HOCoz OCH OOOCHOC 0C3 HOHHO>oz COHOOCm OOHC.H momH.m mCoHoz HOCoCOo Ho HCOCHmOCa .>OQ CCOm COOS (I COMM .CCOHm AOCHHCOV HOOHOCC [ENE .. .. .. 177 NmHO.N NHOm.H om:o.m mme.H mmzm.H mHCN.H MCNN.H mm:m.a SCmm.H owm®.H mHHw.C SocCCom .a HCOCom «Chow SOS COCO HoHOCOm OHHOHoonQ MHH®.: OOCOCCoo Ho COCEOCU OCH Ho HCOOHOOCH HOCOHHOz momw.: AHmmv COHHOOHHOO>CH Ho COOCCm HOCOOOH OCH Ho HOHOOCHQ mpmw.4 COCOCCC CH mOHHHC aso .oeazv CoHHON aHCOwHo SHOOHB OHHCOHHC CHHoz OCH CH OCOOHB OCH Ho HOOCOCCOO mCmm.C OHOCOO cocoon oopHCC OCH Ho Cocooq SbHCoHoz m:mm.: mmaoo OCHCOS OOHOHO OOHHCD OCH H0 HOCOCOO OCHOCOCCOO mCmm.C .CH .OCHC COCHCH CHHCOS .HH .OoCOHOHCoo CHCOHOOOOH COHHOHCCU CHOCHCom OCH ..Ho COOS UCO HOOOOH OHCme HH>HU cthz CmCm.C OCHHmoH osom COOHoCz CmHHoCC op OOHHHC uCoo HOCOHHOZ qOCOm Ho COOS szm.: Cotz .2 CHOCon .OoooHoCco HOHHCOC nHOOHmnO0H> COOHHCCCOm COSMOS mmmm.: SHHU OMCOH o Co Cosom oCH CH Cosossm -Cosoo mOCOHm HHSHO oCmoz C 1|. .>OQ xCOm COOS .>OQ MCOm COOS SOHH .oCOHm .oCOHO AHOEECOV SOCHO CHOS AmCHHva HOOHOHH MHNE ml 178 mmmm.H mm OOCm.C SOCC.H mm OHCO.C HHOHHHHC OCH CH CHOHCOCO C HOOC.H am omam.C mmmm.o om COOO.C OOCOHOC HHSHO Ho HoHOOCHm HOooH HHHC.H mm OOSC.C mmmH.H Hm mmHO.C OOCOHOC HHSHO Ho CoHOOCHC HOCOOOC Smom.H Cm OOCm.C CCOO.H mm OSHH.C OSHH H OCOCC COHOHO HooCom OCH CH HOCHoCHHC C mHom.H Hm CHOC.C SOOC.H mm OSHS.C OOSHH CCoH SOC OCH Ho HOHCouCHuaoHHOm oemm.H Om mmHO.C OCCS.H Cm mmmS.C COCO COOC qOHOHm Ho SHOHOHOOm .m.b mmmm.H OH HOOO.C mCmO.H mm mmms.C HHHCCCCoo as CH COCOOHHoO C OSOO.H Hm Oomm.C mwm0.H Om mmmSC CoHHOHooOOC HOCoHHOoCOm HOCoHHOz OCH Ho OOOC HQHO.H Cm Hmom.C Homo.m Sm mmma.C COzoCCOOHm .C HCOHCC .SCHC OCH Ho HOHOCOO nCO OOHOHm OOHHCD OCH Ho HCOCHOOHH COsHom mHCO.H mm Hamm.C mmOO.H mm mmmS.C HCOOC OOH>COm HOHOOm OOHOHm OOHHCD C ommm.H Om mmOm.C CCmo.H mm SCmS.C OHOHm OCHCoCCOHOz O Ho CoCHO>oc HHHm.o QC Ommw.C mmmo.H 0C mmms.C COCOCOst COHHHCO OCHOOOC C nOOCHH. COUCOH OCH Co HOHHOCOm OSHO.H mHHO.C HOHHHOHC as SoHH OHCHOHOHOOH OHOHm OCH Ho HOCCOS .>OC CCOm COOS .>OC CCOm COOS COHH .OCOHO .OCOHm AHOSSCOV HOCHO CHOz AwCHHmmv HOOHOHH 179 mHOm.H OH mmmH.C OOQC.H SH mmmH.C cams.H mm Hamm.C wmmC.H CH moCo.C mCmm.H OH stm.m mmmm.H om mmmm.C COOO.H OH OCQH.C OOSC.H mm Hmom.C smmw.H mH mmmm.m OCmm.H mm OCSS.C SHHm.H mm OCOO.C momC.H Sm QHNO.C mHom.H mm mmmO.C .>OQ CCOm COOS .OCOHO AHOEECOV SUCHm CHOS Omoo.m OH mHom.C HOCsom OOCoOo .COOHCOHC Ho OHOHO OCH Ho C0CHO>ow wmmm.H NH momm.: 0H ow COHHo HmoC H HOHOOQ HOOHOOS OCH mCmO.H mH momm.: COOCHHQ HHOHO>m «OHOCHHHH COHH HOHOCOm O HOCOOH OHOCOm COOHHCCCOm momQ.H mH mmmm.: >9 Co HOHOHCOECOO szz CHOSHOz C HQOm.H om mmmm.C HOHOOOOst >H OHHHoSOC Hz OmmC.H Hm HmHC.C OOOC SHCOHCOOC H HOCOCOSOZ OCH Co HOHHOCOm mhom.H mm HHNC.C CHHOS SC Ho HOCOOH OCOHOHHOm HOCOHHOz OOHS.H mm mmmm.C COCCOC CCoH .SHHOCOSHC: OHOHO COOHCOHC Ho HCOOHOOCC nmmH.m Cm mmmm.: CHHOH SC Ho HOOOOH OCOHOHHOC HOOOH mmom.H mm mmmm.: OCOHOCH CH HCOOC CHOS OCOHHoom omom.H mm ma:m.: OOCOHOQ HH>HU Ho HOHOOHHQ OHOHw Comm.H Hm omwm.: O>HH H OHOCB HHoCCoo SHHO OCH Ho COCSOC mmmm.H mw mCmm.C CoHHOHOOOOC HOOHCOS COOHHOCC OCH Ho COOS .>OQ MCOm COOS SORW .OCOHO AOCHHQOV HOOHOHH 180 OCOO.H O OOOS.O HCOO.H C OmOm.m OCOHCC HOOO a: OOOO.H O omam.m OOOO.H O mHHO.m oOOOHCO Ho CHoH st Ho OCCHO OCH CH OCH03 0C3 COCHoz HOHoom OOHH.O OH mHHm.m OOOO.O O OOmO.m HOCHoz SC OHOO.H m OOOO.m mHmC.H a Smom.m OOHHOHO OOC oCz soCC H CoOCOa C OOOS.H a mmOs.m COHS.H O HHOm.m CoHHOoCOH oH HHoCCCO HOHO uCOCHm OHoC H0O OHCHOHOHOOH OHOHm OCH OCHHOMOHH OHOCOOOB COOHCOHS Ho CCOHO OCH HOH COCOOxomm mmmo.m Om OOCO.C OOSS.H m OOOO.C OHCOHO HHSHO oCOOz CoH OOHOHHOCoCOQ 0C3 COOHOH OHHCS C OOHO.H HH OOOO.m OOOO.H OH samo.C HHHOCoOHOO soCC H HOCH HCCOO Ho .HOOHHO .COHOHCHC C OOO0.0 O OOO0.0 OOSO.O HH HOOH.C HOCHOH H2 OOOO.O Om OOCm.C mOHO.H OH OOOH.C OCHOCoC CH CoHHOCHsHHo -OHC HOHoOm OOOOHHC HoH HHOC HHHO OCHOCOH HmOm OCH OCHHOMOHm OHCOO -CHO HHHOHOSHCO OHOHO COOHCOHC OHOO.H O moma.m OHmO.H OH OOOH.C OSHH H OHOCC COHOSO HOOCOO OCH CH HOCOOOB C OOSO.H Om OCHO.C OOSO.H CH HOOH.C OCoHHOz OOHHCO OCH H0 HOHOCOO SHOHOHOOO HOOm.H OH OHOO.C OSHO.H OH OOCO.C COHCOS HHOm «HHCOU OCOHCCm OOHOHm OOHHCD OCH Ho OOHHOCO HOHCO .>Om CCOC COOS .>Om xCOm COOS COHH OCOHO .OCOHO AHOEECOV SUCHm CHOS AOCHHCOV HOOHOHH 6. iii‘ 181 mmCC.H m mHmm.N HH©:.H H HmmH.m .>Om CCOm COOS .OCOHO AHOSSCOV SOCHO CHOS mOOC.H H OCOO.O mmmm.H O OOOO.O .>OQ. CCOm COOS .UCOHm AwCHHCmv HOOHOHH OOH>OC HOCOOHOH Com 0H ow 0H SHOCHH HmoS CO H CoOHOm OCH SHHOS 0H OCHOO CC Ho OOHHCOS H CoOHOm OCH COmw 182 APPENDIX C Factor scores for the principal axis factor and the two factor rotated solution for the person factor analysis of the belief-disbelief ratings of the 123 person descriptions (sources) obtained for the pretest. Note:--The method of computing the scores was outlined in the text. A positive standard score indicates that the source was above the mean on the belief-dis- belief rating scale indicating source had beliefs and ideas different from his own. A negative standard score indicates the opposite or that the source had beliefs and ideas similar to his own. Factor 2 scores marked with a single asterick (*) represent the sources which they rated as more similar to their own beliefs and ideas by at least one standard deviation than either Factor 1 or the first principal axis factor. Factor 2 scores marked with a double asterick (**) represent the sources which they rated as more different from their own beliefs and ideas by at least one standard deviation than either Factor 2 or the first principal axis factor. Frfinaxr r': ~ V. Item No. 10 11 12 13 14 15 16 183 Source Description Convicted Killer of Lee Harvey Oswald, Jack Ruby Member of the Communist Party in England Secretary General of the United Nations Scotland Yard Agent in England Heavyweight Boxing Champion, Cassius Clay (Muhammed A11) The Person you are Least Likely to go to for Personal Advice President of General Motors The Mayor of the City I Live in Spokesman for the Group of Michigan Teachers Picketing the State Legislature for More Financial Support to Education A Person Whose Opinion you Neither ReSpect nor DisreSpect Republican Senate Leader and Senator from Illinois, Everett Dirksen Prime Minister of Great Britain Governor of a Southern State A United States Secret Service Agent My Mother Vice-President of the United States A First Two Factor Principal Solution Axis Factor 1 _g__ 1.13 1.09 1.44 1.45 1.48 0.33* -0.97 -0.85 -1.68 -O.73 -O.73 -O.57 1.10 1.15 0.18 1.71 1.67 1.67 -0.50 -0.54 0.24 -0.50 -0.51 -0.27 -1.08 -0.97 -1.78 -0.21 -0.22 -0.23 —0.93 -0.99 -0.17 -0.47 —0.41 —0.79 1.05 0.97 1.64 —0.82 -0.86 0.07 -1.41 -1.42 -0.87 -0.35 -0.25 —0.24 mmgi—w‘ct' 'Il ~w-r.'