SPECIFIED COGNITIVE STRUCTURES AND THEIR EFFECTS ON LANGUAGE ENCODING BEHAVIORS Thesis for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY HERBERT WAYLAND CUMMINGS 1970 This is to certify that the thesis entitled LIBRARY ‘ Michigan State University SPECIFIED COGNITIVE STRUCTURES AND THEIR EFFECTS ON LANGUAGE ENCODING BEHAVIORS Date 0-169 presented by Herbert Wayland Cummings has been accepted towards fulfillment of the requirements for an... Mic-:5:— May 25, 1970 Major professor I BINDING av HUNG & SUNS' BWK BINUENY INC. LIBRARY amps Rs smumn Inn-Ian :11. I o.I‘.4IO‘ ‘ '1" .l. .‘ The It'l’p as ,, ‘9 angu‘wr. 4:. . Wactc’IM. o Inltudv'. . }. I‘. @1105»: ID"! . .' ‘ ‘ h‘..l”1l ruse-"Mr! .-; .4, ,. , . H W of Ind-)pr-n's...“ nIr-.., .3 gr -. .» , - "I n ddltirn, u. . v.2 I; T"~Watructt-I; cv- dmawrnmm I-,_-_~ .- '_ ~ ‘ . lax-«1? m L‘u‘i'f'u. .;.r.-;:. . 2', 5?» ” OOMtth" alturnre. a: w ;_=th 1“.- "lm LNG MIJ’vdcloqy ‘5 a “as“. . _... y 1 I ”I. o! amt! mum lul)0~".: we» "in, I m Mud on m source. 1:: { m an” “0 Wood tree‘- a o! W unclean... thou «:qu at“ I. M chattel 1- mi '« ' '-‘« ABSTRACT SPECIFIED COGNITIVE STRUCTURES AND THEIR arms on LANGUAGE ENCODING BEHAVIORS BY Herbert Wayland Cummings " The purpose of this study was to examine the '.. relationship between certain message variables and inde- ,iimdent characteristics of sources, such as dogmatism, f.Tmi§xiety, attitude, vocabulary skills, spatial abilities, .;~:tant analysis research typically has analyzed messages a h-«'tha absence of independent measures of source charac- .r y’iistics. In addition, such research has been directed I a.\ j? 15%" ' - the construction of dictionaries, which has limited -4“‘1iathodology largely to lexical analysis. This study ¢ >a i;‘ gaosigned to construct alternatives to previously {I U "tied problems using the methodology of content I7 ‘ _. . £8.15"- 'fl¥§”;:‘h random sample of ninety-eight subjects was taken, . 7. {.4 rflmdspandsnt measures obtained on each source. In iation of structural linguistics. These variables =;-tua11y related to extant theories in semiotics Herbert Wayland Cummings ii ‘Al‘, i and semantics. Of these message variables, a subset of forty were evaluated as appropriate for analyses of vari- ance. Seventy per cent of these forty variables yielded significant information about one or more characteristics 'of the sources. The syntactical descriptive system offered in this study had three basic elements: (1) limiters (modifiers); (2) subject words (verb complements); and (3) connectors (verbs). Other distinctions made were the relative fre- quency of occurrence of: (1) past, present, and future tense verbs: (2) associative and dissociative assertions; .' - (3) action and comparison verbs; (4) indicative and sub- *5 j‘fl junctive verbs, and (S) primitive and defined subject "signs and connectors. Each of these variables were ana- -‘.,lysed in terms of how they related to cognitive and demo- :L:figraphic characteristics of the encoder. 'QU . This system allowed for the possibility of sub- Quling a semantic approach to the study of meaning into a p fiflhtactical theory of signs. The study suggested that the Q§3R1ss distinctions made by previous researchers in content I23 ysis have been unfruitful or inconsistent in part be- , they have failed to utilize a highly developed ification system. . Accepted by the faculty of the Department of cations, College of Communication Arts, Michigan ,g._ , University, in partial fulfillment of the require- 1 .‘., ..~. .‘ I _ts for the Doctor of Philosophy degree. ‘ ( DIrechr of Thesis (7 Guidance Comittae : , Chairman SPICIPIBD COGNITIVE STRUCTURES AND THEIR srncTs 0N LANGUAGE ENGODING ssanvmss 3:! Herbert Wayland Cummings Ant II . libutod : fzu . Z‘ .1 ("‘me total-i; V. " A I to bc~ A THESIS Submitted to Michigan State University “-4“ partial fulfillment of the requirements for the degree of .. m . sector or ssnosopmr Deperuent o! Ce-Inicetion {I , 191a . :4 ‘. . u ACKNOWLEDGMENTS ‘ Scholars need intellectual stimulation, but no - "lr1 . ' , ‘, ~lslllan being is all intellect. He is also a human being. f s a hulanness has been maintained through graduate school "by my wife, Patricia Ann, who has had an unlimited reser- I P {voir of love, encouragement, and patience. . . For the intellectual stimulation that has made IL'I ~"7‘_'ald Miller, William Lashbrook, and Murray Hewgill. I" _ s school meaningful, I must thank Erwin Bettinghaus, L I And finally, my sons--Gene, Ron, and Dom-who ,u b -4 el I a part of themselves with the sacrifice they when their father was too tired to be as human as he to."- Chapter I. . I P‘.’ TABLE OF CONTENTS WI“ . O I I O C I I O I O I BACKGROUND AND THEORETIC APPROACHES . . Content Analysis Defined . . . . . Theoretic Approaches: Summary of Literature. . . . . . . . . Theoretic Approaches: General Rationale and Measurement . . . . . . . . amt! O O O O 0 O O s C O O RATIONALB AND HYPOTNESBS. . . . . . The Problem of Meaning 9. . The Problem for a Syntactical Approach to the Meaning of Meaning. . . a Descriptive Classification of Signs. Summary and Permutations of Message variables . . . . . Cognitive Variables, Demographic Characteristics . . . . . . . . Dogmatism . . . . . . . . . miety O O O O I O O 0 O O Attitude. . Spatial Relationships and Vocabulary 8‘11. 0 O O O O s O s O w..e. 0 ~ 0 O O O O 0 O O I C9nclusion. . . . . . . . . . IETHOD OP ANALYSIS. . . . . . . . Sampling Procedure, Description. . . iii Page Page Attitude Concepts . . . . . . . . 54 Sex, Level, and Age . . . . . . . 61 schwl I I I I I I I I I I I I 62 Coding and Scoring of Message Date . . . 62 stage 1 I I O I I I I I I I I 63 stage 2 I I I I I I I I I I I 63 Stage 3 . . . . . . . . . . . 64 sage ‘ I I I I I I I I I I I 64 stage 5 I I I I I I I I I I I 64 stage 6 I I I I I I I I I I I 65 Saga 7 I I I I I I I I I I I 65 Internal Consistency . . . . . . . 66 Derived Variables. . . . . . . . . 67 Analysis. . . . . . . . . . . . 67 Summary . . . . . . . . . . . . 72 ”8st I I I C I I I I I I I I I 73 Afferent-Efferent Distinctions . . . . 73 hypothesised Relationships. . . . . . 74 Unhypothesized Results . . . . . . . 81 Dmti'm. I I I I I I I I I I 82 Anxiety . . . . . . . . . . . 83 Spatial Relationship Abilities. . . . 83 Vocabulary Skills . . . . . . . . 83 Attitude Direction. . . . . . . . 84 Attitude Inten‘ity I I I I I I I I 85 “val I I I I I I I I I I I I 86 schml I I I I I I O I I I I I 86 Errors in Spatial Abilities and vocabulary Skills . . . . . . . . 86 Sex and Age . . . . . . . . . . 87 sun I I I I I I I I I I I I 87 DISCUSSION. . . . . . . . . . . . 89 Summary of Rationale. . . . . . . . 89 Interpretation of Results . . . . . . 92 ', ‘ m1 1 e e e e e e e e e e e 9': J 5. VI]. 2 e e e e e s e e e e e 10 )1 if. ‘7‘ 1 3 e e a e e e e e e I e 10‘ ‘2, I, ‘ m1 ‘ e s o e e e e e e e e 101 " "I ~.‘\ .I iv Page lvaluation as a Syntactical Theory of at”. I I I 115 Recon-andations for Further Study. . . . 119 '."' mm“ o e a e o s s . o a e s s s s 12 1 EAEIIIIDICIB he “m'm a s o o e e s e s e o a 125 ',..‘\. h.*mnmsu_nrynata. . . . . . . . 129 c. aquomirel e e e a a s e e e s s 158 ‘. 3‘. 6‘” 1 ." II -' A I 1 ' ; .I. Sunuw-v. ‘ P‘ 1'35. Senna: ' ' ' Hyr"~ . 3- ‘. bums , "If. \T - ‘ Q av? ‘ M51“. 0.. -3 \ . \I‘Y'qusnfr begun: my Varia" . lnxxecy; 5" "-. . Variance. .h‘_ ' *k“.m10tY: :‘m' s. Varmint: .. , , . tism .mcn- . ., A rid :‘k‘.‘r:1MLJZs‘:L'.". Wt}; ' ty: SW3!" c: i...» v.»- \L .1 ~- :: igflgflance; NeguLi-‘s $‘r . . - - ' 7W‘.‘ . ’ ; —, _ . 2’ -,- .... Intensity; Sana-an n: ' .0 rats of furnace: 7». . . w .; (81D: . . . . LIST OF TABLES Page Summary of Message Variables . . . . . . 39 Summary of Frequency Distribution: Attitude Direction. . . . . . . . . . . . 60 Summary of Frequency Distribution: Attitude Intensity. . . . . . . . . . . . 60 Summary of Derived Variables . . . . . . 68 Summary of Descriptive Statistics: aypothesized Variables . . . . . . . 75 Summary of Correlations: Hypothesized vuiable'. I I I I I I I I I I I 76 Dogmatism: Summary of One-Way Analysis of Variance: Present Tense Connectors (ClPR). 77 Dogmatism: Summary of One—Way Analysis of Variance: Future Tense Connectors (ClFU) . 77 Anxiety: Summary of One-Way Analysis of Variance: Present Tense Connectors (ClPR). . 77 Anniety: Summary of One-Way Analysis of Variance: Future Tense Connectors (ClFU) . 78 Dogmatism: Sumary of One-Way Analysis of . variance: Negation Connectors (NCl). . . 78 Anxiety: Summary of One-Way Analysis of .‘ variance: Negation Connectors (NCl). . . 79‘ .' Attitude Intensity: Summary of One-Way Analysis of Variance: Defined Subject "Old! (81D) 0 s e s o a e s s s e 79 vi 14. 15. 16. 17. 18. 19. 20. 21. I 22. 23. 3-1. 3-2. Attitude Intensity; Summary of One-Way Analysis of Variance: Defined Connectors (C1D) e e e I e e e s e e e s a Spatial Reasoning; Summary of One-Way Analysis of Variance: Comparison Connectors (COMP) . . . . . . . . . Verbal Skills; Summary of One—Way Analysis of Variance: Subject Words (51) . . Verbal Skills; Summary of One-Way Analysis of Variance: Limiters (L) . . . . . Intercorrelation Matrix of Level 1 Variables. Intercorrelation Matrix of Level 2 Variables. Limiters. . . . . . . . . . . . Intercorrelation Matrix of Connectors . . . Intercorrelation Matrix for Afference . . . Intercorrelation Matrix for Efference . . . Summary of Descriptive Statistics . . . . Summary of Simple Correlations: Message Variables with Cognitive Variable. . . . Summary of Analysis of Variance: Independent Variable: Sex . . . . . . Independent Variable: School . . . . . Independent Variable: Level . . . . . Independent Variable: Dogmatism . . . . Independent Variable: Anxiety. . . . . Independent Variable: Vocabulary Skills . Independent Variable: Spatial Errors . . Independent Variable: Vocabulary Errors . Independent Variable: Spatial Abilities . Independent Variable: Attitude Direction . Independent Variable: Age . . . . . . Independent Variable: Attitude Intensity . Page 80 80 81 81 95 102 106 109 110 110 129 131 134 136 138 140 142 144 146 148 150 152 154 156 Figure . 1. Frequency Topic 1 Frequency Topic 1 PrctuBSY Topic 2 Frequency Topic 2 Frequency Topic 3 Frequency Topic 3 Frequency Topic 4 Frequency . Topic 4 ug., Frequency ,' Topics ‘~1:?": iFriquency "+.L. ghn 99P1¢ 5 ‘1 , ,‘4 LIST OF FIGURES Distribution: Distribution: Distribution: Distribution: Distribution: Distribution: Distribution: Distribution: Distribution: Distribution: Attitude Attitude Attitude Attitude Attitude Attitude Attitude Attitude Attitude Attitude 5_Summary of Level 1 Interpretation 'fi'flp. summary of Level 2 Interpretation .gf: JSFFEEIY of Level 3 Interpretation Vq';h!gmmary of Level 4 Interpretation .Ng , viii Page Direction . . . . 55 Intensity C I e e 55 Direction . . . . S6 Intensity I I I I 56 Direction I I I I 57 Intensity . . . . 57 Direction I I I I 58 Intensity I I I I 58 Direction I I I I 59 Intensity e e e e 5.9 I I I I log I e e e - ‘ . 165 e e e e 'v_ 1598' , 114 INTRODUCTION This study explores the relationship between certain specified message characteristics and selected antecedent characteristics of the source. Under the rubric of psycho- linguistics and the methodology of content analysis, this study seeks to systematically and objectively identify some subset of 100 message variables that will explain cognitive characteristics of the source. To facilitate this study, a random sample of 114 subjects was taken from the introductory speech course at Grand Rapids Junior College in the Fall of 1969. Each sub- ject was asked to fill out five instruments which became the criterion measures for this study. Three weeks later, these same subjects were asked to write a speech--a11 subjects writing on the same topic--choosing any position on the topic they wish. Sixteen subjects were lost, due to absence from the classes when the second stage of the study was con- ducted. Criterion measures used in this study included dog- matism, communication anxiety, attitude, vocabulary skills, and spatial relationship abilities. In addition, four r demographic variables were obtained. These included sex, class, level, age, and school. Independent variables were word index scores based on the frequency of occurrence of 100 message characteris- tics for each subject. Examples of the most general message variables considered were the relative percentage of occur- rence of: (1) past, present, and future tense verbs; (2) associative and dissociative assertions; (3) action and com- parison verbs; (4) indicative and subjunctive verbs, and (5) primitive and defined subject signs and connectors. Based upon previous research in the field, it was believed that verb tenses would explain significant vari- ance in the criterion measure of dogmatism. It was also believed that dissociative assertions would eXplain sig— nificant variance in dogmatism, while defined verbs and signs would explain significant variance in attitude inten- sity. The structure for this study is as follows: Chapter I presents a summary of the literature and theoretic ap- proaches in content analysis; Chapter II includes the ra— tionale and hypotheses for the study; Chapter III offers the method of analysis used; Chapter IV includes the results, and Chapter V presents a summary, discussion, and sugges- ' tions for further study. CHAPTER I BACKGROUND AND THEORETIC APPROACHES Content analysis is a multipurpose research method designed to investigate numerous problems in which the data for inference is the content of the communication exchange. This chapter examines content analysis as a method, followed by a review of the literature for three general theoretic approaches used in content analysis. There have been prob- ably thousands of studies using some form of content analy- sis, but many seem to have had little or no theory as a basis for their use. The studies reviewed here have had significant theoretic foundations underlying them which are important to this study. Included in this chapter will be a discussion of some of the general rationales associated with content analytic studies, and issues related to the .measurement procedures used in content analysis. Content Analysis Defined Ole R. Holsti (1969) defines content analysis as . . . any technique for making inferences by systemat— ically and objectively identifying specified character- istics of messages (p. 601). Holsti defends this definition as being less restrictive than those definitions held by others, thus allowing for more inclusiveness of relevant literature. While one can quarrel with any definition as being too broad so as not to define anything, the definition provided by Holsti provides useful distinctions. The method of content analysis has generated differences in the field, based on distinctions made between quantitative and qualitative approaches (Holsti, 1969). Pool (1959, pp. 8-9) provides examples of this distinction. He describes typical studies of qualita- tive analysis as being preliminary readings of communication materials for purposes of hypothesis formation. Another example is the study which makes dichotomous distinctions between attributes which are said to belong or not to belong to an object. He illustrates the quantitative approach as including studies which are a systematic analysis of mes- sages for the purpose of testing hypotheses. Another exam- ple would be a study which looks at attributes of a message which can be rank-ordered, or which permit interval measure- ment. In general, qualitative analysis has depended upon the absence or presence of attributes of messages which were artistically intuited. Pool (1959) argues: It should not be assumed that qualitative methods are insightful, and quantitative ones merely mechanical methods for checking hypotheses. The relationship is a circular one; each provides new insights on which the other can feed (p. 192). While Fool's assertion may seem on the surface to be valid, the same argument might be used for the justification of A the distinction between ”humanistic” and "scientific" means of knowledge acquisition. It is true, of course, that in- sightful methods are used in quantitative approaches, but they are made a priori, not concurrent with the analysis of the data. ( Theoretic Approaches: Summary of Literature Bolsti (1969) reports that three-fourths of the studies utilizing content analysis have been primarily in three disciplines: (1) sociology/anthropology; (2) polit- ical science, and (3) general communication. Although these disciplines are not mutually exclusive, it is possi- ble to structure the literature from these disciplines with Bolsti's category scheme. It is also helpful in denoting common assumptions and problems held by all three classes of scholars. 1. Sociology/anthropology. Much of the research in this area is concerned with the comparison of cultural norms, or the change of norms over time as represented by a content.analysis of certain messages. Some of the best known researchers have looked at themes of folktales of a given culture, or between cultures. Kalin, Davis, and McClelland (1966) were interested in the effect of the general consumption of alcohol and frequency of drunkenness on folktale themes of differing cultures. McClelland (1958), HeClelland and Friedman (1952), and Riesman, Glaser and Reuel (1950) looked at the effects of need for achievement as indexes of cultural change. Dahlberg and Stone (1966) were interested in authoritarianism as a determinant of message characteristics between cultures. The common theme of these scholars is a content analysis of literature and speeches as indicators of cul- turally important characteristics. Few studies have ob- tained direct measures of source characteristics as a means of validating inferences made from the messages. This has been due to the inaccessibility of sources or the huge energy requirements necessary in obtaining samples large enough to generalize to a whole culture. Representative of much of the work of these re- searchers is that of Benjamin Colby (1966), who developed an anthropological dictionary for use with the General In- quirer (Stone, Dunphy, gt_al., 1966). His dictionary was based on data obtained from folktales of Navajo and Zuni Indians. After considerable preliminary work, Colby devel- oped his Third Anthropological Dictionary, which grouped its tags under five main sections: (1) plot structure; (2) behavioral systems; (3) mental processes; (4) analytic—ex- perimental, and (5) second-order tags. The first section deals with characteristics in the plot and action of those characters. Plot characteristics include sex tags, kinship tags, and pronoun tags. Action tags indicate orientation and movement in space and time, communication and work. The second section--behaviora1 systems--deals with personality, including body, emotions and abilities. A social dimension is also included, which subsumes aspects of social relations. A culture dimension includes values and norms. Mental processes--a third main section of the dic- tionary--inc1udes tags of three main subclasses--explana- tion, description, and perception. The fourth section has tags which look at the connective words related to logical structure, while the fifth section refers to those 3,600 selected words in the Harvard III Psychosocial Dictionary (Stone, Dunphy, g£_§1., 1966). Twenty-seven subjects in Colby's study gave verbal responses to the TAT protocols, 15 Navajo subjects and 12 Zuni subjects. Results of the study, Colby says, "tend to demonstrate" that the TAT texts of Zunis are more concerned with moisture than those of Navajos. This was explained-- ex post facto by Colby--as being related to the primary oc- cupations of Zunis as crop growers as distinct from Navajos who were sheep-herders. On the other hand, Navajos who have a cultural theme of travel and movement according to Colby, encoded significantly more travel words than Zunis (p < .05). Navajos also produced more affection words, especially between husband and wife, while Zunis encoded more social power and dominance words. It is significant that the common concern of these researchers is the identification of social and cultural A r antecedent conditions which explain differences in messages as response variables. Some anthropologists who have been psychologically oriented have centered their research on the relationships between culture and personality. The next major section of this chapter--Genera1 Rationale and Meas- urement--wi11 elaborate on this significance. 2. Political Science. The antecedent conditions of messages which are of primary interest to political scientists are those of political party membership. The message may be created by one person, or a group of persons. At times, this group is interested in identifying who wrote a message (Mosteller & Wallace, 1964) or determining salient political themes over time (Smith, Stone, & Glenn, 1966; Namenwirth, 1969). Blumberg (1954) sought to measure political bias in editorials in a presidential election, while Namenwirth and Brewer (1966) were interested in com- paring messages created by editors in different countries. Leites and Pool (1949) also looked at the frequency of oc- currence of political symbols in messages created by edi- tors. The basic concern of the political scientist in us- ing content analysis is: Is there some class of tags in a message which will indicate the identity of a politician and/or the political bias of an encoder? The corpus of messages includes political speeches, newspaper stories and editorials, and political pamphlets. Ole Holsti (1966) provides an illustration of the research concerns of the political scientist. He studied F the political relationships between the Soviet Union and Red China. Using theoretic formulations from balance theory, i.e., Heider (1946), Newcomb (1953), and Osgood, Suci, and Tannenbaum (1957), Holsti hypothesized that Chinese and Soviet attitudes toward the United States would tend to be similar in periods of high interbloc (East-West) conflict, whereas during periods of decreasing tensions, attitudes toward American policy would diverge. Data for Holsti's analysis were 38 Soviet and 44 Chinese documents, totalling nearly 150,000 words written from 1950 to 1963. The periods included two during which East-West relations were relatively calm, four of high tension, and one in major crises. Three coding operations were performed which: (1) separated complex sentences into one or more themes; (2) identified the syntactical position of key words; and (3) characterized the themes in terms of time and mode of expression. Scores on the evaluative, potency, and activ- ity dimensions were tallied for (l) actions in themes in which the United States was the agent, and (2) qualitative characteristics ascribed to the United States. The diction- ary included about 3,600 words with such terms as "nuclear," "blackmail," "industrial," and "achievement," which had been previously tagged and scaled for intensity along the three dimensions. Holsti argued that the results of the study support his hypothesis. During periods of high East- West tension, Soviet and Chinese documents demonstrated similar attitudes. Significant differences between China 10 and the Soviet Union appeared in periods marked by a more relaxed international atmosphere. It is important to note that of the many studies done by political scientists, the Holsti study represents the most sophisticated in terms of theoretic basis. The same problem exists here, as for sociology/anthropology, in that inferences are made from message data about source characteristics, but little has been done to ascertain the validity of these inferences. In most cases it is impossi- ble; in others, the energy requirements are often too high. It is impossible because often the sources of the messages are not living, or are social and political leaders who are not willing or available for measurement of their cognitive processes. The energy requirements are often too high be- cause sample sizes necessary to generalize to a political party are impractical. Further, the time and money for coders to analyze several hundred messages has discouraged many researchers. Although computer technology has improved the speed of the coding task, computers can only read what they are told to read, i.e., a dictionary is required. An alternative is that a coder may tag words in specified ways, but this again adds to the energy load of coders and re- searchers with respect to time and money costs. 3. General Communication. Content analysis re- search in the area of general communication represents a far more heterogeneous grouping of interests than that of the previous two approaches. This reflects the varied 11 theoretic commitments of communication scholars, who seek to test several models of human behavior purporting to ex- plain message production. For example, Bales (1950) has developed a sign process analysis which seeks to infer roles of members of groups in terms of messages encoded. Other researchers have developed models of internal states of affairs based on personality structure (Allport, 1946; White, 1947; Rokeach, 1960; and Paige, 1966). Others have looked at antecedent conditions (internal) such as anxiety (Mahl, 1959), self-evaluation (Kauffman & Raimy, 1949) logical styles (Schneidman, 1963), and drive production- reduction (Dollard & Mowrer, 1947). One of the influential models of human behavior is that of Osgood (1957, 1959). Osgood (1959) reports a method of content analysis--Eva1uative Assertion Analysis--as at- tempting to . . . extract from messages the evaluations being made of significant concepts, with a minimum dependence on the effects of the messages on coders or on their existing attitudes (p. 41). This type of content analysis grew out of Osgood's mediation hypotheses (Osgood, Suci, & Tannenbaum, 1957). One of the basic assumptions of Osgood is that humans are sign-using animals, i.e., that humans acquire and/or create signs that represent elements within the human beings' en— vironments. Following Charles Morris (1946), this approach assumes humans using signs to represent significates, or objects, in the environment, with such sign-significate 12 relationships being considered as the semantic dimension to language, the relation between signs being syntactical, and the relation between signs and the users being pragmatic. The content for Osgood's representational model was described by himself (1957) and provides a useful tool in understanding the assumptions he makes when undertaking his evaluative assertion analysis. Osgood describes two general approaches by which signs become related to referents. He outlines a "substitution” approach as that situation in which a pattern of stimulation which is not the referent be- comes a sign of that referent when it evokes the same re- sponse as the referent did in absence of the sign. It is a classical conditioning (Pavlovian) paradigm of how signs (CS) become related to referents (UCS) or objects. Osgood also describes the mentalistic approach as that situation in which a pattern of stimulation which is not the referent becomes a sign of that referent when it gives rise to the £922 associated with that object. Osgood's mediation hypothesis appears to be an attempt to combine both the substitution and mentalistic approaches in order to escape the apparent over-simplified sign-object relationship implied in the substitution view. Osgood's hypothesis states: A pattern of stimulation which is not the object is a sign of that object if it evokes in an organism a mediating reaction which (a) is a fractional part of the total behavior elicited by that object, and (b) pro- duces distinctive self-stimulation that mediates re- sponses which could not occur without previous We association of non-object and object patterns of stimu- lation (1957, p. 7). 