SOCIAL STATUS AS A PREDICTOR OF COMMUNICATION BEHAVIOR; Thesis for the Degree of M. A. MICHIGAN STATE UNIVERSITY Robert L Van Dam I964. mssxs lllflmllfllflllj’lljllfluflIIIHNIHUNUIUHHIHHIHII 10410 6012 L I B R A R Y Michigan State University SOCIAL STATUS AS A PREDICTOR OF COMMUNICATION BEHAVIOR By. Robert L: Van Dam AN ABSTRACT OF A THESIS Submitted to Michigan State University in partial fulfillment of the requirements _for the degree of MASTER OF ARTS Department of Communication I luli A .\~ E a.“ .fi \ uf.‘ nifl uVI‘ ‘ A . N -C :1 E F. LL i a CL «C .ru a. J I C AL .0.» n t F. t «v C H. s van 1; M. «u e . .5 y . C u. I a C; AV us as at nu .. . :1. .L u .PL rt .1“ .~I§ nu (i\ a “J. r»... ac PU «\u .2: he at. i x 1 \ ABSTRACT SOCIAL STATUS AS A PREDICTOR OF C HHUNICATION BEHAVIOR by Robert L. Van Dam The purpose of this study was to find out what commun- ication variables can be predicted by social status. The instrument used for measuring social status had not been used previously. The present index was designed to measure the relative prestige of persons, based on an occupational rating scale. The types of communication ygxiables that were considered ”‘uo _ m’dfl" ‘- in the present study were: media habits, Opinion leadership, information level, formation of opinions, intensity of attitudes and various channels of interpersonal communication (social interaction with relatives; "friends, neighbors, and people you work with;" and activity in organizations. Also, the relative effect of a mailed message among persons of different status levels was investigated. It was predicted that persons of high social status would be affected more by the message on information level, formation of Opinions and intensity of attitudes than persons of low social status. Social status, as defined for this study, successfully predicted news magazine readership, activity in voluntary organizations and public affairs information level. However, social status failed to predict reading of "main news stories" in new5papers, social interaction with relatives, social interaction with friends, opinion leadership, number of l 13 opinions, intensity of attitudes and kinship orientation. Considering the sample as a whole, the message tended to influence information level and intensity of attitudes. However, the mount of message effect did not vary from status group to status group. In addition, the message had no influence on the formation of opinions that people had. hon-correspondence of the present findings in relation to previous research is discussed. One reason suggested was that measures of social status in earlier studies usually included some index of economic class, as well as prestige. SOCIAL STATUS AS A PREDICTOR OF COMMUNICATION BEHAVIOR By Robert L. Van Dam A THESIS Submitted to hichigan State University in partial fulfillment of the requirements for the degree of MASTER OFIARTS Department of Communication 1964 APPROVED: #3299010, 67/.2' 2/1.“!- Inmr‘rnfj jrxnr‘r-r 1" CI [LULH‘IUAI UJQ'LEs'u Q I wish to express my sincere appreciation to all those who have contributed to the completion of this study. Special thanks go to Dr. Verling C. Troldahl, my thesis adviser, and to Dr. David K. Berle and Cr. Halcolm Kaclean for their assistance. The Department of Communication allowed me to in- tegrate this study into a major field study conducted by the Department, for which I am also grateful. Finally, I want to thank my wife. In addition to typing the manuscript, she was a constant source of understanding and motivation. Chapter I II TABLE OF CONTanf: THE SEA? \CEiT OBLE;I. . . . . . . . Introduction. . . . . . . . . . Communication Behavior. . . . . The Research Problem. . . . . . Problem One: Variety of Definitions 0 O O O O O O O 0 Problem Two: Imprecision of Ins t rument S . Q Q . . . O O 0 Problem Three: Social Status or Opinion Leadership . . . . Problem Four: Lack of Experimental Field Studies . . . . . . . . Hypotheses for Present Study. . Field Experiment Hypotheses . Correlational Hypotheses. . . Rationale for Hypotheses. . . RESEARCH DESIGL‘I o o o o o o o o o 0 Background of the Study . . . . The Sample. . . . . . . . . . . Operationalization of Variables Social Status . . . . . . . . Opinion Leadership. . . . . . Information Level . . . . . . Number of Opinions. . . . . . Intensity of Attitudes. . . . 2:9 d i a o o o o o o o o o o o 0 iii 15 16 16 17 18 20 23 23 24 24 26 28 30 30 30 23 Chapter III IV News Na azine Readership . . . . O' O NeWSpaper Readership . . . . . . Radio and TV . . . . . . . . . . Kinship Orientation. . . . . . Gregariousness . . . . . . . . Organizational Leadership. . . Data Collection. . . . . . . . . Data AnaIYSiS o o o o o o o o o o FIIIDIlQGS O O O C O O O O O C O O O 0 Description of the Sample. . . . Interviewing Success . . . . . Comparability of Experimental and Control Groups . . . . . . Characteristics of the Social Status Groups. . . . . . . . . Tests of Hypotheses. . . . . . . Experimental Hypotheses. . . . Summary of Experimental I‘lypOt-heses o o o o o o o o o o Correlational Hypotheses . . . DISCUSSIO:I o o o o o o o o o o o o o SummarYooooooooooooo DiSCUSSion o o o o o o o o o o 0 Effects of a Single Communication. . . . . . . . . Status and Opinion Leadership . . . . . . . . . . General Communication Behavior iv 31 31 31 32 32 33 33 34 37 39 43 43 48 49 59 62 62 67 69 36 59 Chapter Implications o o o o o o o o o o REFERENCED BIBLIOGRAPHY . o o o o o o APPENDICES o o o o o o o o o o o o o APPENDvaA: Occupational-Prestige Scale Instructions . . . . . . . APPENDIX B: Interviewer-Schedule InStrUCtionS o o o o o o o o o 0 APPENDIX C: Detroit Opinion and Information StUdYo o o o o o o o 73 79 82 86 77 79 LIST CF TABLES Table Page I Information Level and Social Status . . . . . 3 II The Follow-ups Have Huch Higher Occupational Standing . . . . . . . . . . . . 4 III The Flow of Influence in Public Affairs Among Occupational Levels. . . . . . . . . . . 6 IV Distribution of Respondents' Occupational Prestige (Status) Scores . . . . . . . . . . . 25 V Distribution of ReSpondents' Opinion Leadership Scores . . . . . . . . . . . . . . 28 VI Interviewing Success . . . . . . . . . . . . . 36 VII Description of the Sample on Demographic Attributes, by Experimental and Control Groups 38 VIII Description of the Sample on Demographic Attributes, by Status Groups . . . . . . . . . 41 IX Average (Mean) Information Level for ReSpondents at Different Status Levels, by Experimental and Control Groups . . . . . . 45 X Average (Mean) Number of Opinions for ReSpondents at Different Status Levels, by EXperimental and Control Groups . . . . . . 46 XI Average (Kean) Intensity of Opinions for ReSpondents at Different Status Levels, by Experimental and Control Groups . . . . . . 48 XII Media Habits for the Status Groups . . . . . . 51 XIII Information Level for the Status Groups. . . . 53 XIV Social Status Groups and Number of Opinions, Intensity of Opinions, and Kinship Orientation 54 XV Social Status Groups and Social Interaction with Relatives, Social Interaction with Friends, Neighbors, and People Worked With, and Activity in Organizations. . . . . . . . . 56 vi Table XVI XVII XVIII X X Social Status Groups and Officership in Voluntary Organizations. . . . . . . . . . . 57 Average (Kean) Intensity of Opinions for White and Non-White Respondents, by Experimental and Control Groups . . . . . . . . 6o Correlations Between Social Status, Opinion Leadership, and Communication variables 68 Correlations Between Social Status, Education, and Communication Variables . . . . . . . . . . 74 vii Chapter I THE RESEARCH PROBLEM Introduction Previous research has demonstrated that social status is an important factor in determining much of a person's social behavior. For example, Kahl (1961, p. 138) re- ports that, "People tend to associate with persons of similar socio-economic status." Moreover, Pope (1948) states that religion is very closely related to the social stratification patterns of American communities; Kinsey, Pomeroy, and Martin (1948) report that different status groups tend to have different patterns of sexual behavior; and Hollingshead (1950) found that marriage choices tend to occur among prestige-class equals. It has also been found that people who occupy the same relative position on some subjective prestige scale or socio-economic index tend to dress similarly (Katz and Lazarsfeld, 1955, p. 331), have common interests and attitudes (Kahl, 1961, p. 153), and Spend their money in a predictable manner (Kahl, 1961, p. 115). Social status is also a good predictor of voting behavior. As Berelson, Lazarsfeld, and McPhee (1955, p. 333) point out, four studies have shown that the higher the status or occupation, the more Republican people tend to be. Communication Behavior Of more immediate interest, however, social status has also been shown to be a good predictor of several kinds of l .1. . u 2 communication behavior. For example, several studies have shown that social status is a good predictor of the consump- —_—.—-»— __ n ‘ .. ~..--'vu 1,-.9'1'5‘4” tion of print media. Schramm and White (1949) found that urn-n“ 1"" “ ‘___._ persons in the highest economic status group read an average of 22% of the newspaper, while the lowest group read an aver- age of 16%. Moreover, Knupfer (1947) reported that 65% of the low status people in a large sample said that they read magazines, compared to 93% of the high status people. Reissman (1954) found that significantly more people in the high occupa- tional group (68%) read 4 or more magazines regularly than in the low occupational group (39%). The same relationship held for income and educational dichotomies. Reissman also reported that the high occupational group read more books than the low occupational group, but this difference was not statistically significant. The relationship between social status and the electronic media is not as clear. Several studies have reported that higher status people tend to Spend less time listening to the radio and watching television. Others have reported little differences in total time Spent listening and watching, but emphasize the variation in types of programs preferred by the different status groups. The present study will only deal with the habits of people concerning news broadcasts; an area on which the author failed to find any relevant research. In addition to mass media behavior, several studies have shown that social status is related to the amount of infor- mation the people have on various topics. Knupfer (1947) 3 reports on the basis of national public opinion polls that high class individuals are more likely than low class individuals to have knowledge of and interest in national and international affairs. In the two examples cited by the author, the relationship was based on income and educational breakdowns reSpectively. Moreover, Sykes (1951) found that higher class individuals in his sample were more informed about local community affairs than were lower class persons. Further, on the basis of a two-item information test on current events in the news, Katz and Lazarsfeld (1955) found a high positive correlation between social status and infor- mation level.1 (See Table I) Table I * Information Level and Social Status Social status level Information level High Middle and low Total High 72% 28% (197) Middle 52 48 (253) Low 28 72 (246) *Adapted from Katz and Lazarsfeld (1955, p. 274) As readily seen in TABLE I, the higher the social status, the greater the proportion on people who attained a high in- 1 . One of the questions was about the British elections and the other about a proposed Central Highway lan both impor- tant topics in the news at that time. PeOp e who answered both questions correctly were assigned to the high information group. The soc1al status index was based on a combination of rent and education, using the median of each distribution as a cutting point. 4 formation level. However, since only two items were used to index information level, confidence in the results is re- duced. Thirdly, social status has been found to predict channels of personal influence. Katz and Lazarsfeld (1955, p. 283) found that in the area of public affairs, "Persons designated as generally influential are higher-up on the status ladder than the people who named them, and that each successive group of experts stands higher than its predecessor."1 (See TABLE II) Table II The Follow-ups Have Much Higher Occupational Standing Original Expert* Expert* Expert* Sample Group I Group II Group III Business & Professional % 24% 33% 33% High White Collar 21 27 39 47 Low White Collar l9 14 10 l0 Wage Earner 51 35 18 10 Total (=100%) (793) (219) (189) (99) *Expert group I refers to the people who were influentials for the original sample; group II refers to the influentials for the eXpert I group; group III refers to the influentials for the expert II group. 