. .: l7 l8 19 2O 21 22 23 24 25 26 27 28 29 30 31 32 184 Syndicated NeWSpaper Advice Columnist, Ann Landers The Person I am most Likely to go to for Personal Advice My Next Door Neighbor "Yankee Go Home" Demonstrators in a Latin American Country Convicted German War Criminal and one time Adolf Hitler's Deputy, Rudolf Hess Cuban Prime Minister A Teacher in the School System where I Live The Person I know whom I Dislike the Most National Commander of the American Legion Commander in Chief of the Russian Army Member of the Communist Party in Red China Member of the Government of France who is a Communist National President of the Chamber of Commerce Democratic Senator from New York, Robert F. Kennedy British Demonstrators who are Against the U.S. Nuclear Sub- marines Docking in England President of the United States -0.47 -2.03 -0.28 1.17 1.60 1.57 -1.00 0.80 -0-37 1.04 1.85 1-59 -O.68 -0.59 0.46 -0.19 —0.58 -l.90 -.025 1.22 1.54 1.59 —0.98 -0.73 -0.53 0.52 -0.14 0.36** -2.51 -0.09 -0.17* 1.89 0.55* -1.13 0.92 -0.10 -0.97 -0.68* 0.16 -.- v—‘hy . 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 185 The Pope, Spiritual Leader of the Catholic Church Number One Man in Nation—Wide Crime Syndicate, Cosa Nostra Grand Dragon of the Klu Klux Klan Communist Party Leader of North Vietnam Commander of the Troops in the North Atlantic Treaty Organi- zation (NATO, our Allies in Europe) Dictator of a Latin American Country President of France A Person I know who is Making a Career out of Service in the Military (any branch) National President of the United Auto-Mobile Workers, Walter Ruether Leader of the American Nazi Party, George Lincoln Rockwell The Person I Married or am Going to Marry A Young Man who Demonstrates Against the War in South Vietnam and Burns his Draft Card Governor of a Neighboring State The Medical Doctor I most often go to A Chaplain in the Military 0.47 1.28 2.02 1.82 -0.71 0-75 0.86 -0.64 -0.27 1.80 —1.83 0.96 -0.65 -0-95 -0-93 0.38 1.24 1-95 1.84 -0-75 0.70 0.88 -0.74 -0.23 1-75 -1-75 1.06 -0.67 —1.00 —0.98 0.69 1.61 2.51 1.07 0.07 1.32 0.51 0.58** -0.95 2.19 -2.11 —0.80* -0.40 -0.68 -0.28 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 Times 186 Archbishop of Canterbury, Head of the Church of England -O.16 Editorial Writer of Pravda, the Russian National Newspaper 1.31 Police Chief in a Small Country town in Latin America 0.26 Demonstrators Against the War in South Vietnam who Greeted President Johnson when he visited Australia 0.53 Head of the Russian Secret Police 1.20 Former Senator and Republican Presidential Candidate, Barry Goldwater 0.09 Editor-In—Chief of the New York Former President of the United States and General of the Army, Dwight D. Eisenhower —0.86 The Police Chief in the City where I Live -0.53 Former Governor of Michigan and Democratic Senatorial Candidate, G. Memmen (Soapy) Williams 0.08 Local Religious Leader of my Faith —1.09 Member of the State Legislature from my District -O.61 Reporter on the NeWSpaper I Regularly Read -O.87 Head of the National Education Association -0.69 American U2 Pilot who was Convicted by the Russians of Spying on Russia, Gary Francis Powers -0.55 -0.21 0.63 1.20 -0.08 -0.47 -0-95 -0.64 0.16 -1.10 -0.61 -0.84 -0.65 —0.56 0.19 0-95 0.72 -0.85* 1.38 1.31** -0.93 0.14** 0.13 -0.19 —0.65 -0.65 -0.87 -0.52 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 187 My Favorite TV Newscaster My Father Commanding General of All British Military Units A Prostitute U.S. Secretary of Defense, Robert McNamara Evangelist, Dr. Billy Graham A Person whose Opinion you have very little Respect For Member of the Communist Party in Russia Editorial Writer for the Leading Newspaper in France Negro Civil Rights Leader and Head of the Southern Christian Leadership Conference, Dr. Martin Luther King, Jr. My Best Friend Chief Justice of the United States Supreme Court, Earl Warren Former Republican Vice-President and Presidential Candidate, Richard M. Nixon A Policeman in my Community Reporter on the London Times, A Leading British NeWSpaper U.S. Secretary of State, Dean Rusk Buddhist Monk in South Vietnam —0.94 —1.27 -0.49 —.78 -0.56 -0.09 1.09 1.21 0.38 -0.36 -1.70 -0.82 -0.67 -0.76 -0-53 -0.72 0.87 -0.96 -1.30 -0.52 0.86 —0.61 0.19 1.04 1.22 0.42 —0.29 —1.65 —0.80 -0.80 -0.84 -0-53 -O-77 0.94 -0-59 -0.67 —0.01 0.23 0.25 0.86** 1.54 0.62 -1.17 -l-73 -l.54 -0.73 0.05 0.18 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 188 Commander of the North Viet- namese Communist Troops Fighting in South Vietnam Head of a Communist Front Organization in the United States A Strip-Tease Dancer Director of the Federal Bureau of Investigation (FBI Leader of the American Communist Party Chairman of the Joint Chiefs of Staff of the United States Military Services Commanding General of the United States Marine Corps Majority Leader of the United States Senate The Police Chief of Paris, France Governor of the State of Michigan, George Romney Member of the City Council where I Live A Member of the Communist Party in my Hometown The Leader of the Communist Party in the City Where I Live Head of the American Medical Association Communist Party Chairman of Red China Russian Novelist who Refused the Nobel Prize for Literature, Boris Pasternak 1.56 1.67 0.77 —0.80 1.63 -0.69 —0.72 -0.49 0.01 -l.l6 -0.82 1-33 1.23 —0.83 1-79 -0.02 1.56 1.65 0.84 -0.92 1.65 -0.76 -0.80 -0.47 0.01 -1.25 -0.86 1.39 1.28 -0.89 1.75 -0.01 1.01 1.29 0.32 0.71** 0.83 0.46** 0.48** -0.52 0.06 -0.05** -0.43 0.18* 0.31 —0.01 1.62 -0.64 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 189 Spokesman for the John Birch Society, Joseph Welch 1.10 State Director of Civil Defense —0.90 Many times Socialist Candidate for President of the United States, Norman Thomas 0.22 Local Director of Civil Defense -0.74 A Person I Know who was Drafted -l.29 Red