13 In content analysis the mediation hypothesis assumes: . . . (1) that in semantic encoding by the source the occurrence of specific lexical items in his message is indicative of the immediate prior occurrence in his nervous system of the corresponding representational mediation processes; and (2) that in semantic decoding by the receiver the occurrence of specific lexical items in messages are predictive of the occurrence in his nervous systems of those representational mediation processes which he has developed in association with these signs (1959, p. 39). This hypothesis led Osgood to consider sources of variabil- ity in language behavior. He defined them as: (1) avail- ability, which asserts that habits which associate signs with meanings in decoding and meanings with linguistic re- sponses in encoding are variable; (2) the representational process itself, including differences due to acquisition, and (3) associations among representational processes, i.e., variability due to individual differences in the hierarchies of the representational processes. Osgood's Evaluative Assertion Analysis requires four stages: Stage I, the identification, isolation, and masking of attitude objects; Stage II, the translation of the message into assertion form; Stage III, the assigning of directions and intensities to connectors and evaluators, and Stage IV, the evaluative scale of attitude objects. Summarizing these stages, Osgood's approach requires the identification and isolation of attitude objects (normally proper nouns and any pronouns that refer to proper nouns); the identification of evaluative meanings (adjectives); the PVT—’— translation into assertions (actor—action-complement form) which can be associated or dissociated, and the formation of such assertions into a message matrix denoted as being asso- ciated with a plus (+) sign, and dissociated with a minus (-) sign. These relationships are finally assigned mean values on a seven-step semantic differential scale, based on the independent codings of judges. Application of Osgood's analysis has been made on such messages as Goebbels' diary, and a patient urdergoing psychotherapy. It is apparent here, as in the other two general theoretic approaches, that inferences are made from message data about source characteristics. Yet, little has been done to ascertain the validity of these inferences, and for the same reasons: it is impossible (Goebbels), or the energy requirements are too high. Why attempt to infer at- ltitudes of a population from content analysis of their mes- sages, when an attitude scale may be used to measure those attitudes? The answer should become clear in Chapter II-- Rationale and Hypotheses. Theoretic Approaches: General Rationale an Measurement Regardless of the academic training of the research— er, the one common assumption is that verbal behavior re- flects to some degree the condition or state of affairs of the person encoding the message. Jaffe (1966), though pri- marily interested in psychiatric dysfunctions, is represen- tative of this general assumption. He states that 15 'psychiatric disorders, regardless of their etiology, are ultimately manifest as disorders of social communication (p. 689)." As sociologists or anthropologists, the cultural norms may be the antecedent variable assumed to be reflected in verbal behavior. As political scientists, it may be political commitments that are the antecedent conditions. As communication scholars, the antecedent conditions may be the "hierarchy of representational process," i.e., Osgood, or certain personality characteristics of the encoder, i.e., dogmatism (Rokeach, 1960), self-evaluation (Kauffman & Raimy, 1949), or drive (Dollard & Mowrer, 1947). Taken together, the previously described antecedent conditions need represent only the varying and perhaps com— peting theoretic commitments or interests of the researcher. This assumes, of course, some agreement on response meas- urement procedures, if theories would be tested. However, of 23 studies previously cited, 15 different measuring pro- cedures were used in analyzing the messages. One study (Blumberg, 1954) used column inches in newspapers as the response measure. Four studies (Kalin, Davis, McClelland, 1966; Mosteller & Wallace, 1964; Leites & Pool, 1949, and White, 1947) used the frequency of occurrence of one or more tags (words of a specified type). Three others (McClelland, 1958; McClelland & Friedman, 1952; and Reisman, Glazer, & Reuel, 1950) used the frequency of a tag per 100 lines. W 16 Two critical questions arise in choosing the meas- uring procedure: (1) is the raw frequency of some tag the best estimate of the effect of some antecedent condition? and (2) should the researcher control for the varying lengths of message samples by using some common denominator, such as total words or sentences encoded? In two hypotheti- cal messages, where Message A has 100 words with 10 verbs, and Message B has 1,000 words with 100 verbs, an affirmative answer to the first question would show Message B greater than Message A. An affirmative answer to the second ques- tion would show Message A equal to Message B. Fifteen of the 23 studies reviewed here generally use the latter ap— proach. It is important to note, however, that none of the studies report the descriptive statistics used, if any, in making such a decision. Scholars resolve the issue on logical grounds. Stone, Dunphy, et a1. (1966) raise the problem as follows: . . . For each tag concept, four numbers [in the General Inquirer] in this set generally can be computed: (l) The raw frequency of occurrence of the tag concept in the document as measured by the sum of the occurrences of all of the entry words of the tag concept in the doc~ ument; (2) the raw frequency of sentences in the docu- ment which contain at least one of the entry terms of the tag concept; (3) the raw frequency of (1) divided by the number of words in the document (word index score); (4) the raw frequency of (3) divided by the num- ber of sentences in the document (sentence tag tally). The most frequently used of these four numbers is the word index score. The others, though easily available, are not as convenient for between-document comparison: the raw word and raw sent nce scores are difficult to use if the documents vary in length. The sentence index score is not useful if the length of sentences varies considerably between documents. La A 17 As mentioned earlier, the use of the word index score usually means that the researcher has made the assump- tion that relative frequency of mention is a stable index of intensity over documents of varying lengths. Thus, the use of the word index scores equates two docu- ments on the tag SELF if the first document, which has only one hundred words, contains one reference, and the second, which has one thousand words, contains ten ref- ences (p. 227). An important question which should be answered is: by con- trolling for varying lengths of message samples, what at- tributable variance in the message is being removed, leading to the probability of Type II error? At this time, no study to this writer's knowledge has addressed itself precisely to this issue. Summary This chapter described three general theoretic ap- proaches to the study of messages where content analysis was the central methodology. We discussed the common assump- tion of all three approaches--that language reflects ante— cedent characteristics of the encoder. Finally, we raised two critical questions regarding measurement procedures used in the literature reviewed. Chapter II will present the rationale and hypotheses for this study, while Chapter III will deal with the issue of measurement found to be a problem in previous research. CHAPTER II RATIONALE AND HYPOTHESES Two general approaches are open to the researcher who has chosen content analysis as his method of ordering data: (1) analysis of the meaning of specified content words in the message, or (2) analysis of the structural meaning of the message. Osgood's Evaluative Assertion Analysis generally falls into the prior category of deter- mining the meaning of specific content words, although he recognizes that structural characteristics of the message also are sources of variance in meaning. tructural linguistics has largely been concerned with the presence or absence of linguistic phenomena, usu— ally ignoring their relative frequencies. As Saporta and Sebeok (1959) note: . . . we have little reliable information [because of the above stated structural approach] on the frequen- cies of, for example, comparable syntactical patterns. Indeed, even for so well-known a language as English there is probably no definitive information as to what the equivalent patterns are. Presumably these equiva- lences must first be identified (according to some clear-cut criteria) and norms as to relative frequen- cies established before deviations can be determined. Only then can deviations in frequency be correlated with the behavioral states of the producers of the message (p. 139). 18 "VF—— 19 Structural linguists have been most concerned with the comparison of the occurrence of some syntactic form to a criteria of what is "good” or competent form. This in large part explains the general lack of interest in structural approaches to the study of meaning in content analysis. In addition, some have been concerned (Pool, 1959) with the equation made between the occurrence of elements in any stretch of speech, resulting in the equating of, say, the verb "ran" with the verb "ate." It is obvious that the verb ”ran” and 'ate' are not the same in meaning, such as in the sentence, "The dog ran,” versus, "The dog ate." Yet, it is also obvious that there may be some contribution to the variance explainable in the cognitive structure of the en— coder and the frequency of occurrence of such verb types. With a theoretically and empirically meaningful classification of signs and sign-sign relationships, it is possible to determine the extent of meaning attributable to syntactical structure. Before presenting a descriptive classification of signs, this chapter will begin with the problem of meaning. After these two discussions, the mes- sage variables for this study will be summarized. (A copy of the coding instruction booklet is in Appendix A.) The cognitive variables used in this study will then be dis- cussed, followed by hypotheses on the relationship between message behavior and cognitive structures. 20 The Problem of Meaning Once we recognize there is a problem of meaning, at least three possible approaches are available. We can begin with a stimulus-response framework, in which meaning is ex- plained in the classical and/or operant conditioning para- digm (a substitution view). Another way of meeting the problem is to assume meaning is an unobservable, internal construct (a mentalistic approach). A third approach is to form a position which includes characteristics related to both assumptions. Osgood (1957) chose to study the problem primarily from the third approach. He determined that any solution must begin with the self-evident fact that the pat- tern of stimulation which is a sign is never identical with the pattern of stimulation which is the object. The sign- to-significate meaning is the semantic dimension to meaning, and as he points out, the pattern of stimulation for each is never identical. We might also note that the patterns of each element are never the same across individuals, leading us to the obvious assumption that it is not reasonable to , assume that persons "read” the same significates in the same way. Thus, socio-cultural patterns, prior experiences, learning ability, need states, and even physiological states become relevant. Concurrently, we must assume in order for any social communication to take place that there is some degree of overlap of semantic meaning due to these ante- cedent conditions, not merely overlap in signs. It is this assumption that makes communication possible, with behavioral :A 21 consequences in terms of the degree of overlap, social awareness, etc., across subjects. Another important assumption is that semantic mean- ing is not independent of pragmatic or syntactical meaning. This assumption is based on the notion that the source of the meaning of a sign can be the significate (denotative meaning), but also the need state of the individual holding that meaning and the degree of perceived contiguity of some other sign to that sign. Another conceptual discrimination that can be made is that significates for signs can be stimuli, or complexes of stimuli, outside (observable) or inside (hypothetical) of the sign-using animal. Those outside we could term afferent, and those inside we could term efferent. Those referents which are afferent are capable of being seen, heard, smelled, tasted, or touched, while efferent referents are internal states of affairs which cannot be sensed through normal sensory inputs. In addition, we might also note the assump- tion that those significates outside of the sign-using ani- mal exercise higher stimulus control than those significates inside of the sign-using animal. It is also worth observing that internal signifi- cates may be contiguous to external significates, producing an internal response to an internal state of affairs, but conditioned to an external significate. This is seen in the studies of Staats and Staats (1967), where evaluative re- sponses have been conditioned to nonsense syllables. 22 Finally, all meaning (semantic, pragmatic, or syn- tactic) is acquired and maintained by discriminative or associative relationships between significate and signifi- cate, significate and sign, sign and sign, interpreter and interpreter, and/or interpreter and sign. It can be observed that these conceptual distinc— tions between semantic, pragmatic, and syntactic meaning are on a meta-meaning level, i.e., meaning about meaning. It is quite a different situation, however, to translate these conceptual notions into operational definitions for purposes of empirical research, particularly when we are using con- tent analytic methods. Holsti (1969) observed: A . . . source of disagreement among those defining con- tent analysis is whether it must be limited to manifest content, that is, the surface meaning of the content. Or may content analysis be used to analyze deeper layers of meaning embedded in the content. The requirement of objectivity stipulates that only those symbols [signs] actually appearing in the message may be recorded. In other words, the coding process cannot be one of "read- ing between the lines." In this sense, content analysis is limited to manifest attributes of the text (p. 600). Holsti's statement can be interpreted as criticizing those who attempt to use content analysis to answer questions of semantics. Certainly the pragmatic dimension of meaning, within the technique chosen for this study, has had little or no attention. Semantic meaning for Osgood is not semantic meaning as these terms are conceptually defined. Osgood criticized (1957) the sign check lists to measure semantic meaning in favor of polar adjectives because of the problem of r 23 measurement of semantic meaning. He also criticized the message matrix procedure on the same grounds. His major criticisms surrounded the issue of comparability. Osgood argued that the semantic differential provided means of comparisons among different individuals and groups, and among different concepts. Message matrices and sign check lists, Osgood said, are context bound. Osgood claimed that a basic distinction must be made between the meaning of a sign and a sign's associations. The semantic differential provides the means of going beyond the context, and allows for comparison of measures of meaning of a sign. However, the solution he accepted to the problem of meaning has not resulted in a "semantic" meaning independent of syntactic or pragmatic meaning, even though the semantic differential in many ways is heuristically more valuable. He uses the term 'assign' to refer to meanings given to them (signs) via association with other signs rather than via direct associa- tion with the objects signified. He notes that his measure— ment procedure--the semantic differential--is a measurement of the meaning of assigns. However, semantic meaning as we normally have understood it in the theory of signs is changed. We find it possible to operationalize signs from languages in use; it is not possible, when looking at the message corpus, to operationalize significate to sign rela- tionships independent of the other dimensions of meaning. Thus the semantic differential is a syntactic and pragmatic measure, and not a "semantic" measure alone. The semantic r meaning of a sign is assessed by the measurement of the 24 meaning of concepts (represented by signs) and the relation of that concept sign to another sign, i.e., good-bad, etc. Thus it is clear from the content analytic approach that we are looking at syntactical relationships. The Problem for a Syntactical Approach to the Meaning of Meaning The basic problem in a syntactical approach to the study of meaning is not in the conclusion that there has been nothing done in the area. Rather, at least three studies have explored a relatively small subset of syntac- tical meaning to the exclusion of others (eg. McEwen, 1967; Kochevar, 1967, and McEwen, 1969 studied the effect of high intense versus low intense verbs on attitude change). In part this is due to a lack of a more complete explication of syntactical meaning, and in part due to the lack of predict- ability of so many hypothetical constructs. Thus, an opera- tionally adequate but scientifically meaningful classifica- tion of signs and sign-to-sign relationships is required. We do not argue that hypothetical constructs can be avoided, nor should be; we argue for hypothetical constructs that have higher explanatory power. Further, it is argued that the pragmatical and semantical dimension of meaning is heuristically and hypothetically valuable, subsumed within a syntactical framework. A syntactical approach to the meaning of meaning in messages is operationally and theo- retically valuable. 25 Syntactical approaches to meaning have led linguists and philosophers as well as psychologists and communication scholars into a study of the grammatical and logical rela- tionships between signs. A descriptive system which defines linguistic or philosophic syntactical relations may be ade- quate for such scholars. However, communication researchers have sought to go farther by including the characteristics of sources and receivers of messages, not merely the mes- sages themselves. Thus, the category system for describing language syntactically will be different for a communication scholar than for either a linguist or philosopher. Phrased in another way, a descriptive system which is very adequate for a linguist may have little value to a communication scholar. The communication researchers' interests will in- fluence the category of signs and sign to sign relationships provided in this study, and should add to the widely held conclusion that semiotics has provided theoretic impetus for widely divergent academic disciplines. A Descgiptive Classification of Signs The descriptive classification of signs to be used in this study has three basic elements: (1) limiters (modi- fiers); (2) subject signs, and (3) connectors. In addition, the term 'unit sign” will be defined, providing a distinc- tion which allows for the subsuming of semantic research into a syntactical framework. 26 l. Modifiers as Signs. The work of Osgood, et al., (1957) has produced considerable research interest in what we will term here as modifier signs, but what he calls characterizations in the congruity principle. In most cases, the semantic differential has sought to place a con- cept word, such as the name of a person, into some semantic space, based on factor analysis of polar adjectives. Osgood notes that they have been able to obtain "evaluative" fac- tors, and to a lesser degree, "activity" and "potency" fac- tors in the measurement of a concept. The same approach has been used in source credibility research, and in atti— tude change research. Modifiers have also been one compon- ent in Osgood's research dealing with congruity theory in the development of his Evaluative Assertion Analysis, and in Rokeach's belief congruence research. In most cases, relevant modifier signs have been extracted from irrelevant modifier signs through factor analytic procedures. The at- tempt is to be able to identify those relevant and unidimen- sional modifier signs that are used by the normal population in "evaluation" of other concepts. From the viewpoint ex- pressed here, these have been syntactical studies. Osgood's evaluative dimension holds up well in many experiments, and is generally characterized as refering to internal states of affairs (internal significates) of the sign-using animal, i.e., hypothetical responses said to oc- cur in the central nervous system. Other modifiers have broken down into many different factors. These latter A 27 classes of modifiers, such as hot-cold, black-white, etc., have been previously termed in this paper as afferent signs. It is of theoretic interest that the explanation for the evaluative dimension accounting for more variance than the other dimensions may be the result of the effect of the kind of significates on the signs representing them. That is, the response of good-bad may refer to generalized signifi- cates within the sign-using animal (as Osgood, §£_§l,, argue [p. 179]) which are in turn related to the concept being measured. The sign response of black and white, since they demand less generalized significates outside of the sign- using animal, exercise more stimulus control over responses than do internal significates. Thus, the high intercorrela- tion of items on the evaluative dimension may be due to low stimulus control of internal significates, while the low intercorrelation of items on the remaining dimensions may be due to the high stimulus control of external significates. It seems reasonable to assume that two nominal categories important to the classification of modifier signs are the external-internal (afferent-efferent) dichotomy of signifi- cates associated with those signs. In order to make this distinction we will require significantly high percentage of independent inter-coder agreement on the nature of the sig- nificates referred to by the signs. Thus, an afferent modi- fier sign is a sign whose significates can be sensed, di- rectly or indirectly, by the five senses; an efferent 28 modifier sign is a sign whose significate is an internal state of affairs, and cannot be touched, smelled, seen, heard or tasted. An example of an afferent modifier sign would be “black," or, "hot." An example of an efferent modifier sign would be "good," or, "democratic." An exten- sion of the definition of what is meant by modifier is post- poned until a discussion of subject signs and connectors is made. 2. Subject Signs. Rokeach and Rothman (1965) dis- cussed what they called "CS units;" that is, the linking together of subjects (S) with characterizations (C). They were recursively defined, with subjects being those words capable of being characterized. This general approach has been used in much research, with subject signs being those concepts measured, such as attitudinal objects in the seman— tic differential, while the polar adjectives represent the characterizations applied to the concepts. While this is what is usually meant by "subjects," the term subject sign needs a more specific definition since we are generally con- cerned with continuous discourse. By subject signs, we mean those words in a continuous discourse message that are substantives or complements of verbs (Roberts, 1954). Stated another way, they are the subjects and the objects of verbs, the rationale for which will be discussed later under modifier signs as limiters. As in the case of modifier signs, subject signs can also be directly or indirectly experienced; they can be r————'_——————‘ 29 afferent or efferent. However, we can also identify sub— ject signs as being things or places as opposed to people or groups of people. Such a sign may refer to one's self, a specific other person or group, generalized and unspecific other persons, or in the communication event--the other- receiver(s) person. Thus we can categorize subject people signs as having referents to self (source) signs (S-S), specific other signs (A-O), generalized other signs (G-O), and receiver (target) other signs (T-O). 3. Unit Signs. As was pointed out earlier, Rokeach and Rothman note that a characterization and its subject form a CS unit. Since it is considered to be of theoretic interest here to make the distinction between syntactical relationships within a unit and syntactical relationships between units, it is necessary to use the term "unit signs." A unit sign has as its elements a subject sign and all of the modifier signs directly or indirectly related to it. Thus, the ”beautiful girl" statement is a unit sign which has as its elements a subject sign, "girl," and a modifier sign, “beautiful." Every unit sign must have a subject sign but does not require, though it may have, a modifier sign associated with it. When a unit sign has no modifiers it is termed a “primitive unit sign (SlP)." When the unit sign includes one or more modifiers, it is termed a "defined unit sign ($10)." 