1The generally influential peOple were located by probing an affirmative answer to the question, "Do you know anyone around here who keeps up with the news and whom you can trust to let you know what is really going on?" 5 The relationship remained unchanged when income, intuitive rating of socio-economic status by the interviewers, and breadwinner's occupation were used as the basis for the social status scale. Moreover, as the authors also note, "People are most likely to choose their experts mainly from within their own social group. But as we move up the ladder of prestige to each succeeding expert group, we find a tend- ency toward vertical influence--an increasing probability that persons from higher strata will be designated as general influentials." Katz and Lazarsfeld, 1955, p. 286) However, when the data used for this conclusion is reworked, the differences are found to be very slight. That is, the percentage of persons in the original sample who contacted persons of higher status about broad public affairs issues was only Slightly less than the percentage contacted by the expert groups. (See TABLE III) On the other hand, all three groups confirm the notion that when people do go outside of their own status group, they tend to go to people of higher status. Moreover, When Specific instances of public affairs discussions were analyzed, extra-familial advice seeking tended to be directed toward people of higher status. While 46% of the reSpondents indicated that they sought information from people within their own status level, 35% indicated that the person they talked to had higher status than themselves.1 However, the sample size for that analysis was very small (37 peOple) because most of the discussions of public affairs 1 This data was adapted from Katz and Lazarsfeld (1955) p. 277, Table 36. topics were intra-familial. Table III The Flow of Influence in Public Affairs Among Occupational Levels* Status of person Original Expert EXpert talked to Sample Group I Group II Higher than respondent 28% 34% 30% Same as re- Spondent 6O 49 58 Lower than reSpondent 12 17 12 Total (=1oo%) (239) (I94) (99) *Adapted from Katz and Lazarsfeld (1955) p. 285, Table 41. Rogers (1963) also reported a relationship between social status and opinion leadership. On the basis of several studies reported, it was found that influentials for farmers tended to be higher in status than those who sought information, but not a great deal higher. For example, Lionberger (1953) reported that opinion leaders were more likely to own their own farms, to live on relatively larger farms, to earn higher gross farm incomes, and to have higher community prestige than their followers. Moreover, Lionberger and Coughenour (no date) found that farmers tended to seek advice from Opinion leaders who were slightly above them in social status, but not "out of sight.” Status in this study was based on the average ratings of Sixteen judges in the community. Berelson, Lazarsfeld, and McPhee (1954) reported the same conclusion. 7 Although there were no concentrations Of opinion leaders in different sociO-economic levels, more opinion leaders were found in the higher prestige occupations within the broad dichotomies of white-collar and blue-collar occupations than in the lower prestige occupations. Further, “Within each socioeconomic status level the opinion leader was some- what more likely to come from the better-educated members of the group." (p. 112) However, the differences were very Slight, and no Significance tests were reported. Therefore, there seems to be a tendency to seek advice from people with higher social status, but in most cases there was not a great deal of difference between the two individuals. When looking at the concept of Opinion leadership (influ- entials) more closely, we find that Opinion leadership, like social status, can predict media habits and information level. .flwwwww--- - ~- Lazarsfeld, Berelson, and Gaudet (1948, p. 51) fOund that political Opinion leaders exposed themselves more to radio, newipapers, and magazines than non-leaders did. In a different topic area, Rogers (1963 p. 238) cited several studies which indicated that influential farmers subscribed to a greater number of farm magazines and newspapers, watched more TV farm Shows, and made greater use of Extension Service bulletins than their followers did. Carter and Clark (1962) found that public affairs opinion leaders used books and magazines more Often than non-leaders to learn about "major public issues in the news," but there was no Significant difference in the use of television. Katz and Lazarsfeld 8 (1955, pp. 310-12) found that persons who were designated as influentials read more books and magazines than the original sample. Holding education constant, the authors analyzed readership for opinion leaders in the areas of marketing, fashion, movies, and public affairs. The positive relationship between leadership and readership held for all comparisons, although some of the differences were quite small. KO Significant tests were reported. In addition, 22% Of the public affairs leaders, compared to 14% of the non-leaders, reported they read ngg§_magazines. Opinion leadership is also capable of predicting infor- mation level. Berelson, Lazarsfeld and ncPhee (1954, p. 111) found that 50% of the Opinion leaders had a high score on an information index, while 29% Of those not designated as Opinion leaders had a high score.1 horeover, by recasting the data Of Katz and Lazarsfeld (1955), Troldahl (1963) found that 70% of the public affairs opinion leaders had a high information level and 46L Of the non-leaders did.2 1The information index was constructed from a single question asking the respondents to identify any job that each of nine men had held. The number of correct identifications served as the base of the index. Opinion leadership was based on re- Sponses to the following two questions: "Compared with the people you know, are you more or less likely than any Of them to be asked your views about politics?" and "Have you talked politics wit anyone recently?" Those people who answered "Lore" to the first 93 "Same" to the first gag "Yes" to the second were defined as Opinion leaders. 2 . . . . Women who indicated on both interViews that they had recently been asked their advice (about problems in the news), and those who said that they had recently been asked their advice, on one interview, but who considered themselves "more likely" than others to be consulted, were defined as Opinion leaders. The information index is the same as discussed previously. 9 From the foregoing we see that a person who is recognized as an Opinion leader is more eXposed to the mass media and has a higher information level than those who are not rec- Ognized as Opinion leaders. Social status and Opinion lead- ership are also related to the number of Opinions a person has. Knupfer (1947) reported that one of the most consistent findings of public Opinion polls is the higher proportion Of "don't know" and "no Opinion" in the low status group. The author states that this may result from a feeling of incom- petence to judge, as well as the generally accepted notion Of lack of information. Katz and Lazarsfeld (1955, p. 282, note 11) report that on the basis of ten questions about current news events, there was a great difference in the number Of Opinions held between the women in the Original sample and their Opinion leaders. This finding also held up when educational level was controlled. However, no data were presented. horeover, on the basis of reSponses to six questions concerning political issues, Berelson, Lazarsfeld, and McPhee (1954, p. 111) found that 66% of the Opinion leaders compared with 39% of the others, had opinions on all Six issues. Since Opinion leadership and social status are both re- /\Iafed to the same communication variables, and are themselves related, it might be asked whether it is the opinion leader- ship attribute Or the social status of a person, or both, that makes him influential. DO Opinion leaders influence regard- less of status, or do high status people influence whether or not 10 they are categorized as opinion leaders? The present study [gwill attempt to clarify this conceptual problem. Personal influence, as discussed by Hatz and Lazarsfeld (1955), assumed face-to-face contact. However, people vary in the extent and kind of interaction they have with others. In order to better understand the flow of infor- mation, as well as influence, the direction and channels of face-to-face communication must be determined. Therefore, visiting and friendship patterns become important in communication research. Once again social status has been found to be a good predictor of behavior. For example, R.S. and H.H. Lynd (1929) found that % Of the "Business class" women said they had g9 friends whatever, and another 13% said they had no intimate friends. The corresponding figures for the "working class" women were 13% and 34% reSpectively. Horeover, {aufman (no date) found in a study of a rural New York community that the mean number of in- timate interpersonal relationships outside the household was 6 for the highest prestige group and 2 for the lowest prestige group. Studies by Dotson (1951) and Anderson(l946) further substantiate the notion that lower class individuals. appear to have fewer friends and engage in less visiting among friends than higher class individuals. Dotson (1951) reported that although interaction with Siblings was quite frequent, 40% of the working class people had n9.intimate friends other than kin. Kahl (1955) reported the same find- ing. Jhile 10% of the high status group had 3g close friends, ll 30% of the low status group had none. As Useem, et a1 (1942) point out, the lower class person tends to limit his social interaction more to his immediate family, in contrast to the higher class person who more willingly ranges outside of his own community. Therefore, by deter- mining social status, we should be able to predict fairly accurately whether an individual is likely to interact more with friends or within his own family. Similarly, social status has been shown to predict membership in voluntary organizations, another form of face-to-face communication. These voluntary organizations, along with work groups, neighbors, and family, provide the primary Opportunities for meaningful communication between people. The relationship between social status and organ- izational membership has been explored in several studies. For example, Reissman (1954, p. 80) found that 66% Of the people in the high occupational prestige group, compared with 39% in the low occupational group, belonged to 2 or more organizations.1 The difference between the two groups was statistically significant. However, the same relation- ship did not hold for median cuts on income and education distributions. Reissman also found that individuals with higher occupations held significantly more leadership roles in organizations than individuals with lower occupations (53% compared to 33% who have held or holds Office). 1Amedian cut of the occupations coded by the Horth-Hatt distribution distinguished the high group from the low group. 12 The relationship between Officership and occupation, income, and education were all statistically significant. Although Reissman found no significant differences between occupa- tional groups on the frequency of attendance at meetings, higher educated individuals attended significantly more meetings than lower educated individuals. Wright and Hyman (1958), in analyzing data from national probability samples of 1953 and 1955, also reported a correlation between occupational status and the number of organizations belonged to. Further, Mather (1941) found that the percentage Of men in the income class earning less than $100 per month who had no group affiliations at all was eight times as great as that Of the men in the higher income class. "In every type of group without exception--church, fraternal, service, recreational, patriotic, political, cultural--membership on the part of the lower income class was markedly lower." (p. 380) Equally conclusive are the findings Of Warner and Meeker (1949). While the upper and upper-middle prestige classes had 100% that belonged to one or more associations, only 30% of the lower-lower prestige group belonged. Also the upper prestige group had an average Of 3.6 memberships per family, while the lower-lower group had an average of .7. The Research Problem The foregoing would seem to indicate that a great deal of communication behavior can be predicted by knowing a person's social status. However, there are several problems 13 with current knowledge about the relationship between social status and various communication-related variables. First, the general and operational definitions of social status have varied considerably from study to study, making it difficult and sometimes indefensible to try to compare these sets of findings. Second, even when a Single Operational definition Of social status is used, the measuring instru- ments have Often been quite imprecise, making the observed correlation between social Status and the other variables lower then the true correlation between them may be. Third, since social status and opinion leadership are both capable Of predicting the same type of behavior, and are themselves related, it is not clear which one is the better predictor of communication behavior. Fourth, although the communica- tion behavior patterns of persons in different social strata can be predicted (with some non-chance accuracy) from the general research literature on social stratification, many Of these predictions have not been tested directly in communication Situations. Because the present research effort confronts these problems, each will be dealt with individually. Problem One: Variety of Definitions Although several studies have given extensive attention to the relationship between social status and communication variables, the indices used to measure social status have varied considerably. Part of the problem stems from the fact that much of the research was conducted prior to most of the 14 recent efforts to develop quantified measuring instruments of the different dimensions of class, status, and power. Therefore, some studies have emphasized the prestige aspect, others have emphasized the economic aSpect of social status, while still others have used a combination of the two. Moreover, thl.