Chinese Students Demonstrating in Moscow in Protest to Russia's Policies Toward Red China 0.79 Social Worker who Works in the Slums of New York or Chicago -l.l9 A Principal in the School System Where I Live -0.83 National Religious Leader of my Faith -l.03 Michigan State University Students Picketing the East Lansing City Hall for alleged Racial Discrimin- ation in Housing -0.89 Accused Killer of Eight Student Nurses in Chicago, Richard Speck 1.74 Former Russian Permier, Khruschev 1.09 Negro Civil Rights Leader and Head of the Student Non-Violent Coordinating Committee, Stokley Carmichael 0.43 A Hindu Holyman in India 0.82 A Negro Civil Rights Demonstrator in the South or a Large City —0.30 Premier of South Viet Nam, Nguyen Cao Ky -0.ll Teamsters Union President, Jimmy Hoffa 0.40 0-99 -0.92 0.33 -O.77 -1.32 0.81 -1.09 —0.82 -1.03 -O.73 1.72 1.10 0.47 0.85 -0.22 —0.06 0-39 2.00** -O-37 —l.25* -0.48 -0.83 0.44 -1.92 -0.40 -0.84 -1.83* 1.65 0-73 0.02 0.54 -1.06 0.58 0.32 7! Fun—m 15 311i 113 114 115 116 117 118 119 120 121 122 123 190 Member of the British Parliament —0.41 A Network News Commentator on TV -0.86 A Person Whose Opinion I Highly Respect -1.61 A Negro Rioter in the Los Angeles Watts Riot of Last Year -/71 A Convicted Dope Peddler 1.25 Head of Sane, National Committee to Abolish Nuclear Bomb Testing -0.21 A White Person Who Demonstrates for Negro Civil Rights —0.92 A Minister, Priest, or Rabbi that I Know Personally -l.23 Reporter for the Official News- paper of the Red Guard in Communist China 1.25 Federal Director of Civil Defense -0.79 President of Michigan State University, John Hannah -0.94 -0.39 -0.86 -1.52 0.75 1.28 —0.05 -0.76 ‘1021 1.23 -0.80 -1.01 -0.46 —1.00 -2.23 -0.18 0.71 -1.59* -2.01* -1.21 1.04 -0.51 -0.22 —-umam 231-. _ ' _- 4 191 APPENDIX D List of the sources selected to be evaluated in the main study using the Berlo et a1. source credibility semantic differential scales. Descriptions 1-24 are defined as the sources associated with the belief end of the belief-dis- belief continuum. Descriptions 25-48 are defined as the sources associated with the disbelief end of the belief- disbelief continuum. 10 11 12 13 14 192 THE PERSON I AM MOST LIKELY TO GO TO FOR PERSONAL ADVICE THE PERSON I MARRIED OR AM GOING TO MARRY A PERSON WHOSE OPINION I HIGHLY RESPECT MY BEST FRIEND SOCIAL WORKER WHO WORKS IN THE SLUMS OF NEW YORK OR CHICAGO MY MOTHER A PERSON I KNOW WHO WAS DRAFTED SPOKESMAN FOR THE GROUP OF MICHIGAN TEACHERS PICKETING THE STATE LEGISLATURE FOR MORE FINANCIAL SUPPORT TO EDUCATION A WHITE PERSON WHO DEMON- STRATES FOR NEGRO CIVIL RIGHTS A MINISTER, PRIEST, OR RABBI THAT I KNOW PERSONALLY MY FATHER MICHIGAN STATE UNIVERSITY STUDENTS PICKETING THE EAST LANSING CITY HALL FOR ALLEGED RACIAL DISCRIMINATION IN HOUSING A TEACHER IN THE SCHOOL SYSTEM WHERE I LIVE SECRETARY GENERAL OF THE UNITED NATIONS 19 2O 21 22 23 24 25 26 27 28 29 30 31 32 33 A NETWORK NEWS COMMEN- TATOR ON TV MY FAVORITE LOCAL TV NEWSCASTER REPORTER ON THE NEWS— PAPER I REGULARLY READ NATIONAL RELIGIOUS LEADER OF MY FAITH PRESIDENT OF MICHIGAN STATE UNIVERSITY, JOHN HANNAH LOCAL RELIGIOUS LEADER OF MY FAITH A PERSON WHOSE OPINION I HAVE VERY LITTLE RESPECT FOR MEMBER OF THE COMMUNIST PARTY IN RUSSIA GOVERNOR OF A SOUTHERN STATE REPORTER FOR THE OFFICIAL NEWSPAPER OF THE RED GUARD IN COMMUNIST CHINA A CONVICTED DOPE PEDDLER HEAD OF THE RUSSIAN SECRET POLICE EDITORIAL WRITER 0F PRAVDA, THE RUSSIAN NATIONAL NEWSPAPER MEMBER OF THE COMMUNIST PARTY IN ENGLAND SPOKESMAN FOR THE JOHN BIRCH SOCIETY, JOSEPH WELCH l5 l6 17 18 193 CHIEF JUSTICE OF THE UNITED STATES SUPREME COURT, EARL WARREN GOVERNOR OF THE STATE OF MICHIGAN, GEORGE ROMNEY THE MEDICAL DOCTOR I MOST OFTEN GO TO REPUBLICAN SENATE LEADER AND SENATOR FROM ILLINOIS, EVERETT DIRKSEN 34 35 36 37 38 39 40 41 42 43 44 45 46 CONVICTED KILLER OF LEE HARVEY OSWALD, JACK RUBY NUMBER ONE MAN IN THE NATION-WIDE CRIME SYNDICATE, COSA NOSTRA COMMANDER OF THE NORTH VIETNAMESE COMMUNIST TROOPS FIGHTING IN SOUTH VIETNAM CUBAN PRIME MINISTER MEMBER OF THE GOVERN- MENT OF FRANCE WHO IS A COMMUNIST LEADER OF THE AMERICAN COMMUNIST PARTY HEAD OF A COMMUNIST FRONT ORGANIZATION IN THE UNITED STATES CONVICTED GERMAN WAR CRIMINAL AND ONE TIME ADOLPH HITLER'S DEPUTY, RUDOLF HESS THE PERSON I AM LEAST LIKELY TO GO TO FOR PERSONAL ADVICE COMMUNIST PARTY LEADER OF NORTH VIETNAM COMMUNIST PARTY CHAIR- MAN OF RED CHINA ACCUSED KILLER OF EIGHT STUDENT NURSES IN CHICAGO, RICHARD SPECK MEMBER OF THE COMMUNIST PARTY IN RED CHINA 194 LEADER OF THE AMERICAN NAZI PARTY, 47 GEORGE LINCOLN ROCKWELL 48 GRAND DRAGON OF THE KLU KLUX KLAN “IO-Bill . Ehfi 195 APPENDIX E Factor loading matrix for the three factor solution for the re-factor analysis of the Berlo et al. Lansing area data deleting the scale "objective-subjective." l': J g I" . - .‘1 ' a. 196 2 Factor I II III h Purity* 01 selfish---unse1fish 69 15 -02 50 96 02 crue1---kind 84 19 —01 75 95 03 unfriendly---friend1y 83 18 O5 72 95 04 dangerouS---safe 80 19 -02 68 95 05 unpleasant--pleasant 82 19 06 71 94 06 quarrelsome---congenia1 82 20 03 71 94 07 harsh---gent1e 82 18 -10 71 94 08 cool---warm 67 15 10 47 93 09 disagreeable---agreeable 81 22 01 70 93 10 unforgiving---forgiving 76 20 -05 61 93 11 unjust---just 81 26 04 73 91 12 gloomy--—cheerfu1 74 16 18 60 90 13 inhospitable---hospitab1e 76 24 09 64 89 14 upset---calm 68 22 08 52 89 15 unsociable---sociab1e 74 25 10 62 88 16 unfair---fair 80 29 03 73 88 17 impatient---patient 70 26 -10 57 86 18 unethical—--ethica1 73 30 05 63 85 19 inexperienced---experienced 25 80 14 72 88 20 untrained---trained 27 82 14 76 88 21 unskilled---skilled 33 77 18 73 81 22 unqualified---qualified 37 76 10 72 80 23 uninformed---informed 34 74 17 70 79 24unauthoritative--—authoritative 21 67 33 61 74 25 inept---able 44 65 18 65 65 26 unintelligent—-—inte11igent 42 63 23 62 63 27 meek-—-aggressive -08 09 77 61 98 28 reserved---frank 04 -09 55 31 97 29 hesitant---frank 01 14 70 51 96 30 forceless---forcefu1 —04 25 60 43 85 31 passive---active 16 25 61 46 81 32 timid--—bold —32 -02 64 51 80 33 tired---energetic 24 24 64 53 78 34 slow---fast 11 31 50 36 Proportion of Total Variance 34.6 16.0 10.6 61.2 Proportion of Common variance 56.5 26.0 17.2 99.9 Factors: I--Safety II——Qualification III--Dynamism *The scales loading on each factor are ordered within by the factor by the purity index. The index is the ratio of Ehe square of the variable's highest loading to its communality (h ). rum 5% .1. 