4. Connectors. Very little research has been done in which the nature of connectors of signs has been 30 explored. Osgood's (1959) Evaluative Assertion Analysis was concerned with internal responses to connectors, but not with the nature of the syntactical structure beyond associa- tive or dissociative relations. Following Osgood, a re- search project is presently underway at Michigan State Uni- versity in an attempt to measure the evaluative meaning of verbs. At this time, no results have been reported. In another study previously cited (McEwen, 1969), the use of subjunctive and indicative moods of verbs was experimentally manipulated in messages, with the assumption that subjunc- tives are less intense than indicatives. As previously noted, we can conclude we do not have an adequate classifi- cation system for connectors, although the importance of them is pointed out in an explication of the Whorfian Hy- pothesis by Joshua A. Fishman (1966). J. P. Guilford (1967), in an explication of his the- oretic notions on human intelligence, has articulated at least some of the problems in getting such a classification. In his discussion of the relationship between logic and psychology, he notes: Psychology is an empirical science, one of whose aims is to describe in terms of general principles or laws what actually happens when individuals think. Its approach is that of observation and induction; its operation is mainly cognitive; it seeks to understand. Formal logic, on the other hand, is not an empirical science. Like mathematics, it starts with axioms, statements that it assumes to be true, and it aims at sets of rules for thinking whereby it is possible to determine whether inferences are true or false (p. 246). 31 Guilford's statements seem to be made in the context of formalization--or lack of it--in psychological theory. He argues that before formal logic can be "fully" applied to psychology there must be a stating of psychological events in terms of propositions. He notes that not every statement is a proposition, but that only those that can be either true or false. He says: A proposition cannot even be partly true and partly false, for logic operates under the principle of the ex- cluded middle; in the excluded middle, part truths dwell. This poses a problem that has to be overcome or circumvented, or it leaves us with a limited application of logic to psychological events. Piaget points out other difficulties. He comments that even much of the usual thinking of an adult is unformalizable in logical terms. Only mathematical thinking is completely formal- ized. In normal thinking, genetically, propositions come first and axioms last, just in reverse order to that of logic. Furthermore, systems of information do not lend themselves to step-by-step treatment such as is characteristic of logic (p. 247). While it may be true that psycho-logic events in continuous discourse are not easily reducible to formalized relationships, the relationships identified in logic--with- out competency or logical validity requirements associated with the use of truth tables-—can be used in uncovering of how humans relate a sign to another sign. As has been pointed out before, we have no way of verbalizing referents except by the use of signs. Further, it seems to this writ- er that logical relationships in a philosophical context compared to logic in a psychological context are similar to the linguistic-psycholinguistic distinctions. Philosophy looks at whether the relationships are adequate or not; 32 psychology looks at the relationships, without judgements of adequacy, and communication scholars are looking at the relationships in terms of their dyadic demands when two or more people interact in a communication exchange. Assuming the communication scholar's interest in logic, the following list of possible relations or connec- tors is offered as defining the relationships between unit signs (i.e., a subject sign and its modifier(s) signs are related to other subject signs and their modifier(s) signs): a. Equivalence comparison connectors (CE). These connectors denote that a unit sign is the same as another unit sign. It is operationalized as the use of any verb or verb phrase (typically the verb-to-be) such as "is," "was," "will be," as in the example, "Communists are atheists." b. More-than comparison connectors (CM). These connectors denote that a unit sign has more of some attri- bute than another unit sign. Normally, this connector oc— curs in the context of a modifier sign applied to two dif- ferent subject signs, but denotes that one subject sign has more of the characterization than the other subject sign. It is operationalized as the use of the comparison modifier of “more,” or the superlative modifier, "most." It also is denoted by suffixes to a modifier such as "—-er," or '--est.' It can be seen in the sentence, "Jane is more beautiful than Sally." The unit signs are "beautiful Jane," and, ”beautiful Sally,” and the relationship between the unit signs is a "more-than" relationship. 33 The "less than" relationship is considered unneces- sary to include as a separate connector, since the order of signs is not assumed to be relevant for this proposed anal- ysis. If one were to say that Sally is less beautiful than Jane, it is the "mirror" meaning of the statement, "Jane is more beautiful than Sally." The meanings are the same, and by transformation, "less than" relationship can be repre— sented by an inversion of unit signs to produce the "more than” relationship. We need only add to our operational definition, "less," and "least," as indicators of the more- than relation. c. Subset comparison connectors (CS). These con- nectors denote that a unit sign is included within but not equivalent to another unit sign. It can be operationalized as the use of verbs such as, "belongs to," "is part of," where class relationships are denoted. It can be seen in the statement, "The army is part of the military complex." d. Spatial comparison connectors (CP). These con- nectors denote that a unit sign is related, as though in space, in some way with another unit sign. It can be oper- ationalized as the use of verb-type phrases such as "is in front of,’I 'is on top of,” "is to the left of," "is to the right of.“ It can be seen in the statement, "Conservatives are to the right of liberals." It is usually seen as a propositional phrase which modifies a verb, and answers the question, “Where?" o 34 e. Time comparison connector (CT). These connec- tors denote that a unit sign is related, as though in time, in some way with another unit sign. It is operationalized as the use of verb-type phrases such as "George went to the store in the evening." It is usually seen as a preposi- tional phrase which modifies a verb, and answers the question, ”When?“ f. Existence comparison connectors (EXT). These connectors denote the existence of a unit sign, and is operationalized as the use of the verb-to-be such as "is," where a demonstrative pronoun is related to a unit sign as in the sentence, I'There is a book." 9. Descriptive comparison connectors (ADJ). These connectors denote the description of a unit sign, and are operationalized as the use of the verb-to-be such as is, where a unit sign is related to a predicate adjective as in the sentence, ”The book is red." h. Action transitive connectors (T). These con- nectors differ from the comparison verbs in that they show action. They are transitive action connectors if they show action going from one unit sign to another unit sign. Stated another way, it is a verb which has both a subject and an object, and shows action as in the statement, "George dropped the ball.” They are operationalized as the appear- ance of any transitive verb—-past, present, or future. 1. Action reflexive connectors (R). These connec- tors show action of a unit sign, with no complement for the 3S connector denoted (a subject of the verb, but no object of that verb). This can be seen in the statement, “George was dropped,” and is operationalized as the appearance of any intransitive verb--present, past, or future. In addition to the categories just described, there are three general classes of connectors which can be ap- plied to all of the relationships just presented (a-i): (l) negations (N), where any of the comparison or action connectors is a negated relationship, and is operationalized as the presence of the adverb, "not," such as in the sen- tence, ”Jane is 222 more beautiful than Sally," or by the use of such prefixes as "dis--, "un--," etc.; (2) dispo- sitionals (S), where any of the comparison or action con- nectors is a dispositional relationship such as in a dis- position to act or be related, but not necessarily so. Dispositionals are operationalized as the presence of any words in a connector which show a subjunctive relationship, such as "might be--,' ”could be--," "may be--," etc., and (3) tenses, where any of the comparison or action connec- tors can be in the past (PA), present (PR), or future (FU) relationship. Tenses are operationalized as the tense of the verb (simple tense), and excludes distinctions dealing with perfect and plu-perfect tenses. 5. Modifier signs as limiters. It seems reasonable to assume initially that when a person uses a modifier sign in relation to a subject sign, it has been perceived by that person as relevant. The proper question appears to be: What is the syntactic function of modifiers in relation to 36 a subject sign. This is a broader question than that asked by Osgood's congruity principle or Rokeach's belief congru- ence principle. We can look, for example, at the limiting relationship that occurs when a modifier is associated with a subject sign. Consider the following: 1. All men are mortal. 2. Some men are mortal. 3. Evil men are mortal. 4. Black men are mortal. In every case above, we can view modifiers in set theoretic notions as limiting words which change the meaning of the subject sign, ”men.” ”Black men," is not "all men." We can also note this same function when preposi- tional phrases are used. Consider these examples: 5. The winter in Florida is beautiful. 6. The summer in Florida is beautiful. In each of these two cases, the modifiers are limiting words words, so that Florida is a limiting word on winter in sen- tence 5. It is not the same as saying, "The winter is beau- tiful,” or, "The summer is beautiful." We also can distinguish between modifiers that limit the meaning of other modifiers, or of connectors. The fol- lowing examples illustrate: 7. The innately beautiful Jane won the contest. 8. Jane walked beautifully. In each of these two cases, a modifier functioned as a limiter of another word, different from the previous six 37 examples. In the seventh sentence, "innately" is a limiter of the word "beautiful"--also a modifier, and the two modi- fiers together with Jane form a defined unit sign. "In- nately” is usually classified as an adverb, as would "beau- tifully” be in sentence 8. However, sentence 8 demonstrates an adverb--"beautifully"--which limits the verb (connector) "walked,” and forms a defined connector. The important con- clusion is that modifiers have a limiting relationship to the words they modify, and that they can be used to modify subject signs directly, or modifiers of subject signs (in- directly), or connectors. Because of theoretical concerns, and because not all possible meaning is going to be studied here, we would pro- pose that all messages can be looked at in terms of subject signs and their modifier signs, forming unit signs (primi- tive or defined); unit signs connected to other unit signs, and connectors in terms of not only the categories previous- ly listed, but modifiers associated with them. Preposi- tions, other than the denotative function they perform in showing whether a subject sign or a verb is being limited, are held out for separate analysis; articles such as an, "a,” and "the," are also held out. Further, since relative pronouns perform a similar role as do prepositions, they will be used to determine which modifiers belong with which subject signs or connectors, and then summed with inter- rogatives and demonstratives for separate analysis into a message category called "Other." All questions 38 (interrogatives) in a message will be treated as denoting subjunctive, or dispositional, relationships between unit signs. Finally, the function of demonstrative pronouns has been traditionally considered by grammarians as denoting the existence of a person, place, or thing (or a unit sign). Such existence may be general, such as in the sentence, "There is a book," and the existence of a book is affirmed. Existence also may be specific, such as in the sentence just described, where the person encoding the message is refer- ring to location. In either case, the demonstrative pronoun is counted as previously stated. The unit sign is retained, and is considered theoretically as a special case of the use of a reflexive connector, although it will form a separate message variable for our analysis here. Summagy and Permutations of Message Variables In order to crystallize the message variables for understanding and clarity, Table l on pages 39-40 shows the message variables to be identified for this study, and the definitions associated with them. In the following chapter on method of analysis, twenty variables derived from these basic variables will be added for analysis. The operation- al definitions provided in the previous section of this chapter, combined with the code book in Appendix A, repre- sent the sources of data from which inferences will be made about cognitive states of the encoders. 39 conunmuofi COmHummEoo .m>euocsmnsm mum nOHSB muouomccoo Hobos m:BOB\m Sum mocmam>esmm cemeummEoo .m>Huocsnn5m mum roars muouomccoo Hmuoe mueoe\m mum mmcmu mucusm .m>Huocsnnsm mus coflc3 muouooccoo Hmuoa muaoa\w 3mm mmcmu ucmmmud .m>auocsnnsm mum SOH£3 muouowccoo Houoa muBOB\m mom mmcmu ummm .o>euocsnn9m mum nowc3 muouomccoo Hobos muaoe\m «mm o>HumoeocH .m>euemcouuce mum coec3 muouomccoo cOeuom Hobos mneoa\m mH m>eumowocfi .m>euemcmuu mum roars muouooccoo QOHuom Hmuoa mnaoe\m 9H mm>HumuumCOEmo cue3 swam pomnnsm mcHDMHOOmmm muouooccoo Hobos mIBOB\m meH m>euoomom cuwz cmem commosm uneaseoommm muouowccoo Hobos mueoa\w hodH muouomccoo mafia comeumdeoo Hobos mueos\m 90H muouooccoo HMfibMdm comeummEoo Hmuoe mueoe\m moH muouowccoo pomosm coweummeoo Hobos mneoa\m muH muouomccoo cmcu\muoEucomwumoEoo Hmuoa moeoe\m 20H muouomccoo oocmam>fisvmuc0mHHMQeoo Hobos mueoa\m moH mmcmu museum .o>wum0eo2e mum cuec3 muouooccoo Hobos mueoe\m DmH mmcou ucmmmud .m>fiucoeocw mus cowc3 muouomccoo Hmuoa muaoa\m mmH omcmu awed .m>HumoHocH our cowc3 muouomccoo Hmuoa muaoa\w mmH ucmummum common one cowcS muouumccoo mo mumwmwoOE Hmuoe muaoa\m mHUA ucoummmo coupon who nowcz muouooccoo mo mumwuwooe Hobos muaoe\m «IHUA ucmuomwm common was codes mouoz bomnosm mo muowmeoos Hobos mnaoe\m muamq ucoummmo smooch was coec3 mouo3 pomnosm mo mumwuwooe Hobos mueoa\m «lama econommw coupon was roars mouoz uownnsm Hobos mneoe\m muam someones comps“ mus cowc3 mouos uooflosm Hmuoa mneoa\m diam muoemeooe oboe no use m>mc cowc3 muouowccoo Hobos muaoe\m 0H0 mumeuwooe o: o>sn cowc3 muouomccoo Hobos mneoa\m mHO womanhoOE whoa no oco m>mc nuac3 mouos uomnosm Hobos mnaoa\m mam mumsuseos on m>ur roars mono; humans» Hauoe .m-eoe\m mam cocoons mouo3 Hence m HIBOB can cowuwcflmmo sponse: Iowus> “I .mmahswus> mmsmmme mo assessmuu.a mqmda 40 m>fiuocsflnsm mum coacz muouomccoo mmcmu ucmmoum powwow: Hmuoa mIBOB\m mmmz m>euocsflo5m mus cowc3 muouomccoo omcmu ammo omucmmc Hobos muaoe\m euemcmuuse mum scene muouomccou sceuom omummmc Hobos mueoa\m mHz m>Huemcmuu mum cofic3 muouomccoo COeuum omummmc Hobos mlaoe\m 9H2 csocoud o>Humuumc06mo m :DHB cmem has: n someoommm roars muouuoccoo m>wusowocw omummmc Hobos mnaoe\m meHz .m>wuoonom cm cue3 amen vac: m mumwoommm soars muouomccoo o>ebmoHon omummmc Hobos muaoe\m oode use» .m>eumuflocH mum soars muouomccoo coheummEoo omummmc Hobos mneoa\m BOHz Hmaumdm .o>eumoeoce mum coec3 muouomccoo cOmeummeoo omummmc Hmuoa mueoa\m mOHz ammosm .m>«umowoce mum cuec3 muouomccoo coheumdfioo omummwc Hmuoe muaoe\m mOHz cmcunmuoe .m>HumoaoCe mum coen3 muouomccoo COmHumoEoo ooummmc Hobos mueoe\u zqu dreamsom .m>«umoHoce mum cuec3 muouomccoo c0meumdsoo monsoon Hobos mnsoa\m mOHz mmcmu museum .m>HumOeocw mum coHcB muouomccoo oousmmc Hmuoe muaoa\m DmHz mmcmu ucmmmud .m>flusoflocw mum nowcz muouooccoo owummmc Houoa mIBOB\m mmHz mmcwu ammo .m>HumoHocw mum cown3 muouomccoo ooummwc Hobos mueoa\m dez omummoc mouo3 uooflosm ooceuoo Hmuoa mueoa\m oamz omummmc mono; uomnnsm m>nue5euo Hmuoe mueoe\m mamz o>euwmcmuuce .m>wuocsmosm mum cowc3 muouomccou Hobos maeoe\m mm m>euemcmuu .o>wuocsnosm mum cowc3 muouomccoo Hobos mneoa\u em o>wuouumc08mo m cues comm uomnnsm meanneuOmmc .m>euocshnsm one comes muouumccoo Hobos mlaoa\u mem m>wuommoo cm cue3 swam bomnosm mcaumHOOmmc .m>euocsno5m who cuen3 muouowccoo Hooch muaoa\u poem use» coweumaeoo .o>wuocsnn5m mus cueca muouooccoo Hobos nuaoa\u Bum Hmfiummm comauodEou .o>wuocsmnsm was coec3 muouooccoo Hooch nuaoa\m mum ummosm coweudeoo .m>euocsnosm mus cowcz muouooccou Hmuoa muaoe\m mum man c0wuacemmo museum: umwuo> owscfiucoonl.a mamda 41 mcsocoud Hmc0mumd cemumd ocoomm ..m.e .um>wmomu on» Ca some» cuec3 mouo3 somehow Hobos mneoe\w ole mcsocoud Hmcom numm c0muod umuam ..m.fl .muusom one on momma coac3 memos homeosm Hobos mueoa\m mum masocoum HMCOmumm COmuwd than» masoum Ho chmumm cemeommmcs 0» Homes nuec3 mouo3 uomnnsm Hobos mIBOB\m ow dsoum no c0mumd oemwomdm m on momma coec3 mouOB pomnnsm Hobos mueoe\m Cd ucoummmm mum coHc3 mumuHEea uouooccou omummmc Hmuoe mueoe\m mlauqz econommm mum soars mumueEHH uouomccoo omummoc Hobos mueoe\m «Iaoqz economwm mum cofin3 mumueEeH ouos pomnosm omummmc Hobos mseoe\m mnequ ucmummwm mum noec3 muoueaea ouo3 uomnnsm ooummmc Hobos mueoe\m duaqu ownemoo mum coac3 muouumccoo ooummmc Hmuoe muaos\m Qauz m>euw8wud mum coec3 muouooccoo owummoc Hobos mueoe\w mauz ucmumwwm mus coac3 mouo3 coomnsm owuommc Hobos mnaoa\m mlamz ucmummwm mum c0a23 mouos uomnnsm omummmc Hooch mneoe\m duamz m>euemcmuucw .m>Huocshn9m mus nowc3 muouooccou owummmc Hmuoe mueoe\m mmz m>Huemcmuu .m>fluocsnnsm mus coec3 muouomccoo owummmc Hobos mueoe\m amz .csocoud m>wuouumc06mo o cue3 cmem vac: m mumeOOmmm cowcz muouooccoo m>euocsflosm twosome Hobos mneoe\m memz m>fluomnom so cuw3 comm uwcs o upswUOmmm cown3 muouomccoo o>wuucsnoom owuommc Hobos mlaoe\w oodmz use» .m>wuocsnn9m was cuec3 muouomccoo COmwummEOO ooummmc Hobos m:909\m Bumz Hmwusdm .m>wuocsmnsm mum cOHc3 muouooccoo coweumdeoo twosome Hobos mueoa\m momz uomnsm .o>euocsno9m mus nowc3 muouomccoo COmwudeoo owummwc Hobos mnaos\u mumz coco whoa .m>«uocsflosm our goers muouomccoo coweusdeoo omummoc Hmuoa muaoa\m zumz mcwussom .o>euocsnn5m one guess nuouooccoo c0mwumaaoo omummoc Hobos mnaoa\m mumz o>wuocsnosm was nowcz muouooccoo once» mucusw twosome Houoa mneoa\w Dmmz GHQ coduwcwmwa musmomz Iowus> owscwucooll.a Manda 42 .HHH Houdmcu cw owcflmadxo on HHe3 Hmuou menu mo sodom>enoo .1 umcuo Hmuoe mueoe\m meo mcoeuemoomud Hmuoe mueoe\m ammo mmeoauum Hmuoe mueoe\u ems masocoum encamuwm cOmHmd ocoomm ..o.w .uo>emomu ecu ou mommy cufic3 mouoz oomnnsm omummmc Hobos mlsoa\m Ouaz masocoum HMQOmHmd cOmuod amuse ..m.w .mousom on» Oh momma cues: mouoz uummosm omummoc Hmuoa muaoa\m mumz mcsocoud Hmcomuod acmuwd when» no .mdsoum .mc0mumd Oeuwumdmcs on women cownz memos uommnsm omummmc Hobos mueoa\m ouz dsoum no common Oeueomdm o ou momma soars mouo3 uuonosm pounds: Hobos maeoa\u Odz c0wuecwmmo museum: Imwwm> E omscwucooln.a mummy 43 ngpitive Variables, Demographic Characteristics Five cognitive measures were chosen for analysis of their systematic relationship with the message variables previously described. Three of these variables--dogmatism, anxiety, and attitude—-have been of general theoretic in— terest to communication scholars. Two other variables-- spatial relationship abilities and vocabulary skills--were of interest to this researcher because of the intuitive be- lief they may be important in the kind of message variables described. In addition to these cognitive measures, four demographic variables were considered important--age, sex, level (freshman, sophomore, etc.), and school. Dogmatism Milton Rokeach (1960) has described a personality construct which he terms open- and closed-mindedness, or dogmatism. He defines dogmatism conceptually as having three dimensions: (1) a belief-disbelief system, which re- fers to the relative frequency of occurrence of a person's beliefs and disbeliefs; (2) a central-peripheral system, which refers to the relative frequency of occurrence of be- liefs and disbeliefs based on authority, and (3) a time di- mension, which refers to the relative frequency of occur- rence of beliefs and disbeliefs in time--narrowness or broadness. Thus, a closed-minded person should have the following characteristics: (1) more beliefs than disbe- liefs; (2) more beliefs and disbeliefs based on authority, 44 and (3) beliefs centered in one time--past, present, or future tense--to the relative exclusion of the other tenses. Rokeach (1960) reports a study which purports to test this third dimension. He reasoned that closed-minded people are highly anxious, and would be expected to have more of their beliefs and disbeliefs in the future tense, since that tense is supposedly more ambiguous. Attempts to cope with anxiety should involve a de-emphasis of the present, and instead handling anxiety with a preoccupation with the future. It . . . follows that persons characterized as having relatively closed systems should manifest not only more anxiety but also more future-orientation than those with relatively open systems (p. 367). Rokeach tested this notion, using responses on five TAT cards, and counting the frequency of occurrence of past, present, and future tense verbs. He then translated the scores into percentages of the total verbs for each tense. He reported no significant differences between open- and closed-minded persons on the frequency of past tense re- sponses, but did report open groups consistently giving more present tense responses while closed groups consistently gave more future tense responses. Rokeach has developed a 44-item Dogmatism scale, used in a considerable amount of research on dogmatism. Powell and Troldahl (1965) have reduced this scale to 20 items without losing too much reliability. The operational definition of dogmatism used in this study was the subject's score on the Powell—Troldahl scale. 45 Anxiety The theoretic relationship between dogmatism and anxiety has already been suggested (Rokeach, 1960). Within the last three years, an anxiety measure in communication situations has been developed at Michigan State University as part of a project for desensitization to such anxiety. Test-retest reliabilities and split-half reliabilities have been reported (Nichols, 1969). Test-retest on 769 subjects was .83 over a lO-day interval, while split-half reliabili- ties were reported at .92. Our Operational definition of anxiety is that 20-item, Likert-type scale of anxiety used at Michigan State University. Attitude Numerous measures of attitude exist in the field, many perhaps better than the one chosen for this study. However, the type of analysis of attitude chosen allows ease in analyzing separately direction and intensity. The measure chosen was that of Louis Guttman and Edward A. Suchman (1947). Direction of attitude is operationally de- fined as a check-mark response of "agree,' "disagree," or "undecided.” Intensity of attitude is measured as the re- sponse to a second question, "How strongly do you feel about this?” Rated from one to four, the possible responses were: 1--Not at all strongly; 2--Not so strongly; 3--Fair1y strongly, and 4--Very strongly. Although Guttman and Suchman developed the scale to look at intensity and a zero 46 point for attitude analysis, the authors concluded it was usable, but report no reliability scores. They also report a study they conducted (no data in the article) which con- cluded certain personality characteristic inferiority, hypomania, psychopathic deviate, and depression) were sig- nificantly correlated with their measure of intensity. They add: This suggests a rich field of exploration of the role that personality traits play in attitudes and opinions on political, social, and economic matters, and from a methodological point of view in determininggrespondents' verbal habits and modes of expression (p. 67) [under- lining added]. Spatial Relationships and Vocabularnykills Two interesting ability variables--spatial relation- ship and vocabulary skills-~are suggested from some of the research of J. P. Guilford (1959, 1967) related to his structure of intelligence model. He has sought to develop, largely through factor analytic procedures, the dimensions of intellectual abilities. He denotes five: (1) cogni- tion, which is the means of discovery or rediscovery or recognition; (2) memory, which he defines as retention of what is cognized; (3) convergent thinking, which he defines as the "convergence” of diverse information for the purpose of achieving a ”right" answer, or at least the recognized "best or conventional answer;” (4) divergent thinking, which he considers to be the different directions one can 47 go in thinking, and (5) evaluation, which are the decisions one makes as to the goodness and correctness of what he knows. John W. French, Ruth B. Ekstrom and Leighton A. Price have prepared what they term a "Kit of Reference Tests for Cognitive Factors, 1963 Revision," which is the result of factor analysis of about 100 tests of cognitive abilities. Many of these tests have been developed by Guilford as part of his intelligence model. Based on factor analysis of total scores (not item analysis), French, 33 31., claim twenty-four factors emerge, among them such ability dimensions as spatial reasoning, induction, semantic spontaneous flexibility, verbal comprehension, syllogistic reasoning, semantic redefinition, and general reasoning. No data on reliability and validity are available. Never— theless, two of the variables used in this study are those of Spatial Reasoning, and a Wide Vocabulary Test (10-16). It was believed this study could attempt to look at the relationship between these measures and the message varia- bles being analyzed subject to the limitations of inadequate reports on reliability and validity. Since scoring proce- dure varies in these types of tests, and no recommended scoring procedure is provided by French, this study followed the suggestion of Guilford (1954); separate scores of the number correct and incorrect were obtained, each analyzed as a separate variable. 