(l96l, pp. 9-10) lists six separate social status variables that can be defined operationally: personal prestige, occupation, possessions, interaction, class _,—.—o~-~-v_-“. Lsonsciousness and value orientations. Considering Specific studies, Katz and Lazarsfeld (1955, p. 226) used a status index of rent and education in analyzing the variables pertinent to their study of personal influence. On the other hand, Berelson,Lazarsfeld and McPhee (1955, p. 368) used an index of education, bread- winner's occupation and interviewer rating of economic level in their Voting study. Wright and Hyman (1958, p. 289) based their findings on a breakdown of U.S. Census occupa- tional categories, while Reissman (1954, p. 80) compared findings from separate scales Of occupation (based on the Horth-Hatt study), education and weekly income. What is needed is a study that utilizes a single social status index in order to check the relationships already found between social status and the various communication variables. In an attempt to find the basic factors associated with the various methods Of social status classification, (ahl and Davis (1955) took all of the variables commonly thought to be central to social class and submitted them to 15 factor analysis. The authors found two factors that best accounted for the variation among the status variables. The first consisted of those items closely related to occupation, such as source Of income, subject's education, and occupations of best friends. The second factor consisted of those items closely related to the house and residential area Of the subject. Amount of income was found to be a poor measure of either of these two basic factors. They concluded, "If one desired a measure of the over-all complex of class behavior underlying all the variables, a scale of occupations was clearly the most efficient instrument to use." Since many of the previous studies involving communication variables utilized status indices which included some economic aSpects (whether actual income or a similar technique like rent), the Kahl and Davis article provides impetus for conducting a communication study which minimizes the economic aSpect in the status index. The present research effort will use an occupational prestige index to measure the relation- ships between status and the various communication variables. Problem Two: Imprecision of Instruments While the variety of definitions of social status has led to difficulties in generalizing findings, an equally important problem is the lack of precision Of many of the instruments used in previous studies. Recent interest in l6 quantified techniques has led many researchers to re-examine commonly accepted procedures. The present study will check the relationship between social status and many of the variables mentioned above. The social status index to be used should have more precision than the dichotomous and trichotomous status variables generally used to test the relationship between social status and other variables. This precision will also make possible a better check of the linearity of relationships and provide the Opportunity to test hypotheses about the interaction of several of the communication variables in producing certain behaviors. Problem Three: Social Status or Opinion Leadership As pointed out earlier, some question exists about whether social status or Opinion leadership is the attribute which makes a person influential. Is it social status that is predicting the individual's media habits and information level, or is it opinion leadership? The present Study will attempt to clarify this issue by analyzing the relationship between social status and the communication variables, while controlling opinion leadership. For comparison purposes, partial correlations between Opinion leadership and the communication variables, controlling social status, will also be analyzed. Problem Four: Lack Of Experimental Field Studies Although the studies mentioned earlier all dealt with communication variables in one way or another, not one in- l7 corporated an experimental procedure for testing communication- type hypotheses. Since one basic interest in communication stems from the desire to predict behavior in communication situations, it would seem that a study which analyzes the effect of Specific messages on the general pOpulation would be very valuable. One major project which did trace the flow and effect of mediated communication in the general population utilized a technique that was highly unusual. DeFleur and Larsen (1958) studied the effects of various messages dropped from airplanes. However, social status was not one of the variables used in analyzing the flow Of information. Advertising and marketing researchers are also very concerned with the relation between communication and behavior, but the literature on their findings is almost non-existent. The experimental portion of the present study provides an Opportunity to check the effects of mediated communication in a Specific communication situation. Respondents in the sample were mailed a message about civil defense information. The effects Of the message will be measured for each status group. {{ Hypotheses for Present Study The present study will focus on eleven hypogheses. The first three are concerned with the incidence and effect of voluntary exposure to a written message on a public affairs topic. As mentioned above, this type of experimental pro- cedure is seldom utilized in the communication area. The 18 other eight hypotheses have been tested before, in one form or another. However, although social status is believed to be the primary predictor of communication behavior,_ I—‘M _____ , ”H‘s—4. ,_":r:.--o-P|.-- A . - - ~1~-' ‘r _. ..._..- -.. — Opinion leadership will be controlled in all Of the re- maining hypotheses in an attempt to clarify whether a person is asked for advice because Of his social status or because he is an Opinion leader.1 Other control var- iables will be determined on the basis Of zero-order correlations with the social status index. A) Field Experiment Hypotheses: (l) The higher the social status Of a person, the more likely he is to voluntarily expose him- self to a message and learn new public affairs information. ' (2) The higher the social status of a person, the more likely he is tO voluntarily expose himself to a message and develop opinions on public affairs topics. (3) The higher the social status of a person, the more likely he is to voluntarily expose himself to a message and increase the intensity Of his Opinions on public affairs topics. B) Correlational Hypotheses: (1) The higher a person's social status, the more likely he is to be a public affairs Opinion leader. (2) Regardless of opinion leadership, the higher 1 Because of the conceptual problem discussed previously, Opinion leadership will also be correlated with the communication var- iables holding social status constant. These findings will also be reported. (3) (4) (6) (7) 19 the social status of a person, the greater his exposure to print media. (This relation- . -._,~. ,,.1-4:r‘."- -" ' ship should hold for news-magazine reading and for newspaper reading.) Regardless of Opinion leadership, the higher the social status of a person, the higher his public affairs information level. Regardless Of opinion leadership, the higher the social status of a person, the greater the number of Opinions he has on public affairs topics. Regardless of Opinion leadership, the higher the social status of a person, the greater the intensity of his attitudes toward public affairs tOpics. Regardless Of opinion leadership, the higher the social status of a person, the greater his orientation toward non-kin persons. Regardless of Opinion leadership, the higher the social status Of a person, the greater his gregariousness. (The three components Of gregariOusnessé-interaction with relatives, interaction with friends, and organizational activity--will be analyzed separately, in addition to their combined use as the gregar- iousness index.) 20 (8) Regardless of opinion leadership, the higher the social status of a person, thgflgreaterm_ the number of his formal leadership roles ...\ H-———-.. in voluntary organizations. Rationale for Hypotheses The rationale for seven of the eight correlational hypotheses has been discussed in detail previously in this chapter, and will not be covered again. However, the rationale for correlational Hypothesis 5 has not been discussed. Klapper reports on the basis Of pre- vious research that most mass media effect is "reine forcement." That is, mass communication tends to rein- force rather than tO change opinions. Studies quoted earlier in this chapter also suggest that the mass media tend to influence high status persons more than low status persons. Social status is related with media habits, information, and number of Opinions. There- fore, it would seem that the mass media would reinforce the attitudes of high status persons more than low status persons. As recommended by Troldahl (p. 32, 1963), an "intensity" measure was used in this study as a measure of reinforcement: To conclude that reinforcement has taken place, some type of evidence should be provided to indicate that a person's position on the favorable-unfavorable 21 dimension of an attitude is less sus- ceptible to change. For example, if a person feels more ”sure" about the attitude he holds (i.e., attitude "intensity" increases), there would be grounds for saying that the atti- tude was "reinforced" by the message. If this reinforcement effect has taken place, and it occurred because of media consumption, there should be a positive correlation between social status and "intensity" Of attitudes. In addition, the possibility of reinforcement effect has been anticipated as a result of the message on community fallout shelters that was mailed to half of the respondents in this study. This message was pre- dicted to increase the information level of high status persons more than for lower status persons. Likewise, the message was OXpected to increase the number of opinions held by high status persons more than for low status persons. Using a parallel rationale, high status persons should be reinforced more in attitudes they already hold than low status persons if they are exposed to the experimental message. Chapter II R SEARCH DESIGN Background of the Study The Department of Communication at Michigan State Univ- ersity was scheduled to conduct a study for the Office of Civil Defense in the Detroit area in February, 1964. This presented an opportunity to obtain data on the variables involved in the present study. An area probability sample of the general Detroit population was personally interviewed. About half of the people had been mailed a message which contained information about nuclear radiation and community fallout shelters. The people who were mailed the message served as the experimental group and were compared with the people who received no message. For the total sample, the relationship between social status and five dependent variables was studied. Measures of the relationship between social status and three other dependent variables were obtained for the control group only because of the possible effect Of receiving the mailed message. The Sample Respondents for the present study were chosen from the Detroit corporate city and adjacent suburbs by area proba- bility sampling procedures. Thirty-five geographical areas were selected for interviewing. Within each of the 35 blocks or block-clusters, eight households were selected randomly from prelisted addresses, providing a total sample 23 24 of 280 households. Half of the interviews were conducted with the "man of the household" and half with the "woman of the household." When a household contained only men or only women, the interview was Obtained from the person who served as "head Of the household." Operationalization of Variables The variables that were listed in the hypotheses at the end of Chapter I will be defined and Operationalized in the present section. ' Social Status The main variable in this study was social status. Since many Of the previous attempts to index social status lacked precision, a new approach was used in the attempt to obtain better precision. The index was developed by Verling C. Troldahl, who describes the rationale for the procedure as follows: In every-day contact between people, the main stimulus people use for predicting another person's prestige tends to be what that person reports he does for a living. Therefore, a 13-point scale of occupational prestige was develOped by having 21 Lansing, Michigan residents sort 98 occupational descriptions as to their relative "prestige." The occupations had been randomly selected earlier from interviews conducted with a national sample of persons by the National Opinion Research Center. Eleven occupational descriptions which the Lansing re- Spondents distributed about equidistant along the "prestige" scale, and rating,“ w: extreme po with great used as an and instru 25 scale, and for which there was high agreement on ”prestige rating," were used as anchors for a "master scale." The extreme positions on the scale allowed for occupations with greater or less prestige than the 11 occupations used as anchors. (See APPENDIX A for the "master scale" and instructions for use) Occupational descriptions were obtained in the present study with the question, "What kind of work does the main wage earner in yourhousehold do?"1 Three judges coded the occupational descriptions as to "prestige," using the ”master scale" develOped by Troldahl. The prestige rating attributed to each occupation by the judges served as the status value in subsequent analyses. When the three judges differed in their ratings, the average value was used. The resulting distribution is listed in Table IV. Table IV Distribution of Respondents' Occupational Prestige (status) Scores Status % of ReSpondents 00 2% 01 2 02 9 03 24 04 24 05 13 06 9 07 4 08 3 09 6 10 2 ll 2 12 0 00% N=l92 1Instructions for the interviewers and the questionnaire used in the study can be found in Appendices B and C reSpectively. 