'I '1 197 APPENDIX F Questionnaires administered during Summer, 1967 for the main study used to obtain the source sample validation data, the background information, and the dogmatism scale data. ..q-‘du 198 Department of Communication Michigan State University STUDENT NUMBER__> INSTRUCTIONS You are participating in a study of how people evaluate other persons. In this part of the study, you will be asked to look at the brief’description of a number of persons and then to rate each on the degree to which each person has beliefs and ideas which are like your own or are different from your own. Below are a couple examples of the kinds of descriptions of persons you will be asked to look at: YOUR FAVORITE LOCAL TV NEWSCASTER MAYOR OF THE CITY OF EAST LANSING, GORDON THOMAS You are to use the ladder scale on the right of each page to rate each person. Please examine this scale. You are to place each person somewhere along the scale according to which position best describes how much that person has beliefs andfiideaS'which are either like your own or are different from our own. After you have decided what position best describes your feelings about the person described, then you would put the‘number of that position in the blank Space to the left of the description. For example, if you thought that "YOUR LOCAL TV NEWSCASTERH was best described by "HAS BELIEFS AND IDEAS QUITE A BIT DIFFERENT FROM MY OWN," you would mark it as follows: 7 YOUR FAVORITE LOCAL TV NEWSCASTER Another example, if you thought that "MAYOR OF THE CITY OF EAST LANGINC, GORDON THOMAS" was best described as having beliefs and ideas ”Very Much Like" my own, you would mark it as follows: MAYOR OF THE CITY OF EAST LANSING, 2 GORDON THOMAS ‘2...— NAME_, 9 HAS BELIEFS AND IDEAS EXTREMELY DIFFERENT FROM MY OWN 8 Very Much Different HAS BELIEFS AND 7 IDEAS QUITE A BIT DIFFERENT FROM MY OWN 6 Only Slightly Different Can'thecIde‘OF Has About the Same Amount of 5 Beliefs and Ideas Which Are Like My Own as Are Different From My Own 4 Only Slightly Like 3 HAS BELIEFS AND IDEAS QUITE A BIT LIKE MY OWN 2 Very Much Like HAS BELIEFS AND IDEAS EXTREMELY 1 LIKE MY OWN 199 Before you mark the scale for the first description, please glance over all of the descriptions of persons so you will get an idea of the kinds of persons you will be asked to evaluate. Then go back to the beginning of the descriptions and work your way through without looking back and forth through the items. Do not worry over or puzzle about indivi- dual descriptions. It is your immediate feelings about each person described that we want. 0n the other hand, please do not be careless, because we want your true impressions. Also, please write the numbers in the Spaces as clearly and legibly as you can. If you have no questions, please begin. 200 CONVICTED KILLER OF LEE HARVEY HAS BELIEFS AND ___OSWALD, JACK RUBY 9 IDEAS EXTREMELY DIFFERENT FROM MEMBER OF THE COMMUNIST PARTY IN MY OWN ___ENGLAND SECRETARY GENERAL OF THE UNITED NATIONS 8 Very Much SCOTLAND YARD AGENT IN ENGLAND Different HEAVYWEIGHT BOXING CHAMPION, CASSIUS CLAY (MUHAMMAD ALI) HAS BELIEFS AND THE PERSON I AM LEAST LIKELY TO 7 IDEAS QUITE A GO TO FOR PERSONAL ADVICE BIT DIFFERENT FROM MY OWN A WHITE PERSON WHO DEMONSTRATES FOR NEGRO CIVIL RIGHTS SPOKESMAN FOR THE GROUP OF MICHIGAN Only TEACHERS PICKETING THE STATE LEGIS- 6 Slightly LATURE FOR MORE FINANCIAL SUPPORT Different .__;TO EDUCATION THE LEADER OF THE COMMUNIST PARTY Can't Decide or IN THE CITY WERE I LIVE Has About the _-_' 5 Same Amount of REPUBLICAN SENATE LEADER AND Beliefs and SENATOR FROM ILLINOIS, EVERETT Ideas Which Are ___DIRKSEN Like My Own as Are Different LEADER OF THE AMERICAN COMMUNIST __ From My Own PARTY 4 Only Slightly GOVERNOR OF A SOUTHERN STATE Like A UNITED STATES SECRET SERVICE ___;AGENT 3 HAS BELIEFS AND ___MY MOTHER IDEAS QUITE A BIT LIKE MY OWN ACCUSED KILLER 0F EIGHT STUDENT ___NURSES IN CHICAGO, RICHARD SPECK FORMER RUSSIAN PREMIER, KHRUSHCHEV THE PERSON I AM MOST LIKELY TO GO 2 Very Much TO FOR PERSONAL ADVICE Like 201 ' g - A STRIP—TEASE DANCER HAS BELIEFS AND "YANKEE GO HOME" DEMONSTRATORS 1 IDEAS EXTREMELY IN A LATIN AMERICAN COUNTRY LIKE MY OWN CONVICTED GERMAN WAR CRIMINAL AND ONE TIME ADOLF HITLER'S DEPUTY, RUDOLF HESS _ 202 CUBAN PRIME MINISTER HAS BELIEFS AND A TEACHER IN THE SCHOOL SYSTEM 9 IDEAS EXTREMELY WHERE I LIVE DIFFERENT FROM MY OWN THE PERSON I KNOW WHOM I DISLIKE THE MOST LOCAL DIRECTOR OF CIVIL DEFENSE '_—- 8 Very Much COMMANDER IN CHIEF OF THE Different .__;RUSSIAN ARMY _*l— 2 . 1'. BUDDHIST MONK IN SOUTH VIETNAM MEMBER OF THE COMMUNIST PARTY HAS BELIEFS AND ‘___IN RED CHINA 7 IDEAS QUITE A BIT DIFFERENT MEMBER OF THE GOVERNMENT OF FRANCE FROM MY OWN ___WHO IS A COMMUNIST ___A PERSON I KNOW WHO WAS DRAFTED 6 Only Slightly Different RED CHINESE STUDENTS DEMONSTRATING IN MOSCOW IN PROTEST TO RUSSIA'S ___POLICIES TOWARD RED CHINA Can't Decide or Has About the Same SOCIAL WORKER WHO WORKS IN THE Amount of Beliefs ___SLUMS OF NEW YORK OR CHICAGO and Ideas Which 5 Are Like My Own as A PRINCIPAL IN THE SCHOOL SYSTEM Are Different ___WHERE I LIVE From My Own THE POPE, SPIRITUAL LEADER OF THE ___CATHOLIC CHURCH 4 Only Slightly Like NUMBER ONE MAN IN THE NATION-WIDE '___CRIME SYNDICATE, COSA NOSTRA HAS BELIEFS AND ___GRAND DRAGON OF THE KLU KLUX KLAN 3 IDEAS QUITE A BIT LIKE MY OWN COMMUNIST PARTY LEADER OF NORTH ___VIETNAM 2 very Much A MINISTER, PRIEST, OR RABBI THAT Like ___I KNOW PERSONALLY DICTATOR OF A LATIN AMERICAN COUNTRY 203 REPORTER FOR THE OFFICIAL NEWSPAPER OF THE RED GUARD IN COMMUNIST CHINA LEADER OF THE AMERICAN NAZI PARTY, GEORGE LINCOLN ROCKWELL THE PERSON I MARRIED OR AM GOING TO MARRY A YOUNG MAN WHO DEMONSTRATES AGAINST THE WAR IN SOUTH VIETNAM AND BURNS HIS DRAFT CARD GOVERNOR OF A NEIGHBORING STATE THE MEDICAL DOCTOR I MOST OFTEN GO TO A CHAPLAIN IN THE MILITARY A FEDERAL DIRECTOR OF CIVIL DEFENSE EDITORIAL WRITER OF PRAVDA, THE RUSSIAN NATIONAL NEWSPAPER NATIONAL RELIGIOUS LEADER OF MY FAITH MICHIGAN STATE UNIVERSITY STUDENTS PICKETING THE