ABSTRACT EVALUATION OF TELEVISION AS A FUNCTION OF SELF-BELIEFS By Keith William Mielke Several studies have indicated that general favorability toward television covaries with educational level, which is typically indexed by number of years of formal schooling. Since the index itself has no explanatory power, several plausible explanations have been offered, but without being given a direct empirical test. A structural similarity was discerned between two such proposed explanations, one by Steiner and one by Schramm. Both could be rephrased, without doing them violence, into statements that persons least impressed with television in general were persons adhering most strongly to a value for critical evaluation or to a value for industriousness. Viewing TV in general would not be expressive of their values, hence the unfavorable attitude. Self-beliefs in terms of values, with television behavior as manifestations of those values, made up the theoretic framework within which the problem of explaining TV attitudes was approached. The design centered on the indexing of self-beliefs in terms of the two proposed values, the indexing of attitudes toward.TV in general, and the indexing of perceived.relevance between the values and the TV attitudes. The application of the theory demanded that the respondent actually perceive 2 Keith W. Mielke favorability toward TV as appropriate to those not adhering strongly to the two values (critical evaluation and industriousness). TV attitudes were indexed in two ways: with a projective instrument and with an adjective rating instrument similar to the semantic differential. value adherence was tapped.by obtaining the respondents' degree of agreement with a series of opinion statements. These state- ments were designed to typify strong adherence to a value for critical evaluation or for industriousness. Perceived relevance between the general values and television was indexed by describing hypothetical types of people at different levels of value adherence. Respondents then esti- mated how each hypothetical type would evaluate television in general. Perceived relevance was defined as statistically significant differences in mean attitude estimates for these hypothetical types. The instrument designed to index value adherence failed to meet several criteria for acceptability, thus not allowing an unambiguous test of the major hypotheses which predicted a negative relationship between degree of value adherence and favorability toward TV in general. Perceived relevance of education and value adherence to favorability toward television in general was established with significant differences in attitude estimates associated with hypothetical types of persons who had different levels of education and value adherence. The TV attitudes associated with sequentially increasing levels of education decreased in favorability at each level, as predicted, when averaged across all respondents. This linear trend was given more precise analysis by controlling on the respondents' own level of education, yielding a predicted interaction effect: the higher a respondent's own 3 Keith'w. Mielke educational level, the greater the difference in his TV favorability esti- mate between high and low levels of value adherence or education. Other data included an upper-middle class identification measure in terms of TV program preferences and a series of Q block analyses. As predicted, it was found that the more a person identifies with the upper- middle class in program preferences, the less favorable he is toward television in general. The Q block analyses, based on an earlier study by MacLean and Talbott, were pursued for exploratory methodological purposes rather than as tests of the theoretic formulation. It was con- cluded that Q methodology did hold promise in a study of this kind, providing the original Q-sort items were structured around the theme of general values. EVALUATION OF TELEVISION AS A FUNCTION OF SELF-BELIEFS by Keith William Mielke A THESIS Submitted to Michigan.State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Communication 1965 ACKN OW L113 D SEE-TEN TS An expression of appreciation goes to my committee members for aid in the planning of this project. Special thanks are due Dr. Verling C. Troldahl, my thesis advisor, to whom fell the greatest share of this burden. In the collection of data, the assistance of several friends was indispensable. Jim Pruyne, Jay Heston, Bob Mertz, Howard Clegg, Mrs. Otto Mielke, Mrs. Vernon.Fratzke, and the Van Westrienen family have my sincere gratitude for their help in contacting respondents and aiding in the administration of a lengthy questionnaire. My collective thanks to the volunteer respondents. Collected data must be processed. my mother-in-law, Mrs. Alice Harper, toiled.many hours in tedious coding of questionnaires. Al Talbott gave unstintingly of his time and considerable talent in processing the data through Michigan State University's electronic computer. My wife has not only worked very hard and effectively in all phases of this project; she had also been a continuing source of psychological equilibrium to me during a frustrating period when several other forces seemed to be having the opposite effect. I either could not or would not have done it alone. Thank you, Pat. ii CHAPTER I II III IV TABLE OF CONTENTS INTRODUCTION AND RATIONALE . . . . . . . . . . . Interdependence of the Media and Social Science Toward an Explanation of Television.Attitudes . Attitudes Toward TV as Manifestations of More General Values . . . . . . . . . . . . . . . The Research Problem . . . . . . . . . . . . . HE‘S I 3351 O O O O O O O O O O O O O O O O O O O 0 Overview of the Design . . . . O O O O O O O O 0 TV Attitude Measurement . . . . . . . . . . . . value Adherence Measurement . . . . . . . . . . Relevance Measurement . . . . . . . . . . . . . Identification Measurement . . . . . . . . . . . Q Block Typologies of Orientations to Television DataCOlleCtion000.000.00.00... Analysis Procedures . . . . . . . . . . . . . . REES ULTS O O O O O O I O O O O I O O O O O O O 0 Item Analyses and Frequency Distributions . . . Correlates of Attitudes Toward Television . . . Findings Related to Perceived Relevance of Value Adherence to TV Attitudes . . . . . . . . . . Special Subgroup Analyses: "Deviants" and "Non-DeVian‘bS" eeeeoooeooeoooe Q Block Findings . . . . . . . . . . . . . . . CONCLUSIONS, DISCUSSION, AND SUHLARY . . . . . . 00113111510113 0 o o e o e o o o o o o e o o o o e Discus 51011 e o e o e e o o o o o o o e o e e 0 Summary 0 o e e e e o o o e o REFEPHEIqCISD BIBLImRWIH O O O O O O O O O O O O APPENDIX A: Pretest Questionnaire . . . . . . . APPENDIX B: verbal Instructions for the Main Questionnaire . . . . . . . . . . APPENDIX C: Questionnaire . . . . . . . . . . iii Page 101 107 TABLE 1. 2. 10. ll. 12. 13. 15. 16. LIST OF TABLES Sample Description and Comparison with U.S. Population CG n5 us Data 0 O O O O O O O O O O O O I O O O O O O 0 Frequency Distributions on Four value Adherence Pieasures O O O O O O O I O O O O O O O O O O O 0 Frequency Distributions on Two TV.Attitude Measures Frequency Distribution on the TV-Programming Status an Identification Measure . . . . . . . . . . . . Relationship Between TV Attitudes and More General Values..................... Relationship Between TV Attitudes and the Discrepancy Between Self and."Average Person" on More General Values..................... Relationship Between TV.Attitudes and.Upper-Middle Class Identification e e o o e e e o o o e e o 0 Mean Estimated TV Attitudes of Others . . . . . . . Mean Estimated TV Attitudes of Others by Subgroups: Value for Critical Evaluation 0 e e e e o o o e 0 Mean Estimated TV.Attitudes of Others by Subgroups: Value for Industriousness . . . . . . . . . . . Mean Estimated TV Attitudes of Others by Subgroups: EducationLeVel................. Comparison of Two Subgroups on the Relationship Between TV Attitudes and More General Values: Fisher Z-Transformations of Correlations . . . . . . . . . TV Attitudes (Projective) by Q Factor Types . . . . TV Attitudes (Adjective Ratings) by Q Factor Types Self-Other Discrepancy on Adherence to a value for Critical Evaluation by Q Factor Types . . . . . . Self-Other Discrepancy on.Adherence to a Value for Industriousness by Q Factor Type3* . . . . . . . iv Page M: So 51 SI 53 Q\ N 6h 67 68 69 69 INTRODUCTION AND RATIONALE Interdependence of the Media and Social Science In 1955, Frank Stanton, president of the Columbia Broadcasting System, addressed the National Association of Broadcasters on the topic, "The Role of Television in Our Society." In this speech, the point was made that television's role cannot be understood or evaluated independently of its social context: . . . the role of television in our society is never going to be determined just by what we do-dwe in industry--or even by what the FCC does, but by what the whole of society does. The less society does, the less we can do. 'We are its mirror. we are the great extension of almost everything. Since nothing less than society itself is going to determine our broadest future aims, let society get in here and push. If society will push, it will find us most responsive; all we ask is that it push us toward the open spaces and not into a corner. we ourselves still don't know where the unconquered spaces of tele- vision are because we still lack a complete map. Accordingly, the Columbia Broadcasting System will provide the funds and form a major public committee for the purpose of a comprehensive, impartial nationwide study of what the public wants from television. (Stanton, 1955) The announced research project, an audience investigation of major proportions, was unprecedented within the television industry. Gilbert Seldes (1956) hailed the "Stanton plan" as having great significance: For the first time a broadcaster has committed himself to the principle of consultation with the entire public instead of being content with measuring the size of that portion of the 1 2 public which constitutes an audience. For the first time the public has, by implication, received freedom to use the air frequencies in order to criticize those who hold those frequencies. (p. 282) On another occasion, Dr. Stanton reportedly advised a broadcasting educator that the networks had greater need of graduates who could tell television decisionsmakers what they were doing-dwhat happened when programs were aireds-than they had of graduates with technical skills. Television as a multidimensional social phenomenon is extremely complex, and.Stanton's request to "give us men who can tell us what we're doing" is no small assignment. Nor is Stanton's challenge unique with the television industry. The research community in the social sciences has been asking the same basic question: what are the major variables involved in describing, explaining, and predicting mass communication behavior? How are these variables interrelated? In the research sponsored by CBS, the proclaimed emphasis was clearly on the audience--what functions they use television for, what kind of television thgy would like to have. Although taking audience variables into account may seem obvious enough for both networks and researchers, the Stanton plan was a pioneering effort. Similarly, early research in communication, according to Bauer (l96h), tended to ignore or underestimate the role of the audience as an active participant in the communication process: The study of communication has traditionally (although not exclusively) been conducted from the point of view of the effects intended.gy the communicator. From this perspective, the dis- parity between actual and intended results has often been puzzling. The answer has come increasingly to be seen in entering the phenomenal world of the audience and studying the functions which communication serves. The failure in research to this point has 3 been that the audience has not been given full status in the exchange: The intentions of its members have not been given the same attention as those of the communicator. (p. 321) The communicator is making a conceptual error, Bauer would argue, when he only asks what effect his message will have 22 the receiver. The question of what the receiver will do with the message--what behavior he will initiate according to his own "rules"-is equally important in the analysis of a communication situation. This latter concern is exemplified in Schramm's well publicized statement, "It seems clear that in order to understand television's impact and effect on children we have first to get away from the unrealistic concept of what television 'does to children' and substitute the concept of what children 23 with television." (Schramm, Lyle, and Parker, 1961, p. 169). Similarly, Klapper urges an analytical approach that is ". . . a shift away from the tendency to regard.mass communication as a necessary and sufficient cause of audience effects, toward a view of the media as influences, working amid other influences, in a total situation." (Klapper, 1960, p. 5). When policyemakers in television see the value of taking the broad view of broadcasting as a multidimensional social phenomenon, the ex- planatory insufficiencies of surface analyses such as ratings become more apparent. A rating service will estimate the percent of homes viewing a particular program; it attempts no explanation on.why that particular audience is viewing, or what the relative effects of the program on the audience might be within the programsaudience transaction. ‘With the importance of audience variables thus emphasized, the inference may be made that an understanding of human behavior-~especially relating, in this case, to how and why different types of people use the h mass media-~can be useful, both from a communication theory standpoint and from the television industry's standpoint. It is noteworthy that the utility of communication theory proposed here in no way relieves television producers of their responsibilities for deciding what to program or how to program it. As Shakespeare's audience was unlikely to request and give specifications for the writing of "Hamlet," so today's television audience is unlikely to request and specify successful new television programs. The utility exists to the extent that knowledge about the consumer affects the structure and content of the product. The distinction between the utility of knowing principles of audience behavior and the "utility" of uninterpreted programming suggestions by the audience is suggested in the published form of the CBS-sponsored research, The People Look at Television, by Gary Steiner (1963): . . . the results appear quite fruitful and interesting to the social psychologist but perhaps of less direct bearing on how broadcasters should change the way they conduct their business. In the main, the people have had a great deal more to say about themselves than they have about TV, and in so doing, they may have provided the communicators with a clearer picture of the audience(s) they are serving, and how they are serving them. (p. 2h9) Toward a§_Explanation 3f Television Attitudes Though Steiner's work was labeled as an attempt to get at the "why" aspects of the audience's uses of television, the study must be dis- tinguished from explanatory research. There were no hypotheses to be supported or rejected, no theory to be supported or modified. Though he asked the sample in various ways "why" they watched or liked television, no empirical evidence was gathered to explain "why" the observed distri- butions of responses were obtained. In this sense, Steiner's study may S be considered descriptive rather than explanatory. Only after collection and tabulation were the data "explained." As is the general case with fig ES§2.£§EEE explanation, further checks are required for verification. One of Steiner's most marked and consistent findings was the negative relationship between education and attitudes toward television in general: . . . the great mass of the American television audience divides roughly into two major segments according to the social- economic standing of the household, with formal schooling the single factor that makes the most difference. (Steiner, 1963, p. 23h) By "making the most difference," Steiner means that the distribution of attitudes toward television covaried with the education variable, through cross-tabulations, to a greater extent than it did with other personal variables such as age, religion, and income. If one desires to progress beyond description to explanation, he is dissatisfied with adjourning his efforts after ascertaining a reliable education-TV attitude relationship. ‘Without inferential "fill-in," a knowledge of the number of years a person has gone to school has no explanatory power. The variable of educational level is an index of something that should be specifiable within an ex- planatory theoretic schema. What is there about acquiring education that should deflate one's general favorability toward television? This is the kind of question that Steiner must have had in.mind when he suggested that his findings should be "fruitful and interesting to the social psychologist." The question must be grappled with, for example, if one wishes to mpdify the extant relationship between education and TV attitudes. The elements that comprise "education" and the interrelationships of these elements with 6 the attitudes toward television need to be specified.within a theoretic framework before the scientist or the network executive can hope to exert any control over the education-attitude covariation. Using factual bases such as Steiner's as a beginning, the next logical step to a theory (or theories) of mass communication behavior is a series of explanatory studies that begin with a tentative theoretic viewpoint and then gather empirical evidence to support or reject that position. The present study is such an attempt to take a theoretic position, derive testable hypotheses there- from, and gather relevant empirical data in an attempt to explain TV attitudes. The negative relationship between broadcasting attitudes and education was not found for the first time by Steiner, but has been found repeatedly. In l9h5, for example, the National Association of Broadcasters sponsored the first nationwide survey of the public's attitudes toward radio: The survey shows that less educated people are the least critical of radio in all of its aspects. Take for example, the question: 'Do you ever feel like criticizing when you listen to the radio?' Twenty-three percent of the college people, 31 per cent of the high school people, but h9 per cent of the grade school people have no criticisms. And so it goes all through the survey: the lower a person stands in the social pyramid, the less likely he is to express criticism and the more inclined he is to approve of radio as it is now. (Lazarsfeld and Field, 19,46: Pp. 66-67) There is a paucity of empirical television studies attempting to explain the significance of the education or social class variable. Media usage rather than media attitudes was the dependent variable examined in an explanatory study conducted by Samuelson gt El (1963). By statistically partialling out the effects of various role involvements on media usage (weekly hours spent on job-connected activities, weekly hours spent on 7 hobbies, number of organizations in which active, number of dependent children, and time spent on other media), the authors found a weak, positive relationship (.O9h) between education and television ugagg. Before the statistical control, the relationship was somewhat negative (-.llh). The authors' suggestion is that increasing education also in- creases role involvements that preempt time a person would otherwise be inclined to spend watching television; i.e., the very educated man would watch TV a lot if he could, but he can't. Another way of summarizing the data might be this: Relative to their uncommitted time, highly educated people tend to watch more television than do people with less education. In 1956, Geiger and Sokol (1959) conducted an audience survey in the Greater Boston area, contrasting television behavior and attitudes for various social classes, social class being indexed by the monthly rent paid and by self-designations. Middle class respondents reportedly watched less and derived less pleasure from television than did working class respondents. The authors were interested in exploring the possi- bility of social class normative influences on television viewing and attitudes. In their research report, two possible methodologies for evaluating such normative influences were mentioned and then discarded in favor of an indirect measure. The first of these was called the "cultural" method, ‘which meant that the specific television behavior for different social classes would be checked for compatibility or logical consistency with more abstract values held by those social classes. The cultural method was discarded for this reason: 8 . . . the strictly cultural approach, like the assumption that behavior always reflects a cultural norm, while suggestive, is not sufficient to the task. Its more important drawback lies in the fact that a pattern of behavior can serve latent functions and thus eventually become a part of a cultural system, even though, in a formal sense, it does not seem to fit in with it. (p. 176) The second method considered was to demonstrate that the public thinks enthusiastic television viewing is appropriate for the working class. This would precipitate the hypothesis that middle and upper social classes might not view enthusiastically because of their awareness that the working classes dg_view a great deal and do enjoy it very much. The "demonstration of public awareness" method was discarded for this reason: On the other hand, it is also possible that the class- typing of enthusiasm for televisiondwatching does get involve an invidious distinction. Perhaps there is, instead, general admiration for the lower-class ability to relax and be entertained. So this approach does not convince either. (Geiger and Sokol, 1959, p. 177) The indirect method actually used by Geiger and.Sokol will now be described. Respondents who designated themselves as "having more friends than.most people" and as "enjoying life more than most people" were selected from the total sample for this analysis. It was assumed that this sub-sample placed high value on qualities of happiness and sociability. These respondents