26 Classification Of the subjects into low, middle, and high status groups was made by Splitting the obtained distribution as closely as possible into thirds. This was done in order to Obtain stable subgroup estimates in subsequent analyses. On the scale of 0 to 12, persons classified "low status" had a score of 3 or less, ”middle status" persons had a score of 4 or 5, and "high status" persons had a score of 6 or higher. The observed median of 4 was below the theoretical mid-point of 6 on the diStribution. However, this median conformed to the findings obtained in the original sorting of 98 occupations, where the distribution of occupations on "prestige" was clustered at the “low prestige" end. Opinion Leadership Previous research has shown that some individuals are more likely than others to serve as influentials for people. That is, people tend to seek advice from some people more than they do from others. In the present study, only "public affairs" Opinion leaders were dealt with. In order to identify "public affairs" Opinion leaders, a self-designating technique similar to that develOped by Katz and Lazarsfeld (1955, pp. 146-48) was used. In an attempt to develop a more precise measure, reSpondents were asked nine questions about their role in discussions of 27 "topics or issues that have been getting a lot of attention in the news."1 Some of the questions were based on those used by Katz and Lazarsfeld (1955) and Rogers (1962, pp. 230-31), and some new questions were added. As a lead-in question, each respondent was asked, "Can you--Off the top of your head--think Of three or four topics or issues that have been getting a lot Of attention in the news lately?" After naming the topics, each person was asked, "During the past week or two, has anyone asked ygg_for your Opinions on any of these topics in the news?" Other questions tapped such things as how Often the respondent was asked for his opinion on tOpics which get a lot of attention in the news, was the respondent more or less likely than his circle of friends to be asked for such Opinions, and how important it was to the reSpondent to be considered a person whose Opinions on these topics are well-founded. In order to check the "opinion leadership" index, an item analysis was done. Correlations between each item and ’the nine-item index varied from .44 to .72. Because all Of the items were positively correlated, all nine were retained in the final index. 1See APPENDIX C, question 33, 40-47 and APPENDIX B; the instruction for these questions. 28 ”Public affairs" Opinion leadership was operationally defined as the total score of the nine items comprising the index.1 Total scores ranged from 0 to 29, with a median value of 11. (See TABLE V) Table V Distribution of ReSpondents' Opinion Leadership Scores Leadership Scores % of Respondents 0-2 7% 3-5 16, 6-8 16 9—11 13 12-14 17 15-17 12 18-20 7 21-23 6 24-26 3 27-29 3 Information Level Two separate information tests were included in the study. One, a measure of "public affairs" information level, was based on current issues that were receiving a lot Of attention in all of the mass media just prior to the time of interviewing L). One item on the "public affairs" information test was as follows: "Astronaut John Glenn recently announced that he was entering the race for:" l . . . . The ”weights” aSSigned to the SpeCific reSponses for each item are indicated in the questionnaire, APPENDIX C. 29 __M_U.S. Representative from Chio ____President of the United States Governor of Ohio X U.S. Senator from Ohio Other topics which the respondents were asked about include violence in the Detroit schools, the poll tax amendment to the United States Constitution, the Hoffa jury-tamp- ering trial, housing discrimination in Detroit, and the Cuban water crises. The "Civil Defense” information test was based on knowledge of topics pertaining to nuclear radiation and community fallout shelters. For example, the reSpondents were asked, "Fallout from a nuclear explosion is composed mainly of radioactive:" ~“__fragments of the bomb itself. _3L~pieces Of dirt stirred up by the explosion. ____water vapor produced by the explosion. smoke particles caused by fire after the blast. Other questions asked about the current state of the community fallout shelter program in Detroit, how long tO stay in a Shelter after a nuclear explosion, the direction that fallout would travel from the blast site, the protection Offered by shelters, and what must be done in order to Obtain protection from radioactive fallout. The information tests were self-administered to the reSpondents. In each Of the tests, information level was operationalized as the score Obtained on a six-item multiple choice test. Each item consisted of one correct and three wrong alternatives. ReSpondents were instructed to choose 30 only one answer for each item and, if not known, "go ahead and guess." Double answers were considered wrong. Number Of Opinions ReSpondentS were asked to react to seven belief statements on the tOpic of community fallout shelters. Before the statements were read, each person was given a card with the responses "agree," "disagree," and "don't know" on it. After each statement was read, the respondent gave the response which best described how he felt about the statement. "Number of Opinions" was operationalized as the total number of "agree" and ”disagree” reSponses to the belief statements. Intensity of Attitudes After responding to each of the belief statements above, the respondents were asked, "How strongly do you feel about your answer?" 3 very strongly 2 strongly l moderately 0 indifferent "Intensity of attitudes" was operationalized as the total score of the reSponses to the intensity questions for the seven belief statements on community fallout shelters. Media Four questions were designed to serve as indices for media habits. Each of the questions was designated to pertain only to media content relevant to public affairs. J of”) 31 Two of the questions involve print media; the other two involve electronic media. News Hagazine Readership--TO Obtain an index of magazine readership, the following question was asked: ”What magazines do you read regularly...that is, three out of every four issues?“ Two probes were used to obtain a fairly complete list of the magazines read by the re- Spondents. The total number of g93§_magazines read reg- ularly (Time, Hewsweek, and U.S. News and World Report) served as the index. Hewgpaper Readershipr-Each reSpondent was first asked how much time, on the average, he Spent reading the dglly_ neWSpaper. Then each respondent was asked, "About how much of that time do you Spend reading the main ggyg stories of the day?” The responses to the second question were used as the index Of newSpaper readership for this study. Radio and TV--Although radio and television were not in- cluded in the hypotheses on media habits, two questions were included in the study for descriptive purposes. For radio, the question was, "About how frequently do you listen to g§g§_br0adcasts on the radio. . .would it be several times a day. . .once or twice a day. . .every other day. . .about once a week. . .or less Often?" For television, the question asked, "How often do you watch news broadcasts on television. . . would it be.more than once a day. . .about once a day. . . 32 every other day. . .about once a week. . .or less often?" It was Specified in the instructions to the interviewers that "gggg broadcasts" were to be taken as the regularly scheduled dgily broadcasts and not the weekly shows. Kinship Orientation Kinship orientation is defined as the tendency to interact with relatives and family, rather than with friends, neighbors, and "people you work with." The index Of kinship orientation for this study was based on two questions: "About how Often, on the average, do you get together with relatives?" and "About how often, on the average, do you get together socially with friends and neighbors. . .or people you work with?" The responses were coded as follows for each question: 3 several times a week 2 once or twice a week 1 once or twice a month 0 less often Kinship orientation was operationally defined as the amount of social interaction with relatives, mlggg_the amount of social activity with friends, neighbors, and people they work with. Gregariousness The gregariousness index for this study is composed of three separate items: (1) the amount of social activity with relatives; (2) the amount Of social activity with friends, neighbors, and people they work with, and (3) the amount of activity in social clubs and organizations. The weights . -- r W 33 for the responses to the first two items are the same as described in the previous section. In order to deter- mine activity in organizations, the reSpondents were first asked what organizations they were "active" in. Then they were asked, "How many meetings have you attended out Of the last four?" for each of the organizations that they had named. The weights for the number of meetings attended were: 0 - 0 Meetings or no organizations 1 - 1-4 Meetings attended (summed over all organizations) 2 - 5-8 Meetings 3 - 9 or more meetings attended Gregariousness was Operationally defined as the sum Of the scores Of the three items comprising the scale. Organizational Leadership Organizational leadership deals with the number Of Offices currently held in the social clubs and organizations named by respondents. After respondents had named clubs and organizations they were active in, they were asked, "Are you an Officer in ...?", adding the name Of each organization that the reSpondent had named. Data Collection The reSpondents designated by the sampling procedures described previously were personally interviewed in their own homes. Fifteen professional interviewers were hired 34 through an interviewing service. The interviewers attended a two-hour briefing session the day beforegping into the field. Interviewing was conducted from February 10 to February 26, 1964. To maximize the completion rate, three "call-backs" were attempted before interviewers were per- mitted to terminate attempts to talk to reSpondentS. ”Call- backs" were made at least half a day apart. Data Analysis The three experimental hypotheses were tested with identical analysis schemes. Only the dependent variables were changed. Each hypothesis was tested with a treatment- by-levels analysis Of variance design. The people who were sent the one-page message on the fallout Shelter program in Detroit served as the experimental group, while the group which received no message served as the control group. The high, middle, and low sub-groups on the social Status scale were considered as "levels." The six groups in this cross-classification were compared as to their mean "civil defense" information level (Hyp. 1), number Of Opinions (Hyp. 2), and intensity Of attitudes toward fallout shelters (Hyp. 3). Significant interactions were predicted for all three experimental hypotheses. That is, it was predicted that the amount of message effect for each of the dependent 'variables increases with higher social status. Since the cell frequencies were not proportional, an "approximate 35 method" Of analysis of variance was used.1 In the remaining hypotheses, a correlational approach was utilized. First, zero-order correlations between social status and the dependent variables were obtained in order to determine which variables Should be controlled out by partial correlations. A Significant correlation between social status and Opinion leadership was Specific- ally hypothesized. Therefore, each of the remaining seven hypotheses were to be tested by partial correlations, holding Opinion leadership constant. As will be explained in the next chapter, this type of control was not needed. The relationship between social status and each of the dependent variables in hypotheses (l) - (8) was predicted to be significantly greater than 0. Three of the correlational hypotheses pertained to variables which represented dependent variables in the field eXperiment. To avoid contamination, correlational hypotheses involving these variables were tested using data from only the control- group reSpondents. 