EAST LANSING CITY HALL FOR ALLEGED RACIAL DISCRIMINATION IN HOUSING HEAD OF THE RUSSIAN SECRET POLICE EDITOR—IN-CHIEF OF THE NEW YORK TIMES F - 4 .4 - n “1‘ . HAS BELIEFS AND IDEAS EXTREMELY DIFFERENT FROM MY OWN Very Much Different HAS BELIEFS AND IDEAS QUITE A BIT DIFFERENT FROM MY OWN Only Slightly Different Can't Decide or Has About the Same Amount of Beliefs and Ideas Which are Like My Own as Are Different From My Own FORMER PRESIDENT OF THE UNITED STATES AND GENERAL OF THE ARMY, DWIGHT D. EISENHOWER COMMANDER OF THE NORTH VIETNAMESE COMMUNIST TROOPS FIGHTING IN SOUTH VIETNAM HEAD OF A COMMUNIST FRONT ORGANI— ZATION IN THE UNITED STATES LOCAL RELIGIOUS LEADER OF MY FAITH PRESIDENT OF FRANCE A NETWORK NEWS COMMENTATOR ON TV A PERSON WHOSE OPINION I HIGHLY RESPECT Only Slightly Like HAS BELIEFS AND IDEAS QUITE A BIT LIKE MY OWN Very Much Like HAS BELIEFS AND IDEAS EXTREMELY LIKE MY OWN 204 PRESIDENT OF MICHIGAN STATE UNIVERSITY, JOHN HANNAH REPORTER ON THE NEWSPAPER I REGULARLY READ 1 HAS BELIEFS AND IDEAS EXTREMELY LIKE MY OWN 205 HEAD OF THENATIONAL EDUCATION HAS BELIEFS AND ASSOCIATION 9 IDEAS EXTREMELY "" DIFFERENT FROM HEAD OF THE AMERICAN MEDICAL MY OWN ASSOCIATION COMMUNIST PARTY CHAIRMAN OF "“RED CHINA 8 very Much Different . MY FAVORITE LOCAL TV NEWSCASTER MY FATHER A CONVICTED DOPE PEDDLER HAS BELIEFS AND '7‘“ 7 IDEAS QUITE A BIT A PROSTITUTE DIFFERENT FROM MY OWN SPOKESMAN FOR THE JOHN BIRCH SOCIETY, JOSEPH WELCH STATE DIRECTOR OF CIVIL DEFENSE 6 Only Slightly Different A PERSON WHOSE OPINION I HAVE VERY LITTLE RESPECT FOR Can't Decide or Has MEMBER OF THE COMMUNIST PARTY IN About the Same Amount ____RUSSIA 5 of Beliefs and Ideas Which are Like My GOVERNOR OF THE STATE OF MICHIGAN, Own as Are Different ___GEORGE ROMNEY ___ From Mngwn MEMBER OF THE CITY COUNCIL WHERE ___I LIVE 4 Only Slightly Like A MEMBER OF THE COMMUNIST PARTY IN MY HOMETOWN ___ __ MY BEST FRIEND HAS BELIEFS AND 3 IDEAS QUITE A BIT CHIEF JUSTICE OF THE UNITED STATES LIKE MY OWN SUPREME COURT, EARL WARREN A HINDU HOLYMAN IN INDIA 2 Very Much A POLICEMAN IN MY GDMMUNITY Like ’.h.' an 0‘ i b I <1, .-.. ..‘. ,u n 206 REPORTER ON THE LONDON TIMES, A LEADING BRITISH NEWSPAPER U.S. SECRETARY OF STATE, DEAN RUSK HAS BELIEFS AND 1 IDEAS EXTREMELY LIKE MY OWN F‘". ' A . "J < v3,.:.-- .6 207 Please Complete the Following: Age: Sex: Male Female (please circle) Year in School: Freshman Sophomore Junior Senior Master's Ph.D. Major: Do your parents live in Michigan? yes no (circle one) Have you traveled in a foreign country outside the United States and Canada? yes no (circle one) How do you consider yourself politically? liberal (check one) conservative neither Marital Status: Married Single (circle one) On how many days out of the last seven have you looked at or read a neWSpaper? On how many days out of the last seven have you looked at or read a magazine? On how many days out of the last seven have you watched TV for recreational purposes? On how many days out of the last seven have you listened to radio? On how many days out of the last seven have you looked at or read a book for recreational purposes? How many times in the last month have you attended a movie? Please check the income range which you estimate wouldcxnmnryour parent's total income last year: .___under $1,500 8,000 to 9 999 .__l,500 to 2,999 “10,000 to 14,999 .___3,000 to 4,499 ""‘l5,000 to 19,999 14,500 to 5,999 “‘20,000 to 29,999 .___6:OOO to 7,999 '___30,000 and over What is your father's occupation? _.“'_ "DI - 1 ‘ ~ 208 L f: Department of Communication NAME Michigan State University Student Number We are interested now in what the general public thinks and feels about a number of important social and personal questions. The best answer to each statement below is your personal opinion. We have tried to cover many different and opposing points of view; you may find yourself agreeing strongly with some of the statements, disagreeing just asstrongly with others, and perhaps uncertain about others; whether you agree or disagree with any statement, you can be sure that many people feel the same as you do. Mark each statement in the left margin according to how much you agree or disagree with it. Please mark every one. Write +1, +2, +3, or -1, -2, -3, depending on how you feel in each case. +1: I AGREE A LITTLE -1: I DISAGREE A LITTLE +2: I AGREE ON THE WHOLE —2: I DISAGREE ON THE +3: I AGREE VERY MUCH WHOLE -3: I DISAGREE VERY MUCH Please write both the number and the sign in the margin left of each statement. The United States and Russia have just about nothing in common. It is often desirable to reserve judgment about what's going on until one has had a chance to hear the opinions of those one reSpects. Man on his own is a helpless and miserable creature. In this complicated world of ours the only way we can know what's going on is to rely on leaders or experts who can be trusted. I'd like it if I could find someone who would tell me how to solve my personal problems. The highest form of government is a democracy and the highest form of democracy is a government run by those who are most intelligent. While I don't like to admit this even to myself, my secret ambition is to become a great man, like Einstein, or Beethoven, or Shakespeare. 209 ki- The present is all too often full of unhappiness. It is —only uture that counts. To compromise with our political opponents is dangerous because it usually leads to betrayal of our own Side. Of all the different philosophies which exist in this *world there is probably only one which is correct. In a discussion I often find it necessary to repeat myself several times to make sure I am being understood. _Most of the ideas which get printed nowadays aren't worth the _paper they are printed on. It is better to be a dead hero than be a live coward. _My blood boils whenever a person stubbornly refuses to _admit he' s wrong. There are two kinds of peOple in this world: those who are _for the truth and those who are against the truth. _It is only when a person devotes himself to an ideas or _cause that life becomes meaningful. Most people just don't give a ”damn" for others. The main thing in life is for a person to want to do *something important. Most people just don't know what's good for them. Even though freedom of speech for all groups is a worth- ‘while goal, it is unfortunately necessary to restrict the freedom of certain political groups. 