were divided into self-designated categories of middle and working classes. The respondents in the two classes were further divided into categories of "more enthusiastic" and "less enthusiastic" television viewers, thus making a total of four subgroups. The four sub- groups estimated who enjoyed television more, an assumed reference group (sociable people, happy people, and.people of the respondent's own sex) or the opposite of the assumed reference group (people with few friends, 9 unhappy people, and people of the opposite sex). The dependent variable here was a percentage difference score computed for the middle and working class respondents. The difference score was defined as the percent of enthusiastic TV viewers who selected the assumed reference group minus the percent of non-enthusiastic TV viewers who selected the assumed reference group. The findings supported the prediction that the differences among the middle class viewers in selecting the assumed positive reference group ("more enthusiastic" minus "less enthusiastic") would be greater than the differences among the working class viewers. Reanalysis using education level rather than social class yielded the same results. The inferred psychological process behind the findings was this: TV viewing is taboo among the middle class (or the educated) but not among the working class (or the less educated). A middle class person engaged in taboo behavior can relieve his guilt feelings by distorting his per- ceptions of his positive reference groups into accordance with his own deviant enthusiasm for TV. .A middle class non-enthusiast also reinforces hi§_position by perceiving the same positive reference groups as being in accordance with his own anti-TV predispositions. There is assumedly less motivation for autistic perception among the working class, because TV viewing for them is not taboo or guilt-producing. Therefore the working class respondents yielded a smaller difference between enthusiasts and non-enthusiasts than did the middle class respondents in selecting the positive reference group as enjoying TV more. The middle class viewer, whether a TV enthusiast or not, supposedly receives from his positive reference groups "imagined social support" for his own attitudes toward TV. Thus the existence of a normative component of TV viewing was inferred, 10 based on the assumption of self-imposed psychological sanctions and rewards. The Geiger and Sokol study (1959) is certainly a step beyond the mere cross-tabulation of attitudes toward television with easily obtained personal and demographic variables. The study does not, however, offer suggestions as to BEE TV viewing should be taboo among the middle class or what function the inferred norm might serve. Perhaps the two metho- dologies previously dispensed with as being "insufficient to the task" can be modified and combined into a complementary and.more explanatory research problem. Particularly needed.is more direct evidence that persons are aware of the self-imposed "norm." Attitudes Toward TV as Manifestations 2f More General Values In a different article, but using data from the same 1956 survey, Geiger and.Sokol (1960) report that many relationships which hold with commercial television are reversed with educational television. ETV viewing, for example, is directly rather than inversely related to edu- cational level and social class: 'We might propose, then, that the more success-oriented or ambitious person is more likely to watch ETV than the person who is less ambitious. In the hope that he will benefit in the future, he takes a basically utilitarian view toward television watching, using it as a kind of investment to further his aspirations for personal success and higher social status. (p. 55) Schramm (1960) takes a position similar to the above that is reminiscent of the "cultural.method" dismissed earlier by Geiger and Sokol, stating that the middle class adheres to certain norms and values; e.g., being active, striving, achieving, bettering oneself: ll . . . most commercial television encourages chiefly passivity and.minimum effort rather than activity, a minimum of social inter- action, a concern with fantasy rather than real life, and living in the present rather than concerning oneself either with self- improvement or the problems of tomorrow. Therefore, anyone who holds strongly to the norms described above will not be able to express his values through watching commercial television, and will probably feel a certain amount of unease or guilt if he devotes a great deal of time to it. (p. 26, emphasis supplied) In summarizing 13 ETV audience studies, Schramm (1963) became particularly interested in 3 studies whose findings did not fit the positive education-ETV orientation pattern found in the other 10. Blue collar people were watching ETV, and that was "deviant" behavior, according to the established empirical pattern and according to the theory of class values. It was found, however, that the blue collar ETV fan rated him- self in a higher social class than did the blue collar ETV non-viewer. Similarly, the blue collar ETV fan attended more cultural and civic activities; he was more likely to perceive himself as an opinion leader; he read more. The indices of education and social class had failed to predict. ‘Why? The blue collar deviants are deviants by classification, rather than by behavior. By occupation and education they are classified with a social group that does not especially value educational television. But they yearn, many of them, to be upwardly mobile in society. They not only reject their class; they also reject many of its norms and customs. (p. 87) In other words, Schramm is arguing that although education and occupation may usually predict ETV viewing (or commercial TV, for that matter, since Schramm considers the two to be opposites), the success of the prediction depends on a third correlation: education or occupation with general values. The Specific behavior (TV viewing in this case) is a function of {‘b 12 the general values, not of the educational level or occupation, which merely index these values. The values which Schramm attributed.mainly to the middle class may be called a "value for industriousness." If a person adheres to a value for industriousness, and if he sees TV viewing as being relevant to the expression of that value, then it should be possible to predict that person's attitudes toward television, with no knowledge of his occupation or educational level necessary. If television is relevant and compatible with the value, the attitude should be positive. If relevant and in- compatible, a negative TV attitude should be found. Steiner (1963) also felt compelled to get at the meaning of the negative education-TV attitude correlation found in his study, if only conjecturally and after the fact. At one point Steiner presents the same argument as Schramm: "As a middle-class, striving American, he more acutely feels the need to spend time usefully than his less ambitious counterparts; and his formal schooling has placed a high value on reading and serious study." (p. 233). Earlier, however (p. h8), Steiner presents a rationale for the education-attitude covariation that, at least on the face of it, is different from the value for industriousness: 1. The highly educated impress less easily on gnylscore, television included. 2. The highly educated have more discrimination, more critical standards, more stringent criteria for excellence. 3. The highly educated are not inclined to give blanket praise; they feel more need to demonstrate critical reactions. h. The highly educated have more specialized tastes, tastes that are not satisfied by a mass medium such as television. 13 S. The highly educated have a wider range of interests than others, and have the finances with which to pursue these varied interests. These qualities suggest a "value for critical evaluation." Neither Schramm nor Steiner made direct efforts to tap their respondents' self-perceived adherences to a value for industriousness or for critical evaluation. Also, neither Schramm nor Steiner made efforts to assess the perceived relevance of TV behavior to the more general values as perceived by the respondents. Yet both Schramm's and Steiner's expla- nations assume that the respondent psychologically places TV, as a concrete expression, into a relevant, more abstract value category. The possibilities for conjectural explanation of attitudes on the basis of education are almost without limit, which is one reason why it would be desirable to disect the education variable into the properties which make it up, properties that are more meaningful explainers of behaviors correlated with education. In one study only a limited number of alternative explanations can be tested simultaneously. Two values were selected for the present study. "Number of years of schooling" almost surely indexes mare than the two values suggested by Steiner and Schramm. The two value-adherence variables should be considered as only two components of whatever it is that is indexed by the variable of educational level. The Research Problem The general hypothesis is this: A person's over-all evaluation of television is a function of how TV viewing is perceived as implementing or impeding implementation of more general values. This general hypothesis 1h is an illustration of the "value-expressive" function of attitudes suggested by Daniel Katz (1960): A man.may consider himself to be an enlightened conservative or an internationalist or a liberal, and will hold attitudes which are the appropriate indication of his central values. (p. 173) Value expressive attitudes, according to Katz, serve the function of maintaining self-identity and enhancing the self-image. Rosenberg (1956) investigated a similar hypothesis, but included an extra variable in his conceptualization: the importance of the general value to the person. Thus, attitudes were shown to vary as a multi- plicative function of two variables: (1) the importance of the general value, and (2) the potency of the object of judgment for implementing or impeding implementation of that value. By holding importance level constant, attitudes were shown to be related significantly to the sum of the implementation potency ratings. By holding the implementation potency ratings constant, attitudes were shown to be related significantly to the sum of the importance ratings of the general values. In brief, Rosenberg's method was to (1) index the direction and intensity of attitude toward the object of judgment; (2) obtain categorical ratings of importance attached to a series of general values; and (3) obtain categorical ratings of the object of judgments' potency for implementing or impeding implementation of each value. Step (1) was accomplished.with a S-alternative attitude question. Steps (2) and (3) were accomplished with card-sorting tasks. From a sample of adult males, M. B. Smith (l9h9) obtained self- reports of general values with an open-end question. Among the values most frequently mentioned were "liberty and freedom" (25%) and "economic 15 security" (56%). Those who mentioned "liberty" were reportedly more con- cerned than those who did not mention "liberty" about Russia's part in the world today. No significance figures were reported on this comparison. In selecting the most important from a list of 12 characteristics of Russia, those who mentioned "liberty" selected "the lack of freedom and democracy inside Russia" to a significantly greater extent than did those not mentioning "liberty." These results were what the author expected with the hypothesis that general values influence specific attitudes. The author was unable, however, to find any relationship between.mentioning "economic security" as a value and attitudes toward Russia (or verbalized salient characteristics of Russia). The author suggests that a value for liberty does predict attitudes toward Russia while a value for economic security does not because the mass media have developed a stereotype of Russia, almost defining her in terms of assaults on personal freedoms. The mass media presentation of Russia is not cast so much in economic terms, however, and the respondents assumedly had little basis for making a cognitive connection between a value for economic security and an attitude toward Russia. The Smith study, the Rosenberg study, the Katz category of "value- expressive" attitudes, and the general hypothesis in the present study all require a person to be aware of a connection, a relevance, between the general value and the object of judgment. If the object of judgment is perceived as neither implementing nor impeding implementation of the general value, then no prediction of attitudes should be possible on the basis of general values. Stated differently, if perceived relevance could be experimentally strengthened for one group and not strengthened for a control group, then the predictibility of attitudes on the basis of 16 general values should be greater for the experimental group. Although cast in terms of "group norms" rather than "general values," such an experiment was conducted by Charters and Newcomb (1958). In the experi- mental groups, a religious-belief questionnaire was administered only after the experimenter had stressed the fact that all in the group were Catholics (or Jews or Protestants), and that the questionnaire was religious in nature. In the control groups, neither point was mentioned. .At least for the Catholic subjects, the experimental group conformed to the theoretically orthodox position to a significantly greater extent than did the control subjects. An experimental manipulation of perceived instrumentality was accomplished by DiVesta and Merwin (1960). The authors exposed groups of subjects to persuasive communications that were all favorable toward the teaching profession as a career. In one group, the rationale for the favorable assertion was that teaching facilitated satisfaction of the need for achievement. In another group, the rationale for the favorable assertion was that teaching hindered satisfaction of the need for achievement. The effective manipulation of the perceived instrumentality of a teaching career in satisfying a need for achievement was followed by a significant change in attitudes toward teaching asa career. The evidence above suggests that at least four things are relevant to the general hypothesis: (1) the "content" of the value, (2) the level of adherence to the value, (3) the perceived relevance of television to the general value, and (h) the "direction" of the relevance; i.e., whether television is perceived as implementing or impeding implementation of the more general value. 17 If the general hypothesis is appropriate for the context of TV attitudes, TV should be perceived as relevant to more general values. If the explanations of TV attitudes by Steiner and Schramm are correct, TV should be perceived as impeding implementation of the two general values under study. If these two conditions hold, it then follows that there should be a significant, negative correlation between value adherence and attitudes toward television in general. Taking the two general values implied by Steiner and Schramm, then, the following two hypotheses should hold: H1: The favorability of people's attitudes toward television is negatively related to their degree of adherence to a "value for critical evaluation." H2: The favorability of people's attitudes toward television is negatively related to their degree of adherence to a "value for industriousness." By modifying the "self-perceived value adherence" variable to a "comparison of self relative to the average person," the following two hypotheses should also hold: H3: The favorability of people's attitudes toward television is negatively related to the discrepancy between their adherence to a "value for critical evaluation" and their perception of the "average person's" adherence to this value. Hh: The favorability of people's attitudes toward television is negatively related to the discrepancy between their adherence to a "value for industriousness" and their perception of the "average person's" adherence to this value. 18 It was proposed earlier that some peOple adhere to certain values more strongly than do other people. This level of adherence is assumed to be an integral part of the cognitive structure. Self-consistency pressures at this abstract level of values then generate the various attitudes toward concrete objects of judgment, e.g., television, as manifestations of the more general value. A major assumption in this analysis is that tele- vision viewing is indeed a manifestation of the general value within the individual's cognitive structure; i.e., that the individual does see tele- vision behavior as relevant to his level of value adherence. ‘Without perceived relevance, there is no explanatory value in the theoretic proposal above that relates the abstract value to the attitude toward the attitude object, television in this case. Considering the two general values being examined in the present study, the following hypotheses should hold: H5: People see television as relevant to implementing or impeding implementation of a "value for critical evaluation." H6: People see television as relevant to implementing or impeding implementation of a "value for industriousness." Assuming television and values to be perceived as relevant to one another, the problem still remains concerning the consistency of direction of that relationship, since the proposed theory predicts a general inverse relationship between level of value adherence and favorability toward television. The following hypotheses should hold: H7: People perceive that the more a person adheres to a value for critical evaluation, the more negative his attitude toward television will be. 19 H8: People perceive that the more a person adheres to a value for industriousness, the more negative his attitude toward tele- vision will be. It is possible for the perceived relationships stated in H7 and H8 to hold on the basis of central tendency even if some individuals in the group perceive the opposite relationships. For such "deviant" individuals, a critical link in the general theoretic rationale would be incorrect. They would ngt see television as impeding implementation of the more general values, therefore their level of value adherence should get be negatively correlated with their evaluation of television in general. Therefore, if the hypothesized negative relationship between value ad- herence and TV attitudes is stronger with the deviants removed from the analysis, there is added evidence that the theoretic rationale is correct as a generalization when appropriately applied. If removal of the deviants weakens or has no effect on the relationship, then the theo- retic rationale and/or the methodology must be reexamined. This will be tested in the following two hypotheses: H9: Among "deviants" (1.6-, persons who do not see TV as incompatible with a high value for critical evaluation), the correlation between their own TV attitudes and their own adherence to a value for critical evaluation will be less negative than the same cor- relation among the nonpdeviants. H10: Among "deviants" (i.e., persons who do not see TV as incompatible with high adherence to a value for industriousness), the cor- relation between their own TV attitudes and their own adherence to a value for industriousness will be less negative than the same correlation among the non-deviants. 20 In addition to levels of value adherence, another basis for estimating TV attitudes can be used. The possibility, mentioned earlier, that the survey variable "number of years of schooling" is probably indexing considerably more than the two values presently being examined may have its counterpart in the cognitive structure of the respondents. That is, the values associated with the concept "education" by the respondents may encompass considerably more than a value for critical evaluation and a value for industriousness. If respondents fail to see the relevance of value adherence (as it will be operationalized in this study) to favor- ability toward television, but do see the relevance of "education" (undefined) to favorability toward television, then the entire effort would not have to be abandoned. Given such a situation, the researcher would first become suspicious of his method of operationalizing value adherence. Still failing, he would then suspicion the values themselves as being poorly labeled. He need not, however, discard the general hypothesis that adherence to gppg value falling under education's rubric is precipitating the attitude toward television. Because of this eventuality, the following two hypotheses, which are not directly related to the theoretic rationale, will be tested: H11: People see the relevance of education (undefined) to favorability toward television. H12: People perceive that the more education a person has, the more negative his attitude toward television will be. Hypotheses 7 and 8 predict that people will generally attribute a more negative attitude to high adherents on the general values than they will to low adherents on the general values. In addition to the central 21 tendencies tapped by H7 and H8, it may be fruitful to attempt a more precise analysis of this relationship, particularly if a respondent's estimates of the TV attitudes of others is influenced.by his own level of value adherence or his own level of education. There is an argument that would predict an interaction effect between a respondent's level of value adherence and his perceptions of other people's TV attitudes. The importance attached to clearly distinguishing TV attitudes on the basis of value adherence probably increases as self-commitment or adherence to that value increases. If this is so, then high adherents would see an even greater difference between low and high adherents in general favor- ability toward television than would the low adherents. The high adherents, as indexed, would contain those who aspire upward in terms of sophistication and social class. As Katz (1960) writes, "Value expressive attitudes not only give clarity to the self-image, but also mold that self-image closer to the heart's desire." (p. 173). Upward aspirants would have greater motivation than non-aspirants to mold and clarify their self-image by emphasizing the differences in value expressive attitudes distinctive to their ideal self. If this is so, then the following hypotheses should hold: H13: The difference between the perceived TV attitudes of persons with high and low adherence to a value for critical evaluation will be greater for high adherents to that value than for low adherents. th: The difference between the perceived TV attitudes of persons with high and low adherence to a value for industriousness will be greater for high adherents on that value than for low adherents. 