1Walker and Lev (1953, p. 382) Chapter III FINDINGS Description_of the Sample Interviewing Success As previously stated, the original sample consisted of addresses of 280 households in 35 areas of Detroit and adja- cent suburbs. Interviewers were instructed to make four calls at each household before ending attempts to complete the interview. As part Of a field experiment, half Of the people in each area were mailed a message containing information about community fallout shelters (subsequently referred to as the experimental group). The remaining half received no message (subsequently referred to as the control group). A total of 202 interviews were completed. (See TABLE VI) Table VI Interviewing Success Expefimental Control Total Group Group Sample Completed Interviews 69% 68% 69% Unusable Interviews (NO Social Status Information) 2 4 3 Respondent Refused Interview l8 l4 l6 Respondent not contacted ll 14 12 Interviews attempted 100% 100% 100% H=l40 N=140 N=280 36 37 For the experimental group, 71 percent of the attempted interviews were completed; for the control group, 72 percent of the attempted interviews were completed. Three completed interviews from the experimental group and seven from the control group could not be used in this study because Of incomplete social status data. Therefore, the analyses which follow are based on a total of 192 subjects. Comparability of Experimental and Control Groups Table VII presents a description of the sample on several attributes, by experimental and control groups sep- arately. The experimental group was not significantly different from the control group on any Of these attributes.1 Approximately half of the respondents were male, and seven of every ten were White. Three of every ten reSpondents were under 40 years of age, and 8 of every 10 were married and living with their Spouse. Six percent of the respondents were single. While almost half of the peOple interviewed had no children living at home, another fifth Of the re- Spondents had 3 or more children living at home. Half of the total sample had 12 years or more of Schooling. lSex: Chi square = .33, d.f. = 1; Race: Chi square = .01, d.f. = 1; Age: Chi square = 5.63, d.f. = 9; Marital status: Chi Square = 3.16, d.f. = 5, Children at home: Chi Square = 6.02, d.f. = 7; Education: Chi square 4.49, d.f. = 7; and Social status: Chi square = .84, d.f. 2. 38 Table VII Description Of the Sample on Demographic Attributes, by Experimental and Control Groups . . Experimental Control Personal Characteristic Group Group Sex: Male 54% 49% Female 46 _§};_ 100% 100% N=97 H=95 Race: White 72% 73% Other .J§;_ 27 100% 100% Age : 29 or less 10% 9% 30-39 years 21 20 40-49 years 29 27 50-59 years 16 24 60 years and Older 24 _29_ 100% 100% Marital Status: Single % 4% Narried and living with Spouse 79 81 Separated 6 4 Widowed 6 5 Divorced 2 5 Other 0 __l 100% 00% lumber of Children Living at Home: 0 42% 8% 1 l7 l6 2 11 16 3-4 18 9 5 or more 6 8 Single 6 3_ 39 Table VII continued Experimental Control Personal Characteristic Group Group Education: 0-4 years 3% 5% 5-8 years 19 18 9-11 years 27 29 12 years (HS diploma) 29 30 1-3 years Of college 12 12 4 years of college (Degree) 7 2 Over 4 years of college __;i_ 4 100% 100% N=97 N=95 Social Status: Low (0-3) 36% 38% Kiddle (4-5) 35 39 High (6-11) _g_2___ ___.._3__ 100% 100% Characteristics of the Social Status Groups The principle variable in this study was social status. Social status was operationally defined as the rating re- ceived on a 13-point occupational-prestige rating Scale. In testing the experimental hypotheses, and in the presentation of descriptive statistics, reSpondents were split into three status groups and compared. The prestige levels rep- resented in each of these status groups are indicated below by the occupations of some persons at each level: 40 High Lawyer Assistant Public School Superintendent Consulting Chemical Engineer Soil Conservation Worker Credit Manager for Advertising Service Middle Rate Clerk for TranSportation Company Offset Photographer Auto Mechanic Salesman for Thread Corporation Low Boiler Operator for Pickle Factory Service Station Operator Waitress Bean and Berry Picker The occupational composition of the high, middle, and low status groups for this study seems somewhat similar to the descriptions that Warner gave for his upper-middle, lower-. middle and low class categories. Table VIII shows the characteristics of the respondents for the high, middle, and low status groups. The status groups did not differ Significantly as to sex, age, or number Of children living at home. However, significant differences in race, marital status, and education among the status groups were Obtained.1 lSex: Chi square = .96, d.f. = 1; Race: Chi square = 16.64, d.f. = 2; Age: product moment correlation = -.08; Marital status: Chi square = 15.73, d.f. = 2; Children at home: product moment correlation = -.10; and Education: product moment correlation = .55. 41 Table VIII Description of the Sample on Demographic Attributes, by Status Groups Personal Characteristic gfgtus giggi: Sfigus Sex: Male 55% 52% 46% Female 45 48 54 100% 100% 100% {=71 N=71 N=50 Race: White 51% 82% 90% Other 49 18 10 100% 100% 100% Age: 29 or less 4% 13% 16% 30-39 years 20 18 22 40-49 years 39 20 24 50-59 years 19 21 20 60 years and over 18 28 18 100% 100% 100% Marital Status: Single 10% 1% 4% Married and living with Spouse 66 93 8O Separated 11 O 4 Widowed 6 3 10 Divorced 7 2 2 Other 0 l 0 100% 100% 100% Number of Children Living at Home: O 44% 42% 48% 1 13 20 16 2 ll 17 12 3-4 15 13 14 5 or more 9 7 6 Single 8 1 4 100% 100% 100% 42 Table VIII continued p . Personal Characteristic éggtus giggig giggus Education: 0-4 years 7% 4% - 5-8 years 31 17 2% 9-11 years 38 27 14 12 years (HS Diploma) 21 36 32 1-3 years of college 3 12 26 4 years of college (Degree) - 4 12 Hore than 4 years Of college - - _;£_ 100% 100% 100% N=7l K=7l N=50 On race, there was a higher proportion of non-white reSpondents in the lower status group than in the higher ones. Five out of every ten low status persons were non- white, compared with two out of ten and one out of ten for the middle and high status persons. The contingency coefficient for the relationship was .28. The middle status group had the highest prOportion Of persons who were married and living with their Spouse. Ninety-three percent of the middle status persons were married and living with their spouse, compared with 80 percent and 66 percent of the high status and low status persons. It is also interesting to note the relatively high percentage of single and separated persons in the low status group, and widowed persons in the high status group. The contingency coefficient for the relationship between social status and marital status was .28. 43 There were also large differences in education among the three status groups. The high status group had more persons with at least some college education than the other status groups. In the high status group, 52 percent of the reSpondents had at least some college, while the correSponding percentages for the middle and low status groups were 15 and 3. Moreover, only 2 percent of the high status group had eight years Of education or less, as com- pared with 21 and 38 percent for the middle and low status groups. The product moment correlation between status and education was .55 and was highly significant. gggts of Hypotheses Experimental Hypotheses To test three hypotheses concerning the effects of mediated communication in a Specific communication sit- uation, half of the reSpondents in the study were mailed a message about community fallout shelters. They received the message about a week before they were interviewed. The first hypothesis was: H1: The higher the social status of a person, the more likely he is to voluntarily expose himself to a message and learn new public affairs information. The type of public affairs information studied in this case was civil defense information. The "information level" of the subjects was measured by their scores on a six-item multiple-choice test about community fallout shelters. 44 In the analysis, the civil defense information level Of high, middle, and low status persons in the experimental and control groups was compared. If the message induced the respondents to learn new information about community fallout shelters, the experimental group Should have a higher mean information level than the control group. Moreover, if H1 were supported, the difference in means for the high status groups would be greater than the difference in means for the low status groups. The mean information scores for each status level in the experimental and control groups are shown in Table IX. The experimental group had a higher mean information level than the control group, and the differences was statistically significant. This means that the message did induce learning Of new public affairs information. In addition, for both the experimental and control groups, the higher the status, the higher the mean information level. Again 3 the differences were significant. However, the differences in mean information level between the experimental and control aSpondents for each status level were about equal. Therefore, the predicted interaction was not confirmed. It is interesting to note that, among persons who did not receive the message, the low status group had a lower mean information level than the other two status groups.’ About the same pattern is found among persons who received the message, suggesting that the message had just as much effect among lower status people as among higher status people. 45 Table IX Average (Kean) Information Level for ReSpondents at Different Status Levels, by Experimental and Control Groups Low Middle High Status Status Status EXperimental Group 2.1 2.7 2.9 Control Group 1.3 2.2 2.2 Sample Size: Experimental Group 35 34 28 Control Group 36 37 22 Analysis of Variance Summary: Source of Variation S.S d.f M.S F F 95 Between message-no message .8188 1 .8188 14.63 3.90 Between status groups .6286 2 .3143 5.62 3.05 Interaction .022 2 .011 .20 3.05 Error 186 .05595 __.__ H2: The higher the social status of a person, the more likely he is to voluntarily expose himself to a message and to develop Opinions on public affairs tOpics. The index for "number Of Opinions" was based on the total number Of "agree" and "disagree" responses to seven attitude items on community fallout shelters. If the message induced the respondents to develop Opinions about community fallout shelters, the experimental group should have a higher mean number of opinions than the control group. More- over, if H2 were supported, the difference in means for the high status groups would be greater than the difference for the low status groups. Table X presents the mean number of opinions for each status level in the experimental and control groups and the analysis of variance results. The mean number Of Opinions for the six groups Shown in the Table were vir- tually the same. The message did not induce the experi- mental group respondents tO develOp Opinions on community fallout shelters, and the predicted interaction effect for H2 was not confirmed. Table X Average (Kean) Humber of Opinions for Respondents at Different Status Levels, by Experimental and Control Groups Low Hiddle High Status Status Status Experimental Group 6.4 6.4 6.3 Control Group 6.1 6.1 6.5 Sampl§_§iggg Experimental Group 35 34 8 Control Group 36 37 22 Analysis of Variance Spmmary: Source of Variation .8. d.f. I S. F r 95 Between message-no message.0293 l .0293 .68 3.90 Between status groups .0367 2 .0183 .42 3.05 Interaction .0818 2 .0409 .95 3.05 Error 186 .0431 m... — 47 H3: The higher the social status of a person, the more likely he is to voluntarily expose himself to a message and increase the intensity of his Opinions on public affairs topics. "Intensity Of Opinion" was Operationally defined as the sum of the reSponses to questions measuring how strongly the reSpondent felt about his answers to seven attitude items about community fallout shelters. If the message induced the reSpondents to increase the intensity Of their Opinions, the experimental group should have a higher mean intensity than the control group. In addition, if H3 were supported, the difference in means for the high Status groups would be greater than the difference in means for the low status groups. Table XI presents the mean intensity scores for each status level in the experimental and control groups and the analysis of variance results. The experimental group reSpondents had a higher mean intensity of opinions than the control group, and the difference was significant. Therefore, the message successfully increased the intensity Of Opinions for the experimental respondents. In addition, the low status persons in both the experimental and control groups had higher intensity Of Opinions than the higher status groups. However, the relationship was not sig- nificant. The difference in mean intensity between the experimental and control groups was greater for the low status group than the higher status group. The direction of these differences was the opposite of that predicted. Thus, H3 was not confirmed. 