210 APPENDIX C The source evaluation questionnaire used during the main study: includes the instructions with a sample page used to obtain the semantic differential evaluation of the sample of source descriptions. 211 Department of Communication NAME Michigan State University STUDENT NUMBER INSTRUCTIONS You are participating for the second time in a study of how people evaluate other persons. Again, we will ask you to look at the brief descriptions of a number of persons and then mark on a series of scales under each. This time you will be asked to look at fewer descriptions of persons but you will mark each on a number of scales instead of only one. When you evaluate each of the persons described, think of some communication situation in which the person could have some- thing to say to you. This could be a face-to-face conver- sation, a Speech, something the person has written in a letter, newspaper, magazine, or other written document, news film on TV, or some other kind of situation in which the person could communicate something to you. Also, the thing that the person would be communicating should be a reasonable thing for that person to be doing. On each page of this booklet, you will find a different description of a person whom you are to consider as the source of some kind of message or information. You are to judge each with the set of scales beneath it. You are to rate the person on each of these scales in order. An example of such a scale might be: Interesting:~va: -':- :a : z : : : :Boring First, look at the description of the person carefully. Con- sider that person in some kind of meaningful communication situation. Then, if you feel that the person would be ex- tremely interesting, or extremely boring, you should plEEe your checkmarks as follows: Interesting: X : : : :Boring o—OR’ Interesting: : : : : : : : : X :Boring If you feel the person would be very interesting but not extremely so, then: Interesting: : X : : : : : : : :Boring If you feel the person would be fairly interesting, then: Interesting: : : X : : : : : : :Boring 212 If you feel the person would be only slightly interesting, then: Interesting: : : : X : : : : : :Boring The same procedure would be followed if you had thought the person described would be boring rather than interesting, as in this example. The direction you check, of course, depends upon which of the two ends of the scale seem most characteristic of the person you are judging. If you feel that both sides of the scale e uall a l to the person, or if the scale is completely irrelevan¥ an unrelated to the person being judged,7then you Should filace your EhecKL mark in the middle Space: Interesting: : : : : X : : : : :Boring IMPORTANT: (1) Place your checkmarks in the middle of Spaces, not on the boundaries: THIS NOT THIS : X : : : : X : (2) Be sure you check every scale for every person described -- DO NOT OMIT ANY (3) Never put more than one check-mark on a single scale Every description of a person in this booklet was used when you participated previously. But, not all of those you rated before are being used this time. Please do not look back and forth through the booklet. Do not try to remember how you checked similar items earlier. Make each item a separate and independent judgment. Work at fairly high Speed through the booklet. Do not worry or puzzle over individual items. It is your first impressions, the immediate "feeling" about the people described, that we want. On the other hand, please do not be careless, because we do want your true impressions. Every description of a person used in this booklet is listed on the page following this one. Before you start marking the scales for the first description, please glance over all of the descriptions listed on the next page so you will remember the kinds of persons you will be asked to evaluate. Then go on to the first description of a person and mark the scales for him. Please work your way through the booklet fully com- pleting each page in order. Fob...‘ ’Al ’,._o.l . ... 'l.-> u: 213 LOCAL RELIGIOUS LEADER OF MY FAITH ACCUSED KILLER OF EIGHT STUDENT NURSES IN CHICAGO, RICHARD SPECK LEADER OF THE AMERICAN COMMUNIST PARTY A WHITE PERSON WHO DEMONSTATES FOR NEGRO CIVIL RIGHTS A PERSON WHOSE OPINION I HAVE VERY LITTLE RESPECT FOR COMMANDER OF THE NORTH VIETNAMESE COMMUNIST TROOPS FIGHTING IN SOUTH VIETNAM MICHIGAN STATE UNIVERSITY STUDENTS PICKETING THE EAST LANSING CITY HALL FOR ALLEGED RACIAL DIS- CRIMINATION IN HOUSING MY BEST FRIEND HEAD OF THE RUSSIAN SECRET POLICE SOCIAL WORKER WHO WORKS IN THE SLUMS OF NEW YORK 0R CHICAGO A CONVICTED DOPE PEDDLER HEAD OF A COMMUNIST FRONT ORGANI- ZATION IN THE UNITED STATES A TEACHER IN THE SCHOOL SYSTEM WHERE I LIVE MEMBER OF THE GOVERNMENT OF FRANCE WHO IS A COMMUNIST A NETWORK NEWS COMMENTATOR ON TV MEMBER OF THE COMMUNIST PARTY OF RED CHINA PRESIDENT OF MICHIGAN STATE UNIVERSITY, JOHN HANNAH SPOKESMAN FOR THE GROUP OF MICHIGAN TEACHERS PICKETING THE STATE LEGISLATURE FOR MORE FINANCIAL SUPPORT TO EDUCATION LEADER OF THE AMERICAN NAZI PARTY, GEORGE LINCOLN ROCKWELL REPUBLICAN SENATE LEADER AND SENATOR FROM ILLINOIS, EVERETT DIRKSEN MEMBER OF THE COMMUNIST PARTY IN RUSSIA MY MOTHE R COMMUNIST PARTY LEADER OF NORTH VIETNAM A PERSON WHOSE OPINION I HIGHLY RESPECT NATIONAL RELIGIOUS LEADER OF MY FAITH CONVICTED KILLER OF LEE HARVEY OSWALD, JACK RUBY EDITORIAL WRITER OF PRAVDA, THE RUSSIAN NATIONAL NEWSPAPER NUMBER ONE MAN IN THE NATION-WIDE CRIME SYNDICATE, COSA NOSTRA I *4 «M Q, gt..- a 214 GOVERNOR OF A SOUTHERN STATE CUBAN PRIME MINISTER REPORTER FOR THE OFFICIAL NEWSPAPER OF THE RED GUARD IN COMMUNIST CHINA A PERSON I KNOW WHO WAS DRAFTED SPOKESMAN FOR THE JOHN BIRTH SOCIETY, JOSEPH WELCH THE PERSON I AM MOST LIKELY TO GO TO FOR PERSONAL ADVICE A MINISTER, PRIEST, OR RABBI THAT I KNOW PERSONALLY COMMUNIST PARTY CHAIRMAN OF RED CHINA GOVERNOR OF THE STATE OF MICHIGAN, GEORGE ROMNEY CONVICTED GERMAN WAR CRIMINAL AND ONE TIME ADOLF HITLER'S DEPUTY, RUDOLF HESS GRAND DRAGON OF THE KLU KLUX KLAN MEMBER OF THE COMMUNIST PARTY IN ENGLAND THE PERSON I AM LEAST LIKELY TO GO TO FOR PERSONAL ADVICE THE PERSON I MARRIED OR AM GOING TO MARRY MY FAVORITE LOCAL TV NEWSCASTER REPORTER ON THE NEWSPAPER I REGULARLY READ CHIEF JUSTICE OF THE UNITED STATES SUPREME COURT, EARL WARREN SECRETARY GENERAL OF THE UNITED NATIONS NW'FATHER THE MEDICAL DOCTOR I MOST OFTEN GO TO Its—.M l . 