48 Hypotheses The legitimacy of constructing hypotheses in a study which uses one sample, and analyzes the data on the basis to be described in Chapter III is generally conclu- sive: research hypotheses in this design are not testable. This is a descriptive study, intended to generate hypoth- eses. No experimental manipulation is taking place. There- fore, the hypotheses offered here cannot be confirmed or re- jected without a comparison of two or more samples. It is also important to note that scores of hypoth— eses could have been chosen for analysis, but only five were selected. These five were chosen on the basis of theoretic significance, and the conceptual similarities of the category system presented here with previous studies. With the above stated caution and limitation, the following hypotheses are offered: H1: The relative frequency of occurrence of present (C1PR) and future tense verbs (ClFU) will ex- plain significant variance (p < .05) variance in dogmatism and anxiety. A corollary hypoth- eses is that dogmatism and anxiety are signifi- cantly correlated (p < .05). H : The relative frequency of occurrence of verb negations (NCl) will explain significant vari- ance (p < .05) in dogmatism and anxiety. H : Those subjects who hold different attitude in- tensity levels on the topic will show signifi- cant differences (p < .05) in the relative frequency of occurrence of defined connectors (C1D) and defined unit signs (SlD). H4: The relative frequency of occurrence of compar- ison verbs (COMP) will explain significant variance (p < .05) in spatial reasoning. 49 H5: The relative frequency of occurrence of subject (81) and limiter (L) words will explain Signif- icant variance (p < .05) in verbal skills. The first hypothesis is based on the theoretic statements of Rokeach, previously discussed. The second hypothesis is based on the theoretic statements of Rokeach, where beliefs and disbeliefs are operationalized as the relative frequency of occurrence of what one believes and what one does not believe. Specifically, negations should be correlated with open-minded persons, while non-negations should be correlated with closed-mindedness. The third hypothesis is based on the conclusions of Guttman and Suchman, with the extension that the more in- tense a person holds his attitude, the more concerned he is that the receiver will understand what he says. A high frequency of occurrence of limiters is believed to be a verbal expression of high concern on the part of the source that his message be understood, while a low frequency of occurrence of limiters is believed to be a verbal expression of low concern on the part of the source that his message be understood The fourth hypothesis is related to Guilford's definition of cognition of transformations in his Structure of Intellect model. He defines transformations as follows: Transformations are changes of various kinds, of exist- ing or known information in its attributes, meaning, role, or use. The most common transformations in fig- ural information include changes in sensory qualities and quantities, in location and in arrangement ofgparts (p. 100) [underlining added]. 50 It would appear that differential abilities to move objects or symbols of objects in space may be reflected in verbal behavior of a similar type. Comparison verbs, as previous- ly outlined, refer to equating, spatial, time, more-than, and subset relations. All or part of these verb patterns may be related to the differential abilities of subjects in the sample as measured by the spatial reasoning test. The fifth hypothesis is related to some of the pre- vious research in content analysis. The TTR (Type-Token Ratio), which was determined by the number of different words found in samples of standard length, was found by Johnson (1944) to differentiate between authors. One of the possible variables that differentiates sources is the vocab- ulary skill that a source has. A roughly similar measure of vocabulary skill would be the number of content words appearing in a text, counting redundancies. Content words can be defined as both subject words and limiters in the message variables previously described. Different from TTR, this measure would include totals regardless of redundan- cies. Conclusion This chapter has presented a theoretic rationale for the study of syntactical meaning, providing categories for the analysis of a sample of messages. Five cognitive meas- ures and four demographic variables have been presented, all intended to ”systematically and objectively identify specific Sl characteristics of a sample of messages for the purpose of making inferences about certain antecedent mental character- istics of the sources." Five ”hypotheses" were offered, although these hy- potheses are not testable from one sample. However, based on the type of data analysis to be offered in the following chapter, we can make informal predictions about which vari- ables will explain significant variance compared to those which will not. Although many hypotheses could have been made, the five chosen were justified on the basis of theo- retic formulations already extant in the field. Chapter III will present a step—by-step procedure for analysis of the data, leading to a multiple regression analysis of message variables against the criterion measures of dogmatism, anxiety, vocabulary skills, and spatial rela— tion abilities. CHAPTER III METHOD OF ANALYSIS The purpose of this study is to determine, through multiple regression and analysis of variance techniques, what subset of the 100 message variables covary with speci- fied cognitive structures and certain demographic character- istics of the sample. Eighty of the message variables were presented in Table 1. In addition, 20 mathematically derived variables were obtained from the basic 80, and are summarized in Table 4. Since little is known about the distribution of these message variables, considerable atten- tion was given to the assumptions of multiple regression analysis. This chapter will outline: (l) the procedures for gathering data, including characteristics of the sample taken; (2) the coding, scoring, and tag consistency of mes- sage data; (3) a description of the 20 derived variables, and (4) the procedure for data analysis. SamplingAProcedure, Description In the fall of 1969, a random sample of 114 subjects from the basic Speech course at Grand Rapids (Michigan) 52 53 Junior College received a public opinion survey. This sur- vey included measures of attitude, dogmatism, anxiety, vo- cabulary skills, and spatial relationships abilities. Grand Rapids Junior College had an enrollment of 5,283 students in the fall of 1969, with 22 sections of the basic speech course. Although the course is not required, approximately 50 per cent of the student body take the course. Six classes were chosen to complete the survey, all taught by one instructor. The instructor was new to the school, thus allowing for the assumption that no student self-selected himself into the classes on the basis of prior knowledge of the instructor. Two subjects did not complete the questionnaire, leaving a sample of 112 for the first survey. Three weeks later, the same students were asked to complete another public opinion survey, which included an attitude measure and a topic for a persuasive speech. Fourteen subjects were absent, leaving a sample size of 98 for final analysis. The following instructions were given: Write the most persuasive speech you are able to do on the following topic, taking the position you favor most. You have 20 minutes to write. Write as though you were to give this speech before an audience that has many different positions on the topic. The topic chosen was, "Public Aid to Non-Public Schools." It was chosen as that topic, among five pre-tested, which best fit a theoretically expected distribution. The atti- tude measures taken at the second survey included the topic 54 chosen along with three others, used as masks. Debriefing followed the week after the last measure was obtained. Attitude Concepts Guttman and Suchman (1947) argue that attitude has at least two dimensions--direction and intensity. They ex- plain that attitude direction should take the form of a U-shaped curve, i.e., a high frequency of occurrence of sub- jects who agree and disagree, with a relatively low fre- quency of occurrence of subjects are are undecided. In addition, attitude intensity, when plotted against frequency of occurrence, should produce an ascending line, with low intensity producing the lowest frequency of occurrence, and high intensity producing the Inghest frequency of occur— rence. In order to determine on the basis of the two cri- teria stated above which topic would be most appropriate for analysis, five concepts were chosen for measure on the first survey. Figures 1 through 10, which plot attitude direction and intensity for each topic against frequency of occur- rence, show the observed frequencies occurring on the first test. (Solid lines on Figures 7 and 8 are the pre-test distributions; dashed lines are post-test distributions.) Topic 4 generally conformed to the theoretically expected frequencies, as demonstrated by the U-shaped curve of atti- tude direction and the ascending curve of attitude intensity. 55 H OHQOB .H meoa muemcmuae assesses :OAuomueo masseuse “GOeDDQHuumeQ accosquMIl.m musmflm "coeusnwuumwo xocmsvmumll.a madman auemcmuce assessed a m m H o a a j . I . II ON TION . T I T Yd T v m r b 3m . 1.3 e in u I. 0 1| K I .I o... Too rl r: r .. r a Kouenbaia 56 N chaos suemcmuce assessor "coHusneuumflo wocmsqmum|l.q madman xuflmcmucH a m N x. _ . t—r—l o //////I .luov N Oflmoa coauomuflo wosueuud "coHusofluumflQ husmsvoumll.m mesmem meanness a a a _ _ _ o . [non .d I I 8 I: .m a ow u . .. D o 'A '| 1.0m Kouenbaig 57 m OHQOB m OHQOB auemcmucH mosueuua coHuomueo mosuauua "coflusnfluumeo zoomsvmuhuu.w madman "coflusneuumfla mocmsvmum:l.m mssmflm abemcmbcH mcsueuuc v m m a O D 4 _ _ _ H o _ _ r o T I 0 v Kouenbeig T O <- I I T Aouenbeig v Demos e canoe 58 huflmcwucH mosueuue :oHuomHHQ mosuwpue “neeusneuumflo >Ocm5qmumII.m wusmflm ”coflusneuumeo >ocwsvaMII.e musmflm mnemcmucH mongoose v m N H o D < t _ _ m o _ _ e o \e I I \ \ .I F T .I II on IIom . m I . I e I .I I I I b . l 3m l3 \ I u l . o v .A I In om low I F Kouenbeig 59 m memos suamcmncH mssueuua “coeusnfluumfio wocmsvmumll.oa seamen muemcmucH e m m H _ _ e — \\\\\\\\. IIjII I I 1’] i 1 I I I—I O N o v I m Demos cofluomuea mosuwuud ”coausnwuumfla wonosmeMII.m seamen masseuse a s a _ e m o . IIo~ J ' I a '- w . IS 8 u ' 3 A . I IIom Kouenbeig m oedoa m Demos suemcmuce masseuse coeuomueo masseuse "coflusofluumeo hucmsquMII.oa musmflm "coeusnfluumeo accosvaMII.m madman 59 wuflmcmucH mpsufiubd e m N H D _ _ _ t o _ o a II ON 0 TION e I. ”an I: m I I e .m rI .o I 3. n ow n I 8 3 u j w T m - A II om rlom Tl II T I: m H we YOL 60 Tables 2 and 3 present the summaries of the pre-test fre- quency distributions for both attitude direction and intensity (N=112). TABLE 2.--Summary of frequency distribution: attitude direction. (N=112) Topics Direction I II III IV V Agree 54 44 40 59 40 Undecided 16 22 17 5 20 Disagree 42 46 55 48 52 TABLE 3.--Summary of frequency distribution: attitude intensity. (N=112) Topics Intensity I II III Iv v 4 34 46 25 50 31 3 48 44 45 42 54 2 24 22 25 18 19 1 5 7 17 2 8 The five concepts included in the first survey were: (1) Do you believe the United States should withdraw all military troop assistance to South Viet Nam? (2) DO you believe Senator Edward M. Kennedy's recent involvement 61 in the fatal accident disqualifies him for President of the United States? (3) Do you believe churches in general are too involved in politics? (4) Do you believe public aid to non-public schools should be prohibited? and (5) Do you be- lieve the current college student protests are justified? Measures of direction and intensity were as outlined in the previous chapter. On the second survey, prior to encoding, Topic 4 again was measured in order to insure that effects of time would be controlled. The dashed line of Figures 7 and 8 show the results of the second measure, with 44 subjects who agreed; 38 subjects who disagreed, and 16 who were neu- tral. Intensity levels show 3 subjects at 1; 18 at 2; 37 at 3, and 40 at 4. The scores generally conformed to the criteria, allowing for the post—test measure of attitude as appropriate for analysis. Sex, Level, and Age Since it was considered possible that sex, level, and age might be important independent variables explaining encoding behaviors, frequencies were obtained on the pre- test measure. Fifty-three males and 45 females comprised the final sample of 98 subjects in the study. Level was determined as the frequency of occurrence of freshmen, SOpho- mores, juniors, and seniors. The sample included 30 fresh- men, 64 sophomores, 3 juniors, and l senior. The latter two 62 levels were collapsed into the sophomore group for analysis of the data. The range in age was from 17 to 45, with the modal age being 18. Eleven students were 17; 48 were 18; 21 were 19; 4 were 20; l was 21; 6 were 22; 2 were 23; 2 were 24; 2 were 25; l was 30, and l was 45. Age variables retained for analysis were 17 (N=1l); 18 (N=48); l9 (N=21), and 20+ (N=18). School Based upon an interview with college officials prior to gathering of the data, it was discovered that Grace Bible College--a four-year, non-accredited, religious institution in the Grand Rapids area--had a significant number of stu- dents attending classes at Grand Rapids Junior College. Be- cause of possible bias in the sample, each subject was asked which college he attended. Of the final sample, 28 students reported they also attended Grace Bible College, and 70 students said they attended only Grand Rapids Junior Col- lege. Since this variable could prove important in the re- sults of the study, the distinctions were retained for analysis. Coding and Scoring of Message Data Seven major stages in the coding of message data were followed sequentially to insure all words would be tagged. In addition, an internal consistency check was 63 made. Messages were typed from the original handwritten manuscripts, maintaining the language as it was presented without making grammatical corrections. If a word was mis- spelled, it was corrected at the typewriting stage. For example, the word ”thier" (gig) was corrected in order to reduce possible error in coding procedures. Stage 1 Three coders independently tallied the total number of words encoded by each subject. Where tallies differed, coders were asked to recount. This variable became TOT-l, as described in Table 1. Stage 2 Five coders consecutively tagged articles a, "an," and "the," with the best estimate of the total arti- cles used by a subject being that total tagged after the first coder, followed by the second, etc., made the appro- priate tags. These were tagged as ART, specified in Table 1. The same procedures were followed for prepositions, conjunctions, interjections, interrogative pronouns, rela- tive pronouns, and demonstrative pronouns. Prepositions were tagged as PREP, as described in Table 1. All other words tagged at this stage were counted as OTHER, as speci- fied in Table 1. Any differences were resolved by the coders. 64 Stage 3 All words tagged in Stage 2 were masked out, with five coders instructed to tag all subjects and objects of verbs. These words were tagged as 81. In addition, coders tagged these words as having some kind of modifier (becoming Sld in Table l), or no modifier (becoming SlP in Table 1). Differences were resolved by the coders. Stage 4 After all words in Stages 2 and 3 were masked out, five coders were instructed to tag all adjectives, adverbs, and objects of prepositions. Prepositions were specially marked on this version of the manuscript to aid coders in identifying objects of prepositions. These words were tagged as limiters (L), with differences again resolved by the coders. In addition, coders determined which limiters modified verbs (LCl), and which modified subject words (LSl). Stage 5 With all words tagged in Stages 2 through 4 masked out, the remaining unmasked words were assumed to be verbs or verb phrases. Coders tagged each verb (see Table l for summary) as action (both I and 8), comparison (C), existence (EXT), or predicate adjective (ADJ). Further, these same verbs were coded as negation (N); past (PA), present (PR), or future (FU) tense; indicative (I) or subjunctive (S) mood, and transitive (T) or intransitive (R). Finally, 65 all verbs which were said to have a modifier were tagged as defined (D), with all others tagged as primitive (P). As before, differences were resolved by the coders. Stage 6 Subject words (51) tagged at Stage 3 were extracted from the messages and, removing redundancies, were judged independently by five coders. They judged the words as to whether the objects to which they referred could be sensed (tasted, smelled, touched, heard, or seen) or not sensed. Words marked as sensed were coded afferent (Sl-A), while not-sensed words were coded efferent (Sl-E). An inter—coder agreement of 60 per cent was required as minimal for the dichotomous distinctions, according to the following formula: Summed Scores Intercoder Agreement = 2N where 3 out of 5 judges agreeing scored 0; 4 out of 5 agree- ing scored 1, and 5 out of 5 agreeing scored 2. N was the total words judged. Subject words in the messages were then tagged as afferent (sensed) or efferent (not sensed), based upon the majority judgement of coders. Frequency of occurrence of such words for each message was calculated. Stage 7 Limiters tagged in Stage 4 were extracted from the messages, and, removing redundancies, were judged as were the subject words in Stage 6 (LSl-A; LSl-E; LCl-A, and 66 LCl-E). The same intercoder agreement criterion and scor- ing procedure was used. Internal Consistengy A critical concern of the researcher who uses con- tent analysis is the internal consistency of his tags. Based upon the theoretic identification of three basic ele- ments made in Chapter 2 (subject words, connectors, and limiters), equations were developed to determine inconsist- encies in coder tallies. Internal consistency was said to hold if the following equations* for each message were true: 1. f(Sl) = f(SlP)+f(Sld)+f(NSlP)+f(NSld) 2. f(Sl) f(Sl-A)+f(Sl-E)+f(NSl-A)+f(NSl-E) 3. f(Cl) f(ClP)+f(Cld)+f(NClP)+f(Cld) 4. f(Cl) = f(IPA)+f(IPR)+f(IFU)+f(SPA)+f(SPR)+ f(SFU)+f(NIPA)+f(NIPR)+f(NIFU)+f(NSPA)+ f(NSPR)+f(NSFU) 5. f(C1) = f(IT)+f(IR)+f(ST)+f(SR)+f(NIT)+f(NIR)+ f(NST)+f(NSR)+f(ICE)+f(ICM)+f(ICS)+ f(ICP)+f(ICT)+f(IADJ)+f(IEXT)+f(NICE)+ f(NICM)+f(NICS)+f(NICP)+f(NICT)+ f(NIADJ)+f(NIEXT)+f(SCE)+f(SCM)+f(SCS)+ f(SCP)+f(SCT)+f(SADJ)+f(SEXT)+f(NSCE)+ f(NSCM)+f(NSCS)+f(NSCP)+f(NSCT)+ f(NSADJ)+f(NSEXT) 6. f(L) f(LSl-A)+f(LSl-E)+f(LCl-A)+f(LCl-E)+ f(NLSl-A)+f(NLSl-E)+f(NLCl-A)+f(NLCl-E) *The symbols in these equations refer to the fre- quency of occurrence (f) of a message variable in which the code names of the variables are presented in Table l. 67 Whenever any one of the above equations was not true, cate- gories were retallied until the error was found. Derived Variables Recall that Table 1 presented 80 message variables. In addition to these variables, it was considered important to obtain 20 derived variables from these basic 80. A sum- mary of the variables is found in Table 4. To give an example of how these variables were obtained, Variable RC1 was obtained by summing the frequency of occurrence of SR, NSR, IR, and NIR (see Table 1). Another example is the derivation of PRIM, which was obtained by summing SlP, NSlP, ClP, and NClP. Analysis It will be remembered in Chapter I that questions were raised on the procedure for counting data, and what transformations should be used, if any, before analysis. Because most of the recent research (Stone, Dunphy, g£_gl., 1966) has used a word index score, we chose to follow this precedence. It should be understood that this study will not be able to answer the question of which coutning proce- dure--raw frequency or word index score--should be used. However, three variables--TOT-l, TOT-2, and TOT-3 (see Tables 1 and 4)--are raw frequency scores, and may give an indication of whether future studies, or this study, should be analyzed based on raw frequency data. 68 muouomccoo m>eummmc can w>euflmom mo Nocosqwum Hobos mIBoe\w HO muouomccoo Am>fiummmCIcocv m>Huwmom mo mucosqmum Hmuoe mneoe\m Hum muouooccoo can mouos uomnosm omcemmo mo Nocmsqwum Hmuoe mneoe\m ammo muouomccoo can mono: coanSm m>HuHEHud mo wocmsvouw Hmuoe mueoa\m szm muouomccoo wmcmu mucosa mo Nocmsqmum Hobos mneoe\m Dmau muouomccou mmcmu commend mo mocwsqmum Hobos mueoe\m mmau mucoumccoo mmcmu ummd mo hocmsqmum Hobos mIBOB\m dmau mouo3 comnnsm mo wocmsqwum Hobos mneoe\m Hm mumuesea mo Nocmsemum sauce mIeoe\m a mumuHEwH one mouo3 oomnosm economwm mo Nocmsqmum Hmuoe muaoa\m mum mumuHEwH one mouo3 pomMQSm nauseous mo hocosqmum Hooch mleoa\m mmm muouomccoo o>wummmc mo mucosqouw Hmuoa musoa\m Hoz muouomccoo o>wuwmcouucfi mo hocmsmmum Hmuos NIBOB\M Hum muouomccoo m>euwmccuu mo aocwsqmum Hmuoe mlaoe\w HUB whosomccoo o>euocsno5m mo Nocmsqmum amuoa muaoe\m Hum muouomccoo o>euooHocfi mo aocmsqoum Hobos mueoe\m HUH muouowccoo scenes mo mucosqoum Houoe MIBOB\M a0904 muouomccoo cemwucdeou mo Nucmsqoum Hmuoa mIeoa\m ozoo muouomccoo oco .mumuweea .mouo3 pomnnsm mo mocmsqmum Hmuoe w MIBOB umcuo can .chNuemOQwum .mmaoauum mo Eda ecu mama cocoons mouo3 Hobos w Nueoe CONDNcHumo musmmmz manowuc> H E .mmaoowus> oo>wuoo mo wussssmII.v momma 69 Following Stone, et al., the following word index transformation score was used: f(X) f(CT) Word Index = where f(X) is the frequency of occurrence of a tag divided by the total frequency of occurrence of all tags (f(CT)). The denominator was the TOT-3 score, which effectively equates 1 connector in 10 words with 20 connectors in 100 words. It is not reasonable to assume that all the possible tags a researcher can derive are meaningful. It also is not reasonable to assume that one method of analysis is ap- propriate for all meaningful distinctions. The hypotheses have been formed in terms relevant for multiple regression and one-way analysis of variance. McNemar (1962) denotes that the assumptions of a multiple regression analysis are: (1) interval data; (2) normal distribution; (3) homogeneity of variance, and (4) linearity. These assumptions--in mul- tiple regression analysis--apply both to predictor and cri- terion variables. In one-way analysis of variance, assump- tions of intervality, normality, and homogeneity of variance are relevant. Stone, Dunphy, g£_gl. (1966) indicate that the intervality and homogeneity assumptions are justifiable, particularly when using a word index transformations. Based upon the studies they cite in relation to their work on the General Inquirer, it seems justifiable for the current data 70 to accept these assumptions. Both the assumptions of nor- mality and linearity were considered to be critical for multiple regression analysis, and were checked. Given these issues, the following step-by-step pro- cedure for analysis and screening of variables was followed: 1. Any variable which had 0 frequencies for all subjects was eliminated before analysis. If one-fourth of the subjects had 0 frequencies on a variable, it was eliminated before analysis. Basic statistics were calculated for each re- maining variable, with measures of kurtosis and skewness derived in order to determine whether the distributions were normally distributed or not (McNemar, 1962). Any significantly skewed, leptokurtic or platykurtic distributions (p < .01) were eliminated from multiple regres- sion analysis. For those variables retained from Step 3, and which were significantly correlated (p < .05) with any one of the four cognitive variables, etas were calculated to determine whether it was reasonable to assume the variables were sig- nificantly non-linear (p < .01).* *None of the variables were retained for multiple regression analysis after Stage 3. Stage 4 procedures were done, and are reported in Appendix B. Stage 5 procedures were not done. 71 5. For those message variables retained after Step 4, a multiple regression analysis was obtained on each of the cognitive measures (dogmatism, anxiety, vocabulary skills, and spatial rela- tionships abilities).* 6. A frequency distribution of variables retained from Step 2 was calculated, with judgements made as to the appropriateness of one-way analysis of variance. Median splits of the cognitive vari- ables were obtained and, using them as independ- ent variables, a one-way analysis of variance on all message variables (as dependent measures) was calculated. Age, sex, level, school, atti- tude direction, and attitude intensity also formed independent variables for analysis. A significant F was set at p < .05. It will be noted that Step 6 implies less stringent require- ments for normality of distribution than that required for multiple regression analysis. Justification for this pro- cedure is based upon the Norton and Boneau studies (Lindquist, 1953) in which the robustness of analysis of variance was supported. *Ibid. 72 Summary This chapter has presented a sample description and procedure for analysis of the data used in this study. Scoring and coding of message data was explained, and a six- stage procedure for analysis of the data was outlined. Chapter IV will present the results of the analysis, with the summary, discussion, and suggestions for further study to be included in Chapter V. CHAPTER IV RESULTS This chapter will have three main divisions. The ifirst will report the inter-coder agreement scores of the aafferent-efferent distinctions for subject words and lim- The second section will present the results of the Finally, 1 ters. krypothesized relationships made in Chapter II. anhypothesized, but significant, results will be reported. Eliscussion of these results is delayed until Chapter V. Afferent-Efferent Distinctions Chapter III described the procedure for establishing ‘tllee level of reliability of the afferent-efferent distinc- ti(on, both for subject words and limiters. Inter-coder a91::eement for subject words was 66.11 per cent, with 509 words judged and a score of 673. Inter-coder agreement for :LjLIniters was 68.96 per cent, with 1,872 words judged and a score of 1,291. These percentages met the minimum \ *Chapter III presented the formula for determining IFEIe level of inter-coder agreement, in which 3 out of 5 \ldges agreeing on a word was scored as 0; 4 out of S was 8Qored as l, and 5 out of 5 was scored as 2. These scores EEre summed (reported as "score" in this chapter), and were (licvided by two times the words judged. 