48 Table XI Average (Mean) Intensity of Opinions for ReSpondentS at Different Status Levels, by Experimental and Control Groups Low Riddle High Status Status Status Experimenta Group 16.1 15.2 13.7 Control Group 14.2 13.6 13.2 Sample Size: Experimental Group 35 34 8 Control Group 36 37 22 Analysis of Variance §ummaryz Source Of Variation S.S d.f. {.5. F F 95 .___‘ L. ___._— Between message-no message2.9301 1 2.9301 5.62 3.90 Between status groups 2.3298 2 1.1649 2.23 3.05 Interaction .6565 2 .3283 .63 3.05 Error 186 .5216 Summary Of Experimental Hypotheses The message induced an increase in information level, but the effect was just as strong for low status as high status respondents. That is, the low status group was as likely to learn from the message as the high status group. Therefore, the hypothesis was not confirmed. The message also induced an increase in intensity of attitudes. There was a greater difference in the message's effect on the intensity Of attitudes for the low status 49 group than for the middle and high status groups. The difference in effect was Opposite of that predicted, and not significant. Therefore, the hypothesis was not confirmed. The message failed to induce an increase in the number Of opinions. Therefore, it was impossible to test whether high status persons would be more =ffected by the message than lower status persons. The hypothesis was not confirmed. Correlational Hypotheses Eight correlational hypotheses were concerned with the relationship between social status and communication behavior. In each case a significant, positive zero-order correlation was predicted. H1: The higher the social status Of a person, the more likely he is to be a public affairs Opinion leader. A significant relationship between social status and opinion leadership was hypothesized on the basis of pre- vious research. However, the product-moment correlation between status and Opinion leadership obtained in the study was only .03, which was non-significant. Further checking indicated that the low correlation was not due to a lack of linearity. The eta for the relationship was .20, but was not statistically significant.1 Therefore, the hypothesis was not confirmed. Partial correlations controlling Opinion leadership 1F statistic = 1.11, d.f. = 10 and 180. 50 were designed for the analyses because both social status and Opinion leadership have been found correlated with several communication variables. In this way, the re- lationship of social status and the communication var- iables could be determined, holding the effect of Opinion leadership constant. It had also been planned to determine the relationship between opinion leadership and the communication variables, holding social status constant, in order to determine whether social status or opinion leadership was the better predictor of communication behavior. However, since the relationship between social status and Opinion leadership was not significant there was no reason to hold the effect Of Opinion leadership constant. Therefore, the remaining hypotheses were analyzed with zero-order correlations. It had also been planned to control out the effect of any other variable highly correlated with social status. Only reSpondent's education was highly correlated with social status, but, for reasonsto be explained in Chapter 4, education was not used as a control variable. In brief, the rationale involves the logic that education is an inherent part Of social status because it tends to be a prerequisite for attaining certain status levels. H2: The higher the social status Of a person, the greater his exposure to print media. The predicted relationship between social status and IrNadia habits was tested with two forms Of print media,news 51 magazines and new5papers. (See TABLE XII) Table XII Media Habits for the Status Groups . Low Middle Hi h Type Of media Status Status Stgtus News iagazine Readership l or more 10% 8% 32% 0 read 90 92 68 100% 100% 100% 1:71 N=7 H=50 NeWSpaper Readership Main News Stories 0-10 minutes 24% 21% 16% 15-20 minutes 24 27 30 25-40 minutes 21 20 32 45 minutes or more . 31 32 22 100% 100% 100% Radio News Listening Several times a day 49% 45% 54% Once or twice a day 33 32 32 less Often ‘ 18 23 14 100% 100% 100% TV News Viewing _ More than once a day; 48% 455 28% About once a day 44 44 50 Less Often 8 ll _22__ 100% 100% 1005 In the first test, social status and news magazine readership were significantly related.1 coefficient for the relationship was .27. The contingency Almost a third of the high status people read one or more news magazines 1X2: 14.88, d.f. = 2, p(.05, l-tailed test. 52 regularly, compared to a tenth Of the middle and low status people. In the second test, however, amount of time Spent reading the "main news stories" in newspapers was not significantly related with social status. The product- moment correlation was .02. Therefore, H2 was only part- ially confirmed. Although a relationship between social status and the use Of electronic media was not hypothesized, data on exposure to news broadcasts on radio and television were collected for descriptive purposes. Once again the results were varied. There was no significant difference in the frequency Of news broadcast listening for people in the three status groups (r=.04). However, higher status people watched significantly fewer TV news broadcasts than lower status people.1 About a fourth Of the high status people watched news telecasts more than once a day, while almost half of the middle and low status people did. The product- moment correlation was -.18. H3: The higher the social status Of a person, the higher his public affairs information level. Two separate information tests were used tO measure the relationship between social status and public affairs information level. (See TABLE XIII) The first one measured the reSpondent's information level on topics that were lA product-moment correlation of .12 was needed to be significant at the .05 level, one-tailed test. 53 currently receiving a lot of attention in all of the mass media. For this test, the high status group had a mean information level of 4.1 out Of a possible 6, compared to 3.0 and 2.7 for the middle and low Status groups re- Spectively. The product—moment correlation was .35 and was statistically significant. Table XIII Information Level for the Status Groups V _ . Low Middle High nean Information Level Status Status Status Public Affairs Topics 2.7 3.0 4.1 Civil Defense Topics 1.3 2.2 2.2 The second information test measured the respondents knowledge of information about nuclear radiation and community fallout shelters. Since the experimental group respondents were mailed a message which contained the annswers to all of the items used in this test, only the cxnmtrol group respondents were used in this analysis. Once again the difference between status groups was significant. Thezlnigh and middle status groups had a mean information level Of 2.2 out Of a possible 6 compared to a mean Of 1.3 for the low status group. The product-moment correlation was .28. 54 H4: The higher the social status of a person, the greater the number Of Opinions he has on public affairs topics. H5: The higher the social status Of a person, the greater the intensity Of his attitudes toward public affairs topics. Because the reading Of the mailed message might have influenced responses to the questions comprising the indices for the variables in Hypotheses 4 and 5, only the control group reSpondents were used for these two analyses. For H4, the status groups did not differ significantly with reSpect tO the number of opinions about community fallout shelters. The product-moment correlation was .07. (See TABLE XIV) Table XIV Social Status Groups and Number Of Opinions, Intensity of Opinions, and Kinship Orientation . Low {iddle High variable Status Status Status Mean Number of Opinions 6.1 6.1 6.5 Mean Intensity Of Attitudes 14.2 13.6 13.2 Mean Kinship Orientation 2.9 2.9 2.7 However, there was a nggative correlation between social status and intensity Of attitudes, but it was not statis- ‘tically significant. The mean intensity for the low status gyroup was 14.2, compared with the middle and high status Ineans of 13.6 and 13.2. Although the relationship was not ' 55 significant, the resulting correlation was in the Opposite direction Of that predicted. The product-moment correlation was -.l6, but was not significant because only the control group was used in the analysis. H6: The higher the social status Of a person, the greater his orientation toward non-kin persons. The kinship index was Obtained by subtracting the amount of social interaction with friends, neighbors, and peOple worked with from the amount Of social interaction with relatives. Data pertinent to this hypothesis are also presented in Table XIV. As readily seen in the Table, the mean kinship orientation for the three status groups was virtually the same. The product-moment correlation was not significant and was —.03. H7: The higher the social status Of a person, the greater his gregariousness. The index for this hypothesis was intended to be the sum Of the reSponses tO three items: amount Of social interaction with their relatives; amount Of social inter- aaction with their friends, neighbors, and persons they vwark with; and activity in social clubs and voluntary <3rganizations. However, when an item analysis was performed CH1 the data, there was no relationship between the iandividual items. Therefore, the relationship between snacial status and each of the individual items was determined. (See TABLE xv) ITSignificant level was .17 at the .05 level, one-tailed test. 56 Table XV Social Status Groups and Social Interaction with Relatives, Social Interaction with Friends, Neighbors, and People Worked With, and Activity in Organizations . Low Middle High Dependent Variable Status Status Status Mean Interaction with Relatives 1.5 1.5 1.5 Mean Interaction with Friends 1.6 1.6 1.8 Mean Activity in Organizations 1.8 2.0 3.3 There were no significant differences between the status groups for social interaction with relatives (r=.00) or for social interaction with friends, neighbors, and people worked with (r=.05). However, as Table XV shows, there was a significant relationship between social status and activity in social clubs and organizations. On a nine-point activity scale, the high status group had a mean Of 3.3; the middle and low status groups had means of 2.0 and 1.8. The product-moment correlation was .20. H8: The higher the social status Of a person, the greater the number Of his formal leadership roles in voluntary organizations. Formal leadership in voluntary organizations was ‘measured by asking respondents if they were an Officer in each of the clubs and organizations they had named in a previous question. 57 Table XVI shows that there was a significant difference in the number of people who held one or more offices for the three status groups.1 However, contrary to the hypothesis, the relationship was curvilinear. Thirty percent of the high status group reported they held one or more offices, compared with 20 percent and 4 percent for the low and middle status groups reSpectively. The contingency coefficient was .27. Table XVI Social Status Groups and Officership in VOluntary Organizations Low Middle High Status Status Status Officership ' One or more 20% 4% 30% None 80 96 70 100% 100% 100% N=71 N=71 N=50 TO summarize, only one of the correlational hypotheses was confirmed. Social status successfully predicted public affairs information level. Two other hypotheses were partially confirmed. The first was the relationship between social status and media habits. Although social status predicted news magazine readership, there was no significant difference in the amount of time Spent reading the "main news stories" in neWSpapers for the three status groups. 