1 ‘ ' 215 A MINISTER, PRIEST, OR RABBI THAT I KNOW PERSONALLY unselfish: : : : : : : : : selfish dangerous: : : : : : : : : :safe inexperienced: : : : : : : : : :experienced harsh: : : : : : : : : :gentle friendly: : : : : : : : : :unfriendly aggressive: : ° ° : : : : : :meek energetic: : : : : : : : : :tired authoritative: : : : : : : : : :unauthoritative unjust: : : : ° ° : : :just inhOSpitable: : : :____ : : : : :hospitable bold: : : : : : : : : :timid hesitant: : : : ° : ' : : :emphatic cool: : : : : : : : : :warm untrained: : : : - : : : :trained intelligentg___g___: : : : : : : :unintelligent unpleasant: : : : : : ° :pleasant reserved: : : : : : : ' : :frank slow: : : ° : : : : : :fast able: ¥_: : : ° : : : ° :inept cheerful: : : ° ' : : : :gloomy calm: : : ° ° : : : : :upset WM ., . ‘1 r {-.A. . informed: ethical: patient: congenial: fair: agreeable: skilled: active: forgiving: kind: : qualified: unsociable: forceful: 216 A MINISTER, PRIEST, OR RABBI THAT I KNOW PERSONALLY :uninformed :unethical :impatient :quarrelsome :unfair :disagreeable :unskilled :passive O. O. O. O. :unforgiving :cruel :unqualified :sociable :forceless 217 APPENDIX H Overall Factor Analyses of the Source Evaluation Data _ ::fl— ll. \ f I Pun“: 218 The question might be asked: Would the present source evaluation data produce essentially the same results and con- clusions if it were analyzed in the same way Berlo analyzed his data? The following three tables of factor loadings present the results of such an analysis. Three overall factor analyses were performed for this data of the type Berlo used for his data. These included one for all sources, and one each for the 24 belief sources and the 24 disbelief sources. The analysis procedure was identical to Berlo's. Unities were used in the diagonals of the correlation matrix. The principal axis model was used followed by a varimax rotation. The criterion for stopping the rotation procedure was the Kiel- Wrigley criterion set at three. For all three factor analyses, the stOpping point was the three factor solution, identical to Berlo's results. (See Appendix E for the three factor solution resulting from Berlo's analysis.) The grouping of scales on all three factor solutions is nearly identical to Berlo's. The only exception is the unauthoritative-authoritative scale. For the 48 sources and 24 disbelief sources solutions, this scale's primary loadings are split between the qualification and dynamism factors instead of being primarily on only the qualification factor. One minor difference is that for all three of the solutions, the variance accounted for by the qualification and dynamism factors is more nearly equal than 219 it is for the Berlo solution. The two solutions involving only 24 sources have a qualification factor accounting for about the same amount of variance as Berlo's solution. For the solution involving all 48 sources, the qualification factor is somewhat larger than for the other solutions. The two 24 source solutions and the Berlo one accounted for about slightly more than 60% of the total variance. The 48 source solution accounted for about 10% more variance, primarily due to the larger qualification factor. The Berlo solution was compared with the three solutions from this study on a factor to factor basis using the index of factorial similarity. The results are in the following table. It is obvious that all three solutions map into the Berlo results very satisfactorily. When analysis of this study's source evaluation data is made in the same way that Berlo analyzed his data, the results of this study would very satisfactorily replicate the results of Berlo's study. NH HH OH mo mo --- wmm. m:a.- 0mm. Oms. 220 --- mam. mam..mmm. III mmH. MHm. --- nae. 50 :mo. pom. mmm mmm. Him. mo mo ems. OOH. Hmm. mmm. Hmo.uwmd. mmww. mmz. ems..mmm. HmH. HHm. --- was. :0 N00.1 AHA. mmm; mmm. mo anWJ ems. Odo. WWW. mm:. NO 83:. Mmmv mm:. HH OH mo mo no 00 mo :0 mo NO Ho EmHEOChQ mOomsom .MHHODQ MOHHOQ Imam semadm HO Emfiswcmm .CHHOSG mOopsom mOHHOm spmmam HO EOHSOGHQ .HHHODG OOOHsom m: Armada HH< Emflsmchm .wHHOSG mpOwOm OHMOm 01 02 O3 O4 05 O6 O7 08 09 10 11 12 13 l4 15 16 17 l8 19 2O 21 22 23 24 25 26 27 28 29 30 31 32 33 221 Three Factor varimax Rotated Solution for All Sources (34 variable, 4752 observations) I--Safety II-—Oualification III--Dynamism selfish---unse1fish cruel---kind unfriendly---friendly dangerous---safe unpleasant---pleasant quarrelsome---congenial harsh--—gental cool---warm disagreeable---agreeable unforgiving---forgiving unjust---just gloomy---cheerful inhospitable---hOSpitable upset-—-calm unsociable--—sociable unfair-—-fair inpatient---patient unethical---ethica1 inexperienced---experienced untrained---trained unskilled---skilled unqualified---qualified uninformed---informed unauthoritative---authoritative inept---ab1e unintelligent---intelligent meek---agressive reserved---frank hesitant---emphatic forceless---forceful passive---active timid---bold tired---energetic slow---fast Proportion of Total variance Proportion of Common variance I II 80 18 92 15 89 16 86 16 9O 21 85 19 88 13 84 ll 86 19 88 15 88 23 81 22 86 18 64 37 81 21 89 2O 4 27 6 2 24 7 26 82 31 80 41 76 4O 69 -O3 47 41 73 45 69 —25 08 06 00 O4 23 -12 28 O7 24 -20 O9 18 25 —Ol 12 41.0 15.4 58.0 21.8 III ~02 14.3 20.2 70-7 100.0 FACTOR PURITY* 222 Three Factor varimax Rotated Solution for the 24 Belief Sources (34 variable, 2376 observations) I—-Safety II-—Qualification III--Dynamism 01 O2 03 O4 O5 06 O7 08 09 10 ll 12 13 l4 l5 l6 17 18 19 20 21 22 23 24 25 26 27 28 29 3O 31 32 33 34 selfish---unselfish cruel---kind unfriendly---friendly dangerous---safe unpleasant---pleasant quarrelsome---congenial harsh---gental cool---warm disagreeable--—agreeab1e unforgiving---forgiving unjust---just gloomy---cheerful inhOSpitable---hospitable upset---ca1m unsociable---sociable unfair--—fair impatient---patient unethical---ethical inexperienced---experienced