73 74 requirement as specified in Chapter III, and allowed for the maintenance of the distinction in analysis. Hypothesized Relationships Thirteen variables were included in the hypothesized zuelationships made in Chapter II, with twelve of these \rariables submitted for analysis at Stages 1 and 2. The Iaemaining variable--attitude intensity--did not require such {procedures, since no assumption was being made as to its jgntervality, homogeneity, or linearity. Table 5 presents a ssummary of descriptive statistics of the variables, exclud- jmng attitude intensity. While all of these variables were trot eliminated at Stages 1 and 2, only dogmatism and spatial Etbdlities were not significantly skewed, leptokurtic, or platykurtic. Table 6 presents a summary of the results of Simple correlations for the variables, none of which were Significant (p < .05). It was concluded that multiple re- QJIflession analysis was not appropriate, and therefore four of 111112 five hypotheses as stated in Chapter II could not be thted. Hypothesis three still could be tested. Based upon a frequency distribution for the message "Eiriables, a one-way analysis of variance was judged appro- E3111‘iate. A11 distributions were unimodal, with a curve which atE>proached a bell-shape, although the variables were sig- r‘m‘ificantly skewed or kurtotic. As outlined in Stage 6 of Eil‘ialysis in Chapter III, median splits were obtained for (itngmatism, anxiety, spatial abilities, and vocabulary 7S .Ho. v a .memouusx ucmosuecmam44 .Ho. v d .mmocsmxm unmowmacmama saavvv.m «memm.ou ommo. mva. omom. omNN. A savaom.N «mmom.o move. nomN. ovvv. oeHN. am «amomm.N «oaoo.o mmao. emHo. ommo. o mzou «aammo.m aavma.o memo. mmva. oomN. ovoo. Dam «Immoa.m «mmmN.o mmNo. mmmo. ooea. oovo. oau a:amom.n «mmmm.a Heme. mmNo. omqa. o Dmao «aaemm.m «Hmeo.o mmNo. homo. omva. o mmau «amovo.m omnm.o No.n mH.HN oq N mawxm .> mmon.N ova.o mm.Nm em.ova mam «v .mmm.sdm «avvvm.N memo.o mm.HH mm.mm mm mm xcd «soqeo.m omom.on vm.va mo.ve moa mm moo memouusx mmmczoxm .o.m com: coax 304 odomeum> .3342? 83.6285 ”morass; Endgame be h~e§Ie BE 76 TABLE 6.--Summary of correlations: hypothesized variables. .--=—-—-r~ 1~ Hypothesis Variables Correlation H1 C1PR--Dog -.06 H1 C1FU--Dog -.06 H1 C1PR--Anx .13 H1 ClFU--Anx -.11 H2 NCl---Dog .12 H2 NCl---Anx .13 H4 COMP-~Spa ng. -.06 H5 Sl--V.Skls -.14 H5 L---V.Skls -.16 skills. These variables became independent variables, with present (C1PR), future (ClFU), and negation (NCl) connectors as dependent variables. Other dependent variables included defined subject words (SlD), defined connectors (C1D), com- parison connectors (COMP), subject words (51), and limiters (L). Following are the results: Hypothesis 1: The relative frequency of occurrence of present and future tense verbs will explain significant variance in dogmatism and anxiety. Based upon Stage 6 pro- cedures, one-way analyses of variance were calculated with dogmatism and anxiety as independent variables. Anxiety and dogmatism were found to be significantly correlated 77 (p < .05), as assumed in a corollary to this hypothesis in Chapter II. Correlation was .31. Tables 7 through 10 pre- sent a summary of the one-way analysis of variance, with no significant differences found. TABLE 7.--Dogmatism; summary of one-way analysis of vari— ance: present tense connectors (C1PR). Source df SS MS F Between 1 .0002 .0002 .24 Within 96 .0840 .0009 Total 97 .0842 TABLE 8.--Dogmatism; summary of one-way analysis of vari- ance: future tense connectors (ClFU). Source df 88 MS F Between 1 .0002 .0002 .26 Within 96 .0561 .0006 Total 97 .0563 TABLE 9.--Anxiety; summary of one—way analysis of vari- ance: present tense connectors (C1PR). ¥ Source df 88 MS F Between 1 .0002 .0002 .19 Within 96 .0840 .0009 Total 97 .0842 g 78 TABLE lO.--Anxiety; summary of one-way analysis of variance: future tense connectors (ClFU). Source df SS MS F Between 1 .0004 .0004 .72 Within 96 .0558 .0006 Total 97 .0562 Hypothesis 2: The relative frequency of occurrence of verb negations will explain significant variance in dog- matism and anxiety. This hypothesis as stated could not be tested. However, Stage 6 procedures were used, with dogma- tism and anxiety as independent variables, and negative con— nectors as the dependent variable. Tables 11 and 12 are summaries of the analyses, with no significant differences found. TABLE ll.--Dogmatism; summary of one-way analysis of vari— ance: negation connectors (NCl). Source df SS MS F Between 1 .0003 .0003 .69 Within 96 .0370 .0004 Total 97 .0373 g 79 TABLE 12.--Anxiety; summary of one-way analysis of vari- ance: negation connectors (NCl). Source df SS MS F Between 1 .0004 .00044 1.15 Within 96 .0368 .00038 Total 97 .0372 Hypothesis 3: Those subjects who hold different attitude intensity levels on the topic will show significant differences (p < .05) in the relative frequency of occur- rence of defined connectors (C10) and defined unit signs (81D). Stage 6 procedures were used with attitude intensity as the independent variable, and defined words (510 and C1D) as dependent measures. Tables 13 and 14 are the summaries of the analysis, with no significant differences obtained. Hypothesis 4: The relative frequency of occurrence of comparison verbs will explain significant variance in TABLE l3.--Attitude intensity; summary of one-way analysis of variance: defined subject words (SlD). Source df SS MS F Between 3 .0003 .0001 .07 Within 94 .1359 .0014 Total 97 .1362 ¥ 80 TABLE l4.--Attitude intensity; summary of one-way analysis of variance: defined connectors (C1D). Source df SS MS F Between 3 .0016 .0005 .80 Within 94 .0617 .0007 Total 97 .0633 spatial reasoning. Again Stage 6 procedures were used, with spatial reasoning as the independent variable, and com- parison verbs as the dependent measure. Table 15 is a sum- mary of the analysis, with no significant differences ob- tained. Hypothesis 5: The relative frequency of occurrence of subject words and limiter words will explain significant variance in verbal skills. With Stage 6 procedures, verbal skills was used as the independent variable. Dependent measures were subject words and limiters. Tables 16 and 17 TABLE 15.--Spatial reasoning; summary of one-way analysis of variance: comparison connectors (COMP). Source df SS MS F Between 1 .00001 .00001 .08 Within 96 .01708 .00002 Total 97 ’ .01709 81 TABLE 16.--Verbal skills; summary of one—way analysis of variance: subject words (51). Source df SS MS F Between 1 .0023 .0023 1.07 Within 96 .2101 .0022 Total 97 .2124 TABLE l7.--Verba1 skills; summary of one-way analysis of variance: limiters (L). Source df SS MS F Between 1 .0090 .0090 1.96 Within 96 .4391 .0046 Total 97 .4481 are the summaries of the analysis, with no significant dif- ferences found. Unhypothesized Results Based upon procedures for analysis outlined in Chapter III, no multiple regression analysis was performed. However, 40 message variables were judged appropriate for one-way analysis of variance, as specified in Stage 6 of the procedure. Appendix B reports three general tables: a summary of descriptive statistics of all variables surviving Stage 2; a summary of correlations, with etas for those 82 variables significantly correlated, and a summary of one-way analyses of variance of those variables judged adequately distributed. Of forty variables analyzed according to Stage 6, 54 significant differences were obtained out of 480 ANOVAS. The significant results will be reported here ac- cording to independent variable.* Dogmatism Low dogmatic subjects encoded significantly more words, regardless of which measure was used. Significance levels were beyond p < .05 for TOT-1 (F=7.55, p < .007); TOT-2 (F=8.45, p < .004), and TOT-3 (F=1l.20, p < .001). However, high dogmatic subjects encoded significantly more articles (ART) (F=7.63, p < .009); significantly more in- dicative reflexive connectors (IR) (F=7.45, p < .008), and significantly more total reflexive connectors (RC1) (F=7.99, p < .006). *Twelve variables were eliminated at Stage 1. They included: SCM, NS-S, NT-O, NA-O, NICT, NSCT, NSCS, NSCP, NIADJ, NIEXT, NSADJ, and NSEXT. Thirty-four variables were eliminated at Stage 2, and included: NGO, NLCl-E, NICS, NICP, NIR, NSPA, NSPR, NSFU, NSCE, NSCM, NST, NSR, NSl-A, NSl-E, NLSl-A, NLCl-A, SPA, SFU, SCE, SCS, SCP, SCT, SADJ, SEXT, NSlP, NSlD, NIPA, NIFU, NICE, NICM, ICM, ICS, ICP, and ICT. The remaining 54 variables were eliminated at Stage 3. Of these variables, fourteen were judged inappropriate for Stage 6 procedures. They included: IPA, IFU, ICE, ST, SR, NIPR, NIT, NClP, NClD, NLSl-E, T-O, A-O, IADJ, and IEXT. The forty variables judged appropriate for Stage 6 proce- dures included: TOT-l, SlP, SlD, ClP, C1D, SlA, SlE, LSl-A, LCl-A, LSl-E, LCl-E, IPR, IT, IR, SPR, S-S, G-O, ART, PREP, 0TH, TOT-2, TOT-3, NCl, COMP, ClACT, ICl, SCl, TCl, RC1, Cl, AFF, EFF, L, 81, ClPA, ClPR, ClFU, PRIM, DEFD, PC1. Keys to these variables may be found in Tables 1 and 4. 83 Anxiety High anxious subjects encoded significantly more total reflexive connectors (RC1) (F=5.28, p < .02); signifi- cantly more total primitives (PRIM) (F=4.85, p < .03), and significantly less total defined words (DEFD) (F=4.86, p < .03). Low anxious subjects encoded significantly more total words using TOT~3 as the measure (F=5.01, p < .03); signifi- cantly more defined subject words (SlD) (F=4.33, p < .04), and significantly more efferent limiters of subject words (LSl-E) (F=4.78, p < .03). Spatial Relationship Abilities Those subjects who had high spatial relationship abilities encoded significantly more other (0TH) words (F=S.73, p < .02). This category included relative, demon- strative, and interrogative pronouns, as well as exclama- tions. Subjects with high spatial skills also encoded sig- nificantly more negation connectors (NCl) (F=4.64, p < .03). For total words, regardless of how they were measured, sub— jects who were low in spatial abilities encoded more TOT-1 (F=4.59, p < .04); more TOT-2 (F=4.79, p < .03), and more TOT-3 (F=5.94, p < .02). VocabularyVSkills Those subjects who had high vocabulary skills en- coded significantly more TOT-l (F=6.89, p < .01); signifi- cantly more TOT-2 (F=6.15, p < .02); significantly more 84 TOT-3 (F=5.39, p < .02); significantly more efferent subject words (Sl-E) (F=3.96, p < .04); significantly more indica- tive reflexive connectors (IR) (F=4.23, p < .04); signifi- cantly more prepositions (PREP) (F=6.00, p < .02), and significantly more total efferent words (EFF) (F=5.96, p < .02). However, those subjects who had low vocabulary skills encoded significantly more afferent subject words (Sl-A) (F=12.4l, p < .001); significantly more generalized—other words (G—O) (F=8.90, p < .004); significantly more total comparison connectors (COMP) (F=6.59, p < .01), and signifi- cantly more total afferent words (AFF) (F=8.29, p < .005). Attitude Direction Eight significant differences were obtained in the encoding behaviors of subjects, when related to direction of attitude. Significant differences were obtained for primi- tive subject words (810) (F=3.83, p < .03); afferent subject words (Sl-A) (F=3.64, p < .03); indicative comparison equal connectors (ICE) (F=6.89, p < .002); prepositions (PREP) (F= 3.19, p < .05); total comparison connectors (COMP) (F=4.21, p < .02); total efferent words (EFF) (F=4.07, p < .02); total limiters CL) (F=3.1l, p < .05), and total primitive words (PRIM) (F=3.64, p < .03). A Scheffe's t-test was used to selectively compare means of attitude directions in the above differences ob- tained. An F at p < .05 was used as the test of signifi- cance. Those who disagreed encoded significantly more 85 primitive subject words (SlP) than those who were undecided, but not significantly more than those who agreed. Those who disagreed encoded significantly more afferent subjgct words (Sl-A) than those who were undecided, but not significantly more than those who agreed. Those who disagreed encoded significantly more indicative comparison equal connectors (ICE) than either those who were undecided or those who agreed. Those who disagreed encoded significantly fewer prepositions (PREP) than those who were undecided but not significantly fewer than those who agreed. Those who dis- agreed encoded significantly more comparison connectors (COMP) than those who agreed, but not significantly more than those who were undecided. Those who disagreed encoded significantly fewer efferent words (EFF) than those who were undecided, but not significantly fewer than those who agreed. Those who disagreed encoded significantly £3335 limiters (L) than those who were undecided, but not signifi- cantly fewer than those who agreed. Those who disagreed encoded significantly more primitive words (PRIM) than those who were undecided, but not significantly more than those who agreed. Attitude Intensity Only one dependent variable--subjunctive present tense connectors (SPR)-—was found to be significant (F=2.79, p < .05), when attitude intensity was the independent vari- able. A check using Scheffe's t-test, however, showed that 86 low intensity subjects did not encode significantly more subjunctive present tense connectors than those subjects with higher intensity of attitude. Level Two message variables were found to be significant when level in college was the independent variable. Soph- omores encoded significantly more generalized-other words (G-O) than freshmen (F=4.79, p < .03), and significantly more articles (ART) (F-4.44, p < .04). School Five message variables were found to be significant, when college--Grace Bible College or Grand Rapids Junior College--were the independent variables. Subjects who also attended Grace Bible College encoded significantly more primitive connectors (ClP) (F=5.62, p < .02); significantly more efferent subject words (Sl-E) (F=9.55, p < .003); sig- nificantly more articles (ART) (F=9.61, p < .003); signifi- cantly more action connectors (CACT) (F=5.42, p < .02), and significantly more primitive words (PRIM) (F=4.82, p < .03). Errors in Spatial Abilities and Vocabulary Skills Chapter 11 noted that in the absence of any recom- mended scoring procedure by French, errors in spatial rela- tionships and vocabulary skills would be kept as separate variables. Ten significant differences were obtained, when 87 errors were used as an independent variable. Those who made high errors in spatial relationships encoded significantly more primitive connectors (C1P) (F=5.05, p < .03); signifi- cantly more afferent limiters of connectors (LCl-A) (F=5.94, p < .02); significantly more indicative comparison equal connectors (ICE) (F=6.26, p < .01); significantly more com- parison connectors (COMP) (F=8.90, p < .004); significantly more subject words (51) (F=8.24, p < .005), and significantly more total primitives (PRIM) (F=4.73, p < .03). Those sub- jects who made low errors in spatial relationships encoded significantly more prepositions (PREP) (F=7.76, p < .006); significantly more total reflexive connectors (RC1) (F=4.72, p < .03), and significantly more limiters (L) (F=6.98, p < .01). Those subjects who made low vocabulary errors encoded significantly more articles (ART) (F=6.36, p < .01). Sex and Age No significant differences were obtained with any of the message variables as dependent measures, when sex or age was the independent variable. Summary This chapter reported the pre—test results of the distributions of attitude direction and intensity, justify- ing Topic 4 as the best concept for analysis. Further, re- liability scores of the afferent-efferent distinction were reported, demonstrating sufficient agreement between judges 88 to maintain the distinction. Multiple regression analysis was not used due to failure to meet criteria established in Chapter III. However, Stage 6 procedures were used. No significant differences were obtained on hypothesized rela- tionships. However, 54 significant differences were ob- tained on relationships not hypothesized. Chapter V will present a discussion of these findings. CHAPTER V DISCUSSION The discussion of the results of this study will be presented in four parts: (1) a summary of the rationale; (2) interpretation of the results; (3) evaluation of the syntactical theory of signs, both as to its theoretic importance and its place in the study of communication, and (4) recommendations for further study. No formal attempt will be made to eXplain the results in terms of theories associated with the cognitive structures. It is our concern here to see if a syntactical theory of signs is fruitful, and whether it justifies further study. Summary of Rationale One of the important assumptions in content analy- sis stated in Chapter I was that verbal behavior reflects to some degree the condition or state of affairs of the person encoding the message. An argument was presented for a quantitative and syntactical approach to the study of meaning. To make such an approach as meaningful as possible, a broad classification system of signs was con- sidered desirable. 89 90 The descriptive system offered in this study included three basic elements: (1) limiters; (2) subject signs, and (3) connectors. Limiters were described as modifying either subject signs or connectors. In addition, they were said to be afferent (referring to objects that could be sensed through normal sensory inputs) or efferent (referring to objects that could not be sensed through normal sensory inputs, i.e., ideas or internal states). Subject signs were also said to be either afferent or efferent in the same senses as limiters. In addition, subject signs were said to refer to people, which could be one's self (S-S)--the encoder, a specific other person or group (A-O), a generalized other person or group (G-O), or the receiver of the message (T-O). Unit signs were said to include a subject sign and any limiters of that sign. If a subject sign had at least one limiter, it was termed defined (81D); if the subject sign had no limiter, it was termed primitive (SlP). The use of the term unit sign allowed not only for the above distinction, but a category system which allowed for the study of syntactical meaning, both within unit signs and between unit signs. A quantitative, comprehensive study of connectors has been generally lacking in the field, as was pointed out in Chapter II. Borrowing from grammar, linguistics, and philosophy, a comprehensive system of between-unit sign syntactical relationships was offered in Chapter II. Termed connectors, these relationships were said to 91 include distinctions between indicative and subjunctive moods, transitive and reflexive, positive and negative, action and comparison, and, as with subject signs, primi- tive and defined. If verbal behavior does reflect the condition or state of affairs of the person encoding the message, it was considered reasonable to assume that such cognitive variables as dogmatism, anxiety, vocabulary skills, spatial skills, and attitude should explain variance in the message variables. Initially, such relationships were posited in the form of multiple regression analysis with the acception of attitude. A relatively rigorous pro- cedure was designed to determine the justifiability of such an analysis. The results, as stated in Chapter IV, indicated a multiple regression analysis was not appropri- ate. One-way analyses of variance were calculated, with 54 significant differences obtained out of 480 analyses. Since demographic characteristics were considered to be potential sources of variance in the message variables, these also were analyzed using Stage 6 procedures (Chapter III). Five hypotheses--four multiple regression, and one analysis of variance--were presented in Chapter II, with no significant results obtained. It was pointed out, however, that hypotheses might not be appropriate for this study, but should rather be generated as the result of the study. The purpose of the study was not to examine 92 theories associated with the cognitive processes, but to determine the adequacy of the classification system and to develop hypotheses for future studies. The result is that reliability of these findings must await a replication of the study. Interpretation of Results In order to facilitate the goals of this study, an interpretation of results is offered by levels. Based upon the 40 message variables which were analyzed at Stage 6 and the 54 significant differences reported in Chapter IV, this procedure allows an informal comparison of information derivable from the classification system. The procedure begins with examination of the grossest distinctions in the classification system to the finest distinctions. To illustrate, we will ask if Level 2 pro- cedure will evoke more information about antecedent characteristics of the source than Level 1. Level 1 is concerned with the information yield of basic distinctions, including total words (TOT-1, TOT—2, and TOT-3), subject words (81), limiters (L), connectors (Cl), articles (ART), propositions (PREP), and other (0TH). (See Tables 1 and 4 for an explanation of what these vari- ables include.) While the basic elements of the classi- fication system are three [subject words (81), limiters (L), and connectors (C1)], the total message corpus is comprised of these three elements and the frequency of 93 occurrence of articles (ART), prepositions (PREP), and other (0TH). It will be remembered that TOT-1 includes all words encoded regardless of classification; TOT-2 includes TOT-1 minus articles (ART), prepositions (PREP), and other (0TH), and TOT-3 includes the sum of the three basic ele- ments of the classification system. Level 2 is concerned with the yield of information obtained when we know more than simply the frequency of occurrence of subject words (51). For example, we might ask what new information is obtained if we know the rela- tive frequency of occurrence of efferent subject signs (Sl-E) and afferent subject signs (Sl-A). Level 3 is concerned with the yield of information obtained when we know more than the frequency of occurrence of limiters (L). For example, what new information is obtained if we know the relative frequency of occurrence of afferent limiters of subject signs (LSl-A) and efferent limiters of subject signs (LSl-E). Level 4 is concerned with the yield of information obtained when we know more than the frequency of occurrence of connectors (Cl). For example, what new information is obtained if we know the relative frequency of occurrence of reflexive connectors (RC1) and the transitive connectors (TCl). Level 4 will include four variables which could not be analyzed at any of the previous levels because of their lack of mutual exclusivity of the variables. These include total afferent words (AFF), which include both 94 afferent subject words (Sl-A) and afferent limiters (LSl-A and LCl-A); total efferent words (EFF), which include both efferent subject words (Sl-E) and efferent limiters (LSl-E and LCl-E); total primitive words (PRIM), which include both primitive connectors (ClP) and primitive subject words (SlP), and total defined words (DEFD), which include both defined connectors (C10) and defined subject words (810). Level 1 Table 18 is an intercorrelation matrix of the nine elements which comprise the message corpus, including the three elements basic to the syntactical approach of this study. In general, it is reasonable to assume that the TOT-1, TOT-2, and TOT-3 scores are not independent. In addition, other (0TH) significantly covaries with TOT-3. Table 18 also shows that it is reasonable to assume that the three basic elements (81, L, and C1) are independent of either of the TOT scores.* However, prepositions (PREP) sig- nificantly covary with all three basic elements, while arti- cles (ART) significantly covary with only one of the basic elements (Cl), i.e., with prepositions (PREP). Those variables which significantly covary should evoke the same information, i.e., a cognitive structure which shows significant differences on one message variable *While a nonsignificant correlation is a necessary condition for independence, it is not sufficient. There- fore, caution is called for in this conclusion. McNemar (1963) notes other procedures necessary beyond a zero-order correlation before two variables can be judged independent. For this study, all statements of independence are subject to this qualification. 95 oo.H no. 00. mo.| mo. vo.l Nm. mo.l Ho.I mBO oo.H ON. mm. Hm.l vm.l vo. vH.I mo.| mmmm oo.H mo.I mN. no. mH.I Ho.I mo. 8&4 oo.H om.l hm.l mo. Ho.l mo. A oo.H mN. mo.l vo.+ H0. H0 oo.H Ho.l mo. Ho.l Hm oo.H mm. wm. MIBOB oo.H mm. NIBOB oo.H HIBOB mBO mmmm Bmfi A HO Hm MIBOB NIBOB H1809 .mmaoseum> a Hm>oa mo resume cowumHouuooumucHII.ma names 96 should show significant differences on those other message variables which covary with the first. In addition, con- sistent results should show the same pattern (direction of significance) of relative frequency of occurrence of vari- ables which positively covary, while those variables which negatively covary should show the opposite pattern (signifi- cant differences in the opposite tail of the distribution). Where inconsistencies occur, it should be attributable to slight or moderate correlations. Finally, one message variable can be said to gypkg more information than another if the first variable orders more cognitive structures than the second message variable. Results discussed in Chapter IV show that all five cognitive measures (dogmatism, vocabulary skills, spatial abilities, and attitude), the two error scores (vocabulary and spatial, and two demographic variables (school and level) yield information at this first level of interpre- tation. Significant differences were obtained between all three total word scores (TOT-l, TOT-2, and TOT-3) and dogmatism, vocabulary skills, and spatial abilities con- sidered as the independent variables. These results were consistent as would be expected with such a high positive correlation between the variables (see Table 18). TOT-1 and TOT-2 yield the same information, while TOT-3 yields more information. Subjects who were low dogmatic encoded more TOT-l, TOT-2, and TOT-3; subjects who were high in vocabulary skills encoded more TOT-1, TOT-2, and TOT-3, 97 and subjects who were low in spatial abilities encoded more TOT-1, TOT-2, and TOT-3. However, low anxiety subjects en— coded significantly more TOT-3 than high anxiety subjects. We can conclude that, based on the data for this sample, TOT-3 represents the best variable of the three. Connectors (C1) evoked no information about the cognitive structures or demographic variables measured. This suggests that connectors (C1) is either an unimportant variable, or--as will be demonstrated later in Level 4-- some identifiable subset of connectors is important. Significant differences were obtained in the rela- tive frequency of subject words (51) when spatial errors was the independent variable. Subjects who made many spatial errors encoded significantly more subject words than subjects who made few spatial errors. Thus, this variable (51) evokes different information from that of the previously discussed variables. Limiters (L) produced significant differences when spatial errors was the independent variable. Subjects who made few spatial errors encoded significantly more limiters than subjects who made many spatial errors. This result, combined with the result obtained with subject words, suggests there may be differences in encoding behavior when errors in general are the independent variable. At this time, however, these results are difficult to inter- pret, since a theory of "errors" in test taking has not been explored. 