1x2 = 15.16, d.f. = 2, p<.05, l-tailed test. 58 In addition, the relationship between social status and the components Of the gregariousness index was partially confirmed. The status groups did not differ significantly as to interaction with their relatives or interaction with friends, neighbors, and people worked with. However, social status did predict the amount Of activity in organizations. Five hypotheses were not confirmed. Social status did not predict public affairs Opinion leadership,number of Opinions on a public affairs topic, kinship orientation, or intensity of attitudes on a public affairs tOpic. More- over, formal leadership in organizations was significantly related with social status, but curvilinearly. Chapter IV FIHDIHGS Summar The present research effect was designed to evaluate the relationship between social status and several commun- ication variables. Specifically, eleven hypotheses were derived from previous research and tested in a field study in Detroit, Michigan. Half of the reSpondents in the study were mailed a message containing information about community fallout shelters, and three of the hypotheses were based on predictions of the effects Of this message. In all three cases, higher status people were expected to be more affected by the message than lower status people. For the remaining eight hypotheses, social status was expected to be significantly and positively related to the communication variables. A significant correlation between social status and public affairs information level was predicted. Respondent's information level on (1) tOpics receiving a lot Of attention in all of the mass media and on (2) community fallout shelters were both significantly correlated with social status. Since the message sent to half the respondents contained information on community fallout shelters, the relationship between social status and information level also became testable in a specific communication situation. Because, in general, higher status people tend to have a higher information level, it was expected that higher 59 60 status people would learn more in a given communication situation. However, in the present study this hypothesis was not confirmed. Using the control group as a baseline, the message seemed to have just as much effect on information level among low status persons as among higher persons. Similarly, past research has demonstrated that high status persons tend to have a greater number of opinions than low status persons. This relationship was checked in the present study, but the relationship was not confirmed. High status persons did ggE_have a significantly greater number Of Opinions about community fallout shelters than low status persons. The relationship between social status and number Of Opinions could not be tested in the given communication situation. Because the message failed tO induce the people who received it to develop a greater Jaumber Of Opinions, it was not possible to determine xvhether a message induces high status persons to develop ea greater number Of new Opinions than low status persons. Since previous research has demonstrated that high astatus people are more exposed to the print media, have a Inigher information level, and have more Opinions than 1£HJ status people, it seemed reasonable to predict that hjégh status people would also feel more strongly about their (OpiJliOnS. The results of the present study were directly conuuter to this hypothesis. The lower status persons had Inorxa intense attitudes about community fallout shelters 61 than higher status persons. However, the relationship was not significant. For the Specific communication situation, the message successfully induced an increase in intensity of attitudes. However, there were no significant differences among status groups in the amount Of message effect on intensity of attitudes. The intensity Of attitudes toward community fallout shelters among low status persons increased a little more than it did among high status persons, but not significantly. Therefore, the hypothesis was not confirmed. Previous research findings indicated that social status and Opinion leadership were related. Since new measures Of social status and Opinion leadership were used in the present study, it was decided to check the extent of correlation between these two variables with the new measures. The results indicated no relationship between social status and Opinion leadership with the new measuring instruments used. A relationship between social status and exposure to print media was also hypothesized from previous research findings. Significantly more higher status persons read one or more news magazines regularly than lower status persons, but there was no significant difference between the status groups for the amount of time Spent reading the "main news stories" in newSpapers. Therefore, the hypothesis was only partially confirmed. Kinship orientation and formal leadership were also 62 predicted to be related with social status. The status groups did not differ significantly with reSpect to kinship orientation. Thus the hypothesis was not confirmed. The relationship between social status and formal leadership was Significant, but this finding is deceiving because a curvilinear relationship exists. Therefore, again the hypothesis was not confirmed in its present form. Some qualification of the hypothesis, considering the curvilin- earity, is needed. Social status was also predicted to be related with gregariousness. However, since the individual items comprising the scale were not themselves intercorrelated, the relationship with social Status was checked for each item separately. There was no significant relationship between social status and amount of interaction with relatives, or between social status and amount Of inter- action with ”friends, neighbors, and people you work with." However, higher status people had a significantly higher activity level in social organizationsthan lower status persons. Therefore, the hypothesis was confirmed only for this one type of gregariousness. Discussion Effects Of a Single Communication A printed message about community fallout shelters ‘was mailed to half the people in the study. It was expected that the message would induce the people who received 63 it to learn the information on fallout shelters, to develOp Specific Opinions on fallout shelters, and to increase their intensity of attitudes about community fallout shelters. This expectation was fulfilled for information level and intensity of attitudes, but not for number Of Opinions. It seems that the people had already formed Opinions on most of the attitudes tested, Therefore, there was very little variability in the total number Of Opinions that people had about community fallout shelters. Since the people already had a large number of Opinions on the subject, there was a little Opportunity for the message to have an effect. Therefore, a "ceiling effect" may have precluded much message effect on Opinion formation. In addition to the experimental effect Of the message, it was hypothesized that the amount Of effect on persons Of different status levels would not be the same. Higher status persons who received the message were expected to be more influenced by it than lower status persons. However, the amount Of effect induced by the Inessage did not differ significantly from one status level to another for any Of the three criterion variables. Low'status persons were affected by the message about as mulch as high status persons. What might have led to this result? One explaination Inight be that low status persons were more favorable than lrigh status persons toward community fallout shelters, txefore they received the message. However, evidence 64 Obtained in this study from the no-message group, indicates no significant differences between the status groups on favorability of attitudes toward community fallout shelters (r=.10). Therefore, this explanation must be rejected. When a Specific message on civil defense was mailed, the message seemed to have increased the information level of low status persons just as much as it did for higher status persons. If this is what happens whenever a message is sent, why do high status persons tend to have a higher information level than lOW'StatuS persons? (The correlation between status and civil defense information level in the control group was .28). ,This discrepency between the two types of findings could be explained by greater retention of information among high status persons. Because the interviewing for the present study was done - about a week after the message was mailed, there was little Opportunity for high status persons to demonstrate that they retain information for longer periods Of time than low status persons. Since most of the persons already had a large number of Opinions on the subject Of community fallout shelters, the test Of the hypothesis concerning formation Of Opinions 'was not adequate. In order to determine the effect Of a specific message on the formation Of Opinions among people Of different status levels, another study on a topic which allowm room for change will have to be done. The result for intensity of Opinions was totally un- expected. Among persons who did not receive the message, 65 low status people tended to feel more strongly about community fallout shelters than high status people, but the relationship was not significant (r=-.l6). Among persons who received the message, low status people had a sig- nificantly higher intensity than high status people (r=-.26). However, the difference between these negative correlations, presumbly due to the message effect, was not statistically significant.1 Further investigation of the data Showed that there was a significant relationship between intensity Of Opinions and race. Since race was also significantly correlated with social status, it seemed that the message effect on intensity of Opinions across status groups may have been affected by race. In fact, a significant difference in the amount of message effect on intensity of attitudes among Whites and non-Whites was Obtained. The mailed message failed to induce any change Of intensity for non-White persons, but had a marked effect on the intensity Of White reSpondents. (See TABLE XVI) Since there are generally large numbers Of non-White persons in the urban areas, the finding suggests that race should be an important consideration in future field studies vflnich measure the effects Of mediated communication. Because Of these findings on race, an analysis Of 'variance was used to test the amount Of message effect on :Lntensity Of attitudes across status groups for only the 1Z = .7; p = .24, one-tailed test. 66 Table XVI Average (Kean) Intensity of Opinions for White and hon-White ReSpondents, by Experimental and Control Groups Race fiessage HO Message lean intensity for White reSpondents 14.7 12.9 Kean intensity for , , non-White reSpondents 16.0 10.0 Sample Size: r hhite 70 O9 non-White 27 20 Analysis of Variance Summary: Source 8.5 d.f H.S. F F 95 Message 27.8427 1 27.8427 1.78 3.90 Race 124.6162 1 124.6162 7.95 3.90 Interaction 88.5282 1 88.5282 5.64 3.90 Error 2948.1327 188 15.6816 White respondents in the sample. Although the "error" variance was reduced, and the mean square for the inter- action between social status and message effect was increased, the difference in amount Of effect on intensity of attitudes . . . .3. 1 from status level to status level was still not Signilicant. In conclusion, though the findings were not significant, lower status persons generally seemed to be a little more sure of (i.e., intense in) their Opinions, and when exposed lF=2.l6, d.f.=2,l33; a confidence level Of about 12; 67 to a message about the attitude Object tended to increase ’this intensity more than higher status persons. This tentative finding seems worthy of further study under more highly controlled conditions. When studied, race should probably be controlled or held constant. Status and Opinion Leadership The Opinion leadership index which was used in this study was made up of questions derived from Katz and Lazarsfeld (1955) and Rogers (1962), along with some new ones. The individual items correlated between .44 and .72 with the total score for the nine items. Furthermore, the nine items used in this study included most Of the items used in the earlier studies that found a significant correlation between public affairs Opinion leadership and social status. Also, opinion leadership was sig- nificantly related with several other variables commonly referred to as correlates of Opinion leadership in earlier studies. (See TABLE XVII) For example, Katz and Lazarsfeld found a relationship between Opinion leadership and reading Of news magazines. In the present study, a correlation Of .23 between these two variables was found. Moreover, similar to the studies by Berelson, Lazarsfeld, and HcPhee (1954) and Katz and Lazarsfeld (1955), the present study found opinion leadership and public affairs information level to be related (r=.18). Therefore, the (a CC) unexpected findings, zero correlation between opinion leadership and status, must be due to differences in the way status was measured, not in how Opinion leadership was.) Table XVII Correlations Between Social Status, Opinion Leadership, and Communication Variables Social 5 Opinion Communication Variables L// , Status g Leadership News magazine readership~’ , .26 3 .23 Kain news story readership I .02 ; .13 Radio news listening pt ; .04 f .05 TV news viewing ‘ -.18 i -.10 Interaction with relatives" f .00 1 .18 Interaction with friends . 