untrained---trained unskilled---skilled unqualified---qualified uninformed---informed unauthoritative---authoritative inept---able unintelligent---intelligent meek---aggressive reserved---frank hesitant---emphatic forceless-—-forceful passive—--active timid---bold tired---energetic slow---fast Proportion of Total variance Proportion of Common variance *The factor purity index is the ratio of the squ variable's highest loading to its communality (h I II 65 13 82 15 79 18 64 30 80 30 75 23 78 12 73 O9 74 23 77 17 69 36 74 15 77 16 55 39 67 20 70 31 64 31 65 38 3O 79 26 81 29 80 4O 78 33 66 O5 52 46 67 38 67 -O6 09 -Ol 03 11 24 O7 37 22 19 -O6 O9 33 17 14 10 31.9 16.3 51.5 26.3 III h2 10 45 12 71 22 70 -Ol 51 14 75 00 62 -10 64 09 54 03 61 05 62 2O 64 2O 62 20 65 —ll 47 27 57 22 63 -14 53 19 60 16 72 15 74 19 76 18 80 31 63 38 41 26 72 24 65 8O 65 63 39 65 49 65 57 71 58 78 63 61 51 63 42 13.7 61.9 22.2 100.0 FACTOR PURITY* 93 94 89 80 85 90 re of the a). 223 Three Factor Varimax Rotated Solution for the 24 Disbelief Sources 01 O2 03 O4 O5 06 O7 08 09 10 ll l2 13 14 15 l6 l7 l8 19 20 21 22 23 24 25 26 2 25 29 3O 31 32 33 34 (34 variable, 2376 observations) I--Safety II--Qualification III--Dynamism selfish---unselfish cruel---kind unfriendly---friendly dangerous---safe unpleasant---pleasant quarrelsome---congenial harsh---gentle cool---warm disagreeable---agreeable unforgiving---forgiving unjust---just gloomy---cheerful inhOSpitab1e---hOSpitable upset---calm unsociable--—sociable unfair--wfair impatient---patient unethical---ethical inexperienced---experienced untrained---trained unskilled---skilled unqualified---qualified uninformed---informed unauthoritative---authoritative inept---able unintelligent---intelligent meek---aggressive reserved—--frank hesitant---emphatic forceless---forceful passive—--active timid---bold tired---energetic slow—--fast Proportion of Total variance jProportion of Common Variance *The factor purity index is the ratio of the square of the 'variable's highest loading to its communality (h2), I 64 85 82 71 84 77 80 74 78 8O 29 —22 -O2 -06 -14 -04 -18 01 -11 31.0 15.4 15.8 II 18 07 1 III 49.9 24.7 25.4 67 l 54 1 62.2 100.0 FACTOR PURITY* 89 97 00 93 98 95 92 00 98 96 92 9O 96 60 94 96 76 L1 224 APPENDIX I Additional Comments on the Generality of the Berlo Dimensions Across Subjects 225 The emphasis of the conclusions summarized in this section was on the relative degree of generality of the Berlo dimensions across the two kinds of sources. They do not take into account the level of the degree of generality. If the level of generality is examined, one might be able to make some additional comments on the generality across the people of Berlo's three dimensions in the individual dimensions of Ource evaluation. It would seem reasonable that in comparing the in- dividual three factor rotated solutions to the Berlo solution that if no more than half of the individual solutions exhibit the characteristic under consideration, the generality across people of that characteristic of the Berlo dimensions of source evaluation was at an unacceptable level. The con- clusions in the last section can be re—examined applying this gross standard. There were a number of comparisons involving the degree to which all three Berlo factors were represented in the individual three factor solutions. None of these exceeded the gross standard of fifty percent of all solutions for either belief of disbelief sources. One com- parison approached this but did not exceed it. For belief sources only, forty-nine percent of all individual solutions had factors which could be characterized as each repre- senting one of the Berlo dimensions (see Table 45). The 226 similar percentage for disbelief sources dropped to thirty percent. One can not infer a very high degree of generality on a multi-dimensional basis of the Berlo results across the persons in this study. For certain other similar comparisons, the generality across people is even lower. There were a number of comparisons using each Berlo dimension one at a time. The generality across peOple was highest for the safety dimension (78%). This held for both kinds of sources. For the other two dimensions (qualification and dynamism), it was only for belief sources that the gross standard of fifty percent was exceeded (63% and 69%,reSpec- tively). For disbelief sources, the generality across peOple dropped to unacceptable levels for both qualification and dynamism (45% and 39%, respectively). Even comparisons in- volving a single dimension at a time, generality across peOple is lacking in places. Certainly Berlo's overall results based on a factor analysis across all persons in his sample and all sources was not an adequate summarization of the individual factor analysis results obtained by factoring data individually for each subject in this study. Would an overall factor analysis across all persons in this sample across all sources be an adequate summarizations of the individual factor analyses? Such an overall factor analysis was done for this data. It is reported in Appendix H. One was done for all sources as E 227 A, well as one for each of the two kinds of sources. The results of all three analyses were not interpretively different from the Berlo analysis. The implication certainly would be that the overall factor analysis of a set of data of the kind obtained for this study is not necessarily highly repre- sentative or constitutes a very meaningful average of the factor analyses done separately on the data that was obtained for each subject. Such an analysis is planned for the data in this study. The primary purpose for doing the overall factor analysis is for index construction. In other words, what snales should be summed over to produce a small set of indices which are relatively uncorrelated and exhaust most of the variance. There always has been a tendency with the semantic differential studies to make the leap that the overall factor structure is somehow representative of the cognitive structure of the persons participating in the study. This study cer— tainly contains implications for questioning the tenability of such a leap. MICHIGAN 5m: UNIV. LIBRARIES 1|11111111111111111111111111111 31293101819533