22 H15: The difference between perceived TV attitudes of persons with high and low education groups will be greater for those with much education than for those with little education. Up to this point, television has been considered "in general," or "all things considered." No distinctions have been made between different categories of television programs or between specific programs within a category. It can be assumed, however, that people both favorable and unfavorable toward television "in general" do have hierarchies of programming preferences. Self-perceptions on value adherence relative to a reference point ("the average person") have already been proposed aS'a correlate of TV attitudes. It would be interesting to see if these "self- perceptions relative to a reference point" can be cast usefully in terms of television program types. If each value-adherence or educational level of the television audience does have a notion of the attitudes toward television held by every other value-adherence or educational level of the television audience, it might also be the case that each value- adherence or educational level of the audience has a notion of the prggramminggppeferences held by a minority audience segment. This being the case, each respondent should then be able to compare his own programming preferences with such.a.minority audience segment of, say, doctors and lawyers. ‘An index of similarity between the two hierarchies of programming preferences could be used as a measure of identification with the upper-middle class. Given a public awareness that the upper- middle class is least favorable toward television in general, which is indirectly tested in the "relevance" hypotheses, the similarity index should also be related to personal attitudes toward television in general. In other words, the more similar a respondent sees his programming 23 preferences to the preferences of doctors and lawyers, the less he should like television in general, assuming the previous rationale to be sound. The following hypothesis, then, should hold: H16: The favorability of people's general attitudes toward tele- vision is negatively related to their degree of identification with the upperemiddle class TV programming preferences. The 16 hypotheses just described constitute the primary research problem structured for the present study. In addition, a secondary investi- gation is proposed that is related to a factor analytic study by MacLean and Talbott (1963). It was felt that this secondary investigation was justified for two reasons: (1) The MacLean-Talbott study suggested possible new avenues for approaching the primary research problem; and (2) the necessary data for the secondary investigation could be obtained quite easily. MacLean and Talbott (1963) conducted a study designed to isolate some factors that might explain the apparent discrepancy between stated prefer- ences in television.programming and actual viewing behavior. In this study, a series of statements, "representing a variety of orientations and beliefs about television and other leisure time activities," was Q sorted (see Stephenson, 1953) by 18 adults and 8 children and subsequently factor analyzed. The three basic types of orientations to TV that emerged as factors are reminiscent of (a) a high adherent to a value for critical evaluation and to a value for industriousness, (b) a low adherent to a value for critical evaluation, and (c) a low adherent to a value for industriousness. Factor type A was "basically unfavorably predisposed toward Spending much 2h leisure time watching TV"; he preferred social activities and visiting. TV appealed to him less than magazines, radio, and recorded music. Factor type B was characterized as a "loyal TV fan." He enjoyed TV a lot, spent a lot of time viewing, and had no guilt feelings about it. Factor type C watched TV quite a bit, but felt guilty about it; he used TV as an escape mechanism. It may aid in keeping these three factor types straight if.an abbreviated label is added to each. Therefore Factor A will be called a "sophisticated consumer." Factor B will be called a "loyal TV fan." Factor C will be called an "escape viewer." By assigning each respondent in the present study to one of these three factor types (or to a "mixed" type), the following hypotheses should hold, based on MacLean's and Talbott's factor descriptions: H17: Factor type A persons ("sophisticated consumers") are less favorable toward television than are persons in factor types B ("loyal TV fans") or C ("escape viewers"). H18: Factor type A persons ("sophisticated consumers") adhere more strongly to a value for critical evaluation than do persons in factor type B ("loyal TV fans") or C ("escape viewers"). H19: Factor type A persons ("sophisticated consumers") adhere more strongly to a value for industriousness than do persons in factor type B ("loyal TV fans") or C ("escape viewers"). CHAPTER II DESIGN Overview pf the Design The general hypothesis and the 16 subsequent hypotheses that consti- tute the primary research problem indicate the nature of the variables to be specified in the design: attitudes toward television in general, self- perceived value adherence, perceived relevance between general values and television, and identification with the upperemiddle class. Attitudes toward television in general were indexed in two ways: with a projective instrument and with an adjective rating instrument similar to the semantic differential. Value adherence was tapped by obtaining the respondents' degree of agreement with a series of verbal statements. These statements were designed to represent very strong or weak adherence to a value for critical evaluation or to a value for industriousness. Perceived relevance between the general values and television was indexed by describing hypothetical types of people at different levels of value adherence; respondents then estimated how each hypothetical type would evaluate television in general. Relevance was defined as statistically significant differences in mean attitude estimates for these hypothetical types of persons. The "identification" instrument consisted of a comparison of two hierarchies of TV program-type preferences: the respondents' own prefer- ences and the respondents' estimates of preferences appropriate to the upper-middle class. 25 26 The secondary investigation described at the end of the first chapter relates to earlier factor analytic work designed to isolate typologies of people with distinctive orientations to television. Respondents in the present study were assigned to the earlier factor types by a method of Q block indexing. These instruments were integrated into a questionnaire. Lack of sponsorship for the present study necessitated the utilization of volunteer intact groups to serve as respondents. The questionnaire was given to these groups on a "group administered" basis. Recognizing the non-random character of intact groups, there were two elements at the researcher's control used in an attempt to obtain as broad a cross-section of people as possible: (1) a variety of groups, rather than one or two, were used so a bias in one group could be cancelled out by a bias in another group; and (2) representation along the entire range of education was actively sought, since education was known to be related to the variables under study. The completed questionnaires yielded quantitative data on the The data were then submitted to appropriate variables being investigated. statistical tests of significance. TV Attitude Measurement Bernard Berelson, in a foreword to The People Look _a_t_ Television, describes one projective device for measuring TV attitudes used by Steiner (1963) as an "ingenious innovation." Respondents were asked to match positive and negative connotative captions to the single most The captions were to represent appropriate of seven line drawings. The "thoughts" that the figures in the drawings might be thinking. drawings included situations of TV watching, theater attendance, playing 27 golf, reading, playing with a child, visiting with "friends," and sitting at a bar. Steiner has this to say about his instrument: The 'thoughts' were designed to incorporate both the favorable feelings of interest, involvement, and relaxation, and some negative counterparts--boredom and shame or guilt about laziness. Obviously the device does not provide an exhaustive measure of the range of feelings associated with the various activities; it merely attempts to quantify the specific attitudes built into the statements. (p. 56) For each of 17 captions, Steiner ran percentage breakdowns, indicating the prOportion of his sample that selected each drawing. Ten of the captions had negative connotations. At this rather gross level of analysis, several apparent inconsistencies of response were attached to television: "The depicted viewer feels, first of all, lazy but (or 22d) relaxed; interested, but he really ought to be elsewhere; fascinated and bored; it is 'real pleasure' but also 'another evening shot'; and so on." (p. 56). Steiner presents evidence suggesting that only part of the apparent ambivalence is due to "among people" rather than "within people" variability. The attraction and novelty of the projective instrument, combined with the limitations of item-by-item description, engendered the decision to use the instrument in the present study, but to modify the analytic procedures. Fourteen of Steiner's captions were used, with equal numbers of positive and negative connotations.l A scoring method was designed to yield a single evaluative index for each respondent. Consider first the 1 For Steiner's 17 items, see The People Look at Television, pp. 5h-55. For the items used in the present study, see Appendix C. 28 "positive" captions: If the television drawing was selected as being most appropriate, a value of 2 was entered for the respondent on that caption. Given a choice of seven alternatives, plus an eighth "none" alternative, it is assumed that the respondent selecting the television drawing over all the rest has a more favorable TV attitude than the respondent selecting another drawing. The choice of anything other than television was thus indexed as a value of 1. Consider now the negative captions: With logic similar to the above, a choice of television was indexed as a zero; anything other than television was indexed as a 1. With the seven negative items, a respondent's sub-total score could range from 0 to 7. With the seven positive items, the sub-total score could range from 7 to 1h. Combining the 1h items into a single instrument, the total score could range from 7 (extremely unfavorable) to 21 (extremely favorable).2 The scores on the 1h captions were intercorrelated to see if the use of a summary index could be justified on the basis of unidimensionality. 0f the 91 obtained correlations, 16 were slightly negative. By removing one caption from the analysis ("This really does you good"), 8 of the 16 negative correlations were eliminated. The revised matrix then contained 78 correlations, 8 of which were slightly negative. Intercorrelation values for the retained 13 items ranged from -.10 to +.27, with a median correlation value of +.08. 2The maximum range was reduced slightly after the caption "This really does you good" was dropped from the analysis. In correlating this item with the other 13, 8 correlations were negative and 5 were insignificantly positive, so there was no justification for its inclusion. 29 In Steiner's study and in the present study, this instrument was administered early in the schedule to eliminate or minimize sensitization toward television. Females were issued a sheet of drawings in which the central figures were women, in order to aid identification and projection. The males saw men in their drawings, but otherwise the two sets of drawings were identical. Steiner also administered an attitude instrument similar in format to the semantic differential, developed by Osgood 32,2; (1957). A typical semantic differential scale is bounded by adjectivial opposites, separated by seven steps, representing a "neutral," "don't know," and "doesn't apply" position in the middle, plus three degrees of intensity of applicability toward each polar adjective. By rating a concept like "Television" on a large number of these scales, a quantitative, connotative profile of the concept may be obtained. In most of the research cited in The Measurement pf Meaning (Osgood, 1957), factor analyses of semantic differential data have isolated three factors of connotative meaning. These three factors have been named (l) evaluative, (2) potency, and (3) activity. Relating the semantic differential to the concept of attitude, Osgood writes: It seems reasonable to identify attitude, as it is ordinarily conceived in both lay and scientific language, with the evaluative dimension of the total semantic space, as this is isolated in the factorization of meaningful judgments. (p. 190) Steiner (1963, p. h5) used 17 six-point scales, several of which (e.g., "too simple-minded, too high-brow"; "stays the same, keeps changing") have no clear evaluative interpretation. As with the projective instrument, Steiner's scales were reported item-by-item on a percentage breakdown basis. Again, a summary evaluative index would be useful. 30 In spite of the methodological limitations of Steiner's instrument, the scales all seem quite relevant to the concept of television in general. Some of Osgood's scales that load heavily on the evaluative factor (e.g., "grateful-ungrateful") do not. The problem of integrating known evalu- ative scales with scales of unknown, but apparent evaluative relevance to a particular concept is discussed by Tannenbaum and McLeod (1963) as they report their procedure in deciding which scales to use for newspaper image research: This raised the question of which SD scales would be most appropriate to use, leading, in turn to examining the connotative dimensions involved in the judgment of a newspaper institution. Then, as now, the procedure to meet such a problem was to construct an SD form including scales previously identified as representing the major dimensions of connotative judgment as well as new scales judged on E. priori ground to be representative of the class of concepts being judged. After initial selection, data were to be obtained and subjected to factor analytic treatment to determine a more parsimonious set of scales. (p. 52) With the exception of the factor analytic pretest, this was the pro- cedure used in determining the eight scales for the present study. The basic "good—bad" scale of the evaluative dimension was modified for the sake of specificity to read "doing a good job-doing a bad job." A second variant of the "good-bad" scale, also used by Steiner, was "in good taste- in bad taste." Two other scales used ("interesting-boring" and "important- unimportant") were common to Steiner's study and to the evaluative factor of media judgments as reported by Tannenbaum and McLeod (1963, p. 56). The remaining )4 scales ("informative-noninfomative," "imaginative- unimaginative," "upsetting—relaxing," and "exciting-dull") were selected from the Steiner scales on the basis of face validity in containing evaluative content for television. 31 No pretest effort was made to refine a special set of scales because of Osgood's earlier work. As with the projective instrument, however, unidimensionality of items was checked via an intercorrelation.matrix before summing across scales for a single summary index. For the 28 independent correlations in the matrix, correlations ranged from -.02 to +.52, with a median correlation of +.25. All 8 scales were retained for the summary attitude score. By coding each scale from 1 to 7, the possible range of attitude scores for each respondent was from 8 to 56. Value Adherence Measurement The problem of indexing each respondent's level of adherence to a value for industriousness and for critical evaluation required the construction of a new instrument. Due to the length of the questionnaire, it was decided that approximately eight scales would be optimum for indexing value adherence, four scales for each value. In preparing to construct the instrument, this basic question was contemplated: ‘What kind of general opinion statements would a person adhering strongly to a value for industriousness or for critical evalu- ation agree with and disagree with? On this basis, a large pool of items was generated. From this pool, hh items were culled out for pretesting in June, l96h. In the pretest, as in the final instrument, respondents described themselves on each item in terms of a lO-step agree-disagree ladder scale. Thus, for any value item, the level of self-perceived adherence was defined as the level of agreement indicated on the lO-step ladder scale. 32 To control for response set, the 22 pretest items representative of each value were subdivided equally into positive and negative items. The pretest schedule was composed as follows: 1. Eleven items with which a person who adheres strongly to a value for industriousness should agree completely; 2. Eleven items with which a person who adheres strongly to a value for industriousness should disagree completely; 3. Eleven items with which a person who adheres strongly to a value for critical evaluation should agree completely; 1:. Eleven items with which a person who adheres strongly to a value for critical evaluation should disagree completely. The pretest questionnaire (see Appendix A) was administered to three adult church groups in Lansing and East Lansing, Michigan. About 50 usable questionnaires were completed. An early check of the educational and occupational distributions in the pretest sample indicated a bias in which high education and occupational-prestige persons were over- represented. It was therefore necessary to discard about 20 questionnaires to get a more even distribution on education and occupation. With data from the reduced pretest sample of 30, all items were intercorrelated. The goal was to find a matrix of four highly inter- correlated items for each value. Of the four items in each matrix, two should be positive and two should be negative statements of the value. The six correlations in the best pretest matrix for a value for industri- ousness ranged from +.OS to +.58, with a median correlation of .33. The six independent correlations in the best pretest matrix for a value for critical evaluation ranged from 4-.26 to 4.52, with a median value of +.32. On the basis of the pretest, these are the items constituting the instrument for measuring value adherence: 33 1. Value for industriousness: a. b. C. d. (positive item) "You can't stand still in this life; you either move up or you move down." (positive item) "You have to keep learning to stay ahead in today's world." (negative item) “I take competition rather lightly." (negative item) "I can 'goof-off' with no guilt feelings whatsoever.“ 2. Value for critical evaluation: a. b. C. d. (positive item) "I am hard to please with anything because I have high standards of excellence." (positive item) "One of the main functions of an intellectual is to serve as critic for the rest of society." (negative item) "You've got to give in now and then to get along in the world." (negative item) "I don't need much information before I can make a decision on something." By reverse coding the two negative items, the strongest possible adherent would receive a coded score of 9 for each item (each ladder scale was coded from 0 to 9). For the summary index of value adherence, then, the maximum range for each value would be 0 to 36. It may be questioned whether self-placement alone on the ladder scale is the most appropriate information that each subject could yield with the instrument. In a typical percentage score in a classroom examination, for example, the score gives the student his achievement relative to the maximum achievement possible, but he usually wants to know, in addition, how he stands relative to the rest of the class, how far above or below the mean he is. Similarly, in the present study, the value index yields data on self-perceived value adherence relative to the maximum adherence possible with the instrument, but does not yield self-perceptions on how 31L far above or below the mean or the "average person“ this adherence level is. With one simple addition to the instrument, this added information may be obtained. In the final study, not in the pretest, each respondent was instructed to also mark where he thought the "average person" fell on each of the eight ladder scales. A second index of value adherence was thus obtained in the form of a difference score ("self" minus "average person"). In order to eliminate negative values, the constant of 10 was added to each difference score. For each summary index of relative value adherence, then, the maximum range for each value would be from h (0 minus 9, plus 10, taken four times) to 76 (9 plus 10, taken four times). Relevance Measurement Perceived relevance is a critical variable in the present study, and must be measured rather than assumed. As with any cognitive variable, perceived relevance must be inferred from some overt behavior. The oper- ational method used in the present study was to present respondents with a variety of hypothetical types of people, requesting them to estimate the favorability toward television of each type on a lO-step ladder scale. Each type of person described was illustrative of high, medium, or low adherence to a value for industriousness or to a value for critical evaluation. Each of the eight items used in the value adherence instrument became the basis for the description of three hypothetical types (a high, medium, and low adherent) in this instrument designed to measure perceived relevance. For example, one item in the value adherence instrument was "I am hard to please with anything because I have high standards of excellence." For 35 the instrument designed to tap perceived relevance, this item became the basis for the three following items: 1. (representative of high adherence) "One who can't really enjoy something unless it's almost perfect." 