98 Direction of attitude was also a significant predictor of differences in the encoding of limiters (L). Subjects who disagreed with the topic encoded significantly fewer limiters than subjects who were undecided, although not significantly fewer than subjects who agreed. We can generally conclude that limiters (L) evokes different information from either of the previously discussed message variables. Three additional variables (PREP, ART, and OTH)-- completing the elements of the message corpus--were analyzed. Subjects who made few spatial errors encoded significantly more prepositions; subjects who made many spatial errors encoded significantly more prepositions; subjects who had low vocabulary skills encoded signifi- cantly more prepositions, and subjects who disagreed on the topic encoded significantly more prepositions. These results were consistent, and prepositions (PREP) evoke more information than subject words (81) and limiters (L), but not more information than either of the total word scores (TOT-1, TOT-2, and TOT-3). Subjects who made few vocabulary errors encoded significantly more articles (ART) than subjects who made many vocabulary errors. Also, high dogmatic subjects encoded significantly more articles than low dogmatic subjects. These results, however, evoke no new infor- mation. However, it was also found that subjects who 99 were sophomores encoded significantly more articles than freshmen, and subjects who attended Grace Bible College encoded significantly more articles than those who attended Grand Rapids Junior College only. While articles (ART) evokes more information, caution should be exercised since the level category of sophomores represents a collapsing of subjects who were juniors and seniors--a11 attending Grace Bible College, which is a four-year institution. Figure 11 presents a summary of the information yielded with Level 1 interpretation procedures. It shows that at the most gross distinctions, eight of the nine message variables yield some information about all of the cognitive characteristics of the encoders, and two of the four demographic characteristics of the encoders. With all nine variables at Level 1 taken together, Figure 11 shows that total words encoded (TOT-3) elicits the most information-~being related to dogmatism, anxiety, vo- cabulary skills, and spatial abilities. Prepositions (PREP) and limiters (L) provide information about the encoder's attitude direction; articles (ART), prepositions (PREP), limiters (L), and subject words (51) provide information about errors in vocabulary skills or spatial abilities, and articles (ART) evokes information about level and school. If this study had stopped at this point, we would have obtained little new information about messages. DEMOGRAPHIC CHARACTERISTICS 100 COGNITIVE CHARACTERISTICS Sex Dogmatism TOT-1 TOT-2 TOT-3 ART Anxiety TOT-3 Age Vocabulary Skills TOT-l TOT-2 TOT-3 PREP Level Spatial Abilities Art TOT-l TOT-2 TOT-3 0TH Attitude Direction. PREP L — — ————-— ——. ~— School Intensity Art Errors Vocab. ART PREP _— 35;: 121— L 81 Figure ll.--Summary of Level 1 Interpretation. 101 Researchers (see Chapter I) have known for some time that total words encoded, articles, and prepositions have been useful in inferring differences between encoders. This study, at this point, only adds to the validity of these findings with measures of specified cognitive and demo- graphic characteristics of the source. The next three levels seek to provide more information about the encoder, and at the same time, help to determine the adequacy or inadequacy of the classification system. Level 2 Table 19 presents an intercorrelation matrix of those message variables which were analyzed using Stage 6 procecures (see Chapter III). In general, it is reasonable to assume that afferent subject words (Sl-A), efferent subject words (Sl-E), source-specific words (S-S), gener- alized other words (G-O), primitive subject words (SlP), and defined subject words (SlD) are pgp_independent of the total number of subject words (51). Only defined subject words (SlD) is independentfof afferent subject words (Sl-A), while defined subject words (51D) and source specific words (S-S) are independent of efferent subject words (Sl-E). Only generalized other words (G-O) is independent of source specific words (S-S), and defined subject words (SlD) is independent of generalized other words (G-O). All results were consistent. *Ibid. 102 TABLE l9.--Intercorre1ation matrix of level 2 variables. S1 Sl-A Sl-E S-S G-O SlP 81D 51 1.00 Sl-A .57 1.00 Sl-E .25 -.51 1.00 S-8 .20 .24 -.10 l 00 G-O .47 .70 -.22 -.12 1.00 SlP .70 .51 .08 .44 .32 1.00 81D .31 .03 .21 -.33 .17 -.47 1.00 Of the seven message variables at Level 2, no significant information about either cognitive or demo- graphic characteristics was obtained from source Specific (S-S) words. It has already been noted that total subject words (81) evokes information about spatial errors of the encoder. It was found, however, that subjects who had low vocabulary skills encoded significantly more afferent subject words (Sl-A) than those who had high vocabulary skills. This suggests that subjects who have low vocabu- lary skills may be more dependent upon sensory (afferent) inputs for their language than non-sensory (efferent) in— puts. Subjects who disagreed with the tOpic encoded significantly more afferent subject words (Sl-A) than subjects who were undecided, but not significantly more than subjects who agreed. These two findings suggest the possibility of an interaction effect between subjects who 103 have low vacabulary skills and disagree with a topic in terms of the relative frequency of occurrence of afferent subject words (Sl-A) and dependence upon sensory inputs for their language. In addition to the above findings, subjects who were high in vocabulary skills encoded significantly more efferent subject words (Sl-E) than those who were low in vocabulary skills. This is consistent with the results of the previously stated results regarding afferent subject words (Sl-A) and vocabulary skills. It was also found that subjects who attended Grace Bible College encoded significantly more efferent subject words (Sl-E) than those who attended Grand Rapids Junior College. This also suggests a possible interaction effect between religiosity, i.e., subjects attending (or not attending) a religious school and vocabulary skills in terms of the relative frequency of occurrence of efferent subject words (Sl-E) and an inferred dependence upon non-sensory inputs for their language. Subjects who disagreed with the topic encoded significantly more primitive subject words (SlP) than those who were undecided, but not significantly more than those who agreed. On the other hand, low anxiety subjects encoded significantly more defined subject words (810) than high anxiety subjects. This, too, suggests an empirical question: Are subjects who disagree with a topic more anxious? If so, are there interaction 104 effects between attitude direction and anxiety on the primitive-defined distinction in their encoding behavior. Level 2 procedures show that five variables (Sl-A, Sl-E, SlP, G-0, and 81D) evoke more information than the total subject words (51), while one variable (S-S) evoked no information. Figure 12 is a summary of the infor— mation yielded, taking Level 1 and 2 procedures together. We can see that more information is obtained about such antecedent characteristics of the encoder, i.e., anxiety, vocabulary skills, attitude direction, level, and school. No new information was obtained at this point with respect to dogmatism, spatial abilities, attitude intensity, errors, age, or sex. Level 3 Five variables which related to classifications of limiters were analyzed using Stage 6 procedures (see Chap— ter III). Of these variables, only three were found to evoke information about the antecedent characteristics of the encoders. In addition to total limiters (L)-—discussed at Level l--afferent limiters of connectors (LCl—A) and efferent limiters of subject words (LSl-E) evoked infor- mation. Table 20 presents an intercorrelation matrix of the five variables. It can be seen that afferent limiters of subject words (LSl-A), efferent limiters of subject words (LSl-E), afferent limiters of connectors (LCl—A), and efferent DEMOGRAPHIC CHARACTERISTICS 105 COGNITIVE CHARACTERISTICS Sex Level 1 TOT-1;TOT-2 TOT-3;ART Dogmatism Level 1 TOT-3 Age Anxiety £231.}. 810 Level 1 TOT—l;TOT-2 TOT-3;PREP Vocab. Skills Level 2 Sl-E;Sl-A G-O Level Level 2 G-O Level 1 ART Spat. Level 1 TOT-1;TOT-2 TOT-330TH Abilities Level 1 Le Attitude vel 2 Direction School Level 2 Sl-E Level 1 ART szLSLBSL-z.) Intensity E Level 1 ART ___ Level 1 PREP;L Sl II'OI‘S Figure 12.—-Summary of Level 2 Interpretation. 106 TABLE 20.--Limiters. L LSl-A LSl-E LCl-A LCl-E L 1.00 LSl-A .31 1.00 LSl-E .37 .15 1.00 LCl-A .39 -.12 -.27 1.00 LCl-E .65 -.11 -.12 .20 1.00 limiters of connectors (LCl-E) are not independent of total limiters (L). However, afferent limiters of subject words (LSl-A) are independent*of efferent limiters of subject words (LSl-E), afferent limiters of connectors (LCl-A), and efferent limiters of connectors (LCl-E). In addition, efferent limiters of subject words (LSl-E) were independent of efferent limiters of connectors (LCl-E), but not independent of afferent limiters of connectors (LCl-A). Finally, afferent limiters of connectors (LCl—A) were not independent of efferent limiters of connectors (LCl-E). Level 1 procedures showed that limiters (L) were able to evoke two antecedent characteristics of the source ——attitude direction, and spatial errors. It was found that subjects who made low spatial errors encoded more total limiters (L) than subjects with high spatial errors. Also, subjects who disagreed with the topic encoded *Ibid. 107 significantly fewer limiters (L) than those who were un— decided, but not significantly fewer than those who agreed. When the category of limiters (L) was subjected to a more detailed analysis, it was found that subjects who made many spatial errors encoded significantly more afferent limiters of connectors (LCl-A) than those with few spatial errors. Finally, subjects who were low in anxiety encoded significantly more efferent limiters of subject words (LSl-E) than those who were high anxious. This last result is peculiar, in that we would intuitively expect that high anxious peOple would be less concerned about sensory inputs as a basis for their language. It suggests further study. Nevertheless, Figure 13 presents a summary of the infor- mation yield obtained, including Levels 1 thru 3 interpre— tation procedures. We have evoked more information about anxiety and spatial errors using this element of the classi- fication system, but no new information was obtained for the remaining coqnitive or demographic characteristics. Level 4 It will be remembered that at Level 1, total con— nectors (C1) evoked no information about the antecedent characteristics of the encoders. It is considerably different, however, when subsets of total connectors are examined. Table 21 (Tables 22 and 23 are separate matrices, showing efferent and afferent interrelationships) is an DEMOGRAPHIC CHARACTERISTICS 108 COGNITIVE CHARACTERISTICS Sex Level 1 TOT-l;TOT—2 TOT-3;ART Dogmatism Level 1 TOT-3 Age Anxiety Level 2 SlD Level 3 LSl-E Voc Level 1 TOT-l;TOT—2 TOT-3;PREP G—O ab. Skills Level 2 Sl-E Sl-A Level Level 1 ART G-O Spat. Level 1 TOT-l;TOT-2 TOT-3;OTH Abilities Attitude Level 1 Level 2 Direction PREP;L SlP 3Sl-A School Level 2 Sl-E Level 1 ART Intensity E Level 1 ART rrors Vocab. Level 1 Le vel 3 PREP;L 81 L Cl-A Spat. Figure 13.--Summary of Level 3 Interpretation. l()9 +1 ll 00. mo. V a 00.5 m0. em. 00. 50.- 00.- N0. 50. 00. 00. 0a. ~00.- 0~.- N0. 00. 00. 50. em. on. 00. 5.0.5 00.5 ~0.- 00. m5.- 00.- 05. 50.- 00.- 50.- 00. e5.- 55.- 50.- 00.- 00.- 50. 50. 0~.- 00. some 00.5 50.- 00. 50. N0. 00. mm. 50. 00. 05. NN. mm. 05. mm. 0m. 5m.- ms. 00. 2505 00.5 ~5.- 00.- 50. 00.- 00. 50.- 00.- 50.- e0. 50. 00.- m5.- m0.- 000.- 50. 55. 005u 00.5 05. N0.- 0N. 0a. 0~.- 50. 00. 5m. M0. on. e5.- 00.- “5.- 05. 00. ma5u 00.5 00.- 50. N0. ~0.- 00. 05. 55.- 500,- 500.- 00. em.- 000. 50.- 000. 4050 00.5 mm.- 00.- em. 00. 05.- 55.- 50.- 50. 05.- mm. 00. 55. 05. 5cm 00.5 05. mm. 50. s~.- 05. 05. 05. 0h. 0m. 05.- 00. 5m. 50.0 00.5 5m.- 50. N0.- 0N. N0. 0~.- 05.- 05.- a0.- 00. 50. 5cm 00.5 00. 00. 00. 5m.- 00. 00. 0s. 05. 0m. om. 505 00.5 0~.- e~.- 00. an. 00. 00. mm. an. em. 50004 00.5 50. 00.- 05.- 05.- N0. 05.- 05. 50. azoo 00.5 05. N5.- m~.- 05.- 05.- m~.- m0.- 5oz 00.5 0~.- 00.- 00.- 00.- 55. 00. mam 00.5 05.- en. en. 05. 05. «5 00.5 00. 50. 50. 05. 95 00.5 mm. 00. 00. «05 00.5 N0.- 05. 050 00.5 50. 050 00.5 50 500 came z5ma :a5o «050 «E50 500 5o» 500 505 50904 5:00 5oz mam m5 e5 «05 a5u 550 5c .muOuooccou uo x5uusfi scabsaouuoououcHII.HN mange 110 TABLE 22.--Intercorre1ation matrix for afference. AFF Sl-A LSl-A LCl-A AFF 1.00 Sl-A .79 1.00 LSl-A -.21 -.51 1.00 LCl-A -.10 -.21 -.12 1.00 TABLE 23.--Intercorre1ation matrix for efference. EFF Sl-E LSl-E LCl-E EFF 1.00 Sl-E .52 1.00 LSl-E .49 .09 1.00 LCl-E .42 -.30 -.12 1.00 intercorrelation matrix of those variables analyzed with Stage 6 procedures (see Chapter III). A survey of the matrix, assuming that non-significant correlations indi— cate independent relationships,*shows 94 of the 184 re- lationships represented as being independent. Of the 90 dependent relationships, the significant correlations vary from -.51 to +.92. Of the 20 variables analyzed, 11 evoked information about antecedent characteristics of the en- coders. In general, high dogmatic subjects encoded sig- nificantly more indicative reflexive connectors (IR) and *Ibid. 111 total reflexive connectors (RC1). High anxious subjects also encoded significantly more total reflexive connectors (RC1). This is not surprising, since dogmatism and anxiety were positively correlated (.31). It may indicate the verbal behavior of high dogmatic and high anxious subjects is more concerned with reflexive (passive voice) action instead of the more assertive, i.e., possibly more intense, language behavior reflected in transitive action connectors (TCl). In addition to these findings, high anxious subjects encoded significantly more total primitives (PRIM), but significantly less total defined (DEFD) connectors. These results are consistent, and indicate the possibility of less concern on the part of high anxious subjects to denote limitations of the meaning of signs than low anxious subjects. Subjects with high spatial skills encoded signifi- cantly more total negation connectors (NCl), a surprising result since no previous studies have linked dissociative assertions to spatial abilities. Vocabulary abilities yielded information in four message variables: subjects who were high in vocabulary skills encoded more indicative reflexive connectors (IR) and total efferent (EFF) words. It is a reasonable question to ask: Since the results indicate that subjects with high vocabulary skills and who are high dogmatic, encode significantly more indicative reflexive connectors (IR), is there an interaction between the two independent 112 variables? Such a possibility requires further study. In addition to these findings, subjects who were high in vocabulary skills encoded significantly lggg total com- parison connectors (COMP) and total afferent (AFF) words. Five significant results were obtained, when atti— tude was the independent variable. Subjects who disagreed with the topic encoded significantly more indicative com- parison equal connectors (ICE), significantly more total comparison connectors (COMP), significantly more total efferent words (EFF), and significantly more total primi- tive (PRIM) words. These significant differences were found (using Scheffe's t-test) between those subjects who disagreed, and those who were undecided in all cases accept indicative comparison equal connectors (ICE). It was found that subjects who disagreed with the topic en- coded significantly more indicative comparison equal con- nectors (ICE) than either those subjects who were undecided or those who agreed with the topic. This message variable suggests strong, assertive, and perhaps "simple-minded" c0gnitions in the equating of two unit signs. Whether this generalization about subjects who disagree can go beyond the topic studied is an empirical question. It suggests an interesting hypothesis. No information was evoked by the message variables concerning sex and level, but total subjunctive present tense verbs (SPR) was related to information about attitude intensity. The trend was for low attitude intensity 113 subjects to encode more subjunctive tense verbs (SPR) than higher attitude intensity subjects, although not signifi— cant. Subjunctive verbs are, logically, dispositional verbs, which denote an action may happen. The relationship between attitude intensity and "dispositional" connectors merits further study. Subjects who attended Grace Bible College encoded significantly more primitive connectors (ClP), action con— nectors (ACTCl), and total primitives (PRIM). If attend— ance at Grace Bible College is an indicator of religiosity, there may be an interaction between religiosity, vocabu- lary skills, and dogmatism where primitive connectors (ClP), total primitives (PRIM), and action connectors (ACTCl) are dependent measures. This is suggested by the Level 1 interpretation and the results of the Level 4 interpre- tation. Spatial errors yielded information on five message variables. Subjects who made high spatial errors encoded significantly more primitive connectors (ClP), more indi- cative comparison equal connectors (ICE), more total com- parison connectors (COMP), and more total primitives (PRIM). Subjects who made high spatial errors encoded significantly less total reflexive connectors (RC1). These results are difficult to interpret since no theory of errors on such tests is extant. In summary, we can see from Figure 14 the infor- mation evoked from Level 4 interpretation. It is readily 114 DEMOGRAPHIC CHARACTERISTICS COGNITIVE CHARACTERISTICS Sex Dogmatism Level 1 TOT-l;TOT-2 TOT-3;ART Level 4 IR TCl Anxiety Level Level Level Level 1 2 3 4 Age TOT-3 SlD LSl-E RC1 DEFD* PRIM* Skills Level 4 IR;COMP AFF** EFF** Vocab. Level 2 Sl-E Sl-A G-O Level 1 TOT—l;TOT-2 TOT-3;PREP Level Level 2 G-O Level 1 ART Abilities Level 4 N01 Spat. Level 1 TOT-1;TOT-2 TOT-3;0TH School Level 1 Level 2 Level 4 Attitude Direction Level 1 Level 2 Level 4 PREP;L SlP;Sl-A ICE;COMP PRIMtiEFF** Level 4 Intensity SPR ClP ACTCl PRIM* ART Sl-E “Level 1 Level 3 Level 1 Errors Vocab. ART Spat. Level 4 PREP LCl-A C1P;ICE;COMP L;Sl RQIIRRIM1_____ *Primitive (PRIM) words include both primitive subject words (SlP) and primitive connectors (ClP). De- fined (CEFD) words include both defined subject words (81D) and defined connectors (ClD). **Afferent words (AFF) include both afferent subject words (Sl-A) and afferent limiter words (LSl-A and LCl-A). Efferent words include both efferent subject words (Sl—E) and efferent limiter words (LSl-E and LCl-E). Figure l4.--Summary of Level 4 Interpretation. 115 noted that if the analysis had stopped at Level 1, we would have found 21, or about 38 per cent, of the signifi- cant differences obtained by making a more detailed analy- sis. Level 1 procedures allowed us to add 8, or 14 per cent, significant results; Level 3, two (3 per cent), and Level 4, 23 (42 per cent). Evaluation as a Syntactical Theory of Signs An evaluation of a descriptive classification of signs necessarily begins with the problem of measurement. The results demonstrate that this problem has not yet been met. The only message variables which were normally distributed were the total word scores (TOT-1, TOT-2, and TOT-3). The fact that the other variables were not norm- ally distributed may be a statistical artifact, due to the transformation (word index score) performed on the data. Two critical questions were raised in Chapter I: (1) Is the raw frequency of a tag the best estimate of the effect of some antecedent condition, and (2) Should the researcher control for the varying lengths of message samples by using some common denominator, such as total words or sentences encoded? It was noted that these two questions could not be directly answered in this study. Control for the vary- ing lengths of messages was used because of precedence in the field. However, the lack of normality for so many of the message variables suggests that either a different 116 transformation would be appropriate, or that raw fre- quencies would be best. At any rate, no definitive re- sults can be obtained until this problem is solved. Another measurement problem is that associated with the issue raised in Chapter IV, i.e., it is not reasonable to assume that one method of analysis is appropriate for all meaningful distinctions. It is de- sirable in content analysis to use parametric analysis. This study was designed to use multiple regression and analysis of variance techniques to analyze the data. The fact that 12 variables were eliminated at Stage 1 (0 fre- quencies for all subjects) may be a function of the popu- lation from which the sample came, or the limitations of the laboratory. It may also be that these variables would be relevant, using analysis methods for nominal data. At any rate, no judgment can be made without other samples, gathered under similar and different conditions. If these variables continue to lack significance, the variables are probably meaningless. Thirty-four variables were eliminated at Stage 2 (Chapter IV). These variables are candidates for analysis, using either nominal or ordinal methods of analysis. Of the 54 variables which were eliminated at Stage 3, four- teen were not analyzed because of the shapes of the distri- butions following plotting of a frequency distribution. These variables also are candidates for non-parametric analysis. 117 Among the 40 message variables submitted for one- way analysis of variance with each of the antecedent vari- ables, 28 were found to significantly yield information about a cognitive or demographic characteristic of the en- coder. The result is that a definitive evaluation of this theory of syntactical relationships awaits a solution to the problem of measurement. With 70 per cent of the mes— sage variables submitted to Stage 6 procedures producing significant results, it is reasonable to assume that this syntactical approach has merit. In addition to the problem of measurement, the question of reliability is critical. Again, this question cannot be answered without a replication of the study on a sample from the same pOpulation and with similar controls. One needs to ask at this time: What are the pro- jected benefits of such an approach to the study of com- munication? We see at least four values: 1. This approach allows for a syntactical analy— sis of how people process information. It goes beyond that conceptualization of syntactics offered by Osgood and his congruity principle in which syntactics between signs, i.e., unit signs in this analysis, was little more than associ— ative (positive) or dissociative (negative) assertions. Furthermore, the congruity principle of Osgood's and the belief congruence principle of Rokeach (with respect to syntactics) is little more than a within—unit sign analysis. 118 frhis approach, if it meets the theoretical problems dis- <:ussed earlier in this section, provides a means of gener- alizing the findings of Osgood and Rokeach to include the findings of this study. 2. This approach offers a distinction regarding the nature of referents which had not been tested previ- ously in the continuous discourse situation. This study demonstrated that the nature of the referent, i.e., afferent or efferent, is important in explaining variance. 3. This approach goes beyond any study this writer is aware of in making between-unit sign distinctions, (allowing for the possibility of attributing variance not Otherwise possible if the total connectors is all that is kncnni. It was demonstrated in the Level 4 procedure of ShTterpretation, where the frequency of occurrence of total connectors (Cl) evoked no information about the antecedent Characteristics of the encoder. Even the gross distinction 0f positive connectors (PCl) and negative connectors (NCl) evoked only one distinction, and that with respect to Spatial skills. This suggests, though only in a prelimi- nary way, that the message matrix of Osgood's (1959) in which associative and dissociative assertions were denoted hold little value in attributing variance. By making finer distinctions, 42 per cent of the 54 significant differences obtained were attributable to characteristics of connectors, and this in spite of the fact that total connectors (Cl) 119 did not add one piece of new information about any of the antecedent characteristics. 4. This study-—other than case studies, and very small samples--is the first of its type to look at the validity of drawing inferences about the cognitive struc- tures of encoders. The problems which have been related to this issue were outlined in Chapter I, i.e., costs in energy and/or the inaccessibility of the sources of the messages. Recommendations for Further Study The most critical study at this point in time is the one that attempts to achieve a solution to the problem of measurement. As stated before in this chapter, an answer to the question of what transformations, if any, should be performed on the message data needs to be ob- tained. A second recommendation would be the careful repli- cation of the study, with formal hypotheses based on the results of this study in order to determine the relia- bility of the results. Following this, questions of generalizability to pOpulations can be empirically studied. A third recommendation is a study which would determine the effects of feedback on these message vari— ables. This would provide a significant bridge between this study and numerous studies which have captured the interest of communication scholars. 