2. (representative of medium adherence) "One who likes good quality but does not demand perfection." 3. (representative of low adherence) "One who isn't hard to please with anything." Four "critical evaluation" items and four "industriousness" items thus became the basis for 2h "perceived relevance" items. For the sample as a whole, then, summary estimates of favorability toward television would be yielded on the following six categories of people: high, medium, and low adherence to (l) a value for critical evaluation and (2) a value for industriousness. In addition, respondents were requested to estimate the TV favorability of hypothetical types of persons with "above average," "average," and "below average" education levels. If the two general values (or educational level) are perceived as irrelevant to television, one would expect the TV favorability estimates representative of different adherence levels to the values to be the same, within the limits of random variability. Therefore, perceived relevance is defined here as statistically significant differences between adherence levels in estimated favorability toward television in general. The favorability estimates were made for three levels of value adherence (and education) when the hypothesized inverse relationship might have been supported with estimates at only two levels. The middle level was included to check the precision and linearity with which the perceived relevance is cognized. That is, do the respondents only see differences between extreme polarizations or do they perceive a continuous, linear 36 relationship? At a dichotomy level, respondents might have been asked, "Who likes TV more, high adherents to X or low adherents to X?" Had the respondents been presented with three levels simultaneously, however, this would no doubt have encouraged a linear estimate of TV favorability, whether the respondents had previously considered the relationship as linear or not. Therefore, a mechanical control was exerted to minimize the effect of one estimate on another related estimate. The relevance measurement portion of the questionnaire was printed in a small, separate booklet, containing only one estimate per page. The three categories (critical evaluation, industriousness, and education) and the three levels within categories (high, medium, and low) were well intermixed throughout the 27-page booklet.3 Identification Measurement In addition to general values, it has been proposed that identifi- cation with the upper-middle class also mediates attitudes toward tele- vision in general, and that it should be possible to state this identifi- cation in terms of television programming preferences. A new instrument is thus required to yield a "similarity index" between a respondent's own preferences and his conception of upperdmiddle class preferences. For this study, "upper-middle class" was typified occupationally as "doctors and lawyers." To yield such a similarity index, an instrument was devised in which respondents sorted 16 TV program types into seven categories ranging from 3Both values were indexed by four items each. Each item was described at three levels. ‘With one description per page, this accounts for 2h pages. The other three pages were the three levels of education. 37 "like the best" to "like the least." A forced unimodal frequency distri- bution was imposed, which meant that the first and seventh categories had one entry, the second and sixth categories had two entries, the third and fifth categories had three entries, and the fourth category had four entries. The respondents first sorted the 16 program types according to their personal preferences; then they sorted again according to their estimates of the preferences of doctors and lawyers in general. With each program type thus assigned two preference scores between 1 and 7, the absolute difference was calculated. The smaller the absolute dif- ference, the greater the perceived similarity. The similarity index was defined as the sum of the 16 absolute differences for each respondent, and has a maximum range of O to hO. The selection of 16 program.types meaningful to laymen and also having a fair degree of exclusivity and scope was no small task. At least two efforts have been made recently to overcome the gross imprecision in such categories as "westerns" and "soap operas" (MacLean and Talbott, 1963; Hazard, Moriaty and Timmons, l96h). In both of these papers, a series of ordinal dimensions was prOposed, along which several facets of a program could.be described, thus escaping the "all or none" character of the single, nominal category. Laymen, however, still use the imprecise category system, and.meaningfulness to the respondents had to take priority over linguistic precision in this study. The set of 16 categories finally used (see Appendix C) was a combination of types suggested.by Thayer (1963) with those used by Steiner (1963) in the coding of his data. 38 ‘Q Block Typologies of Orientations to Television The secondary investigation in the present study requires that the respondents be assigned to previously established factor typologies. Talbott (1963) describes an economical method (called "Q Block indexing") of assigning respondents to factor types, given "detailed knowledge of a stable Q typology factor structure." Here, in general terms, is the procedure: The pattern of responses (level of agreement with the TV orientation items, in the present case) that a pure factor type would make is estimated by weighting the patterns of responses given by those persons most highly associated with that factor. For each item, the weighting value is the person's loading on that factor. The higher the loading, the greater the factor representativeness, thus the greater consideration of that person's pattern of responses as being what the pure factor type would have selected. The weighted response patterns for each factor are summed and converted into z-scores. (In the present case, there would be three hierarchies of TV orientation items, each being representative of the response pattern for a factor type. Each item in each hierarchy would have a z-score value.) There are as many statements in a Q block as there are factors, three in this case. The statements for each block are selected from the z-score arrays (three hierarchies of TV orientation items). The criteria for the first statement are that it be strongly accepted by Factor.A and rejected (acceptance at least 1 standard deviation less) by Factors B and C. The second statement should be accepted by Factor B but be rejected bynA and C. The third statement should be accepted by Factor C but be rejected by A and B. This particular sequencing is, of course, arbitrary. The number 39 of Q blocks that can be abstracted is dependent on how long the criteria can be met for one standard deviation of discrepancy (both vertical and horizontal within a Q block) between accepted and rejected items. In this case, three Q blocks were assembled. The respondent reacts to the cluster of items in each block, rank ordering the three (or however many) items in terms of his agreement with them. The least agreeddwith statement is coded as a l, the most agreed- with as a 3. The rank codes for each factor type are then summed across all Q blocks. The factor acquiring the highest sum of ranks is the factor to which that respondent is assigned. when the high sum of ranks score is a tie, a mixed factor assignment is made. Such.was the method followed in the present study. For the Q block items used, see Appendix C. Data Collection The questionnaire, which.may be seen in Appendix C, was divided into seven sections. The first section was the projective instrument to measure indirectly attitudes toward television. Two loose sheets with line drawings of the seven leisure activities were paper-clipped to the questionnaire for easy removal by the respondents. One of these sheets featured a female in the drawings, the other sheet had a man as the central character. The second section contained the semantic differential type of attitude instrument. A vertical "ladder" scale was used rather than the typical 7-point horizontal scale, thus allowing the degree of intensity to be labeled at each point of each scale. Respondents did not use the same middle scale position to indicate the three different responses of "neutral or 50-50," "don't know," and "doesn't apply." Instead, the ladder scale ho contained only the "50-50" alternative. Special blanks were provided at the right of the ladder scale for "don't know" and "doesn't apply." For this study, however, the coding was done with the typical procedure: all three responses were coded as a h. The "positive" adjective of the word pair was alternately placed at the top and the bottom to cancel out acquiescent response-set. The ladder steps were numbered from 1 to 7, which served as direct preceding when the "positive" and was up, but required reverse coding when the "positive" and was down. Section Three contained three Q blocks on a single page. Examining a single Q block of three statements each, the respondent was required only to pick the statement he agreed with most and the statement he agreed.with least. ‘With this information, the rank order of preferences for the three Q blocks was later entered in a specially prepared coding form on which the factor assignment was calculated by hand. Section Four, the 27-page small booklet containing the relevance- measuring instrument, described earlier, was also paper-clipped to the regular questionnaire for easy removal. The TV attitudes of hypothetical others were estimated on lO-step ladder scales, numbered from O to 9, which served as precoding. The ends of the ladder scale were anchored with "extremely favorable" and "extremely unfavorable," with two additional anchors ("slightly favorable" and "slightly unfavorable") at the fourth and seventh ladder steps. As a matter of curiosity, each page of the small booklet also contained a S-point confidence rating ("just guessing" to "I'm.almost positive") for the TV attitude estimate made on that page. There are no hypotheses in the present study relevant to these confidence ratings. hl Section Five contained the value adherence instrument, which again utilized lO-step ladder scales numbered 0 to 9. These ladder scales were anchored at the ends ("agree completely" and "disagree completely") and at ladder steps 6 and 3 ("agree somewhat" and "disagree somewhat"). The "rungs" of the ladder were made wide enough so the respondent could make two written entries on a single ladder step: "ME" for his self-assessment, and "A.P." for his assessment of the average person. The "agree completely" and of the ladder was always on top because the value statements them- selves alternated from positive to negative versions. Section Six, the identification.measure, consisted of two pages, similar in format. On the left side of each page, the structure of the forced.unimodal distribution was indicated by a symmetrically stacked series of 16 empty "boxes." The top of the structure was labeled "like the best"; the bottom was labeled "like the least." On the right side of the page, the 16 TV program types were listed and numbered 1 to 16. The respondents' task, then, was to place the 16 numbers (representing program types) into the 16 empty boxes. The task was completed twice: first for personal preferences and second for the estimated preferences of doctors and lawyers in general. Section.Seven, the final portion.of the questionnaire, consisted of questions on demographic and personal data that are useful in describing the sample. These data included.media consumption habits; self-perceived selectivity in TV viewing, evidence of active criticism of TV, age, sex, income, education, occupation, and specification of two TV programs, the most and least liked. The length of the questionnaire was.lS pages, plus the 27-page booklet, plus:the.two loose pages of line drawings. Three of these pages hz contained only practice items designed to familiarize respondents with the mechanics of marking the scales. The questionnaire was designed to be group administered. Printed instructions were kept to a minimum because a written description of pro- cedures adequate for self-administration would have appeared too fore~ boding and lengthy. The questionnaire administrator, therefore, read aloud a specially prepared set of instructions (see Appendix B) prior to each section of the questionnaire, with the exception of Sections Three and Seven which were self-explanatory. This method had the advantage of allowing the respondents to ask procedural questions before continuing with a new section. It also had the advantage of allowing the administrator, when the group consisted of about 12 or less, to "keep an eye" on the marking behavior of the entire group; several instances of misunderstood instructions were detected in this manner. The method did have the dis- advantage, however, of detaining the speedier respondents until the slower ones had caught up. This disadvantage was explained in advance to all groups in order to keep frustration at a minimum. In some instances, it was possible to give individual attention to the respondents, allowing them to finish at their own speed. When large groups were assembled, up to four administrators were sometimes used, each being responsible for a subgroup. Twelve groups, mostly occupational and religious groups, took the questionnaire during June and July, l96h. Three churchprelated groups were located in Lansing, Michigan; one in Paw Paw, Michigan; one in Dover, Illinois; one in Enid, Oklahoma; and one in New Orleans, Louisiana. Two occupational groups consisted of persons at a school custodian's conference and a petroleum marketing engineer's conference, both held at Michigan ,1 h3 State University in East Lansing, Michigan. A group of h-H mothers took the questionnaire at Michigan State University. Two classes of students cooperated at Andrews College in Berrien.Springs, Michigan. Four questionnaires were completed by disinterested acquaintances. From this conglomerate assortment, about 2hO questionnaires were turned in, 215 of which were usable. In Table 1, this sample of 215 is described in terms of age, sex, income, and education, and is compared with United States census data for 1960. Examination of Table 1 indicates that the sample is under-represented h on low education, old age, and low income. The sample is over-represented on high education. The interpretive qualifications necessitated by these discrepancies from known parameters will be discussed in Chapter IV. Analysis Procedures Hypotheses l, 2, 3, and h predict negative relationships between TV attitudes and value adherence. The two attitude measures and the value adherence instrument yield quantitative data that may be tested for signifi- cance with a Pearson productdmoment correlation. Hypothesis l6 predicts a negative relationship between TV attitudes and degree of identification (perceived similarity) with the upperdmiddle class. A product-moment correlation is again appropriate. Since perceived similarity increases hStatistically significant (p less than .001) chi-squares were computed on the variables of education, age, and income, using the census data for expected values. ht TABLE 1 SAMPLE DESCRIPTION AND COMPARISON WITH U .S. POPULATION CEL‘JSUS DATA (N = 215) _ VI 1960 U.S. Population per cent per cent Education * 0-8 years % 37% 9-11 years 11 19 High School diploma 30 26 1-3 years college 12 10 College degree 12 5 Beyond college degree 29 3 100% 100% Age 15-19 years 13 -- ** 20-29 23 20% 30-39 26 22 hO-h9 2h 20 50-59 17 16 60 and over 9 22 100% 100% Family Income $0-$3,ooo % 217*** mew-$6,000 35 33 $6,001-s10,ooo h3 31 Over $15,000 3 h 100% 100% Sex Male hey: 1:3ng Female 52 S2 100% 100% %%%%P *Percentages based on adults 20 years old or more. All census figures in Table l are derived from data listed in the following source: U. S. Bureau of the Census. Vol. I, Characteristics of the Population. Summary (washington, D. C.: Education data are on p. U .8. Census of Population: 1960 Part I, United States U.S. Government Printing Office, 196h). 1“h0ho ercentages based on adults 20 years old or more, p. l-lh6. '*** Percentages based on 1959 income of U.S. families, p. 1-226. It should be pointed out that the census categories begin rather than and on even thousands of dollars. ercentages based on adults 20 years old or more, p. 1-580. hS as the similarity index score decreases, it should be pointed out that a positive correlation between attitude scores and similarity index scores is being predicted in H16. Hypotheses S and 6 predict that people will see television as relevant to implementing or impeding implementation of a value for critical evalu- ation or a value for industriousness. Similarly, H11 predicts that people will perceive educational level as relevant to TV attitudes. If the mean estimated TV attitudes for hypothetical types at different levels of value adherence or education are significantly different from each other, this will be interpreted as perceived relevance. No linearity of effect is implied in this analysis. The most powerful statistical model.for this problem is a treatment by subjects analysis of variance (see Lindquist, 1953, p. 156)- Hypotheses 7 and 8 predict that people will associate increasing adherence to values for critical evaluation or industriousness with decreasing favorability toward television in general. Similarly, hypothesis 12 predicts that people will associate increasing levels of education.with decreasing favorability toward television in general. No significance testing will be done for these three hypotheses. The treat- ments by subjects analysis of variance called for in H5, H6, and H11 will yield the mean estimated TV attitudes for hypothetical.types at different levels of value adherence or education. Hypotheses.7, 8, and 12 will be considered confirmed if the means are ordered in this way: Mean esti- mated TV attitudes for low levels of value adherence (or education):> medium levels of value adherence (or education):> high levels of value adherence (or education). 146 Hypotheses 13, lb, and 15 predict that the difference between esti- mated TV attitudes appropriate to hypothetical types with high and low levels of value adherence (or education) will be greater for people who themselves have a high level of value adherence (or education) than for people who have a low level of value adherence (or education). In analysis of variance terms, this predicted "difference of differences" is an interaction effect. This predicted interaction is not incompatible with the predicted over-all directionality of mean attitude estimates described above in H7, H8, and H12; it merely predicts that the EEEES of mean attitude estimates should be greater for persons with high levels of value adherence (or education) than for persons with low levels of value adherence (or education). The additional control of the respondents' own level of value adherence (or education) may be inserted into the treat- ments by subjects analysis of variance called for in H5, H6, and H11, described above. This results in a treatment by subjects by levels analysis of variance (see Winer, 1962, pp. 306-308). Three of these complex analyses of variance can check nine hypotheses simultaneously: H5, H6, and H11 (mean over-all estimated attitudes will be different); H7, H8, and H12 (mean over-all estimated attitudes will be ordered in a linear fashion); and H13, th, and H15 (attitude estimates will interact with respondents' own level of value adherence or education). Hypotheses 9 and 10 predict that the negative correlation between respondents' value adherence and TV attitudes will be stronger for those who associate low value adherence with favorable TV attitudes ("non- deviants") than for those who associate high value adherence with favorable TV attitudes ("deviants"). A deviant is defined as a respondent who estimates that hypothetical types with high adherence to the general h? values are more favorable toward television in general than are hypo- thetical types with low adherence to the general values. All others are defined as non-deviants. By dividing the total sample into deviant and non-deviant groups, eight sets of two independent correlations between value adherence and TV attitudes may be computed: both values on both adherence measures for both attitude measures. By submitting the sets of two independent correlations to Fisher Z-transformations, it is possible to test whether or not a "deviant" correlation is significantly different from its corresponding "non-deviant" correlation (see walker and Lev, 1953, pp. 25h-256). The attitude-value adherence correlations for the sub- groups will also be submitted to the standard product-moment correlation test of significance to see whether or not the subgroup correlations are significantly different from zero. Hypotheses 17, 18, and 19 predict that favorability toward tele- vision and value adherence will differ significantly among the three factor types to which respondents are to be assigned on the basis of Q block indexing. Specifically, it is hypothesized that Factor A respondents ("sophisticated consumers") will (H17) be less favorable toward TV than will Factor B respondents ("loyal TV fans") or Factor C respondents ("escape viewers"); (H18) adhere more strongly to a value for critical evaluation than Factors B or C; and (H19) adhere more strongly to a value for industriousness than Factors B or C. The appropriate statistical model for H17, H18, and H19 is a simple randomized analysis of variance, followed by individual t-tests for selected contrasts. CHAPTER III RESULTS Item Analyses and Frequency Distributions In Chapter II, the pretesting procedure was described whereby eight items were originally obtained for the value adherence instrument: four "critical evaluation" items and four "industriousness" items. These eight items were again submitted to an item analysis with the sample of 215 in the main study. In the pretest, the six "critical evaluation" intercorrelations ranged from +.26 to +.52, with a median correlation of +.325. ‘With the main sample, this degree of inter-item correlation collapsed: the six intercorrelations ranged from -.25 to +.22, with a median correlation of +.03. By eliminating the items "You've got to give in now and then to get along in the world" and "I don't need much information before I can.make a decision on something" from the analysis, the one remaining correlation was +.22 for the "self only" measure of adherence to a value for critical evaluation and +.15 for the "self-other discrepancy" measure. The two items finally used to index adherence to a value for critical evaluation are as follows: 1. (positive item) "I am hard to please with anything because I have high standards of excellence." 