120 Finally, a study which manipulates selected message variables from this classification system to determine their effects on receivers is needed. This study repre- sents only an attempt to explicate the source-message aspect of the communication paradigm. It is not reason- able to assume that all significant differences in the source-message aspect are perceived as relevant to a re— ceiver, i.e., influencing that receiver's behavior. BIBLIOGRAPHY BIBLIOGRAPHY Allport, G. W. Letters from Jenny. Journal of Abnormal and Social Psychology, 1946, 41, 315-350, 449-480. Bales, Robert F. A set of categories for the analysis of small group interaction. American Sociological Review, 1950, 15, 257-263. Blumberg, N. B. One party press? Coverage of the 1952 presidential campaign in 35 daily newspapers. Lincoln: University of Nebraska Press, 1954. Cartwright, & Harary. Structural balance: A genera- lization of Heider's theory. Psychological Review, 1956, 63, 277-293. Colby, Benjamin. Deve10pment and applications of an anthropological dictionary. In Philip J. Stone, Dexter Dunphy, Marshall Smith, & Daniel Ogilvie, Theygeneral inquirer: A computer approach to content analysis. Cambridge: The M.I.T. Press, 1966. Dollard, J., & O. H. Mowrer. A method of measuring tension in written documents. Journal of Abnormal and Social Psychology, 1947, 42, 3-32. Fishman, Joshua A. A systematization of the Whorfian hypothesis. In Alfred G. Smith (Ed.), Communi— cation and culture. New York: Holt, Rinehart and’Winston, 1966. Guilford, J. P. Psychometrics. New York: McGraw-Hill, 1954. Guilford, J. P. The nature of human intelligence. New York: McGraw-Hill, 1967. Guttman, Louis, & Edward A. Suchman. Intensity and a zero point for attitude analysis. American Sociological Review, 1947, 12, 57-68. 121 122 Heider, F. Attitudes and cognitive organization. Journal of Psychology, 1946, 21, 107—112. Holsti, Ole. Content analysis. In Gardner Lindzey and Elliot Aronson (Eds.), The handbook of social psychology. Vol. II. 1969. Pp. 596-692. Holsti, Ole. External conflict and internal consensus: The sino-Soviet case. In Philip J. Stone, Dexter C. Dunphy, Marshall 8. Smith, & Daniel M. Ogilvie (Eds.), The general inguirer: A computer approach to content analysis. Cambridge: The M.I.T. Press, 1966. Jaffe, J. The study of language in psychiatry: Psycho— linguistics and computational linguistics. In S. Arieti (Ed.), American handbook of psychiatry. Vol. 3. New York: Basic Books, 1966. Pp. 689- 704. Johnson, W. Studies in language behavior: I. A program of research. Psychological Monographs, 1944, 56, No. 255. Kalin, Rudolf, William N. Davis, & David C. McClelland. The relationship between use of alcohol and the- matic content of folktales in primitive societies. In Philip J. Stone, Dexter Dunphy, Marshall Smith, 5 Daniel Ogilvie (Eds.), The general inguirer: A computer approach to content analysis. Cambridge: The M.I.T. Press, 1966. Kauffman, P. E., 8 V. C. Raimy. Two methods of assessing therapeutic progress. Journal of Abnormal and Social Psychology, 1949, 44, 83-112. Kochevar, J. J. The effects of message intensity on evalu- ations of the source, the message, and the topic. Unpublished Master's thesis, Michigan State Uni- versity, 1967. Leites, N. C., 8 I. deSola Pool. On content analysis. Washington, D.C.: Library of Congress, Experi- mental Division for Study of War-Time Communi- cations. Document No. 26, 1942. Lindquist, E. F. Design and analysis of experiments in psychology and‘education. Boston: Houghton Mifflin, 1953. McClelland, D. C. The use of measures of human motivation in the study of society. In J. W. Atkinson (Ed.), Motives in fantasy, action and society. Princeton: Van Nostrand, 1958. Pp. 518-552. 123 McClelland, D. C., 8 G. A. Friedman. A cross-cultural study of the relationship between child-rearing practices and achievement motivation appearing in folktales. In G. E. Swanson, T. M. Newcomb, 8 E. L. Hartley (Eds.), Readings in social psychology. (2nd ed.) New York: Henry Holt, 1952. Pp. 243-249. McEwen, William. Counter-attitudinal encoding effects on message style and performance. Unpublished doctoral dissertation, Michigan State University, 1969. McNemar, Quinn. Psychological statistics. (3rd ed.) New York: John Wiley 8 Sons, 1962. Morris, Charles. Signs, language and behavior. New York: Prentice-Hall, 1946. Mosteller, F., 8 D. L. Wallace. Inference and disputed authorship: The federalist. Reading: Addison- Wesiéy,'l964. Namenwirth, J. Zvi. Some long- and short-term trends in one American political value: A computer analysis of concern with wealth in 62 party platforms. In George Gerbner, Ole R. Holsti, Klaus Krippen- dorff, William J. Paisley, 8 Philip J. Stone (Eds.), The analysis of communication content. New York: ddhn Wiiéy 8 Sons, 1969. Namenwirth, J. Zvi, 8 Thomas L. Brewer. Elite editorial comment on the European and Atlantic communities in four countries. In The general inquirer: A computer approach to content anaiysis. Cambridge: The M.I.T. Press, 1966. Newcomb, Theodore M. An approach to the study of communi- cative acts. Psychological Review, 1953, 60, 393-404. Nichols, Jack G. An investigation of the effects of varied rates of training on systematic desensiti- zation for interpersonal communication apprehen- sion. Unpublished doctoral dissertation, Michigan State University, 1969. Osgood, Charles, George Suci, & Percy Tannenbaum. The measurement_of meaning. (2nd ed.) Urbana: Uni- versity of Iliinois Press, 1967. Osgood, Charles. The representational model and relevant research methods. In Ithiel De Sola Pool (Ed.), Trends in content analysis. Urbana: University of"I11indis_§iess, 1959. 124 Paige, Jeffery M. Letters from Jenny: An approach to the clinical analysis of personality structure by computer. In Philip J. Stone, Dexter C. Dunphy, Marshall S. Smith, 8 Daniel M. Ogilvie (Eds.), The general inquirer: A computer approach to contedt— analysis. Cambridge: The M.I.T. Press, 1966: Riesman, D., N. Glazer, 8 D. Reuel. The lonely crowd. New Haven: Yale University Press, 1950. Roberts, Paul. Understanding grammer. New York: Harper 8 Brothers, 1954. Rokeach, Milton. The open and closed mind. New York: Basic Books, 1966. Rokeach, Milton, 8 G. Rothman. The principle of belief congruence and the congruity principle as models of cognitive interaction. Psychological Review, 1965, 72, 128-142. Saport, Sol, 8 T. A. Sebeok. Linguistics and content analysis. In Ithiel De Sola Pool (Ed.), Trends in content analysis. Urbana: University of Iiiinois Press,’1959. Pp. 131-150. Smith, Marshall 8., Philip J. Stone, 8 Evelyn N. Glenn. A content analysis of twenty presidential nomi- nation acceptance speeches. In Philip J. Stone, Dexter C. Dunphy, Marshall S. Smith, 8 Daniel M. Ogilvie (Eds.), The general inquirer: A comwuter approach to content analysis. Cambridge: The M.I.T. Press, 1966. Staats, Arthur W. Learning, language, and cognition. New York: Holt, Rinéhart, and Winston, 1968. Stone, Philip J., Dexter Dunphy, Marshall Smith, 8 Daniel Ogilvie. The general inquirer: A computer approach to content analysis. Cambridge: The M.I.T. Press, I966} Troldahl, Verling C., 8 Fredric A. Powell. A short-form dogmatism scale for use in field studies. Social White, R. K. 'Black boy': A value-analysis. Journal of Abnormal and Social Psychology, 1947, 42, 440-461. APPENDICES APPENDIX A CODE BOOK APPENDIX A CODE BOOK Each coder is instructed to do, in successive order, the following tasks and subtasks. You may work as a group with differences of opinion resolved at each step or sub- step. If you are unsure whether a word should be circled, ask other coders in order to resolve any differences. At each step, record the frequencies of each tag on back of message. Any totals not agreeing should be recoded until agreement is achieved. Step 1 a. Circle all articles "a," "an," and "the." b. Circle all conjunctions "and," "but," "or," "so," "since," etc., where a conjunc- tion is understood to mean a word which links other words to groups of words of the same class (linking clauses, nouns, verbs, adjectives, prepositional phrases, etc., such as in the statement, "Edgar apd Stanley. . . ." c. Circle all relative or interrogative pro- nouns such as: who whoever whose whosoever whom whomsoever which whomever that whichever as whichsoever but whatever what whatsoever d. Circle all interjections, or words which normally are one-word sentences which usu- ally express emotion. They are often 125 Step 2 Step 3 Step 4 126 followed by an exclamation mark in normal literary discourse. Examples of such words are "Oh!" "Ouch!" "Whew!" "tsk-tsk." Circle all demonstrative pronouns, meaning those words which point out persons or things, such as "This book . . . ," "That person . . . ," "These cars . . . ," and, "Those hooks . . . ." Circle all words which are the subjects or objects of verbs. If the word has a word which modifies it, mark the subject or object of the verb as "D;" if not, mark it with a "P." Circle all words which are adjectives, ad- verbs, or objects of prepositions. Words already circled are prepositions, and there- fore you should determine what is the object of that preposition. If the word you circle modifies a subject word, mark it as "LSl," and if the word you circle modifies a verb, mark it as "LCl." All words circled as objects of prepositions should be marked as "LSl" if the preposi- tional phrase describes the subject or ob— ject of the verb. If the prepositional phrase describes some kind of action, i.e., modifies a verb, then mark the object of that preposition as "LCl." All words you have on this version of the manuscript are verbs or verbal types. If a verb shows action, mark it as "A." If the verb is a linking verb, i.e., "is," "be," etc., mark that verb as "C." All verb phrases which have a negative in it, i.e., the word "not," or a prefix such as "dis . . . ," ”un . . . ," etc., and which you believe to negate the verb, mark that verb as "N." Mark all verbs which you believe to be past tense as "PA," all verbs you believe to be 127 present tense as "PR," and all verbs you be- lieve to be future tense as "FU." d. Some verbs have words in the phrase which denote that the action might, or could hap- pen, but did not or will not necessarily occur. Where you see the words "might," "may," "could," "should," "ought," etc., mark that verb as "S." All other verbs should be marked as "I." e. Of those verbs you marked as "A," if the verb has both a subject and object, mark it "T." If the verb has no object to it, mark it "R.“ Step 5 a. Of those verbs marked as C, if the verb says the object of the verb is the same as the subject of the verb, mark it "E." i.e., "Republicans are Fascists.” b. Of those verbs marked as C, if the verb says either the subject or object of the verb has more or less of some quality, mark it as "M," i.e., "Jan is more beautiful than Sally." c. Of those verbs marked as C, if the verb says either the subject or object of the verb "belongs to," "is part of," etc., the other (subject is part of the object, or object is part of the subject), mark that verb as "S.” It can be seen in the sentence, "Girls are part of the human race." d. Of those verbs marked as C, if the verb has a subject, and the verb is followed by a prepositional phrase which answers the ques- tion where the subject is located, mark it as "P.” It can be seen in the sentence, "Conservatives are to the right of liberals." e. Of those verbs marked as C, if the verb has a subject, and the verb is followed by a prepositional phrase which answers the ques- tion when, mark it as "t." It can be seen in the sentence, "George was on :time." f. Of those verbs marked as C, if the verb links a demonstrative pronoun with a noun, mark that verb as "EXT." It can be seen in the sentence, "There is the book." Step 6 a. 128 Of those verbs marked as C, if the verb links a noun with an adjective, mark it as "PRED." It can be seen in the sentence, "The book is Egg." Of those words marked as "L," if the word has a negative, i.e., "no," or "none," modi- fying it, mark that word as "N." Of those words marked as "51," if the word has a negative, i.e., "no," or "none," modi- fying it, mark that word as "N." APPENDIX B TABULATED SUMMARY DATA 129 ..mmma.~ .omsm.o mmao.o mmao.o omvo.o o musoqz .«smvv.m .oomm.o Nomo.o m-o.o oemo.o o m-quz ..qmmm.v .mmmo.a voao.o moao.o ommo.o o aduz .«Hmmv.s .masm.a omao.o «ado.o ommo.o o aaoz ..ssm~.m .mmme.a amao.o mmao.o osmo.o o 9H2 ..oaqm.m «mamo.a nmao.o smao.o osmo.o o maaz ..aa~o.m .Homm.a mmoo.o ano.o omao.o o mm .«maom.m .oamm.a saao.o amao.o oomo.o 0 9m «.Hoom.m .mmvm.a mmmo.o smuo.o omma.o o mam ..~mmm.s .mmmm.a mmmo.o amao.o omma.o 0 8H ..msom.m .vwmm.o mmmo.o omaa.o omH~.o oamo.o 9H ..Hmsm.m «mamm.o Hmoo.o ooao.o omvo.o o moH .«sams.ma .mmmm.~ «Hmo.o vao.o omaa.o o 38H .«mqma.m .mmom.o oevo.o msma.o oasm.o oomo.o maH ..aa~m.v «NHH~.H oaao.o mmao.o osmo.o 0 «EH .«vmam.m mmoo.o aemo.o amms.o ommm.o omoo.o muaoq ..ssao.m .v¢s~.o omeo.o Homa.o oomm.o ommo.o muamq .«sosa.m .mmmm.o ammo.o ammo.o omma.o o a-aoq ..ms~m.m .omvm.o ommo.o memo.o omfla.o o anamq ..smmm.m .mmma.o Hmao.o meaa.o omm~.o omoo.o muflm .«mmmo.m .nmmm.o mamo.o mmsa.o ooam.o omoo.o a-Hm .«mmoa.m .mmm~.o mmwo.o ammo.o omaa.o oovo.o ado .«smoo.m .mms~.o ommo.o ooma.o oo¢~.o omqo.o ado .«Hmmo.m «Hema.o mnmo.o mmva.o oom~.o oaoo.o cam ..am-.m .mmqm.o mome.o omva.o omam.o oomo.o mam mmmv.~ mmmo.o- Na.om sm.m- sum as H-909 ..o~w~.ma mmmv.m ma.oa m~.m as o uouum Hmflummm mama.~ ovma.o mm.~m sm.ova MAN «4 Hafiumam ..vm~m.~ mmes.o mm.s Hm.~H mm H nouns sumasbmoo> ..mo¢o.m omhm.o No.5 mH.H~ as m sumasbmoo> ..a¢em.~ ammo.o mm.aa mm.am Gm mm sumflxca oeso.m omom.ou vm.va mo.v> aoa mm smflbmsmoa ouHsM .3mxm .o.m cmmz now: 309 mannawm> .m0flumaumum m>wumflwomoo mo mwmfisdmnl.alm mqmda 130 mo. v Q mwmouwsx .mflm88 Ho. v Q mmwc3mxm .mflme «ammoa.m «mva.o mmmo.o ommm.o omvm.o oama.o Hum «8mmom.v «mmmm.o mmmo.o Hmmm.o ommm.o ommH.o ammo aamom~.m asvmm.o mono.o Hemm.o omvm.o omma.o EHmm «8Hmo~.> «mmmm.a avmo.o mmmo.o omva.o o omHO «8Hnmm.m «Hmno.o mmmo.o nooo.o omva.o o mmau «aamo~.v «omoa.a mmao.o mnao.o onwo.o o doao «evaom.~ «mwoo.o movo.o mom~.o ovvv.o onam.o Hm «aaavv.m «mnmm.on ommo.o mmmv.o omoo.o ommm.o q «aommv.~ «moma.ou mnmo.o mmov.o omoo.o omm~.o mmm aamvom.m «omm~.o mamo.o omh~.o ommv.o omma.o mum 88>HHH.v aooov.o vmeo.o ~mmm.o onav.o ovma.o ABOBVHO «cvoom.m «hmam.o ammo.o mmmo.o omoa.o o Hum «8HmHm.~ «momv.o ommo.o mova.o omv~.o omho.o HOB 88>¢0H.n «comm.a onmo.o momo.o omma.o o How «ammuo.m «moam.o vmmo.o ~ma~.o onwm.o omHH.o aUH «aoaam.~ «mmmv.o ammo.o -ma.o omm~.o oama.o HUBU¢ «cooom.~ «oaoo.o mmao.o hmao.o ommo.o o mZOU «8mmH~.v «qNNH.H mmao.o mmmo.o oamo.o o Hoz mnvv.~ mmoo.o Ho.ov hm.~ma mmm hm mneoe mmmm.m mmma.o om.mv mm.ooa mum mm Niece «ammaa.m «Hmam.o mmvo.o mho~.o ommm.o oaoa.o $90 «aommm.~ «mano.o Hmeo.o woma.o o>o~.o ovmo.o ammo aammam.~ Nooo.o mmvo.o mmma.o ovm~.o osmo.o 8&4 «emhh>.m «oomm.a mnoo.o mooo.o oovo.o o meH aemavm.m «wmom.o mmoo.o Hmoo.c ovmo.o o ammH «amavv.m umaom.o mmvo.o Hmma.o o~m~.o oamo.o Ono aamoo~.m amnmm.a vmao.o vmoo.o ommo.o 0 03¢ «evaav.ma «ammm.m vmao.o mmoo.o oo~H.o o one acmvmm.m 8mmmv.a mono.o mmmo.o omma.o o mum TABLE B-2.--Summary of simple correlations: 131 ables with cognitive variable. message vari- Var. Corr. Var. Corr. Var. Corr. DOGMATISM ANX .305* IT .085 TOT-2 -.312* V-C -.031 IR .169 TOT-3 -.342* V-I .147 SPR -.124 NCl .119 S-C .255* ST -.133 COMPAR -.135 S-I -.001 SR -.070 ACTCl .111 TOT-1 -.310* NlPR .089 IC1 .207 SlP .028 NlT .180 SCl -.164 SID -.072 NClP .138 TCl .090 C1P .110 NClD .023 RC1 .125 C10 .001 NLSl-E .076 C1(TOT) .143 Sl-A .034 NLCl-E -.035 AFF .052 Sl-E .031 S-S -.003 EFF -.076 LSl-A .050 T-O -.117 L -.079 LCl-A -.017 A-O .046 $1 -.030 LSl-E -.097 G-O .035 ClPA -.010 LCl-E -.060 IPRD .116 ClPR -.061 IPA -.067 IEX -.063 ClFU -.059 IPR .170 ART .185 PRIM .099 IFU —,011 PREP .044 DEFD -.057 ICE -.107 0TH .028 PC1 .120 ANXIETY DOG .305* IT -.023 TOT-2 -.216* V-C -.208* IR .080 TOT-3 -.258* V-I .111 SPR .121 NCl .133 S-C .120 ST .167 COMPAR -.151 S-I .137 SR .098 ACTCl .144 TOT-1 -.230* NlPR .131 IC1 .031 SlP .224* NlT .075 SCl .147 SlD -.l75 NClP .196 TCl .102 C1P .150 NClD -.035 RC1 .134 C1D -.058 NLSl-E .178 C1(TOT) .178 Sl-A .141 NLCl-E .075 AFF .169 Sl-E -.014 S-S .137 EFF -.217* LSl-A -.008 T-O -.131 L -.169 LCl-A .094 A-O -.044 $1 .097 LSl-E -.289* G-O .214* ClPA -.118 LCl-E -.094 IPRD .078 ClPR .126 IPA -.111 IEX -.118 ClFU -.114 IPR .104 ART .055 PRIM .258* IFU -.169 PREP -.094 DEFD .214* ICE -.051 0TH .151 PC1 .120 TABLE B-2.--Continued. 132 Var. Corr. Var. Corr. Var. Corr. VOCABULARY SKILLS ANX -.208* IT -.001 TOT-2 .247* COG -.031 IR .238* TOT-3 .198 V-I -.096 SPR -l.75 NCl -.091 S-C -.126 ST -.166 COMPAR -.270* S-I -.141 SR -.023 ACTCl .071 TOT-1 .245* NlPR -.112 IC1 .032 SlP -.104 NlT -.050 SCl -.159 SlD -.035 NClP -.110 TCl -.114 C1P -.122 NClD -.017 RC1 .248* C1D .092 NLSl-E -.113 C1(TOT) -.111 Sl-A -.286* NLCl-E -.051 AFF -.184 Sl-E .214* S-S -.085 EFF .224* LSl-A .168 T-O .055 L .155 LCl-A .031 A-O .108 81 -.137 LSl-E .179 G-O -.257* ClPA -.046 LCl-E .014 IPRD .113 ClPR -.200 IPA -.032 IEX .088 ClFU .198 IPR -.040 ART .004 PRIM -.l49 IFU .252* PREP .214* DEFD .021 ICE -.151 0TH -.047 PC1 -.062 SPATIAL ABILITIES ANX .120 IT .106 TOT-2 -.128 DOG .255* IR -.180 TOT-3 -.l71 V-C -.126 SPR -.006 NCl .108 V-I .105 ST -.032 COMPAR -.056 S-I -.123 SR .064 ACTCl -.019 TOT-1 -.131 NlPR .034 IC1 -.003 SlP .165 NlT -.004 SCl .076 SlD -.023 NClP .053 TCl .116 C1P -.004 NClD .125 RC1 -.157 C1D .074 NLSl-E .051 C1(TOT) .096 Sl-A .127 NLCl-E .141 AFF .139 Sl-E -.005 S-S .008 EFF -.182 LSl-A .052 T-O -.076 L -.l69 CLl-A -.021 A-O .086 51 .167 LSl-E -.203* G-O .209 ClPA -.039 LCl-E -.105 IPRD -.031 ClPR .033 IPA -.031 IEX -.004 ClFU -.052 IPR .054 ART .012 PRIM .115 IFU -.l31 PREP -.064 DEFD .069 ICE -.050 0TH .206* PC1 .049 133 TABLE B-2.--Continued. Var. Corr. Var. Corr. Var. Corr. VOCABULARY ERRORS ANX .111 IT -.050 TOT-2 -.l38 DOG .147 IR -.021 TOT-3 -.121 V-C .096 SPR -.055 NCl .100 S-C .105 ST -.055 COMPAR .094 S-I .219* SR -.156 ACTCl ~.133 TOT-1 .144 NlPR .130 IC1 .008 SlP .058 NlT .220 SCl -.ll6 SlD .018 NClP .085 TCl .013 C1P .033 NClD .070 RC1 -.108 C10 .137 NLSl-E -.049 C1(TOT) -.117 Sl-A .121 NLCl-E -.156 AFF .180 Sl-E .218* S-S .185 EFF -.ll8 LSl-A .089 T-O -.094 L .068 LCl-A .030 A-O -.178 81 .032 LSl-E .058 G-O .089 ClPA -.158 LCl-E .032 IPRD -.064 ClPR -.073 IPA .170 IEX -.177 ClFU -.188 IPR .087 ART -.248* PRIM -.038 IFU .180 PREP -.038 DEFD -.047 ICE .031 0TH -.138 PC1 -.133 SPATIAL ERRORS ANX .137 IT -.028 TOT-2 .140 DOG .001 IR .113 TOT-3 .180 V-C .141 SPR -.090 NCl -.075 V-I .219* ST -.119 COMPAR .367* S-C .123 SR -.042 ACTCl -.024 TOT-1 .172 NlPR .056 IC1 .143 SlP .034 NlT -.068 SCl -.152 SlD .081 NClP -.022 TCl -.l35 C1P .216* NCld -.108 RC1 .067 CID .084 NLSl-E .166 C1(TOT) .195 Sl-A .012 NLCl-E .026 AFF -.073 Sl-E .101 S-S -.035 EFF -.022 LSl-A .011 T-O .153 L -.137 LCl-A .119 A-O .024 $1 .114 LSl-E .203* G-O -.093 C1PA .008 LCl-E .008 IPRD .072 ClPR -.058 IPA .036 IEX -.082 ClFU ~.178 IPR .179 ART .070 PRIM .125 IFU .117 PREP -.036 DEFD -.008 ICE .251* 0TH .064 PC1 .170 *Significant correlation p < .05. 134 am. am. mo. NH. vHvN. HHoN. mooo. mBO ha. mm. mm. mo. ohm. mood. moooo. ammo vH. ma. so. m~.m mha. mom. ooo. axe om. ma. mm. ov.m mam. mom. moo. Ono mo. mo. on. mo. mmoo. ammo. ooooo. mum mo. mo. ma. m>.a Hmo. omo. mooo. mom oo. oo. oo. mm. mmo. mmo. Hooo. mH Ha. NH. «o. oo. mom. mod. hooo. 9H Ho. Ho. mm. om. woo. woo. moooo. moH om. ma. mm. on. mmH. nod. mooo. mmH om. ma. mm. mo. now. now. ooooo. mHOA ma. ma. om. om.a nod. «ma. moo. Mama oo. oo. oo. om. who. who. mooo. «Hun mo. oo. om. om.H woo. woo. oooo. damn NH. Ha. mm. mo.a mod. and. moo. Mlam ha. ma. mo. NH. mmm. mmm. mooo. «Hm mo. oH. ma. o¢.m moo. moo. Hoo. 0H0 ma. ma. om. mm. ova. mma. oooo. mHO ea. ma. ma. Hm.a oma. «ma. moo. on om. «a. HH. nm.m mom. omm. ooo. mam m.mmm m.mm~ om. mo. m.omovmv o.~momm¢ m.ova Huaoa Ameuzv Ammuzv Hm>mq m mm mm mm .um> monEom x moan: x mam m Hmuoa cwzumz comsuom .ocmmoo .Hom\a u up axwm "manomum> ucmocomoocHH mocmwwm> mo mammamcm mo hnmfifidmln.mlm mqmda 13S mo. v a on ucmoauacmam. mm. om. om. mo. mo. mo. om. me. no. mm. mm. mo. ma. vo. aN. ma. No. mo. a.mma m.ooa mm. mm. mm. mo. mo. No. mm. me. so. mm. om. mo. ma. mo. mm. ma. No. No. m.ama v.ooa mo. mo. am. on. mm. mm. mm. on. aw. mm. mm. vm. mm. on. me. mm. mm. mm. om. mm. ma. mma. ao.v ova. hm.a ohm. aa. mono. on. mooo. am. ommo. av. «Nam. mo. amvv. on. mmo. om. amN. mo. mesa. ao. mmamo. vo. omma. mo.a who. om. ama. ao. mama. am. onao. vo. tho. co. a.mommma aoo. a.vm~om~ «Na. ooa. mom. memo. ommo. mmmo. maam. mmvv. omv. omm. ovma. mmaoo. mmma. moo. oma. mama. ohao. tho. v.0vmmma m.om~o- mooo. aum woo. ammo moo. 2Hmm ooooo. Dhau mooo. mmau moooo. dmau mooo. am vooo. A aoo. mum mooo. mmé mooooo. a0 vooooo. aUm ooooo. aUB aoo. aUm mooo. aUH aoooo. aUBUd moooo. @200 aoooo. aUz h.No MIBOB a.m NIBOB 136 am. am. om. vo. oavm. mavm. aooo. mBO ha. ha. mm. oo. obma. mbma. moooo. mmdm ma. ma. moo. 8ao.m moa. mma. oao. 9mm ma. ma. mp. no. amam. omam. mooo. ono mo. mo. om. om.a mooo. mooo. mooo. mum mo. mo. mm. mo.a momo. mooo. mooo. mom vo. mo. mm. mm.a ammo. ovmo. mooo. _ ma aa. ma. av. mm. mooa. mmoa. mooo. 9H aao. moo. om. am.a aooo. omoo. aooo. mOH ma. ha. mm. om.a mma. ooa. moo. mmH ma. va. ma. mv.~ mom. omm. boo. mnaOA ma. ma. on. aa. ooma. moma. mooo. mnamq no. mo. om. ov.a moo. who. aoo. duaoq mm. ovo. mm. on. woo. woo. aooo. «lama aa. va. moo. «mm.m ama. voa. oao. muam ma. oa. oa. mm.a mmm. omm. moo. «lam oa. mo. mo. om. mmoo. mmoo. vooo. oaO Na. va. mo. *mo.m ooa. mma. moo. maO ma. «a. mo. mo. mma. mma. mooo. oam va. oa. ma. oo.m mom. aom. moo. mam m.~mm m.omm mm. mm. m.omoo~v m.momom¢ a.mmom alaoa osmox omox ao>oq m mm mm mm .wm> I I mam m amuoa casuaz cmm3umm .ocommo .Hmm\a u an “828 "mannaum> ucmoamamocH_ wmscauaoouu.mum mamas 137 .mo. v a on unmommmammmm mm. om. oo” mo.m mmm. omm. moo. mom om. om. mm. me.m oom. mom. moo. ammo mm. mm. mo. .mm.o mem. mom. emo. 2mmm mo. mo. mm. om.m mooo. mmmo. mooo. ammo oo. oo. me. mm. moo. moo. mooo. mmmo mo. mo. mm. mm.m ommo. memo. mooo. ammo om. om. em. om. mmm. mmm. moo. mm om. mo. mo. mm.m moo. mmo. mmo. m ea. eo. oo. mm. memo. eomo. mooo. mam mm. om. mo. mo.m mmm. mmm. moo. moo om. em. mm. em.m mem. mem. moo. mo mo. oo. mm. mm. moo. moo. oooo. mom om. mm. mm. mm.m omm. omm. moo. mos mo. oo. om. mo.m oeo. meo. moo. mom mm. mm. mm. mm.m mmm. mom. moo. mom mm. om. mo. .mo.m mmm. mmm. eoo. moeom mo. mo. om. mo.m memo. momo. mooo. mzoo mo. mo. me. mo. memo. memo. moooo. moz o.mmm m.mmm mm. mo. m.oommmm m.memmmm o.mm muses m.emm ~.mom mm. mm. m.om~om~ ~.omom~m m.mom~ muses 138 am. om. om. mm. avm. mmm. moo. mBO oa. ma. ma. oa.m mmm.. mma. aoo. mmmm oa. ma. vo. «oo.o ma. ea. ao. Ema ea. ma. mo. «me.o am. om. ao. Ono mo. oo. oo. em.m moo. ooo. moo. mum mo. mo. mm. mm. momo. oomo. mooo. mmm oo. oo. mm. mo. ammo. vmmo. aoooo. ma ma. ma. mm. mo. mooa. mmoa. voooo. EH ao. mo. mm. oo. aooo. omoo. moooo. MOH oa. oa. mmm. oo.o ma. ma. oo.o mom ma. ea. ma. em.m mmm. mmm. eoo. mtaum oa. oa. mo. am. ema. ema. aooo. muamm oo. oo. eo. ma. who. eeo. aooo. «Iaom oo. vo. oo. ea. evo. evo. aooo. «smog ma. ma. oe. oa. aoa. oma. mooo. muam ma. ea. om. oo.a mmm. mmm. moo. «lam mo. oa. mo. eo. moo. moo. vooo. oau ma. ma. ma. oe.a ova. ema. moo. mau va. ma. mo. mm. oma. oma. mooo. mam ma. ea. mm. mv.a ovm. mom. voo. mam e.omm a.avm om. oo.m m.omovmo m.mmvmmv m.meom aIBOB .zmomx .nmmumx mm>mm m mm mm . mm .um> I l mam m amuoe casuaz coozumm .ocomon .Hom\a u no «ao>mq "manuaum> ucmocmmmocHu omscaucoonl.mum mmmdfi .mo. v a um ucmoawacmamm 139 mm. mm. mv. oo. mvma. mmma. mooo. aom mm. om. mm. hv.a ova. mva. moo. ammo om. mm. ma. om.a ohm. mom. aao. Zamm mo. mo. vv. ao. momo. ommo. vooo. Dhao oo. oo. om. om. woo. moo. mooo. mmau mo. ao. mm. mm.a ommo. mmmo. mooo. «mau om. mm. om. vm. mam. aam. moo. am mv. ov. mv. oo. mvv. mqv. moo. A ho. mo. mm. mm. mmv. vmv. voo. mum mm. mm. om. mo.a amm. oom. moo. mhd om. mm. am. oo.a mha. mma. moo. aU mo. mo. mm. om. oaoo. vaoo. mooo. aux ma. va. mo. oa. omma. mmma. mooo. aUB co. «0. we. mo. mmho. mmmo. ooooo. aUm am. am. om. moo. mooma. mooma. moooo. aUH ma. ma. mm. ma. mama. mama. mooo. aUBU¢ mo. mo. mm. mm. ahao. obao. aooo. mZOU mo. mo. ov. om. mhmo. ommo. mooo. aUZ m.mma a.mva aa. oo.m a.mommma m.mmaama o.aoav mIBOB m.mma m.ama va. mm.m a.vmmomm m.maaamm m.avam mIBOB 140 am. am. mm. oa. vavm. mavm. mooo. mBO oa. ea. om. mo.a oma. vma. moo. mmmm ma. va. moo. «mo.e mea. moa. mao. Bad ma. ma. oo. om. amam. eaam. vooo. Ono mo. mo. mo. mm. mooo. mooo. mooo. mum mo. mo. mm. vooo. mmomo. mmomo. moooooo. mam vo. mo. moo. «mv.e vmmo. vamo. ovoo. ma aa. maa. ao. vo. mmoa. mooa. moooo. EH oao. oao. om. mo. aooo. aooo. mooooo. MOH ma. oa. mm. om.a ooa. moa. moo. mma oa. ma. ev. mm. eom. mom. moo. muaom ea. oa. em. vm.a ema. vma. moo. mnamq oo. oo. mo. oa. eoho. ooeo. aooo. «lava vo. vo. mm. mooo. movo. movo. moooooo. mnama aa. mm. mm. ma.a aoa. mea. moo. Miam ma. ea. am. ao. mmm. mmm. moooo. «lam mo. oa. mm. am. mmoo. amoo. mooo. aao ma. ma. av. mo. ova. mma. aoo. mau ma. va. mo. mo. moma. moma. ooooo. oam ma. ma. oo. ma. mmvm. mmvm. mooo. mam a.mvm o.mom eoo. «mm.e m.omovmv m.omammm a.ovmom anaoa oooomx moommx mm>mm a mm mm mm .um> I I mam m amuoa casuaa coozumm .ocommo .moo\m u no usmmumsmoo “mmbmmum> unmoammmocmo omsamucoouu.mum mamas .mo. v Q um acmoamacmama 141 mm. mm. ma. mh.a mma. mma. moo. aum mm. mm. mm. mooo. mmva. emva. moooooo. ammo mm. om. am. oo.a mmm. mom. ooo. EHMQ mo. mo. ao. om. momo. aomo. mooo. DmaU oo. oo. mo. vm. mvom. ovmo. mooo. mmau mo. mo. mm. mm. ommo. vmmo. mooo. «mau om. om. no. mo. mam. mam. mooo. am ov. mv. mm. mm.a mvv. mvv. ooo. a mv. hv. mm. ma. mhmv. onmv. oooo. hmm mm. mm. mm. mo. amm. amm. aooo. and mm. om. ma. mb.a mha. mha. moo. aU mo. oo. ooo. «mm.h moo. hmo. moo. aux ma. ma. mm. ao. mmmma. mmmma. moooo. aUB vo. mo. mm. mm. mmmo. mmho. mooo. aUm am. mm. oa. oo.m ama. ova. voo. aUH ma. ma. va. ma.m mma. maa. moo. aUBUfl mo. mo. mv. mv. aoao. omao. moooo. AZOU mo. mo. av. mo. mhmo. ohmo. mooo. aUZ v.mva v.maa aoo. «om.aa a.mommma m.mmomma v.ommoa mIBOB m.vha m.mva voo. «ov.m a.vmmomm h.ommhom v.mmmma mIBOB 142 om. mm. mo. ma.m avm. vmm. moo. mBO oa. ea. mo. mm. oema. aema. mooo. mmmm ma. va. ma. am.a mea. mea. moo. 9mm ma. va. om. ov. mam. aam. aoo. olo mo. mo. mm. mo. mooo. vooo. moooo. mum mo. mo. ao. om. momo. vomo. aooo. mmm vo. mo. oa. oe.m mmo. vmo. moo. ma ma. aa. mv. mm. mmoa. mmoa. eooo. 9H aao. moo. ma. om.m aooo. meoo. aooo. mum ma. oa. ma. oe.a oma. vma. moo. mma oa. ma. mm. mv.a mmm. mmm. voo. mnaom ea. ma. mo. «me.v ema. eoa. moo. muamm oo. eo. ma. me.a heo. meo. aoo. duaom vo. vo. oo.m oo.o mo. mo. oo.o mnamm ma. ma. em. mo. oooa. mooa. moooo. mnam ea. ma. oo. ma. ammm. ovmm. mooo. «Iam oa. oa. mm. mooo. mmoo. mmoo. aoooooo. oao ma. ma. mm. am.a ova. oma. moo. mao ma. va. vo. «mm.v oma. oma. ooo. Dam va. oa. eo. om.m ovm. mmm. moo. mam m.mmm m.mam mo. om.m m.omovmv m.meoaav m.mavma aneoa xcmomx xcdamx ao>oa m mm mm mm .um> I I mam m aouoa casuaE coo3uom .ocommo .moo\m n ma reummxaa "mmbmmum> ucmocmmmocmo omscmpaoouu.mum mamas .mo. v a so homomomcmmm. 