2. (positive item) "One of the main functions of an intellectual is to serve as critic for the rest of society." h8 h9 In the pretest, the six "industriousness" intercorrelations ranged from +.05 to +.58, with a median correlation of +.33S. With the main sample, this degree of inter-item correlation also collapsed: the six intercorrelations ranged from -.10 to +.25, with a median correlation of -.O35. By eliminating the items "I take competition rather lightly" and "I can 'goof-off' with no guilt feelings whatsoever" from the analysis, the one remaining intercorrelation had a value of +.25 for the "self only" measure of adherence to a value for industriousness. For the "self- other discrepancy" measure of adherence to a value for industriousness, the one remaining intercorrelation had a value of 4.36. The two items finally used to index adherence to a value for industriousness are as follows: 1. (positive item) "You can't stand still in this life; you either move up or you move down." 2. (positive item) "You have to keep learning to stay ahead in today's world." It can be seen that the elimination of the negative items from both value adherence measures removed the control for response set that had originally been designed into the instruments. Frequency distributions for the main variables are presented in Tables 2, 3, and h. The deletion of the items that intercorrelated on the pretest but not in the main study decreased the maximum range on the value adherence measures, shown in Table 2. There are indications that the measuring instruments for value ad- herence failed in their function. On the basis of the inter-item cor- relations indicating a certain lack of homogeneity, the negative items had to be dropped. This left only two positive items, which would make the measuring instrument susceptible to response set biases. Note, for TABLE 2 FREQUENCY DISTRIBUTIONS ON FOUR VALUE ADHEREHCE NEASURd3* Value for Critical Evaluation Value for Industriousness Self-Other Self-Other Self Only Discrepancy Self Only Discrepancy (Max. Range (Max. Range (Max. Range (Max. Range 8 00-18) = 02-38) 8 00-18) - 02-38) Score Freq. Score Freq. Score Fr 9. Score Freq. 00-01 17 8-9 R 0b 1 07-08 2 02-03 26 10-11 3 05 1 09-10 h 0h-05 1h 12-13 8 06 3 11-12 0 06-07 59 111-15 18 07 1 13-1t 2 08-09 39 16-17 hl 08 0 15-16 2 10-11 19 18-19 hO 09 7 17-18 1h 12-13 23 20-21 h5 10 8 19-20 35 1h-15 12 22-23 35 11 0 21-22 35 16-17 1; 211-25 10 12 13 23-211 116 18 __2_ 26-27 S 13 8 25-26 Sh 215 28-29 3 1h 13 27-28 10 30-31 2 15 29 29-30 9 32-33 .__1 l6 12 31-32 __g. 21 17 18 21 s 18 .121 s 215 * The higher the score, the greater the self-reported adherence to the value. [J 2 1 TABLE 3 FREQUENCY DISTRIBUTIONS ON Two TV ATTITUDE unasunss* Projective Measure (Max. Range = 06 to 19) Adjective Ratings (Max. Range 8 08 to 56) §22£§ 3323- §22£S Ere“: 07 5 08 1 08 2 17-18 2 09 11 20-21 3 10 3h 22-23 3 ll hl 2h-25 3 12 b8 26-27 7 13 33 28-29 16 1h 23 30-31 20 15 13 32-33 32 15 3 3h-35 31 l7 1 36-37 30 18 1 38-39 27 21‘? 110-111 19 hZ-h3 ll hh-hS 6 h6-h7 ___g 215 *The higher the score, the greater the favorability toward TV in general. TABLE h FREQUENCY DISTRIBUTION ON THE TV-PROGRAMMING STATUS IDENTIFICATION fiFQSURE* (Maximum Range 8 00-h0) .2223: £323 §E££E EEES- Score Frag- 00-01 h lh-IS 10 26-27 22 02-03 1 16-17 11 28-29 15 06-07 1 18-19 26 30-31 15 08-09 3 20-21 30 32-33 9 10-11 3 22-23 21 3h-3S 6 12-13 7 2h—25 28 36-37 3 2T? *-Scores denote perceived similarity between personal programming preferences and perceived programming preferences of doctors and lawyers. The lower the score, the greater the perceived similarity. S2 example, the distribution of self-placement scores on the "value for industriousness" instrument (see Table 2). Almost half of the respondents (101) claimed the maximum adherence to a value for industriousness on the summary index. To get this maximum score on the summary index, respondents had to claim maximum adherence on both of the component items. Correlates pf Attitudes Toward Television Hypothesis 1 predicts that the favorability of people's attitudes toward television is negatively related to their degree of adherence to a value for critical evaluation. The data relevant to H1 are presented in Table 5. In the case of both attitude measures, the relationship between value adherence and TV attitudes is not linear, and neither correlation is significant. H1 is not confirmed. Hypothesis 2 predicts that the favorability of people's attitudes toward television is negatively related to their degree of adherence to a value for industriousness. Data relevant to H2 are presented in Table 5. Because 101 respondents claimed the maximum adherence to this value, it was necessary to illustrate this relationship with only two subdivisions. With the projective attitude measure, the relationship was actually statistically significant in the direction opposite the one predicted. H2 is not confirmed. Hypothesis 3 predicts that the favorability of people's attitudes toward television is negatively related to the discrepancy between their adherence to a value for critical evaluation and their perception of the "average person's" adherence to this value. Data relevant to H3 are 53 TABLE 5 RELATIONSHIP BETWEEN TV ATTITUDES AND MORE GENERAL VALUES Value for Critical Evaluation* TV Attitude TV Attitude Projective Measure Adjective Ratings N Adherence Level: Mean S.D. Mean S.D. Low 12.1 1.9 3h.9 h.9 57 Medium low 11.8 1.6 35.5 h.6 59 High 12.0 2.2 35.2 5.7 ha Product-moment Product-moment r g -009 r 3 -0014 Value for Industriousness TV Attitude TV Attitude Projective Measure Adjective Ratings N Adherence Level: Mean S.D. Mean S.D. LOW 11.5 1.8 Bhoé 5.11 11).]. High 12.2 2.0 3h.6 6.3 101 Product-moment Product-moment r - +.l6 r - +.0h *A minimum correlation of .lh is needed for significance at the .05 level, 2-tailed test. TAlthough a slight curvilinearity is apparent, the same probability is obtained when the data are analyzed in terms of F-ratios, indicating no advantage in terms of ascertaining relationships to be gained by converting to a correlation ratio. 5h presented in Table 6. With the projective attitude measure, the relation- ship is significantly negative, as predicted, but not linear. With the adjective rating attitude measure, the relationship is neither significant nor linear. H3 is not confirmed. Hypothesis h predicts that the favorability of people's attitudes is negatively related to the discrepancy between their adherence to a value for industriousness and their perception of the "average person's" adherence to this value. Data relevant to Hh are presented in Table 6. The interesting thing here is the return of variability to the "industriousness" adherence measure. ‘With half of the respondents placing themselves at the "ceiling" of the instrument, they placed the "average person" at differing distances from themselves. Along with this necessary variability came a slight but insignificant negative relationship (non-linear) between TV attitudes and value adherence. Hh is not confirmed. A third correlate of TV attitudes was proposed in H16: The favor- ability of people's attitudes toward television is negatively related to their degree of identification with the upper-middle class. Because the scores on the identification.measure go down as perceived similarity or identification goes up, a positive correlation with TV attitudes indicates the negative relationship proposed in H16. Relevant data are presented in Table 7. For the indirect measure of TV attitudes, the predicted relation- ship is statistically significant. The relationship between upper-middle class identification and the adjective ratings of TV attitudes is not significant, however, and H16 cannot be confirmed. 55 TABLE 6 RELATIONSHIP BETWEEN TV ATTITUDES AND THE DISCREPANCY BETWEEN SELF AND "AVERAGE PERSON" ON MORE GENERAL VALUES Value for Critical Evaluation TV Attitude TV.Attitude Projective Measure Adjective Ratings N Adherence Level: Mean S.D. Mean S.D. LOW 11.9 1.? 3h03 5.0 by? Medium low 12.5 1.8 36.0 5.1 h9 High lloh 109 Bhoo 5.6 63 Product-moment Product-moment r . -Olh* r ' '009 Value for Industriousness TV Attitude TV Attitude Projective Measure Adjective Ratings N Adherence Level: Mean S.D. Mean S.D. LOW 11.7 20h 3306 509 29 Medium low 12.1 1.7 35.7 6.3 65 Medium high 11.9 1.9 3h.3 5.2 6h High 11.5 109 Bhoo 507 S7 Product-moment Product-moment r . -.10 r I -.Oh * p less than .05, 2-tailed test. 56 TABLE 7 RELATIONSHIP BETWEEN TV ATTITUDES AND UPPERéHIDDLE CLASS IDENTIFICATION TV Attitude TV Attitude Level of high-status Projective Measure Adjective Ratings N identification: Mean S.D. Mean S.D. LOW 12.2 200 3503 5.2 ’48 Medium low 11.8 1.7 35.0 S.h 51 Medium high 1201 2.1 3308 7.0 50 High 11.h 1.7 3h.h 5.5 66 Product—moment Product-moment r ' +olh* r . +005 *p less than .05, 2-tailed test. Findings Related to Perceived Relevance of Value Adherence to TV Attitudes Three "relevance" hypotheses apply to a value for critical evaluation. Hypothesis 5 predicts that people see television as relevant to imple- menting or impeding implementation of a value for critical evaluation. "Relevance" was operationally defined as statistically significant dif- ferences in mean attitudes attributed to hypothetical types representative of different levels of adherence to the value. The data in Table 9 indi- cate that the differences in these estimates are highly significant. Therefore, H5 is confirmed. Hypothesis 7 predicts that people perceive a negative linear relation- ship between favorability of TV attitudes and degree of adherence to a value for critical evaluation. The data in Table 8 indicate the predicted relationship between high and low adherents, but the respondents did not differentiate between the estimated TV favorability appropriate to low and 57 TABLE 8 MEAN ESTIMATED TV AT TITUDES OF OTHERS (N = 215) When respondents are judging the TV attitudes of persons with . . . High adherence to a value for critical evaluation Medium adherence to a value for critical evaluation Low adherence to a value for critical evaluation (maximum range I 00-18) When respondents are judging the TV attitudes of persons with . . . High adherence to a value for industriousness Medium adherence to a value for industriousness Low adherence to a value for industriousness (maximum range I 00-18) When respondents are judging the TV attitudes of persons with . . . Above average education Average education Below average education (maximum range I 0-9) Mean 7.1 12.2 12.2 10.9 lh.l 12.3 N91? room h.0 3.1 3.0 h.l 3.9 NHN 0 W000 58 TABLE 9 MEAN ESTIMATED TV ATPITUDES OF OTHERS BY SUB3ROUPS: VALUE FOR CRITICAL EVALUATION On a value for critical evaluation, persons judged by respondents have . . . Self-Other High Adherence Med. Adherence Low Adherence fl_ Discrepancy: Mean S.D. Mean S.D. Mean S.D. LOW 80]. 3.8 1200 300 12.1 3.1 ’47 Medium Low 7.h 11.3 12.6 2.8 12.2 2.8 19 Medium High 7.1 3.8 12.9 2.h 12.h 2.8 56 High 6.0 11.0 lloh 3.6 12.2 302 63 ANALYSIS 9;; VARIANCE* Source of Variation SS DF MS F F .95 Between Subjects 2,927.2h (21h) Levels of Self- other discrepancy 109.36 3 36.h5 2.73 2.65 Error (between) 2,817.88 211 13.35 within Subjects 8,17h.00 (1130) Levels of Value Adherence 3,755.63 2 1,877.82 183.56 3.00 Interaction: Discrepancy'by Value Adherence 100.16 6 16.69 1.53 2.10 Error (within) 11,318.21 1122 10.23 Total 11,101.2h 6hh *The analysis of variance model used here and in Tables 10 and 11 is from Winer, Statistical principles in experimental design, pp. 306-307. New York: McGraw Hill, 1962. 59 medium adherence to a value for critical evaluation. In general, H7 is confirmed, but the difference in rating is entirely between the high adherence group and the lower two groups. Hypothesis 13 predicts that the differences between estimated TV attitudes appropriate to high and low adherence to a value for critical evaluation will be greater for high adherents to that value than for low adherents. Examination of Table 9 will show that this is indeed the case for the present sample. This predicted interaction effect did not reach the traditional level of statistical significance, however. Therefore, H13 cannot be considered confirmed. The "high" self-other discrepancy level distinguished most clearly between the three hypothetical types in the predicted direction. The differences in TV attitude estimates between high and low adherence to a value for critical evaluation increase with the respondents' own level of self-other discrepancy in a linear fashion: h.O for the low discrepancy level, h.8 for the medium low discrepancy level, 5.3 for the medium high level, and 6.2 for the high level. Three "relevance" hypotheses apply to a value for industriousness. Hypothesis 6 predicts that people see television as relevant to implementing or impeding implementation of a value for industriousness. Table 10 indi- cates that the differences in mean attitude estimates for hypothetical types with varying levels of adherence to a value for industriousness are statistically significant. Therefore, H6 is confirmed. Hypothesis 8 predicts that people perceive a negative linear relation- ship between favorability of TV attitudes and degree of adherence to a value for industriousness. Relevant data are in Table 8. "Low adherence" is estimated as more favorable than "high adherence," as predicted, but a 60 TABLE 10 MEAN ESTIMATED TV ATTITUDES OF OTHERS BY SUBGROUPS: VALUE FOR INDUSTRIOUSNESS On a value for industriousness, persons judged by respondents have . . . Self-Other High Adherence Med. Adherence Low Adherence N_ Discrepancy: Mean S.D. Mean S.D. Mean S.D. LOW 12.2 309 114014 109 1109 305 29 Medium Low 10.8 8.0 13.h 2.6 11.2 3.9 65 Medium High 10.6 3.9 1h.h 2.6 13.1 3.6 6h High 10.5 h.6 1h.5 2.h 12.9 h.2 57 AMALISIS 9E VARIANCE Source of Variation SS DF LB F F .95 Between Subjects 2,586.71 (21h) Levels of‘Self- other discrepancy 209.66 3 69.89 6.20 2.65 Error (between) 2,377.05 211 11.27 Within Subjects 6,7h7.33 (ABC) Levels of Value Adherence 1,116.33 2 571.67 113.37 3.00 Interaction: Discrepancy by value Adherence h0.h3 6 6.7h .51 2.10 Error (within) 5,563.57 A22 13.18 Total 9,33h.0h 68h 61 curvilinearity of estimate is introduced with the "medium adherence" category, contrary to the hypothesis. Therefore, H8 is not confirmed. Hypothesis lb predicts that the difference between estimated TV attitudes appropriate to high and low adherence to a value for industrious- ness will be greater for high adherents on that value than for low adherents. Relevant data are in Table 10. The interaction effect is insignificant; therefore th is not confirmed. Though not directly related to the theoretic rationale, three "relevance" hypotheses apply to the variable of education (see p. 20). Hypothesis ll predicts that people see the relevance of education (undefined) to favorability toward television. Data in Table 11 indicate that the differences in mean attitude estimates for hypothetical types with varying levels of education are highly significant, confirming H11. Hypothesis 12 predicts that people perceive a negative relationship between level of education and favorability of attitudes toward television. Data in Table 8 indicate that this is so, thus confirming H12. Examination of Table 8 indicates that the attitude estimates for the "middle" groups are quite similar to the attitude estimates for the "low" groups. Examination of Table 11 indicates that the predicted linear estimates held across three of four subgroups, with only the low education group deviating from the predicted pattern. Hypothesis 15 predicts that the difference between estimated TV attitudes appropriate to high and low education groups will be greater for those with much education than for those with little education. Relevant data are presented in Table 11. In using the undefined concept of "education" rather than value adherence, the results fit the predicted model very well. The differences in estimate between high and low 62 TABLE 11 MEAN ESTIMATED TV.ATTITUDES OF OTHERS BY SUBGROUPS: EDUCATION LEVEL Persons judged by respondents have . . . Respondents' level High Education Med. Education Low.Education :N_ or education Mean S.D. Mean S.D. Mean S.D. 0-11 yrs. school 5.0 3.0 6.7 1.9 6.0 2.8 36 High school diploma h.6 2.8 6.6 2.0 7.1 2.h 6h l’h yrs. 0011839 307 205 7.0 10h 709 2.0 53 beyond college degree 307 207 7.0 1.3 8.1 1.8 62 ANALYSIS gr; VARIANCE Source of Variation SS DF MS F F .95 Between Subjects 1,085.27 (21h) Levels of respondents' education h.52 3 1.51 .29 2.65 Error (between) 1,080.75 211 5.12 ‘Within Subjects 3,566.00 (A30) Levels of others' education 1,256.7h 2 628.37 12h.68 3.00 Interaction: Respondents' by others' education level tholO 6 30.69 6.09 2010 Error (within) 2,125.10 h22 5.0h Total h,651.27 6hh 63 education increase with the respondents' own level of education in a linear fashion: approximately 1.0 for those with 0-11 years of schooling, 2.5 for those with a high school diploma, h.2 for those with l-h years of college, and h.h for those with education beyond college. This inter- action is statistically significant, and H15 is considered confirmed. Special Subgroup Analyses: "Deviants" and "Non-Deviants" Hypothesis 9 predicts that among "deviants" in the TV attitude esti- mates of others with high and low adherence to a value for critical evalu- ation, the correlation between their own TV attitudes and their own adherence to a value for critical evaluation will be less negative than the same correlation among the non-deviants. A deviant was defined as a respondent who estimates that hypothetical types of persons with high adherence to the general values are more favorable toward television in general than are hypothetical types with low adherence to the general values. Deviants and non-deviants were separated into two groups. There were h3 deviants and 172 non-deviants in the case of a value for critical evaluation. Relevant data are presented in Table 12. Three of the four correlations became more negative for non-deviants and more positive for deviants, as predicted. Differences between correlations for the deviant and non-deviant groups were computed for both attitude measures and for both adherence measures via Fisher Z-transformations (see Walker and Lev, 1953, pp. 25h-256). None of these four comparisons are statistically significant, however, and H9 is not confirmed. In the Analysis Procedures section of Chapter II, it was stated that the attitude-value adherence correlations for the subgroups would also be 6h TABLE 12 COMPARISON OF TWO SUBGROUPS ON THE RJLATIONSHIP BETWEEN TV ATTITUDES AND MORE GENERAL VALUES: FISHER Z-TRANSFORMATIONS OF CORRELATIONS Value for Critical Evaluation TV Attitude TV Attitude Projective Measure Adjective Ratings Adherence Non-Deviants Deviants Non-Deviants Deviants Measure: (N=172) (N=h3) (N=l72) (N=h3) Self 01113, ‘008 -011 ".09 4'316 Self-other discrepancy -.15 -.11 -.13 +.06 Value for Industriousness TV Attitude TV Attitude Projective Measure Adjective Ratings Non-Deviants Deviants Non-Deviants Deviants (N=109) (NIlO6) (N=109) (N=106) Self only +015 +020 +003 +007 Self-other discrepancy' -.15 .00 .00 .00 65 submitted to the standard product-moment correlation test of significance to see whether or not the subgroup correlations were significantly dif- ferent from zero. This attitude-value adherence correlation indexes the same relationships tested earlier in H1, H2, H3, and HA. The only dif- ference here is that subgroups are being used, rather than the entire sample. In the case of a value for critical evaluation, where no cor- relation reached statistical significance in H1 or H3 (see Tables 5 and 6), one correlation (r I -.l5) was significant at the .05 level, 2-tailed test, and another correlation (r I -.13) was significant at the .10 level. Both of these correlations involved the self-other discrepancy measure of value adherence. Hypothesis 10 predicts that among "deviants" in the TV attitude estimates of others with high and low adherence to a value for industrious- ness, the correlation between their own TV attitudes and their own adherence to a value for industriousness will be less negative than the same correlation among the non-deviants. There were 109 non-deviants and 106 deviants. Relevant data are presented in Table 12. Comparing Table 12 with Table 5, it can be seen that the two positive correlations for the total sample subdivided into two slightly less positive correlations for the non-deviant subgroup and two slightly more positive correlations for the deviant group, which is the direction predicted by H10. This differentiation is not significant, however. Comparing Table 12 with Table 6, it can be seen that the stronger of the two negative correlations for the total sample subdivided into a more negative correlation for the non-deviant group, while the deviant group correlation collapsed to zero. This is compatible with H10. The weaker of the two negative correlations for the total sample collapsed to zero for both deviant and non-deviant 66 subgroups. In no case is the difference in correlation between deviant and non-deviant subgroups statistically significant with the Fisher Z-transformations. H10, therefore, is not confirmed. In Table 6, the correlation of -.10 approaches statistical significance. In Table 12, these same variables (self-other difference on a value for industrious- ness and projective TV attitudes) are significantly correlated (r I -.15) in the predicted direction for the non-deviant group. ‘g Block Findings Hypothesis 17 predicts that factor type A persons ("sophisticated consumers") are less favorable toward television than are factor types B ("loyal TV fans") or C ("escape viewers"). Relevant data are presented in Tables 13 and 1h. The predicted direction of differences can be seen in all four comparisons. Three of these four comparisons are statistically significant; therefore, H17 is considered confirmed. Hypothesis 18 predicts that factor type A persons adhere more strongly to a value for critical evaluation than do persons in factors B or C. The data in Table 15 show that the mean adherence levels for factors A, B, and C are almost identical. H18 is not confirmed. Hypothesis l9 predicts that factor type A persons adhere more strongly to a value for industriousness than do persons in factors B or C. The data in Table 16 indicate that this is not so. H 19 is not confirmed. 67 TABLE 13 TV ATTITUDES (PROJECTIVE) BY Q FACTOR TYPES Mean Attitude S.D. N Type A 11.2 1.9 88 Type B 12.9 1.8 58 Type 0 11.7 1.8 A7 Mixed 11.8 10,4 22 Source of Variation ANALYSIS 91: VARIANCE SS DF MS F F Between Factor Types Within Factor Types 10A.52 3 3h.8h 10.90 2.65 67b.11 211 3.19 INDIVIDUAL t-TESTS Comparison t t 0 95’ 2.1331]. TYPc B - A 5.70 1.97 Type C - A 1.h6 1.97 68 TABLE 1h TV ATTITUDES (ADJECTIVE RATINGS) BY Q FACTOR TYPES Mean Attitude S.D. N Type A 31.0 5.7 88 Type B 38-h h-h 58 Type C 3507 [$03 117 Mixed 36.h 3.8 22 ANALYSIS OF VARIANCE Source of Variation SS DF MS F F .95 Between Factor Types 2,090.82 3 696.9h 28.97 2.65 Within Factor Types 5,075.67 211 2h.06 INDIVIDUAL t-TESTS Comparison t t 095, 2.1331]. Type B -A 8.91 1.97 Type C -A 50211- 1.97 69 TABLE 15 SELF-OTHER DISCREPANCY ON ADHERENCE TO.A VALUE FOR CRITICAL EVALUATION BY Q FACTOR TYPES Mean Adherence S.D. N Type A 19oh hob 88 Type B 18.h h.0 58 Type C 18.9 303 h? ANALYSIS 9E VARIANCE Source of Variation SS DF MS F F .95 Between Factor Types 36.53 2 18.26 1.12 3.06 Within Factor Types 3,088.h7 190 16.26 TABLE 16 SELF-OTHER DISCREPANCY 0N ADHERENCE TO A VALUE FOR INDUSTRIOUSNESS BY Q FACTOR TYPES Mean Adherence S.D. N Type A 22.9 3.h 88 Type B 22.2 hol 58 Type C 23.3 3.2 A7 ANALYSIS 9g VARIANCE Source of Variation SS DF MS F F .95 Between Factor Types 32.61 2 16.30 1.27 3.06 Within Factor Types 2,h38.51 190 12.83 CHAPTER IV CONCLUSIONS, DISCUSSION, AND SUL-E'ARY Conclusions 0n the basis of the indexing method for value adherence used in the present study, the broad hypothesis that value adherence is negatively related to favorability toward television for a population typified by this sample cannot be confirmed. In eight correlations of value adherence measures with TV attitudes (see Tables 5 and 6), six were negative, but only one of these six correlations was statistically significant. Evidence to be discussed later in this chapter will indicate that this lack of support for the major hypothesis is probably due to faulty operationalization of value adherence rather than to the inapplicability of the value-expressive theory of attitudes to a television context. It is concluded that people do see value adherence as being relevant to the evaluation of television in general. In three different tests (see Tables 9, 10, and 11), the mean estimated TV attitudes which subjects attributed to hypothetical types of persons who have different levels of value adherence or education were significantly differentiated. Perceived relevance of general values to an object of judgment (e.g., television) is a prerequisite for the applicability of the "value-expressive" theory of attitudes. This criterion was met. The proposition that there is a negative relationship between value adherence and TV attitudes assumes, within the value-expressive theory of 70 71. attitudes, that favorability toward television will be perceived as negatively related to value adherence. Contrasting the mean estimated TV attitudes attributed to hypothetical types of persons at high and low levels of value adherence or education only, this negative relationship was also verified in three different comparisons (see Table 8). Only when the hypothetical types differed in amount of education, however, were the estimates of TV favorability linear across all three levels, as predicted. In the case of attitude estimates for hypothetical types at different levels of adherence to a value for critical evaluation, no differentiation was made for the total sample between the perceived medium and low adherents (see Table 8). In Table 9, however, it will be seen that the predicted differentiation did occur among two of the four subgroups. For hypothetical types at different levels of adherence to a value for industriousness, medium adherence was attributed the greatest TV favorability, not only for the sample as a whole (see Table 8), but also for all four subgroups considered individually (see Table 10). It can be concluded that a negative, linear relationship is perceived between education level and favorability toward television. It can also be tentatively concluded that the perceived relationship between value adherence and favorability toward television is negative, considering only the extremes of the ranges. There is no evidence in this study, however, to indicate that this perceived generally negative relationship is linear. 0n the basis of prior evidence cited in Chapter I, it is to be antici- pated that the more a person identifies with the upperemiddle class, the less favorable he will be toward television in general. This expectation ‘was partially supported with a significant correlation between the 72 projective TV attitude measure and the program preference "similarity index" specially constructed for this study. Identification was defined as the similarity perceived between a personal hierarchy of preferences in TV program types and an estimate of the hierarchy of preferences held by doctors and lawyers in general. Even the partial evidence in support of the "identification" hypothesis must be considered tentative because of the slight curvilinearity apparent in Table 7 and because of some Observed difficulties with the instrument which will be discussed later in this chapter. In exploring the public image of the correlates of favorability toward television, data in the present study are somewhat more precise than the compatible data obtained by Geiger and Sokol (1959), who write: . . . the respondent's own economic and educational level and feelings about televisionswatching affected his answer, but not sufficiently to change the generality of the image of the television fan as poor and relatively uneducated. It may also be concluded from this study that the "TV fan" is Characterized as a person with a below-average education. In addition, however, the effect of the respondent's own position on his attitude estimate for others is obtainable in the present design by doing an inter- action analysis (see Table 11). As an expansion on the earlier quotation from the Geiger and Sokol study (1959), the following may be stated for the education variable: Respondents with low education estimate that the highly educated are more favorable toward television than do respondents who have high education. Respondents with high education estimate that the lesser educated are more favorable toward television than do respondents with low education. Both extremes are "overestimating" the TV favorability 73 of the other. Another way of stating the pattern that emerges in Table 11 is this: The more educated a person is, the lower his TV favorability estimate for highly educated persons will be, and the higher his TV favorability estimate for lesser educated people will be. Regardless of his own educational level, a person will still estimate that the lesser educated are more favorable to television than are the highly educated. The highly educated, however, make sharper distinctions than the lesser educated on the degree of TV favorability to be attributed to high and low educational groups in general. Though the pattern of interaction effects was more clear cut using education as a basis for the two-way analysis, a somewhat similar pattern is apparent when a value for critical evaluation is the basis (see Table 9). The more a person sees himself above the average person in adherence to a value for critical evaluation, the lower his TV favorability estimate for high adherence types will be. Strong adherents to a value for critical evaluation.make sharper distinctions than the weak adherents on the degree of TV favorability to be attributed to high and low value adherents in general. This pattern of interaction effects is only sug- gestive, however, since the interaction did not reach significance. The analysis of data on a value for industriousness did not fit the pattern established by the variable of education and approximated with the variable of adherence to a value for critical evaluation (see Table 10). Medium adherence was attributed the greatest TV favorability not only for the sample as a whole, but also for all four subgroups considered indi— vidually, making tenuous any conclusions based on "high vs. low" comparisons only. 7h The secondary investigation consisted of some Q block analyses which are only tangentially relevant to the theoretic rationale in the present study, but are methodologically suggestive for future studies dealing with general values and TV attitudes. Q block indexing can isolate those who are favorable toward television (see Tables 13 and 1h). In this case, the Q block items were themselves a form of attitude statement, so this may amount to a positive correlation of one attitude measure with another attitude measure. The Q blocks did not successfully separate different levels of value adherence, however (see Tables 15 and 16). This is probably due to the fact that the original pool of Q-sort items in the MacLean-Talbott study (1963) were not consciously structured around the theme of value ad- herence, but were a collection of opinion statements derived primarily from remarks made in a series of focused interviews. The fitting of the factors to the value adherence types is therefore "forced" and after the fact. If the original items were such that a Q typology would have a clear interpretation as a value adherence level, then Q block indexing 'would indeed be an economical method of analyzing variability in attitudes toward television. Discussion With all the visual acumen afforded by hindsight, three major problem areas come into focus in.the present study. Consider first the problem of measuring one of the critical variables: value adherence. The method Irtilized here began with opinion items indicative of the two values on a .face validity basis. From that point of origin, the utility and defensi- Ixility of the newly constructed instrument designed to measure value 7S adherence deteriorated through the pretest and final study stages. Even the best four items selected on the basis of a pretest sample contained disappointingly low correlation values. This generated some suspicion that the values suggested by Steiner and Schramm might not have the uni- dimensional character which had been tentatively attributed to their explanations of variability in TV attitudes. The decision was made to go ahead with the instrument into the main study, being cautious about interpreting the results of the instrument as a single value unless the correlations held up with the later sample containing more variability in age and education. When the instrument was reanalyzed using value adherence data from the main study, two more setbacks were sustained. The negative items, which had been selected as a control for acquiescent response set, did not correlate positively with the positive items, and had to be dropped from the final analysis scale. Not only did this restrict the range of the value adherence scale, it also removed the response set control, inextricably confounding response set influences with "true" value adherence responses among the remaining two positive items. The positive items themselves did not correlate as strongly in the main study as they did on the pretest. The first difficulty, then, was low inter-item cor- relation values on the value adherence instrument, indicating a lack of unidimensionality in the value construct as operationalized in this study. Non-homogeneity in scale items has the effect of increasing error vari- ability, which works against the precision of measurement needed to reject the null hypothesis. The second problem to be considered is the effects of the possible acquiescent response set biases introduced as a result of the low over-all .1 L. 76 inter-item correlations. Assuming response set biases did exist in the value adherence instrument, they should have differing effects as this instrument is correlated with other measures in the study, if these other correlates differed in degree of response set biases of their own. A major correlate of value adherence investigated in this study was attitudes toward television. Two measures of attitudes were given, one of which was also susceptible to acquiescent response set, and one of which wasn't. Any two scales should correlate positively if respondents simply acquiesced by agreeing with or being favorable toward all items on both scales. The adjective rating attitude measure controlled for "mechanical" responses and position effect by alternating ends of the scale on which positive adjectives were placed, but the possibility of acquiescent response set was still present. The projective attitude measure, however, would appear to be free of acquiescent response set biases. With seven drawings to choose from in matching captions to drawings, there was no obvious choice if a respondent wished to acquiesce. The projective measure was administered first in the questionnaire schedule, and television was not emphasized before that in order to ‘minimize sensitization. If both attitude measures are in fact tapping the same variable of .favorability toward television, and if both attitude measures are cor- related with a value adherence measure beset with response set biases, 'the attitude measure containing similar response set biases should cor- relate more positively (or less negatively, as the case may be) than the vattitude measure free from response set biases. Thus, projective attitude zneasures should correlate more negatively with value adherence than ‘adjective rating attitude measures. Such was the case in three of four 77 comparisons (see Tables 5 and 6). This evidence, plus the fact that it was the positive items that were retained, supports the conclusion that the value adherence measure was indeed impaired with response set biases. A third problem in the present study to be considered is the way in which this sample deviated from the general pOpulation by being over- represented in high education and under-represented on low education, old age, and low income (see Table 1). If this sample restricted the repre- sentation of the education variable, and if education is the best index of value adherence, as the data of Steiner suggest, then the range of value adherence was also restricted in the present study. If true, this means that the sample variance on value adherence is less than the variance in the general population on value adherence. Such statements are necessarily conjectural, since no parameters on value adherence are available for comparison. To the extent that the range or variance of value adherence is restricted relative to the population variance on value adherence, however, the effect of the restriction will be to lower the sample correlation value from the parameter value (see McNemar, 1962, p. lbh). This effect, if present, was another influence working against the support of the major hypothesis proposing a negative relationship between TV attitudes and value adherence. If a satisfactory value adherence instrument can'be devised overcoming the present weaknesses of lusterogeneity and lack of response set controls, a stratified sample, 3proportional to the general population on relevant characteristics should 'be utilized to eliminate this source of restriction on the magnitude of 'the correlation. One can only conjecture as to why the main sample was so different ;firom the pretest sample in the cohesiveness shown in marking the value 78 adherence items. Contextual differences between the pretest and the actual study may have had an effect. The pretest consisted primarily of the hh value adherence items. In the main study, the eight selected value adherence items were surrounded by attitude scales, estimates of others' TV favorability, and the like. There is, of course, a possibility that the pretest correlation matrix was so large that the two smaller abstracted.matrices of four items each consisted of correlations that were high because of chance variability. In any event, the incorrect assumption was made that a pretest would yield sets of items that would hold together for another sample. ‘When four of the eight items selected had to be dropped because of no correlation or even negative correlation, the utility of the value adherence measure was greatly impaired. If this same instrument design is to be used in a future study, it is recommended that the pretested items be cross—validated on a second pretest sample. A larger number of items for the final scale is also recommended, so a few items could still be dropped at the final stage without destroying the utility of the instrument. A factor analytic method could also be used to devise a value adherence instrument, using a two-stage pretest. The first stage would be exploratory, going to the field with open-end questions on likes and dislikes regarding television and the underlying reasons. Every time a respondent would mention as a reason something that sounded like a general value, this would be recorded for the second pretest stage. The entire pool of general values thus recorded would then be administered to a second pretest sample for "agreement" or "applicability to me" ratings. These ratings would then be submitted to factor analysis. Here the 7 .factors would represent a parsimonious set of independent value clusters. 79 By discerning an underlying commonality of content, each value "cluster" could be called a single general value. The items most highly loaded on each factor would become the basis for the value adherence instrument in the main study. Whereas the present study used values suggested by other researchers, this method would allow people in the field to articulate the values in terms that are meaningful to them. This should improve the value adherence instrument a great deal. The form of the final value adherence instrument could be identical to the form used in the present study: a lO-step ladder scale of agreement. Summary scores for each respondent on each factor would then be computed and entered into a multiple regression formula, predicting TV attitudes on the basis of the respondent's position on all factors. The work of Rosenberg (1956) suggests that not only should the iso- lation of value types be accounted for, but also the relative importance of these general values for the individual respondent, and the relative potency of television to implement or impede implementation of these general values. If the two-stage factor analytic pretest described above would isolate, say, six general values, respondents could rate (or rank order) them in terms of personal importance. Similarly, the perceived potency of television to facilitate or hinder expression of each general value would be rated. An algebraic swm of cross products (importance times potency) would be computed and entered into a correlation with TV attitudes for all respondents. In the present study, the indexing of relative importance of values did not seem to be warranted because only two values were being investigated. All of the suggestions above have applied to television in general. The research picture becomes simultaneously more precise and complicated 80 if several £32322 of television could be isolated and considered separately. Quality "A" of television would then be perceived as having "B" amount of potency for implementing value "C," which had "D" amount of importance for the individual respondent. Attitudes might be indexed separately for each facet or quality of television, in which case elements "B" and "D" should predict them, using Rosenberg's formula. Or, the average algebraic sum of cross products (importance times potency) could be used to predict "over-all" TV attitudes, which would be averaged across all facets. The concept of "facets" of television is probably related closely to the concept of "functions" of television. The assumption is that dif- ferent facets serve different functions. Charles wright (1960), building upon earlier work by Lasswell and Merton, suggests several categories of functional (or dysfunctional) communication activities applicable to the individual. There are subdivisions of four general categories: (1) surveillance (news), (2) correlation (editorials, interpretation, and prescription), (3) cultural transmission, and (h) entertainment. 