143 mm. mm. ma. om.a mma. mma. moo. aom om. vm. mo. «om.v ova. mma. moo. ammo mm. am. mo. «mm.v vhm. ovm. mmo. Zamm mo. mo. ov. mm. momo. ommo. vooo. Dmau oo. oo. oo. mm. mvoo. ovoo. mooo. mmao mo. mo. om. mo.a ommo. mmmo. mooo. mmau om. om. um. oa.a vmam. mmam. mooo. am ov. mv. mm. ov.a ovv. avv. moo. a mv. ov. aa. mo.m mmv. omv. mao. mum mm. mm. om. mo.a amm. mmm. moo. mud mm. mm. oa. vm.m mma. oma. moo. a0 mo. oo. mo. «om.m oaoo. mmmo. mmoo. aum ma. ma. om. mo. mmmma. vmmma. voooo. aoB mo. vo. mm. on. mmoo. mmoo. oooo. aUm am. mm. mm. mm. moma. moma. mooo. aUa ma. ma. om. hm.a mma. oma. moo. aUBUd mo. mo. mm. vm.a ahao. moao. mooo. @200 mo. mo. mm. ma.a mhmo. momo. vooo. aoz m.ava o.mma mo. aao.m a.mommma m.ooomva o.oomh mIBOB o.moa m.ama mo. om.m a.vmmomm m.ommmam m.mmom mIBOB 144 am. om. mm. mm. vavm. oovm. mooo. moo oa. ma. mo. «oo.o mma. oma. aao. mmmm ma. ma. am. ao. mamha. vamea. moooo. Bm< ma. ma. voo. «om.m mam. ,vma. mao. olo mo. mo. mm. ao.a mmoo. mmoo. eooo. mum mo. mo. mm. ee. momo. aomo. vooo. mom vo. mo. vo. 8mm.v mmo. mmo. moo. ma ma. aa. om. mm.a moa. moa. aoo. 9H ao. ao. ma. mm.a aooo. meoo. mooo. mum oa. ma. ov. me. mma. oma. aoo. mma ma. oa. mv. mv. mmm. mmm. aoo. mnaom ma. ea. oo. vo.m ema. mma. eoo. ulama oo. oo. mm. oo.o eeo. eeo. oo.o dlaoa mo. vo. ev. mm. movo. eovo. mooo. «lama aa. ma. mo. «om.m aoa. mea. eoo. muam ma. oa. aoo. «av.m mmm. omm. mmo. «lam mo. oa. mv. em. mmoo. mmoo. vooo. oaO va. ma. am. ao.a ova. ema. moo. mao ma. mm. mm. oo.o oma. oma. oo.o oam ma. va. am. mo.a ovm. mvm. moo. mam m.mom m.mvm ao. «mm.o, m.omovmv v.moommm m.movom alaoa oo>0Am oo>amm wwmow m mm mm mm .Hm> . aouoe canvas coozuom .ocoooo .mom\a u no “maaaxm whoasnmoo> “manuauo> ucoocoomocau omscaucouul.mlm,mam¢9 145 .mo. v a pm ucmomomcmmmm mm. mm. mv. mo. mma. vma. aoo. aom mm. om. mm. ov. ova. mva. aoo. tha am. mm. ma. ov.m hm. om. ao. Eamm omo. mmo. om. mo.a omo. mmo. aoo. DmaU moo. omo. ma. mm.a voo. moo. moo. mmao mao. oao. mm. aa.a ommo. mmmo. mooo. «mau om. mm. om. mo.a mam. oam. moo. am vv. ov. ma. om.a mvv. mmv. moo. a mv. mv. mo. «om.m omv. mav. omo. mum mm. om. moo. «mm.m amm. mom. mmo. mud mm. mm. oa. mo.m mma. oma. moo. aU mo. oo. oo. mo.m moo. mmo. moo. aux ma. va. ma. oa.m oma. mma. moo. aUB vo. mo. mm. mm. mmmo. ammo. mooo. aUm mm. am. om. mm. moma. aoma. mooo. aUH ma. ma. mm. mo. moama. amama. ooooo. aoaud mmo. mao. ao. «mm.o hao. oao. aoo. mZOU mo. mo. mm. mm. mmmo. momo. mooo. aUz o.mma a.ava mo. «mm.m a.mommma m.mvomva o.mmmm mIBOB m.mva m.ama mo. «ma.o a.vmmomm m.vmomam v.mmoma mIBOB 146 am. am. vo. aa. vavm. aavm. mooo. mBO ma. oa. ooo. *oe.e om. ma. ao. mmmm ma. ma. om. moo. moamoa. moamba. mooooo. 9mm ma. va. mv. em. mam. aam. aoo. Ono mo. mo. me. ma. mooo. vooo. aooo. mum mo. mo. mo. vo. amomo. mvomo. moooo. mam mo. vo. ma. am.m mmo. vmo. aoo. ma aa. ma. va. mm.m moa. ooa. moo. 9H eoo. mao. ao. «om.o aooo. omoo. mooo. MOH ma. oa. aa. vm.m mma. mma. moo. mmH ea. ma. mo. mo.m mm. mm. ao. mlaoa oa. oa. ov. mm. mma. mma. aoo. ulama no. oo. vo. «vm.m moo. moo. voo. mlaoa vo. vo. om. mo. mmovo. mmov. aoooo. «lama aa. ma. om. om.a ama. mma. moo. mlam ea. ma. vm. om.a mmm. mmm. voo. diam oa. mo. mm. mo.a moo. moo. aoo. oau ma. va. mo. «mo.m va. ma. ao. mau va. ma. om. mo.a oma. vma. moo. 0am va. oa. ma. om.m mm. vm. ao. . mam o.mam m.mmm va. am.m m.omovmv m.mvmvav o.mvmm anaoa oa.m am.m mv. mv. mm.mo mo.eo vm. and muouum mwouum ao>oa mm mm mm .Hm> momoa woman mam m m amuoa casuaz com3uom .ocmmoo X x .mom\a u no emwowum amaumom ”mannawm> ucoocoooocao ooscaucoonn.mlm mamde .mo. v m on ucmomomcmmmm 147 mm. mm. va. om.m mma. mma. moo. aum mm. mm. om. oo. omva. omva. aooo. ammo om. am. mo. «mo.v om. mm. mo. Zamm mmo. omo. oa. ao.m omo. mmo. aoo. Dmau oo. oo. mo. oa. mvmo. avoo. aooo. mmau mo. mo. oo. ma. ommmo. ammmo. voooo. dmaU mm. am. moo. «vm.m am. om. mo. am ov. vv. ao. amm.o mv. mv. mo. a mv. ov. mv. om. mmv. mmv. moo. mum mm. mm. am. ao. omomm. ooomm. voooo. mmd mm. om. oa. mo.m moa. aoa. voo. a0 oo. mo. mo. «mo.v moo. mmo. moo. aom va. ma. mm. ma.a oma. vma. aoo. aUB vo. mo. om. vo.a voo. moo. aoo. aow am. mm. va. om.m ama. ova. moo. aUH ma. ma. mm. vo. omama. mmama. moooo. aUBU< ao. mo. voo. «om.m oao. oao. aoo. @200 mo. mo. vo. aa. mmomo. maomo. voooo. aUZ o.vma o.oma aa. om.m a.mommma v.mvmama o.mmov mIBOB m.ama o.ooa ma. om.m a.vmmomm o.oooomm v.oovm mIBOB 148 om. mm. mo. mmm. avm. vmm. ooo. 380 oa. oa. oo. oa. ooma. voma. mooo. mmmm va. ma. ao. eom.o moa. voa. aao. 9&4 ma. va. mm. mm.a mam. mom. moo. Ono mo. mo. mm. mm.a woo. woo. aoo. mum mo. mo. mo. mm. momo. vomo. aooo. mam vo. mo. vv. oo. vmmo. ommo. mooo. ma aa. aa. om. mm. mmoa. omoa. vooo. 8H ao. ao. ao. om. aooo. oooo. moooo. MUH ma. oa. vm. ov.a mma. mma. moo. mmH oa. ma. mo. oo. oomm. momm. mooo. MIaUa oa. oa. mv. mo. oma. mma. aoo. mnama oo. oo. mm. oo.a ooo. ooo. aoo. «lava vo. vo. mm. mo. mmovo. omovo. moooo. diama ma. aa. oo. av.m ma. oa. ao. mnam oa. ma. mm. mm.a mmm. amm. voo. mlam oa. mo. mm. mm.a moo. moo. aoo. QaU ma. ma. ao. va. mmma. omma. mooo. mao va. ma. am. vv. oma. oma. aoo. mam ma. ma. mv. mv. ovm. mvm. aoo. mam m.mmm m.omm mm. mm. . m.omovmv v.ommmav m.ovav aIBOB muouum muouum ao>ma mm mm mm .Hm> oo>oax oo>amx mam m m amuoa canuaz coo3uom .ocoooa .aom\a u no emwowum onwasnmoo> ”mannawm> #coocomoocao concauCOUII.mlm mmmds 149 .mo. v a no unmommmcmmme mm. mm. mm. vmo. oo. mo. om. mv. mv. om. om. mo. ma. vo. am. ma. mo. mo. o.oma m.moa mm. mm. mm. mmo. oo. mo. om. mv. ov. mm. om. mo. ma. mo. mm. ma. mo. mo. o.mma m.mma oo. om. mm. oo. mm. ao. om. om. om. av. mo. vm. om. mm. ov. mv. mm. oa. mm. vm. ma. voo. aoooo. om.m ao. om. mo. moo. mo. mo. mo. ao. mo. mm. ao. mo. mv. mm.a ma.a mm. mvma. mvma. oomva. oomva. mmmmomom. mvmmomom. omo. vmo. mmvmo. amvmo. ommo. mmmo. vmam. vmam. moovv. moovv. mmv. vmv. amm. mom. oovoa. movoa. ammaoo. mvmaoo. mmmma. vmmma. voo. moo. ama. oma. mma. ama. aoao. ooao. omo. omo. a.mommma a.vmmomm a.mommma m.voavmm mooo. aoooo. vooooooo. moo. aoooo. aooo. aooo. aoooo. voo. moo. voooo. mooooo. voooo. aoo. aoo. aoo. aooo. aoo. o.ooma m.mvam aom Damn Zamm Dhao mmau ¢maU am mum and a0 aux aUB aUm aUH aUBU< ASOO aUZ mIBOB mIBOB 150 om. vm. mm. ao. moo oa. mma. oma. aoo. mmmm ma. oamoa. oamoa. aoooo. Bud ma. am. om. ao. ouo mo. mooo. mmoo. aooo. mum mo. oamomo. vomomo. aooooo. mom mo., mmo. vmo. aoo. mm am. ovmmoa. mvmmoa. aooooo. 9H ao. ooooo. momoo. moooo. mum oa. omoma. mmoma. aoooo. mom oa. omm. vmm. moo. mnamo oa. oma. ama. ooo. muaom oo. aooooo. mooooo. oooooo. «Iaom vo. mmovo. omovo. ooooo. duaom ma. amm. moa. aoo. mnam oa. mmm. mmm. moo. duam oa. mmoo. amoo. aooo. aao ma. aoooo. amooomma. omooomma. aooooooo. mau ma. mmoma. vmoma. mooooo. cam ma. ovm. mvm. aoo. mam o.mvm m.omovmv o.mvovov m.mvmma ansoe omwoa mm mm .Hm> x aouoa canvas coosuom .ocomoo mo ummmummmba mmmummm “mannaum> ucoocomoocHo ooscaucoonl.mlm momma 151 .mo. v m on ocmommmcmmmo mm. mm. me. om. momm. oomm. mooo. mom mm. mm. mm. mo. eooom. mooom. ooooo. noon mm. om. oo. oo. oeo. oem. ooo. ammo mo. mo. mm. oo. mooo. oooo. mooo. ammo moo. ooo. om. mm.m ooo. moo. moo. mmmo mm. em. om. mo. ommmo. oommo. moooo. «mmo om. om. oo. mo. mmm. mmm. moo. mm oo. oo. om. mo. moo. ooo. ooo. m oo. oo. em. mo.m mmo. omo. moo. mom em. om. em. mo. mmm. mmm. moo. mom mm. om. om. oo.m oem. mem. moo. mo oo. oo. em. om.m moo. moo. moo. mom mm. mm. mo. om. oomm. momm. mooo. moo mo. oo. mo. mo. mmeo. mmeo. mooo. mom mm. mm. mm. mo. ooomm. eooom. moooo. mom om. om. em. om. mmm. mmm. moo. momom mo. mo. me. oo. ooemo. ooemo. moooo. ozoo mo. mo. mo. .oo.o emo. omo. moo. moz o.mom o.mmm mo. .oo.o m.momoom o.ommoom e.omom muses o.mem m.oom mo. .me.o m.oomomm o.mooomm m.moeom mueoa 152 om. om. mm. om. mm.a vm. mm. ao. moo oa. va. oa. mo. ama.m om. ma. ao. mMMm ma. ma. mm. mm. oooo. maoamoa. moamoa. mooooo. Bad vm. va. ma. ov. vm. mam. mom. voo. Ono mo. vo. mo. mm. mm.a moo. ooo. moo. mum mo. mo. mo. mm. ao. mmomo. omomo. moooo. mom mo. mo. vo. vm. oo. vmmo. mmmo. aoooo. mm ooa. oma. maa. mm. mm.a moa. oom. moo. 9H moo. oao. moo. moo. omm.o moo. ooo. aoo. mua ma. oa. oa. vm. mo.a mma. vma. voo. mom ooa. mva. mma. mm. om. mm. mm. ao. mumoq oa. ma. oa. mm. om. oma. mma. ooo. maawm moo. mmo. ooo. mm. oo.a ooo. moo. moo. «uaom omo. mmo. mmo. oo. mv. movo. movo. vooo. «Imam ma. . aa. ma. mv. am. ama. moa. moo. mnam oa. am. oa. mo. «oo.m om. vm. mo. diam mo. oa. oa. om. mm. moo. moo. aoo. oaU ma. ma. ma. am. mm.a ova. mma. voo. mao om. va. va. mo. mo.m va. ma. ao. cam ma. oa. ma. mo. cmm.m ovm. omm. mao. mam a.omm o.mmm m.vmm om. aa. m.omovmv v.moammv m.omm aIaoa ao>om mm .uo> oooaooocpm mvmmomaom ooumam mam m m aouoa cannaz coo3uom .ocomoo .amm\m u no «:0auoouao oosuauud «mandauo> ucoocomoocao noscaucooia.mlm mamdfi 153 .mo. v a no unmoamacmama om. om. om. mo. vao. mm. ov. om. mm. mo. va. om. ma. mo. mo. a.mma v.mma mm. mm. mm. mo. oo. mao. am. mv. vv. mm. mm. mo. oa. vo. mm. ma. mo. mo. m.mma m.ooa om. mm. om. mo. oo. omo. mm. mv. ov. mm. om. mo. ma. vo. mm. ma. mo. mo. o.mma m.mma mo. mo. mo. mo. va. vm. mo. mo. om. ma. mm. ma. mm. oo. mm. mo. mo. mm. mm.m mm. «vo.m mm. vv. mm. ov.a oaa.m aoo.v mm. am.a mo. oo.a mo. vm.m mv.a aam.v mo. mm. oa. mma. maa. ooo. aom ova. mva. moo. ammo om. mm. vo. Zamm momo. oomo. mooo. Dhao voo. mmo. aoo. mmau mmo. mmo. aoo. om mm mm mm .un> +omm omx om emx 0mm m m memos amaums ammsuwm .oaodoo .mooxm a mo «on: “wandauo> unoocomooca_ ooflcaucounl.mlm manta .mo. v a no ucooawacmmm 155 mm. mm. mm. mm. mm. oa. mma. vma. oooo. aum om. om. mm. mm. mm. mv.a ov. mma. ooo. ammo om. mm. mm. am. vm. mm. vom. mom. moo. Same mo. mo. mo. mo. mm. mm. omo. omo. mooo. DmaU oo. mo. oo. oo. mm. om.a vmo. amo. moo. mmau mo. ao. mo. ao. ma. mm.a mmo. amo. aoo. «mau om. om. om. mm. oo. mm. mam. oam. moo. am vv. ov. mv. ov. am. ma. mvv. mvv. moo. A ov. ov. ov. ov. om. mo. mmv. omv. aoo. mom om. mm. mm. mm. mm. mm. amm. mmm. moo. mod om. vm. om. om. om. mm.a moa. moa. ooo. a0 vo. mo. oo. oo. om. oo.a moo. ooo. moo. aum oa. va. va. va. mm. om.a oma. ama. moo. ave mo. vo. vo. vo. mo. mm. voo. moo. aoo. aum am. am. am. am. mm. am. ama. oma. aoo. aUa ma. ma. ma. oa. mm. oa. mma. ama. oooo. aUBUd mo. mo. mo. mo. ma. vo.a oao. oao. mooo. QEOU mo. mo. mo. mo. mo. mm. omo. omo. oooo. auz a.mma o.mma m.oma o.mma om. mm. a.mommma m.mmovma m.moma mIBOB o.moa mma m.mma m.moa om. vm. a.vmmomm m.mvmmmm a.aavm mIBOB 156 om. mm. am. oa. oa. ma.m vm. mm. mo. :80 em. em. em. om. me. om. mmm. mmm. moo. ammo om. om. mm. mm. mm. me. mem. mem. ooo. 9mm mm. om. mm. em. mm. om.m mm. om. mo. o-o mo. mo. mo. mo. mo. oo. ooo. eoo. . moo. mum mo. mo. mo. oo. mo. .oe.m mmo. ooo. ooo. mom oo. oo. mo. oo. mm. mm. ommo. ommo. oooo. mm mm. mm. mm. mm. om. om. mom. mom. moo. em omo. mmo. omo. ooo. mo. om. mmoo. oeoo. mooo. mom mm. om. om. om. mm. mm. mom. emm. moo. mom om. mm. mm. om. mm. mm. emm. oom. moo. mumom em. om. mm. om. om. me. emm. mom. ooo. m-mmm oo. eo. oo. eo. mo. mo. eeo. meo. moo. «-mom oo. oo. mo. mo. oe. mo. eoo. ooo. moo. «-mmm mm. mm. mm. mm. mm. om. mom. oom. moo.. m-mm em. mm. mm. em. me. oo. mmm. mmm. ooo. «-mm mo. om. om. mm. om. oo. . moo. moo. moo. omo mm. mm. mm. mm. eo. mm. oom. emm. moo. mmo om. mm. mm. om. om. eo. momm. . ommm. mooo. omm mm. om. om. em. om. mo.m mm. om. mo. mmm o.omm m.mmm o.omm e.omm me.. oo. o.oooomo o.mmoomo o.meoo mueoe o m m m mm>wm a mm mm mm .um> m m . m, m mmm a memos cmnums amozuwm .ocmdmo o H¢m\m nu NO uNUfiWGOflfl—H Omvfiufiuud «OHDMNHMNV HGOUGUQQUSHH @OQQflUCOUllomIm Waugh. .mo. v a no unavawacmama 157 mm. mm. mm. om. ov. vm. mma. ama. moo. aum mm. om. mm. mm. am. oo. ova. mva. moo. ammo mm. mm. am. mm. mv. mm. om. om. mo. 2Hmm mo. mo. mo. vo. mm. mm. momo. mmmo. vooo. DhaU oo. oo. oo. oo. mo. ov. vmo. mmo. aoo. mmau mo. mo. ao. ao. om. mm.a mmo. mmo. aoo. mmau om. mm. am. am. mv. mm. am. am. ao. am ov. ov. mv. mv. om. oo.a mv. mv. ao. a mv. ov. ov. mv. mo. mm. vv. mv. ao. mum om. mm. om. om. vo. mv. amm. omm. voo. and mm. om. om. om. mo. oo. moa. aoa. moo. aU mo. mo. mo. oo. mo. mv. moo. aoo. aoo. aux va. ma. ma. ma. am. mm. oma. vma. aoo. aua vo. mo. vo. oo. oo. am.m oo. oo. ao. aUm am. mm. mm. om. mv. am. ama. ova. voo. aUH ma. ma. ma. om. am. mo. mma. maa. moo. aUBQd mo. mo. mo. mo. mm. mo. aoao. ooao. mooo. AZOU mo. mo. mo. mo. mv. vm. omo. omo. aoo. auz m.mma m.mma m.mma o.vma am. mo. a.mommma o.mamama v.omom mIBOB o.ooa o.oma a.mma o.ooa mo. mv. a.vmmomm o.avommm v.mamm mIBOB APPENDIX C QUESTIONNAIRES APPENDIX C Do Not Write In This Space Michigan State University Subject No. East Lansing, Michigan Group No. Project No. PUBLIC OPINION SURVEY ( College-University) Test 1 Name Age 2 ’1 (years) Sex F Social Security or 3 Student Number 9 Year in "__.Freshman College Sophomore 5 Junior College attending Senior 6 _ Date 7 e I Michigan State University's College of Communication Arts is conducting an exploratory study on personal and social opinions. This test booklet has several short blocks or groups of questions which are essential in the carrying out of this project. Please remember there are no right or wrong answers. You are asked to give your frank and honest Opinions at this time. The school in which this test is being conducted is not sponsoring this survey, and neither the school, its administration, the instructor or anyone associated with the school will have usage of this infonmation. He ask for your name, et. al., for identifying purposes only. Your anonymity_is guaranteed. At a later date, a Michigan State University staff member will return to answer any questions you might have about the project. Please do not open this booklet until you have received appropriate instructions from the project leader. Thank you for your cooperation. 158 159 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 pgrsonalpgpinion. He have tried to cover many different and opposing points of view; you may find yourself agreeing strongly with some of the statements, disagreeing just as strongly 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. Hark 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 WHOLE +3: I AGREE VERY MUCH -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 hiShGSt form 0f 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. 160 Continue marking your answers in this manner: +1: I AGREE A LITTLE -l: I DISAGREE A LITTLE +2: I AGREE ON THE WHOLE -2: I DISAGREE ON THE WHOLE +3: I AGREE VERY MUCH -3: I DISAGREE VERY MUCH The present is all too often full of unhappiness. It is only the future that counts. To compromise with our political opponents is dangerous because it usually leads to betrayal of our own side. It is only when a person devotes himself to an ideal 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 worthwhile goal, it is unfortunately necessary to restrict the freedom of certain political groups. In a discussion I often find it necessary to repeat myself several times to make sure I am being understood. Host 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 to be a live coward. Hy 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. Of all the different philOSOphies which exist in this world there is probably only one which is correct. 161 Name: CARD ROTATIOHS TEST -—-S-l This is a test of your ability to see differences in figures. Look at the 5 triangle-shaped cards drawn below. h a 7 V A All of these drawings are of the swme card, which has been slid around into different positions on the page. Now look at the 2 cards below: These two cards are not alike. The first cannot I: il be made to look like the second by sliding it around on the page. It would have to be flipped over or made differently. Each problem in this test consists of one card on the left of a vertical line and eight cards on the right. You are to decide whether each of the eight cards on the right is the same as or different from the card at the left. Put a plus (+) or cross (X5 on-the card, if it is the same as the one at the beginning of the row. Put a minus (——Q on the card, if it is different from the one at the beginning of the row. Practice on the following rows. The first row has been correctly marked for you. If) G7 (5] <9 <3 <7 G El- [‘9 @@©©©©©© D 6 [5 053900 {:3 Your score on this test will be the number of cards marked correctly minus the number marked incorrectly. Therefore, it will not be to your advantage to guess, unless you have some idea whether the-card is the same or different. Work as quickly as you can without sacrificing accuracy. You will have h minutes for each of the two parts of this test. Each part has 1 page. When you have finished Part 1, STOP. Please do not go on to Part 2 until you are asked to do so. DO NOT TURN THIS PAGE UNTIL ASKED TO DO SO. Convrizht (a 1962 by Educational Testing Service. All rights reserved. mgimmvéaw v.00 m.mgg a sesame see/asp an“; m®b© QQO\M§0 msafiae d© a deified Namesake» @ fi0€i%§g e32 Bebe as seeds: Eng-remodms QééaEah PAGE UNTIL ASKED TO DO SO. DO NOT TURN THIS eeeezdsoneao as Ms EQEVEUQWQEQ m& was levees sees as ansEGAoa GQQPWM .WQ mGa®ADflOTflflG beam messhegflwoaeflo saw a mGeeGQQAZUEQ a} «Ukoaemgqmvw Gamma Ev a seeoneokeieam minutes) OANY OTHER TEST UNTIL ASKED TO DO 80. m m 164 This section of the survey seeks to determine your own personal opinion on certain topics. You will be asked your opinion on five statements, stating whether you agree, disagree, or are undecided, and then you will be asked to state how strongly you feel about your position. For example: Do you believe a college education is worthwhile? .__X__ Agree Disagree Undecided How strongly do you feel about this? Not at all strongly Not so strongly .__1L_ Fairly strongly Very strongly Thus, in the above question, the X mark denotes you agree a college education is worthwiile, and that you feel fairly strongly about your belief. Now proceed with the following five statements, giving your own personal opinion. Please answer every question, with one answer only. Do you believe the United States should withdraw all military troop assistance to South Viet Nam? Agree Disagree Undecided How strongly do you feel about this? Not at all strongly Not so strongly Fairly strongly Very strongly 165 Do you believe Sen. Edward M. Kennedy's recent involvement in the fatal accident disqualifies him for President of the United States? Agree Disagree Undecided How strongly do you feel about this? Not at all strongly Not so strongly Fairly strongly Very strongly Do you believe churches in general are too involved in politics? Agree Disagree Undecided How strongly do you feel about this? Not at all strongly Not so strongly Fairly strongly Very strongly Do you believe public aid to non-public schools should be prohibited? Agree Disagree Undecided 166 How strongly do you feel about this? Not at all strongly Not so strongly Fairly strongly Very strongly Do you believe the current college student protests are justified? Agree — Disagree Undecided How strongly do you feel about this? Not at all strongly Not so strongly Fairly strongly Very strongly 167 wmmcsvocmxnsr -- v-3 Me is e. test of your knowledge of word meanings. Look et the eemple below. (be of the five numbered words has the same meaning or neerly the some meaning he the word above the numbered words. lurk your enewer by putting an X through the nmnber in front of the word thet you eelect. Jovial l-refreshing 2-ecnre 3-thickeet hawiee X—Jolly the enewer to the sample item is number 5; therefore, an X hee been wt through nude: 5. Your ecore will be the number marked correctly minus e traction of the ntmer marked incorrectly. Therefore, it will pat be to your edventece to gueee unleee you are able to eliminate one or more of the enewer choicee ee 'wrong. You will have 6 minutes for each of the two perte of thie teet. hch pert hee one page. When you have finished Part 1, STOP. Pleue donotgoontoPerthtilyouareeekeitodoeo. DONOP‘IURNTIIISPAGEUNI'ILASKEDTODOSO. Copyright 1962 by Educational Testing Service cottontail l-squirrel 2—poplar 5-boa hemarshy'plant 5-rabbit earteteble l-partiean 2-Jocular B-marriageable k-aalable 5~eesential hose! l-afraid 2-false b-marshy h-dense fieblack gruesoleness l-blacknese Z-flalsenese B-vindictivenese h-drunkenneae S-ghastlineae loathing l-diffidence 2-lazinees j-abhorrence h-cleverneee 5-comfort bantam. l-fowl 2-ridicule B-cripple h-vegetable S-ensign 10. 11. 168 Part 1 (6 minutesl evoke l). l-wake up 2-surrender S-reconnoiter h-transcend 5-chl forth unobtrusive 1h. l-unintelligent 2-cpiloptic 5-illogical h-linesl 5-modest terrain l5. l-icc cream 2-ftmal test b-trsctor h-area of ground 5-weight capriciousncss l6. l-stubbornness 2-co e 5-whimsicality h-nnazenent S-greediness melstrom 17 . l—slander 2-whirlpool i-enmity h-armor 5-majolica tentative 18. l-criticnl 2-conclusive b-cuthentic h-provisional S-cpprehensive placate l-rchabilitate 2-plagiarize S-depre date h-cpprise 5-conciliete BUICCSBS l-enlightennent 2-cecsation B-incttention h-censor 5-substitution apathetic l-v:nderinc 2-izpcssive B-hstcful h-prOphetic 5-overflowing paternostcr l-paternalism 2-patricide 5-mslediction h-benediction 5-prayer opalescencc l-opulence 2-senescence 5-bankruptcy h-iridescence 5-assiduity lush l-stupid 2-luxurious S-hazy h-putrid S-lansuishins no NOI‘ TURN TE PAGE UNTIL ASKED TO DO SO. 19. 21. 22. 23. 2h. curtailment l-expenditure 2-abandohmcnt 3-abridgment h-inprovcmcnt 5-forgery perversity l-adversity 2-perviousness 5-trcvesty h-wnywardnesc fi-gentility calumnious l-complincntary 2-analcgcus B-slsndc I'GZLS h-tenpestuous 5-msgnanir3us illiberality l-blgotry 2-1mbecility 3-illcgibility h-ceutery 5-inmaturit clabber l-rejoice 2-gossip B-curdle h-crow 5-hobble sedulousness 1-diligcnce 2-credulousness 5-seduction h-perilousnesa 5—frcnkncss STOP. ehortcehs )1. l-condieent 2—peetry S-truit beast-set Q-veaeteble hlrdtack 32. l-neil 2-textile S-Ietpon h-wnod 5ebiecuit commendable 3). l-pleesureble 2-cheritable Salucrative h-proecriptive 5-leudsble nonchelant 5h, l-eerceetic 2-diecourteoue D-noble h-unoonoerned , S-uneophieticeted coloration l-pianntetion 55° 2-elternetion B-configuretion h-prevention 5-taint eridity l-bitternese 2-eurfhce b-eonority b-dryneee 5-torridity )5. 169 Part 2 (6 minutele demoniacal 37. leeloof Zheythioal B-thoughtful h-fiendish 5—eccentric highrosd 38. l-mountain road 2-right of way )-main road h-roadbed 5-concrete road befog 39. l-dampen 2-forget 54min h—myetify 5-belittle platoon ho. l-tebleland 2-bridge of boats j-body of soldiers hocommnnplece remark 5-trigete bl dullsrd l-peon 2-duck b-brssssrt h-thief 5-dunos he. momentously l-frivolouely 2-Ioderately jaweightily hdmomentarily 5-modiehly corroboratory l-plausible 2-enticipstory B-confirlatory h-explanatory 5-esoteric figurine l-metsphor 2-wine 3-poem h-organ 5-statuette MCOI‘OLII l-malignant 2-Jubilant B-abashed h-inglorious 5-careless inveteracy l-habitualness 2-migrat ion 3—bravery h-covering 5-hatefulness choler l-anger 2-chorister B-guard h-saliva 5-refrigerator vacillation l-purification 2-wavering 5-expulsion h—tempting 5-roolishness DO HOT 00 BACKZTO BERT 1 AND A}. “5. h6. h7. DO IUT'OO'OI'TO'AIY’OTEIR'TIBT UNTIL AINIDTTO DO 80. easrendizement l-thett 2-ilpeachment B-derision h-amazement 5—enlsrgement effulgence l-prominence 2-outline 3-chsnge h-radiance 5-energy aphasia l-loes of speecr: 2-drunkenness B-anemia h-lose of memory 5-rash panoplied l-philosophical 2-dressed in . armor 5-pant1n8 h-frenzied 5—etavi stic sacrossnct l-eecrificial 2-dorment 5-1nviolable h-superficial S-Eullible prurience l-modesty 2-eapience 5-provender h-lust S-security BTOP. 170 P I C A - form 168 'ihis instrument is composed of 20 statements regarding feelings about ell-niacin. with other people. Indicate the degree to which the statements apply to you by marking whefier you (1) strongly agree. (2) agree, (3) are undecided. (A) dieelrse. or (5) stroqu disagree with each statement. work quickly, just record your first iqrsssion. b m mark on this page. Please use the answer sheet provided. 1 2 l. I look forward to an opportunity to speak in public. 8A A 2. It hands tremble when I try to handle objects on the platform. 8A A 3. I dislike to use my body and voice sxpressivsly. 8A A A. W thoughts become confused and jubled when i speak before an audience. as A 5. I have no fear of facing an audience. M A 6. Although I a nervous just before getting up. I soon forget my fears and enjoy the experience; 8‘ g > 7. I face the prospect of making a speech with complete confidence. 0. Although I talk fluently with friends I am at a loss for words on the platform. Q. I feel relaxed and comfortable while speaking. 10. 1 always avoid speaking in public if possible. ll. I enjoy preparing a talk. 33:33 >>>>> ll. lly postuu feels strained and unnatural. 13. l a fearful and tense all the while i am speaking before a a group of people. it. I find the prospect of speaking mildly pleasant. :3 m PLEASE fill one: a upcu come- 83 .U U 13. 16. 17. 18. 19. 20. 171 I look forward to enpressing my opinion at neetings. “tile participatiu in a conversation with a new acquaintance I feel very nervous. Conversing with people who hold positions of authority causes me to be fearful and tense. I wold enjoy presenting a speech on n local television show. I feel that I - nore fluent when talking to people than nost other people are. M I am tense and nervous while participating in group discussions. 8A 8D 5D SD 8D 8D 172 NAME STUDENT OR SOCIAL SECURITY NUMBER Michigan State University's College of Communication Arts is conducting an exploratory study on personal and social opinions. The school in which this test is being conducted is not sponsoring this survey, and neither the school, its administration, the instructor or anyone associated with the school will have usage of this information. We ask for your name, et. al., for identifying purposes only. Your anonymitygis guaranteed. After you have completed this exercise, a Michigan State University staff member will answer any questions you might have about the project. 173 This survey seeks to determine public opinion on certain topics. You are being asked your opinion on three statements stating whether you agree, disagree, or are undecided, and then you will be asked to state how strongly you feel about your position. For example. Do you believe Richard Nixon is a good President? x Agree Disagree Undecided How strongly do you feel about this? Not at all strongly Not so strongly x Fairly strongly Very strongly Thus, in the above example, the x mark denote you agree Nixon is a good president, and that you feel fairly strongly about your belief. Now proceed with the following three statements. 174 Do you believe Sen. Edward M. Kennedy's recent involvement in the fatal accident disqualifies him for President of the United States? Agree Disagree Undecided How strongly do you feel about this? Not at all strongly Not so strongly Fairly strongly Very strongly Do you believe churches in general are too involved in politics? Agree Disagree Undecided How strongly do you feel about this? Not at all strongly Not so strongly Fairly strongly Very strongly Do you believe public aid to non-public schools should be prohibited? Agree Disagree Undecided 175 How strongly do you feel about this? Not at all strongly Not so strongly Fairly strongly Very strongly Write the most persuasive speech on thv following topic, taking the position you favor most. You have 20 minutes to write. Write as though you were to give this speech before an audience that has many different positions on the topic. PUBLIC AID TO NON-PUBLIC SCHOOLS '6 '4 ~. I .sl i.l4.“1$hn‘flno .