'Wright's schema might be used as a basis for isolating facets of television which could be used in the more precise attitude study described above. For example, a respondent might be asked the following: "In terms of how it serves the function of keeping you abreast of the news, how would you evaluate television?" After close observation of a variety of respondents as they marked the instrument designed to measure identification with the upperdmiddle class, it is felt that the instrument is not entirely satisfactory in its present form. Older people and lesser educated people had a difficult time trying to "empathize" with the general category of "doctors and lawyers" in their program preferences. A few refused to cope with this 81 assignment at all, and several gave indications that the task was threatening to them: e.g., "Why should they be any different from us?" The techniques used in the instrument seem quite attractive, but different content is needed that is less demanding on empathic skills. MacLean and Talbott (1963) report no difficulty in getting respondents to sort program titles as they thought their spouse would.prefer them, but apparently a general- ized preference hierarchy appropriate to a category of persons is more difficult to cope with. Perhaps the lows; class as a reference point would have been more satisfactory. Relationships were consistently stronger when the self-other dis- crepancy measure rather than the self-placement alone was used in indexing value adherence. It is therefore suggested that future studies skip the self-placement analysis completely and go directly to the discrepancy measure, which contains the greater amount of information. The cor- relation between "self only" and "self-other" in a value for critical evaluation was +.7l; in a value for industriousness, this correlation was +.61. Though not necessarily required by the design of the instrument, these strong correlations are understandable. The higher a respondent places himself in an "absolute" sense, the higher he places himself relative to the average person. If a respondent puts himself at the bottom of the scale, he £32233 put the average person.much lower than himself. If a respondent puts himself at the top of the scale, however, he negd get place the average person any lower than himself, but he in fact tends to do so. There were only a few instances where the respondent placed the average person higher than himself. It is interesting to note the results of the Steiner projective TV attitude measure as it was specially scored for this study. Correlations 82 with the projective measure were consistently stronger than correlations with the adjective rating scales (see Tables 5, 6, and 7). The advantages of this instrument in terms of freedom from acquiescent response set have been mentioned earlier. The projective measure correlated significantly (r 8 .1h) with the identification or similarity index while the adjective rating measure did not (r - .05). It correlated higher with income (-.15 vs. -.OS), and education (-.32 vs. -.23) than did the adjective ratings. Its correlation with the adjective rating measure itself was +.38. As a purely exploratory venture, this new attitude scale yielded surprisingly rewarding results. The respondents enjoyed marking this instrument. Its simplicity and graphic appeal made it easy to administer compared with, say, the self-other comparisons on value adherence. Mechanically, this instrument is "infinitely" expandable in terms of the variety of drawings used and the content of the captions; it is deserving of further exploration and use. Education correlated significantly with the two TV attitude measures (-.32 and -.23). In the relevance section where respondents estimated the TV attitudes of hypothetical types, the predictions were most clearly con- firmed when education, not value adherence levels, was used for the hypo- thetical types. Education still predicts TV attitudes better than value adherence as operationalized in this study. The early conceptualization of education as a multi-dimensional variable, encompassing more than the values for critical evaluation and for industriousness, is felt to be justified. Even the significant relationships found in this study are weak in terms of variance accounted for. In the present study, an effort was made to predict and explain TV attitudes on the basis of two psychological 83 variables. The baseline for evaluating the success of these efforts was the predictibility afforded by educational level, which correlated -.32 and -.23 with the attitude measures in this study. In terms of variance accounted for, this is hardly an overwhelming relationship to begin with. Assuming value adherence to be a component of education, it is not sur- prising that the value adherence-attitude relationship was also weak. A multivariate approach may be necessary to accumulate enough power to surpass education with psychological variables as predictors of TV attitudes. Summary Several studies have indicated that general favorability toward television covaries with educational level, which is typically indexed by number of years of formal schooling. Since the index itself has no explanatory power, several plausible explanations have been offered, but without being given a direct empirical test. A structural similarity was discerned between two such proposed explanations, one by Steiner and one by'Schramm. Both could be rephrased, without doing them violence, into statements that persons least impressed with television in general were persons adhering most strongly to a value for critical evaluation or to a value for industriousness. Viewing TV in general would not be expressive of their values, hence the unfavorable attitude. Self-beliefs in terms of values, with television behavior as manifestations of those values, made up the theoretic framework within which the problem of explaining TV attitudes was approached. The design centered on the indexing of self-beliefs in terms of the two proposed values, the indexing of attitudes toward TV in general, and 8b the indexing of perceived relevance between the values and the TV attitudes. The application of the theory demanded that the respondent actually perceive favorability toward TV as appropriate to those not adhering strongly to the two values (critical evaluation and industriousness). The instrument designed to index value adherence failed to meet several criteria for acceptability, thus not allowing an unambiguous test of the major hypothesis which predicted a negative relationship between degree of value adherence and favorability toward TV in general. Perceived relevance of education and value adherence to favorability toward tele- vision in general was established.with significant differences in attitude estimates associated.with hypothetical types of persons who had different levels of education and value adherence. The TV attitudes associated with sequentially increasing levels of education decreased in favorability at each level, as predicted, when averaged across all respondents. This linear trend was given more precise analysis by controlling on the respondents' own level of education, yielding a predicted interaction effect: the higher a respondent's own educational level, the greater the difference in his TV favorability estimate between high and low levels of value adherence or education. Other data included an upperémiddle class identification measure in terms of TV program preferences and a series of Q block analyses. As predicted, it was found that the more a person identifies with the upper- Iniddle class in program preferences, the less favorable he is toward tele- vision in general. The Q block analyses, based on an earlier study by IflacLean and Talbott, were pursued for exploratory methodological purposes :rather than as tests of the theoretic formulation. It was concluded (I 85 that Q methodology did hold promise in a study of this kind, providing the original Q-sort items were structured around the theme of general values. 2. 3. h. 5. 9. 10. ll. 12. 13. REFERENCED BIBLIOGRAPHY Bauer, R. The obstinate audience: the influence process from the point of view of social communication. American Psychologist, 196,4, l9, ppe 319“28e Charters,'w., and Newcomb, T. Some attitudinal effects of experi- mentally increased salience of a membership group. In Maccoby, E., Newcomb, T., and Hartley, E. (eds.), Readin s is social psychology. New York: Holt, Rinehart and Winston, 1958, pp. 276531. Divesta, F., and Mervin, J. The effects of need-oriented communi- cations on attitude change. Journal‘s: Abnormal and Social Psychology, 1960, 60, pp. 80-35. Geiger, K. and Sokol, R. Educational television in Boston: memorandum #3. In Schramm, W. (ed.), The impact 2f educational television. Urbana: University of Illinois Press, 1960, pp. 39-67. Geiger, K., and.Sokol, R. Social norms in televisionawatching. American Journal 2f Sociolo , 1959, 65, pp. l7h-81. Hazard,'w., Meriaty, D., and Timmons, V. A nontopical system of TV program categories. Audio Visual Communication Review, l96h, 12, pp. 1146-63. Katz, D. The functional approach to the study of attitudes. Public Opinion Quarterly, 1960, 2h, pp. 163-20h. Klapper, J. The effects gf_mass communication. Glencoe, Illinois: Free Press, I960. Lazarsfeld, P. and Field, H. The people look 33 radio. Chapel Hill: University of North Carolina Press, l9h6. Lindquist, E. Design and analysis 2: experiments in psychology and education. Boston: Houghton Mifflin Company, 1953. MacLean, M., and Talbott, A. Prediction of individual patterns of television viewing. Paper presented at the August, 1963, American Sociological Association Convention. East Lansing: Communications Research Center, Michigan State University (mimeographed). McNemar, Q. Psychological statistics. New York: Wiley, 1962. Osgood, C., Suci, G., and Tannenbaum, P. The measurement 2f meani g. Rosenberg, M. Cognitive structure and attitudinal effect. Journal 2£.Abnormal and Social Psychology, 1956, 53, pp. 367-72. 86 15. 16. 17. 18. 19. 20. 21. 22. 23. 2b. 25. 26. 27. 28. 87 Samuelson, M., Carter, R., and Ruggels, L. Education, available time, and use of mass media. Journalism Quarterly, 1963, no, pp 0 h91'Ll-96 O Schramm, W. The audiences of educational television. In Schramm, w. (ed.), The impact 3; educational television. Urbana: University of Illinois Press, 1965, pp. 18-38. Schramm, w., Lyle, J., and Parker, E. Television ip‘phg lives pf our children. Stanford, California: Stanford University Press, 1961. Schramm, W., Lyle, J., and Pool, I. The peOple 100k.EE educational television. Stanford, California: Stanford University Press, 1963. Seldes, G. The public arts. New York: Simon and Schuster, 1956. Smith, M. Personal values as determinants of a political attitude. Journal pf Psychology, l9h9, 28, pp. h77-86. Stanton, F. The role of television in our society. N.A.B. address, 1955. In Fellows, J. (ed.), Television news and public affairs programming. Syracuse, New York: Syracuse University Radio- Television Center, 1959. Steiner, G. The people look 23 television. New York: Knopf, 1963. Stephenson,'w. The stu ‘3: behavior. Chicago: University of Chicago Press, 19 3. ‘ Talbott, A. The Q block method of indexing Q typologies. Paper presented at the 1963 Association for Education in Journalism convention. East Lansing, Michigan: Communications Research Center, Michigan.State University (mimeographed). Tannenbaum, P., and.McLeod, J. Public images of mass media institutions. In Danielson, w. (ed.), Paul J. Deutschmann memorial papers in mass communication research. Cincinnati: Scripps-Howard Research, 1963. Thayer, J. The relationship of various audience composition factors to television program types. Journal pf Broadcasting, 1963. 7. 217-25. U.S. Bureau of the Census. 23§3 Census pf Population: 1960. Vol. I, Characteristics 23 the Pppulation. Part I, United States Summary. Washington, D.C.: U.S. Government Printing Office, 196h. Walker, H. and Lev, J. Statistical inference. New York: Holt, Rinehart, and Winston, 1953. 88 29. Winer, B. Statistical principles in experimental design. New York: McGraw Hill, 1962. 30. 'Wright, C. Functional analysis and mass communication. Public Opinion Quarterly, 1960, 2h, pp. 605-20. PPEN DIX A PRETEST QUESTIOBLMIRE Opinion study Department of Communication Michigan State University June, 196h Thank you very much for cooperating in this research; this short question- naire will not take long to complete. There are three sections to the questionnaire. If you have time to complete all three sections, the information will be very useful. If your time is quite limited, any portion of the questionnaire that you pap complete will be greatly appreciated. Please try not to look at the responses of others in your group. Avoid discussing the questions until after completion. Note: This information will be grouped together with all the rest of the questionnaires for analysis in an electronic computer. Your particular questionnaire, which you will not sign, will be treated as strictly confidential information. 89 90 SECTION ONE In this section, you will be asked to react to a number of opinion statements. Notice the lO-step ladder scale below. Your response to each statement may run from step 1 (disagree completely) to step 10 (agree completely). Read each statement, then look at the ladder scale. Decide which ladder step best describes your personal opinion, then place the ladder step number in the blank in front of the statement. Don't spend a lot of time on any one item; we want your offhand re- actions. This is not a test; there are no "right" or "wrong" answers. Just give your frank opinion. E 332162 soapleielt __.__10 9 “.3 7 I agree somewhat I disagree somewhat vaH .3. _...._1.._. Insisaarss_92m212tslz Q-9: STEP NUMBER . I am interested in a wide variety of topics. Q-lO: STEP NUMBER . Life is too short to be "at the grind- stone" all the time. Q-ll: STEP NUMBER . Mass production has raised the general quality of goods. Q-l2: STEP NUMBER . I like to be active in organizations and groups. Q-lB: STEP NUMBER . I rarely give compliments. Q-lh: STEP NUMBER . Most people are too pushy in their ambitions to get ahead. Q-lS: STEP NUMBER . The average person enjoys the same kinds of things I enjoy. Q-l6: STEP NUMBER . Five years from now, I expect to be a lot better off than I am now. PLEASE PM SURE YOU HAVE RESPOI-IDED TO ALL ITEMS, THEN GO TO TEXT PAGE. Section One continued. Read each item, look at the ladder scale. 91 Enter the ladder step number beside each item that best describes your personal opinion. Q-l7: Q-18: Q-l9: Q-20: Q-2l: Q-22: Q-23: Q-2h: 0-25: Q-26: Q-27: STEP STEP STEP STEP STEP NUMBER I asses soapleieln _ .____.1O 9 _.__B I agree somerset- _ _. 7 ‘5‘ .3. I'disa iiia NUMBER NUMBER 1M4 BER NUMBER NUMBER NUMBER NUMBER NUMBER UMBER NUMBER ree somewhat 3 2 aree_92m2lstskz ._;L_ I am hard to please with anything because I have high standards of excellence. Success comes mainly to those who get the lucky breaks. I can usually judge whether something is good or bad "right off the bat." You only have a few years in this world, so you've got to work very hard to accomplish anything. Such descriptions as "wonderful,""sensa- tional," and the like are used so much, they are meaningless. You can't tell much about a man by the kind of job he has. The average person and I would probably agree in deciding the quality of most things 0 When.a man is serious about his work, he doesn't stop thinking about it at five o'clock. Things of beauty and cultural value take a lot of training to be understood. I'm not a perfectionist. I don't have the time, interest, or the money for the "finer things of life." PLEASE MAKE SURE YOU HAVE RESPONDED TO ALL ITEMS, THEN GO TO NEXT PAGE. Section One continued. Read each item, look at the ladder scale. Enter the ladder step number beside each item that best describes your personal opinion. Q-28: Q-29: Q-303 Q-32: Q-33: Q-3h: Q-BS: Q-36z Q-37: Q-38: Q-39: STEP STEP STEP STEP STEP STEP STEP STEP STEP STEP STEP STEP NUMBER I 2329.9. aoapIeieIL .. M10 9 ___8 I gazes soaeahit ._ _ 7 ‘_3_ .3. I _° iafirsefiemswllai _ __.__E _II_ 2 I Eiiasraefemalstslx _.__1 NUMBER NUMBER NUMBER NUMBER NUMBER NUI'IBER NUMBER NUMBER NUMBER NUMBER The people who matter to me are quite proud of the success I have achieved. In such terms as "mass culture," "mass production," "mass media," mass every- thing, the word "mediocre" could be substituted for "mass." You'll never change the world, so why try? I am really interested only in a few subjects. You can't stand still in this life; you either move up or you go down. One of the main functions of an intellectual is to serve as a critic for the rest of society. A lot of competition is a bad thing. You've got to give in now and then to get along in this world. I've got to do a job right if I do it at all. Real excellence is a virtue that few under- Stand. People in upper management positions pay too high a price for their status. The general public understands most world issues. PLEASE MAKE SURE YOU HAVE RESPONDED TO ALL ITEMS, THEN GO TO NEXT PAGE. Section One continued. Read each item, look at the ladder scale. 93 Enter the ladder step number beside each item that best describes your personal opinion. Q-hO: Q-hl: Q-h2: Q-bS: Q-hé: Q-h7: Q-b8: Q-h9: Q-SO: Q-Sl: Q-SZz STEP STEP STEP STEP STEP STEP NUMBER NUMBER NUMBER NUMBER NUMBER NUMBER NUMBER NUMBER NUMBER NUMBER NUMBER NUMBER NUMBER I disagree somewhat I agree sowIeIeIsL _ ___10 9 __._B Igmsrm@r___ 7 14444404 I feel guilty when I waste time. On nearly everything, quality is more important to me than quantity. When the day's work is over, I forget my work problems until the next day. I don't need much information before I can make a decision on something. It's always harder to lead than to follow, but the rewards are worth it. When "just everyone" is in favor of some- thing, I become suspicious. I take competition rather lightly. I frequently describe things I like as "wonderful," "sensational," and the like. You have to keep learning to stay ahead in today's world. The things I really appreciate are too specialized for most people. I can "goof-off" with no guilt feelings whatsoever. In almost every case, "mass culture," "mass production," "mass media," mass anything, means the better things of life have been made available to everyone, not just to a privileged few. You tell me what a man does for a living, and I'll tell you a lot about the man. 9h SECTION TWO In this section, you will be asked to compare two types of people, esti- mating which type is more favorable toward television in general. Simply place a checkmark by the best alternative. Q-SB: Who likes TV more, very active people or very inactive peOple? very active people They're about the same. T5- TIT very inactive people I have no idea whatsover. 31 (05 Q-Sh: Who likes TV more, lawyers or truck drivers? truck drivers They're about the same. (15 lawyers I have no idea whatsoever. Z05 Q-SS: Who likes TV more, educated people or uneducated people? educated people They're about the same. 215 uneducated people I have no idea whatsoever. 0 Q-Sé: Who likes TV more, people with low standards of excellence or people with high standards of excellence? eople with low standards They're about the same. I 5 o f excellence Z15 eople with high stand- I have no idea whatsoever. I 5 ards of excellence Z05 Q-ST: Who likes TV more, people on their way up in the world or people on their way down in the world? eople on their way up They're about the same. I 5 in the world 1 eople on their way I have no idea whatsoever. down in the world 0 Q-SS: Q-S9: Q-éO: Q-6l: Q-62: Q-63: 95 Who likes TV more, people who don't read very much or people who read a lot? eople who don't read They're about the same. 2 5 very much (15 people who read a lot I have no idea whatsoever. T7 T5)— Who likes TV more, people with high income or people with low income? neople with high income They're about the same. eople with low income I have no idea whatsoever. “(TD “(6? Who likes TV more, people of low social status or people of high social status? people of low social They're about the same. status (15 people of high social I have no idea whatsoever. status (05 Who likes TV more, sophisticated people or unsophisticated people? sophisticated people They're about the same. i d unsophisticated people I have no idea whatsoever. ii a Who likes TV more, negroes or whites? negroes They're about the same. i i whites I have no idea whatsoever. 1 a Who likes TV more, city residents or rural residents? city residents They're about the same. 1 4 rural residents I have no idea whatsoever. (05 i 96 SECTION THREE In this section, we would like some general information about yourself. This will go very quickly, because most items only require a single check- mark on the one alternative that best describes yourself. Q-6h: First, about how much time do you spend, on an average day, reading your daily newspaper? (Please place a checkmark by one of the alternatives below.) 0 minutes 1 hour 2 hours 15 minutes 1 hr., 15 min. 2 hrs., 15 min. or 9 more 30 minutes 1 hr., 30 min. iii 2 hS minutes 1 hr., hS min. (7) Elias Q-65: Next, about how much time do you spend on an average day listening to the radio? (Please place a checkmark by one of the alterna- tives below.) 0 minutes 2 hours h hours 1% l/ I 1” ’2 hour 212 hours he hours or more 1 hour 3 hours 11;» hours 3% hours flifld M 97 Q-66: About how much time do you spend on an average day watching television? 0 minutes 2 hours h hours (05 (55 o 1/2 hour 21:» hours 1ng hours or more (15 ($5 9 1 hour 3 hours 737— 757— 15 hours 3% hours 737— 777- Q-67: How many newspapers and magazines combined do you subscribe to? none four eight (05 (H5 0 one five nine or more W 737' 9 two six (25 (55 three seven "or 777' Q-68: Do you subscribe to TV Guide or buy it regularly? No Yes “(57" TI)- Q-69: Do you usually read TV program information in the papers and magazines? No Yes 75)— "T17" 98 Q-70: In general, do you make a point of planning an evening TV viewing schedule ahead of time, or do you tend to "take your chances" with what- ever TV programs are on whenever you have the time to watch? I don't watch TV. 0 I usually don't plan a viewing schedule ahead of time. 1 Yes, I usually plan a viewing schedule ahead of time. 2 Q-7l: Do you usually watch TV alone or with someone else? I don't watch TV. I usually watch TV by myself. I usually watch with someone else. Q-72: Do you select most of the TV programs you watch or does someone I don't watch TV. I usually select the TV programs I watch. Someone else usually selects the TV programs I watch. I select about half the programs I watch. Q-73: Have you ever written a letter of praise or criticism about tele- vision to: a TV station? No Yes i i 7 O M (D U) a TV sponsor? i i :3 o I-<: a) m a TV network? i i a TV star? *3 O i—< (D U) a newspaper or magazine? 6‘ “ H: «D U) a government agency or official? 2 O E? U) i i Q-7h: What is your approximate age? 20 years or less 36 to hO years 56 to 60 years a 21 to 25 years hl to hS years 61 years or more (3) (9) '21 26 to 30 years hé to 50 years 51 a 31 to 35 years 51 to 55 years (3) (7) Q-YS: What is your sex? Male Female 757' TE— lOO Q-76: How many years of formal schooling have you completed? elementaqy high school undergraduate graduate 2 or less 7-8 years l3-lh years 17-18 yrs. {05 (35 (65 3-h years 9-10 years 15-16 years 19 yrs. or '—CEI' _I57_ "I77- _T§U—hore 5-6 years ll-l2 years 2 (S5 Q-77: About how much does the main wage earner in your household earn per year? $2,000 or less a8,001-a10,000 $16,001-s18,000 (B 5 a a $2,001-su,000 $10,001-$12,000 $18,001 or more (9) 31 a $h,OOl-$6,000 $12,001-Slh,000 $4 a $6,001-$8,000 $lh,OOl-$16,000 (35 a Q-78: ‘What is the occupation of the main wage earner in your household? (If applicable: What kind of industry is that?) Please write in your answer below. Q-79: What is your favorite TV program? What, in your opinion, is the worst TV program on the air? THIS CONCLUDES THE QUESTIONNAIRE. Thank you very much for your co- operation! If you have any comments you would like to make about this questionnaire or about television in general, please feel free to write them on the back of this page. 101 APPENDIX B VERBAL INSTRUCTIONS FOR THE MAIN QUESTIONNAIRE Background This study is being conducted by Keith Mielke from the Department of Communication at MSU. It is a study of opinions about the mass media, including television. All questionnaires are considered confidential, the responses will be coded by number for analysis in the University's electronic computer. This is a group-administered questionnaire. You will be given instruc- tions section by section, so please be patient if you finish ahead of other people and have to wait a few minutes. You might use such time to check back and make sure you did not leave anything out. Don't spend a lot of time on any one item. Please don't look at your neighbor's questionnaire or discuss the questionnaire until after you're finished. It will take about us minutes to complete. Please be sure to keep all sections of your questionnaire together. They should all be handed in at the same time. Thank you very much for cooperating with us on this research. 102 Notice the loose green and yellow sheets with the drawings on them. The men are to take the green sheets and the ladies are to take the yellow sheets. There are seven drawings on these sheets, and each drawing has a number beside it. You will be reading some thoughts the man or woman in the drawings might be having, and I'd like you to pick out which picture each thought belongs with . . . in which situation he or she is most likely to be feeling that way. You may select any drawing as many times as you want to. First, you will read the "thought" that the man or woman in the drawings might be thinking. Then you are to quickly scan over the seven drawings and select the one drawing most appropriate for that thought. On your questionnaire, you will then put an "X" beside the drawing number that fits each thought best. If the thought doesn't seem to fit agy drawing, check "NONE" on your questionnaire. Let's work the practice item on the cover page now to see exactly how this section goes. The thought that the man or woman.might be having is: "I should be going home soon." Look at your sheet of drawings and choose the one that goes best with that thought, then put an "X" in the appropriate blank on your questionnaire. Do this now. (PAUSE) If you have any questions about how this section is to be marked, please ask them now. If you have no questions, you may begin on the next page and complete Section One. 103 Verbal Instructions for Section Two In this section, you will be asked in several different ways how you feel about television in general. You will be given a series of word pairs that are opposites, such as "noisy-quiet." A 7-step ladder runs between these extremes, and you are to put an "X" on the ladder step that you feel best describes TV in general. You should not spend a lot of time on any one item, because we just want your offhand impressions. Look at the practice item on the cover sheet which uses the words "noisy- quiet." Reading from top to bottom, this scale would read like this: TV is generally extremely noisy, very noisy, fairly noisy, SO-SO noisy and quiet, fairly quiet, very quiet, or extremely quiet. Now place an "X" on the ladder step that you feel best describes TV in general. If you just can't decide on a ladder step, check the "don't know" blank on the right. If you think that the words "noisy-quiet" do not apply to television, check the "doesn't apply" blank on the right. If you have any questions about how this section is to be marked, please ask them now. If you have no questions, you may begin on the next page and complete Section Two. 10h Verbal Instructions for Section Four Section Four is the booklet with half-size pages. Please turn to this booklet now. In this section of the questionnaire, we are interested in how much you think different types of people like or dislike television in general. If you will turn to the second page of the booklet, we will work a practice item to see how this section goes. On each page of the booklet, one type of person will be described briefly. You are to estimate how favorable this type of person is toward TV by placing an "X" on the appropriate ladder step. In the practice item, the type of person is "one who lives on a farm." If you think people like this are favorable toward television, you would place your "X" somewhere on the top half of the ladder. If you think people like this are unfavorable toward television, you would place your "X" somewhere on the bottom half of the ladder. Mark the Practice Item now. After you have estimated how much each type of person likes TV, we would like to know how sure you are of your estimate. Simply put an "X" by one of the alternatives on the bottom scale. Don't spend a lot of time on any single page. If you have any questions on how the booklet is to be marked, please ask them now . . . . If not, please complete the booklet. This should only take from 10 to 12 minutes. 105 Verbal Instructions for Section Five As soon as you've finished filling in the booklet, turn to page 9, the pink page, (PAUSE) in your questionnaire, Section Five. Public Opinion polls often tell us that the average person has this opinion or that opinion. In this section, we are interested in your estimate of the average person's opinion on several topics, and how you would compare yourself with the average person on these topics. __....._9_ Look now at the practice item on the pink page. Think of how much the average person might agree or disagree with the statement, "I like to stay up late and then sleep late." When you have decided, write the letters "AP" . . . . which stand for AVERAGE PERSON . . . . on that step of the ladder. Now indicate how you personally feel about the statement by writing the word "MB" on the ladder step that expresses your own agreement or dis- agreement with the statement. The letters "AP" and the word "ME" can be written either on the same ladder step or on different ladder steps. Remember, you will write two things on each ladder in Section Five: "AP" for the average person's agreement and "ME" for your own agreement. If you have any questions on how Section Five should be marked, please ask them now. (PA SE) If there are no questions, go ahead and fill out the next two pages. 106 Verbal Instructions for Section Six-A 329.2 ‘ — In Section Six-A, we are interested in the types of TV programs that you like and dislike. Look over the list of 16 types of TV programs and pick out the three program types you like best . . . . As soon as you have chosen them, draw a circle around them so you can find them easily . . . . do this now . . . . Now, . . . from these three . . . . pick out the single program type you like the very best and enter the number for that type in the box at the Egg . . . . Then put the numbers for your other two choices in the two boxes next to the top. Now scratch a line through those three program types E6"§bu will know which numbers have already been entered. (PAUSE) As you can see, you have 16 numbers to put into 16 boxes. Now look over the remaining types of TV programs and pick out the three program types you like the least and circle them. From these 3, pick out the program type you like the very least and enter its number in the bottom box . . . . Then put the numbers for your other two choices in the boxes next to the bottom. As you enter a number in a box, scratch through that program type. Now, from the remaining program types, pick out the three you like best and enter their numbers in the three empty boxes at the top. (PAUSE) Next, pick out the three program types you like least and enter their numbers in the three empty boxes at the bottom. Be sure to mark a line through each program type as you enter the numbers in the boxes. And last, enter the numbers of the four remaining program types in the four middle boxes. When you have finished p. 12, continue with page 13, which is worked the same way, only this time sort out the program types as you think doctors and lawyers in general would prefer them. (Read instructions through again for Section 6-B if necessary.) When you have finished page 13, read the instructions carefully and work on to the end of the questionnaire. APPENDIX C QUJSTIOHNAIRE Michigan State University Department of Communication Project 152 SECTION ONE PLEASE WAIT FOR INSTRUCTIONS BEFORE BLiINNING THE NEXT PAdE. HEN: Please take out the green sheet with the drawings on it. WOMEN: Please take out the yellow sheet with the drawings on it. You will be given instructions on how to choose a drawing to match the various statements. EXAMPLE: "I should be going home soon." DRAWING N O . : 1 E NONE 107 152 01 {groups 00 (85‘) 01 108 _ 109. llO CHOOSE THE. ONE DRAZ‘IING THAT FITS BEST. I‘TARK THE. BLANK WITH AN "X". 1h: "Boy, this is fun!" OMB-TING IO. : l 2 .— 3 II " ‘ 2191?: .g...... 5...... 7...... 15: "I'm a little ashamed of myself for spending my time like thig." DRAL'JII-IG NO.: 1 2 v. 3 .. I; EOI‘IE ‘§‘_-——- ........ 5...... 1...... 16: "This fascinates mg." DRAWING NO.: 1 3...... 3 H I‘IONE .3...... ._......_. 32...... 1...... 17: fl; really_should}b§ doing something else." N ONE HWY 111 13: "Thi_s_ 13 flat _I_ Sill real nleagmmg." BEAT-‘11 N '1‘: N 0.: 1 ... 2...... 3 1i NONE 3...... ......... .....5.. 7 19: "1'. wish L could give this 32." DRJXYJ'JIIQG N00: 1 2...... 3.... H Nor-IE 5...... 7 20: "it meally makgg me £33]; good 39. 32339 my time like this." DPJLEHNG NO.: l ._ 2 3 NOTE 3’ _ 5...... 7 21: "What a; waste_ gf time." Dibl‘iING NO. : l 2 3 .. I; I-IONE 3...... _......._. .......S 1.... 22: "This really does you good." DRAL‘IING NO. : NOT-TE 112 23: "I'm getting pretty bored with this." DJLA'L‘J INS NO. : l 2 3 17““ NONE 3...... .5...... 7 2h: "g perfect way 23 relax." DRAWII‘IG NO . : l 2...... 3 E l‘AIO‘l‘TE g...... 5...... 7 2S: "Tmaoped again!" DRAWING N O. 2 l 3...... 3 1| NONE g...... AE...... 1...... 26: "This ii really interesting." DRANING NO.: 1 2 3 NONE 3...... ‘5...... 7 27: "£5.2 lazy!" DRA:~JING N O . : 1 9...... 3 1| I‘lOI‘TE .3...... mg...... 7 28-29: 113 SECTION TWO PLEASE WAIT FOR INSTRUCTIONS BEFORE BEG NNING THE NEXT PAGE. In this section, you will be asked in several different ways how you feel about television in general. You will be given instructions on how to mark the scales. An example of the kind of scales used in Section Two is given below as a practice item. NOISY 7 extremely 5 very 3 fairly don't know TV is generally: H 50-50 3 __ fairly doesn't apply 2 very l extremely QUIET 11h PLACE AN "X" ON TEE LADDER STEP YOU FEEL BEST DESCRIBES TV IN GENERAL 30: TV is generally: EXCITING 7 extremely 5 very 5 -— fairly don't know H 50-50 3 fairly doesn't apply 2 very 1 extremely 8 F 31: TV is generally: IN BAD TASTE extremely very fairly don't know doesn't apply WW IN GOOD TASTE IMPORTANT 7 extremely 5 very 3 fairly don't know 32: TV is generally: 5 50-50 3 fairly doesn't apply 2 very 1 extremely UNIMPORTANT UPSETTING extremely 6 very E fairly don‘t know 33: TV is generally: N 50-50 3 fairly doesn't apply 2 very I extremely 115 PLACE AN "X" ON TIDE LADDER STEP YOU F.LL BEST DESCRIBES TV IN GENERAL. 3h: TV is generally: IN TERES TING 7 extreme 1y 6 very 3 fairly don't know E 50-50 3 fairly doesn't apply 2 very 1 extremely UN INTEREST ING 35: TV is generally: DOING A POOR JOB 7 extremely 6 very E fairly don't know IT"— 50-50 3 fairly doesn't apply 2 very 1 extremely DOING A GOOD JOB INF OM~'BITIVE 7 extremely 5 very 3 fairly don't know 36: TV is generally: 5 50-50 3 fairly doesn't apply 2 very 1 extremely NON-INF OBI-EATIVE UNIMAGINATIVE 7 extremely 5 very E fairly don't know 37:'TV is generally: H 50-50 3 fairly doesn't apply 2 very I extremely IMAGINATIVE 38-39: 116 SECTION THREE From each cluster below, pick the statement you agree with most and the statement you agree with least. Go ahead and work this section. CLUSTER ONE A. Too many people are too eager to criticize TV. I think it is actually very good. B. I get more relaxation and enjoyment from the radio or a record player than I do from TV. C. ‘When I am with friends, and I know my favorite TV program is coming on soon, I'm tempted to suggest that we all watch it. I AGREE MOST with statement . (A: B: or 07 I AGREE LEAST with statement . (A, B, or C) CLUSTER TWO A. I'd much rather read a magazine or a book than watch TV. B. Sometimes I like to just drift away into the fantasy world of a TV program for a while and forget all about my own concerns and the things around me. C. I consider TV a real friend and companion. I AGREE MOST with statement . (A: B: or W I AGIWE LEAST with statement . (A: B: or 67 CLUSTER THREE A. I sometimes feel uneasy about watching TV when I know there are other things I should be doing. B. I find that I'm a better person, more informed and up-to-date on things, because of TV. C. I'm so busy with other things, I have little time to watch TV. hO: factor I AGREE MOST with statement (A, B, or C} I AGREE LEAST with statement (A, B, or C7 117 (Half-Page Separate Booklet) SECTION FOUR PLEASE WAIT FOR INSTRUCTIONS BEFORE BEGINNING THIS SECTION. PRACTICE ITEM TYPE OE PERSON: One who lives on a farm. HOW FAVORABIE TOWARD TV ARE PEOPLE LIKE THIS? 9 extremely favorable a ____________ 7 5 sygflzgmg@a__ A... 3 2 l O sygflz@@@¥@a_ extremely unfavorable HON SURE ARE YOU OF HON YOU RATED THEM? I'm almost positive fairly sure a hunch fairly uncertain just guessing Ol—‘NWS" 118 TYPE OE PERSON: One who can't really enjoy something unless it's almost perfect. hl: HON FAVORABLE TOWARD TV ARE PEOPLE LI’E THIS? 9 _e§t£emely_favorahle_ _ 7 5 -Slifihill 3330233319. _ __ .g.__.__ 3 .Slishilx aniaxoaahla _ 2 l O _e§t£emely_unfavgrahle_ 142: HOW SURE mu; you or 11cm YOU RAT"D Twz? h I'm almost positive 3 fairly sure 2 a hunch I fairly uncertain 0 just guessing TYPE OE PERSON: One who would like to make progress in life so long as it doesn't require too much effort. DB: HOW FAVORABLE TOWARD TV ARE PEOPLE LIKE THIS? 9 _e§t£emely_favorahle_ _ .§______ 7 5 -Slishilx 910261212 _ .. .§______ {E_____. ______3 filiahilx 9910261212 _ 2 .1_.___.. O _e§t£emely;unfav2r§hle_ hh: HOW SURE ARE YOU OF HOW YOU RATED THEM? I'm almost positive fairly sure a hunch fairly uncertain just guessing Ol-‘NUJ-E" 119 T;PE OF PERSON: One who easily gives in on issues just to "get along." hS: HOW FA"ORABLE TOHARD TV ARE PEOPLE LIKE THIS? 9 _e§t£emely_favgrable_ _ 7 3L sliahilz iazoaahle. _ _ E__.___ --—--3 -slizhilz Bfiazoaahla _ _1__.___ O _e3c_t£emely_u_nfa_v_orable_ 1R5: HO’J SURE ARE YOU OF ZION YOU RATED THERE? I'm almost positive fairly sure a hunch fairly uncertain just guessing Ol—‘NUJP" TYPE 93 PERSON: One who feels he must be "top dog," first in everything. ’47: HCM FAVOMBLE TG'IARD TV AW PEOPLE LIKE THIS? 9 _e§t£emely_fa_vgrab_le_ _ 7 5 .Sliahilz iazoaaals _ _ -Slishilz 9910361212. _ WW hB: HOW SURE ARE YOU OF HOW YOU RATED THEM? I'm almost positive fairly sure a hunch fairly uncertain just guessing Ol—‘NULF: 120 TYPE OF PERSON: One who has an average education. h9: How malomm TOLJARD TV ARE PEOPLE LIKE THIS? 9 _e§t3emely_favgrable_ _ 7 5 -Slishilz 333038312 _ _. $125-$312 Eniaxoaahla _. WIT! 50: HOW SURE ARE YOU OF HGJ YOU RATED TERI-I? I'm almost positive fairly sure a hunch fairly uncertain just guessing OHMWC’ TYPE 93 PERSON: One who does not like "eggheads" telling other people what is "good" for them. 51: HOW FAVORABLE TOMARD TV ARE PEOPLE LIKE THIS? _e§t£eaely_favarable_ _ -Sliahilz iazoaahls _ ._ -slishilz Eniazoaaéla _. IOIHINIWTIT 52: HOW SURE ARE YOU OF HON YOU RATED TEE}? I'm almost positive fairly sure a hunch fairly uncertain just guessing Oi-‘NUJJ-T' I21 TYPE 9: PERSON: One who always keeps trying to learn new things in order to stay ahead. 53: HOW FAVORABLE TOWARD TV ARE PEOPLE LIKE THIS? 9 _e3_c_t_r_eme_ly_fa_v_o_rab_le_ _ .§______ 7 5 -Slishilx 91033219. _ .. §_—_——_ E_____. .___._..._3 same gniaxoaahlsz _ 2 .1._..__.__ O _e§_tremely_u_r_1fav2rabl_e_ 5h: HOW SURE ARE YOU OF HOW YOU RATED THEE? A I'm almost positive 3 fairly sure 2 a hunch l fairly uncertain 0 just guessing T 9_ PERSON: One who prefers not to make decisions until he has at least some information on a subject. 55: HCM FAVORABIE TCI'IARD TV ARE PEOPLE LIKE THIS? 9 _e§tremely_favgrable_ _ _____5 7 5 -Slishilz iazoaaéls _ _ g..— _.____E 3 1 O -Slishilx mama-212 - 56: HCN SURE ARE YOU OF HON YOU RATED THEM? I'm almost positive fairly sure a hunch fairly uncertain just guessing OI—‘mmfl" 122 TYPE 93 PERSON: One who can completely "goof-off" with no guilt feelings at all. 57: How FAVORABLE TOWARD TV ARE PEOPLE LIKE THIS? _e§t£emely_f2verable_ _. filiahilz Eniazozaala ._ .Sliahilz 9910281219. _. lolplmlWlT 58: H011 SURE ARE YOU OF HOW YOU RATED THEM? I'm almost positive fairly sure a hunch fairly uncertain just guessing OI—‘NWC’ TYPE OF PERSON: One who likes good quality but does not demand perfection. 59: HOW FAVORABLE TOE‘IARD TV ARE PEOPLE LIKE THIS? 9 _e§t£em_e_ly_fa_vgra_b_l_e_ _ 7 5 _Sli£h.’°_11 231025219. _ _ .Slishilx 9910281212 _. WIT! 60: HOW SURE ARE YOU OF HOW YOU RATED THEM? I'm almost positive fairly sure a hunch fairly‘uncertain just guessing OHNWE‘ 123 TYPE OE PERSON: One who would rather "take it easy" than worry about "getting ahead." 61: HOW FAVORABLZ TOMARD TV ARE PEOPLE LIKE THIS? 9 _e£t£e_mely_favgrable_ _ 7 5 -5liihilx 232028212. _ _ :3...... EL...... ‘2______._$l;£hihxTEPERXQESRlE._ 2 .3;____ O extremely unfavorable 62: HOW SURE ARE YOU OF HOW YOU RATED THEM? I'm almost positive fairly sure a hunch fairly uncertain just guessing OHMWC‘ TYPE 93 PERSON: One who is willing to stand up against all odds for what he believes in. 63: HOW FAVORABLE TONARD TV ARE PEOPLE LIKE THIS? 9 _e§tremely_fav2rable_ _ 7 3“— -Slishilz Taxoaagls _. _ _S.l.ish£1z Buazoaahla _ lolplmlwfl“! 6h: How SURE ARE YOU OF How YOU RATED THEM? I'm almost positive fairly sure a hunch fairly'uncertain just guessing OHNWS" 12h TYPE 93 PERSON: One who likes to win, but doesn't get upset if he loses. 6S: HCM FAVORABLE TOUJ'ARD TV ARE PEOPLE LIKE THIS? 9 _e§t£emel_y_fa_v2rgbl_e_ _ 7 5 .Slisihilz 910261212 _ _ 3...... EL...._ 3...... slizhilz amiazoaaals _ 2 l... O _e§t£emely_unfav2rable_ 66: HOW SURE ARE YOU OF HOW YOU RATED THEM? I'm almost positive fairly sure a hunch fairly uncertain just guessing Ol-‘MWJI" TYPE OE PERSON: One who has an education far below average. 6?: HOW FAVORABLE TOWARD TV.ARE PEOPLE LIKE THIS? 9 _e§tr_e_mel_y_fav_or_able_ _ 7 .5,__5 .Slishialz $810921: _ _ _.._____3 .slishialx Eniazozahls _ 2 1 O 68: HOW SURE ARE YOU OF HOW YOU RATED THEM? I'm almost positive fairly sure a hunch fairly uncertain just guessing or-mexr' 125 TYPE OE PERSON: One who acts as a critic for the rest of society because he considers himself an intellectual. 69: HCM FA klORABLE TOWARD TV ARE PEOPLE LIKE THIS? 9 _e§t£emely;favgrable_ _ 7 5 _%%@u£u%@%-_ filiahflz 992033219. _ FIRINIWICH 70: HOW SURE ARE YOU OF HOW YOU RATED THEM? I'm almost positive fairly sure a hunch fairly uncertain just guessing OHMWP’ TYPE 93 PERSON: One who sometimes makes an effort to keep up by learning new things. 71: HOW FAVORABLE TOWARD TV.ARE PE PLE LIKE THIS? 9 _e_)gt£emel_y_f_a_v_o_r_abl_e_ _ 7 5 .sliahfilz 9203611212 _ _ .sliahilz wazoaaéls _ WW 72: HOW SURE ARE YOU OF HOW YOU RATED THEM? I'm almost positive fairly sure a hunch fairly uncertain O_____' just guessing H NU): 126 TYPE 93 PERSON: One who will make snap decisions even if he has very little information on a subject. 73: HOW FAVORABLE TOWARD TV ARE PEOPLE LIKE THIS? _e§tzeme;y_f2v2r2b_l.e_ _ -Sliahilx Eazoaaéls _ _ _Sliah.t.lz 2119109212 _ IOIHINIUWTIT 7h: HOW SURE ARE YOU OF HOW YOU RATED THEM? I'm almost positive fairly'sure a hunch fairly uncertain just guessing Ol-‘NWC' TYPE 9_ PERSON: One who feels very guilty unless he's doing something worthwhile all the time. 75. HOW FAVORABLE TOJARD TV ARE PEOPLE LIKE THIS? _e§t£eTe.1_y_feversb_l.e_ .. filiahilz 910361212 _ _ .8lizhilz Rn£aX°£a212 _ WWW?” 76: How SURE ARE YOU OF HOW YOU RATED THEE/1? I'm almost positive fairly sure a hunch fairly uncertain O just guessing I-‘NUJLT' 127 TYPE OF PERSON: One who is easy to please with most things. 77: HOW FAVORABLE TCMARD TV ARE PEOPLE LIKE THIS? 9 _ez<_t£eTely_faV2rsble_ _ 7 5 -Slishflx Eaxozahls _ _ -Siishflx Eniazoaaéls _ IOIHINHTI 78: HCM SURE ARE YOU OF HOW YOU RATED THEM? I'm almost positive fairly sure a hunch fairly uncertain just guessing OI—‘NUJLT' TYPE 91: PERSON: One who constantly drives himself and pushes to get ahead in this world. 79: HON FAVORABLE TOWARD TV ARE PEOPLE LIKE THIS? 9 _e§t£em_el_y_fav_or_a_b_le_ _ 7 §__5 -Siishflz iazoaahls _ _ 5...... 3 11.135215: 991025212. _ L... O - — - _ 80: HOW SURE ARE YOU OF HOW YOU RATED THEM? I'm almost positive fairly sure a hunch fairly uncertain just guessing OI-‘NW-IZ" 128 TYPE 93 PERSON: One who will sometimes give in on a few points in order to "get along." 1h: HOW FAVORABLE TOWARD TV.ARE PEOPLE LIKE THIS? 9 extremely favorable g.__ ____________ 7 6““?— _Slishilx aazozaals _ _. 11:. 3 2 1 O _slish£lx 2n£a202a212 _ 15: HOW SURE ARE YOU OF HOW YOU RATED THEM? A I'm almost positive 3 fairly sure 2 a hunch l fairly uncertain 0 just guessing TYPE OE PERSON: One who won't compete if it requires a lot of effort. 16: HOW FAVORABLE TOWARD TV ARE PEOPLE LIKE THIS? fistaemelyjavereble. _ .Sliahilz 9109212 .. _ -Slishilx Eniazoaaéls _ ITINIWW 17: HOW SURE ARE YOU OF HOW YOU RATED THEM? I'm almost positive fairly sure a hunch fairly uncertain just guessing 129 TYPE 9: PERSON: One who has an education far above average. 18: HOW FAVORABLE TOWARD TV ARE PEOPLE LIKE THIS? 19: HON SURE ARE YOU OF HOW YOU RATED THEM? h I'm almost positive 3—_—_—- fairly sure a hunch l fairly uncertain O just guessing TYPE OE PERSON: One who listens to the Opinions of critics of our society, but who often disagrees with what the critics say. 20: HOW FAVORABLE TOWARD TV ARE PEOPLE LIKE THIS? 21: HOW SURE ARE YOU OF HOW YOU RATED THEE I? I'm almost positive fairly sure a hunch fairly uncertain just guessing Ol—‘NWC‘ l3O TYPE OE PERSON: One who doesn't see how learning new things will help him advance at all. 22: HOW FAVORABLE TOWARD TV ARE PEOPLE LIKE THIS? 23: HOW SURE ARE YOU OF HON YOU RATED TEETH I'm almost positive fairly sure a hunch fairly uncertain just guessing OHNUJJI“ TYPE 93 PERSON: One who won't make a decision until he has all available information on a subject. 2h: HOW FAVORABLE TOWARD TV.ARE PEOPLE LIKE THIS? 25: HOW SURE ARE YOU OF HOW YOU RATED THEM? I'm almost positive fairly sure a hunch fairly'uncertain O just guessing l-"lNUJE’ 131 TYPE Q: PERSON: One who feels slightly uneasy when he wastes time. 26: HOW FAVORABLE TOHARD TV ARE PEOPLE LIKE THIS? _e1