1 .05 g .15 Officership 5 .15 } .23 Organizational activity Q .20 g .23 Opinion leadership 1 .03 ; -_- Kinship orientation 2 -.03 .04 Intensity of attitudes* 1 -.16 1 .22 Number of opinions* ; .07 g .18 Social status 1 --- n .03 Civil defense information* 1 .28 , .04 Public affairs information g .35 2‘ .18 .1 K Correlations reported are for the control group only. How do the social status measures of past studies compare with the new index used in this study? Eany Of the previous studies have used social Status indices that included economic attributes as well as education. For example, Katz and Lazarsfeld (1955) used an index of rent and education. Berelson, Lazarsfeld, and HcPhee (1954) included interviewer ratings Of economic level, in addition to education and occupation Of breadwinner. Xany of the studies cited by Rogers (1962) used either the size 69 Of farm or annual income as part or all Of the social status index. Since the index for social status used in the present study was based on "prestige" ratings Of occupations, the economic aspect (i.e., class) Of socio- economic status was minimized. The Katz and Lazarsfeld study and the Berelson, Lazarsfeld and McPhee study included education as another attribute in their social status index. Although education is Often a prerequisite for certain occupations, the "prestige" aSpect Of the index used in the present study should include attributes in addition to education. Opinion leadership was correlated with social status in several studies. Host of the studies used some sort Of economic index for measuring social status. Therefore, Opinion leadership may be related to the economic attribute Of social class, not to the "prestige" dimension. However, since the present study did not include any measure Of economic standing, that possibility cannot be investigated here. Identification Of the attributes of social status which are related to opinion leadership would be a good topic for future research. General Communication Behavior Independent Of the effect of the mailed message, positive relationships between social status and public affairs infor- rnation level, number Of Opinions about public affairs topics, .and intensity Of attitudes about public affairs tOpics were 70 predicted. Only the control group was used to analyze the data for these three hypotheses because Of possible changes induced when the message was received. Only the relation- ship between social status and information level was confirmed. A negative relationship between social status and intensity of attitudes approached significance. Lower status persons tended to feel more strongly about community fallout Shelters than higher status persons. Since non- White respondents had a higher mean intensity Of attitudes than White reSpondents (16.0 compared with 12.9), and were heavily represented in the low status group, the tendency for a negative relationship is partially explained by race. Whether the negative relationship between status and intensity will hold for other public affairs issues will have to be checked in further research. Although there was some difference in intensity Of attitudes between the status groups, there was no difference in the mean number Of Opinions that people in different status groups had about community fallout shelters. Only six percent Of the total sample had fewer than 5 Opinions out of a possible 7. Therefore, the low correlation between status and number Of opinions may have been due ‘to the low variance caused by a "ceiling effect." The ‘relationship between social status and number of Opinions should be tested on a different topic for which people luive fewer formulated Opinions. 71 A relationship between social status and exposure to print media was also hypothesized. The hypothesis was confirmed for mean number of news magazines read. However, the relationship for amount of time Spent reading the main news stories in newspapers was not. The index of neWSpaper readership was admittedly gross. The reSpondents were asked how much time they Spent reading the "main g§g§.5tories of the day." Since there are many non-public affairs tOpics in the news section, the question may not have been able to differentiate people who were oriented toward public affairs news stories from those who were not. A relationship between social status and kinship orientation was also hypothesized. Previous research indicated that low status persons had stronger kinship ties than high status persons. Iowever, the status groups in this study did not differ with reSpect to amount Of inter- action with relatives rather than friends. This finding may be due to the type Of index used in measuring the attribute. Only two items were involved in the index. These items were designed to Obtain rough estimates of the ratio of amount of interaction with family to that of friends. NO attempt was made to ascertain their preference on who tO interact with. Since this study was done in an urban area, the preference for social interaction with relatives may not have coincided with availability Of these persons. This may have caused the lack of kinship orientation for the lower status persons. 72 It also seems possible that the predicted pattern of interaction no longer holds. Television and other activ- ities may have reduced the frequency of interaction with relatives and friends, and also the difference between status groups in their preference for interaction with relatives or friends. Further study using better instru- ments might be useful in clarifying this relationship. A relationship between social status and gregariousness was hypothesized. Since the combined gregariousness index did not hold up, the three components Of the planned scale were analyzed separately. Social status did predict activity in clubs and organizations, but failed to predict interaction with relatives or interaction with “friends, neighbors, and people you worked with." In addition tO the comments above, it is quite possible that social interaction with relatives and friends is predicted by the economic aSpect Of status rather than prestige. The final hypothesis predicted a relationship between social status and formal leadership in social organizations. Although the relationship was significant, it was curvi- linear. Thirty percent Of the high status group currently held one or more Offices, compared with 4 percent of the Iniddle group and 20 percent of the low status group. One ‘possible explanation for this finding might be that the low status persons held offices in different kinds Of 73 organizations. However, the data on type of organization named did not Show a predominance Of low status reSpondents in any one type of organization. On the other hand, the high status group reSpondents were predominantly members of public service organizations (30 percent)1 and church groups (38 percent). It might be noted, however, that the finding would have conformed to that of Reissman (1954) if the data had been analyzed for dichotomous groups as it was in the previous study. Implications The findings Of the present study did not support the notion that social status is a good predictor Of many types Of communication behavior. Only one prediction was fully supported, and two others were partially supported. On the other hand, eight hypotheses were not confirmed. A social status measure of occupational prestige ‘was utilized in order to Obtain an index which provided something more (prestige) than just the educational contribution to certain behaviors. Since education is inherently part Of occupational success and prestige, 130 attempt was made to control out the effect Of education in measuring the relationships between social status and 'the communication variables. The product-moment correl- 1ation between social status and education was .55. However, :as Table XVIII shows, education was a better predictor tflnan social status for most Of the communication variables :studied. Therefore, it seems possible that education may '71Examples include? Red Cross, civic clubs and charities. 74 have been reSponsible for the significant relationships between status and communication behavior found in the present study. In addition, while social status and opinion leadership were not significantly correlated in this study, education and opinion leadership were. Since the Katz and Lazarsfeld (1955) study used a status index including education, the findings would be comparable. Education seems to be a better predictor of communication behavior than prestige, and is much easier to Obtain than the occupational prestige index used in the present study. Table XVIII Correlations Between Social Status, Education, and Communication Variables Communication Variables Social Education Status News magazines read .26 .34 Main news story readership .02 .00 Radio news listening .04 .08 TV news viewing -.18 -.ll Interaction with relatives .00 -.Ol Interaction with friends .05 .15 Officership .15 .20 Activity in organizations .20 .27 Opinion leadership .03 .18 Kinship orientation -.03 -.11 Intensity Of attitudes -.16 -.09 Number Of Opinions .07 .04 Education .55 --- Social status --- .55 Civil defense information .28 .28 Public affairs information .35 .42 The findings Of the present study indicate that social status (prestige) is capable Of predicting public affairs 75 information level,news magazine readership, and amount Of activity in voluntary clubs and organizations. These findings support previous research. However, education was a better predictor of each of these variables than status was. The insignificant findings in the present study for the relationship between social status and communication behavior does not necessarily contradict the findings of previous research. host Of the studies involving commun- ication variables used a status measure that included economic class. The present study attempted to relate the "prestige" attribute Of status to behavior. Evidently the "prestige" aSpect of social status adds little predictive power beyond that explained by education. It is possible that the economic aSpect of social status is a better predictor Of communication behavior than education or "prestige." However, a study comparing the separate contributions Of the various attributes Of status for predicting communication behavior must be done before any such generalization can be accepted. Several studies are suggested by the findings of the present research effort. For example, more study Of the effect Of Specific communications seem necessary. Find- ings from the present study suggest that lower status persons learn as much from a mailed message as higher status persons do under voluntary exposure conditions. These results 76 were different from correlational findings between Status and information level. What is needed is a multi-phase study designed tO measure both learning and retention Of information. If the Specific message involves an area for which people have few formulated Opinions, the effect of the message on number of Opiniormifor different status levels can also be measured. Regardless Of the topic used in the study, the intensity Of attitudes can also be studied. Since there had been little research on intensity Of attitudes, eSpecially involving Specific messages, rep- lication that takes account of other personality and social structural variables is needed to clarify the concept Of intensity. Results of the present study indicate that high status people have a higher information level than low status people, but low status people have a higher intensity Of attitudes than high status people. Is this negative relationship between status and intensity Of attitudes result simply from a tendency among low education (thus generally low status) persons to be more sure Of themselves generally, or is something more Specific involved? A study is needed to explain this negative relationship. 10. 11) 12. 1.3. 77 REFERENCED BIBLIOGRAPHY Anderson, W.A. "Family Social Participation and Social Status Self-Ratings," American Sociological Review, June, 1946, 11, 253-258. Berelson, Bernard, Lazarsfeld, Paul F. and HcPhee W. Voting: A Study of Opinion Formation in a Presidential Campaigg, Chicago: University of Chicago Press, 1955. , Carter, Roy E. and Clark, Peter "Opinion Leadership Among Educational Television Viewers " 59ericap Sociological Review, December 196 , 7, 792-799. De Fleur, Kelvin L. and Larsen, Otta W. The Flow of Information, Hew York: Harper and Brothers, 1958. Dotson, Floyd "Patterns Of Voluntary Association Among Urban Working-Class Families," American Sociological Review, October, 1951, 16, 687-693. Hollingshead, August B. "Cultural Factors in the Selection Of Marriage hates," American Sociological Review, October, 1950, 25, 619-627. Rahl, Joseph A. The American Class Stgucture, new YOrk: Holt, Rinehart and Winston, 1961. Kahl, Joseph and Davis, James A. "A Comparison of SociO-Economic Status," American Sociological Review, June, 1955, 20, 317-325. Katz, Elihu and Lazarsfeld, Paul Personal Influence, Glencoe, Illinois: The Free Press, 1955. Kaufman, H.F. Prestige Classes in a New York Rural Community, Cornell University, Agricultural EXperiment Station, hemoios 200. Kinsey, Alfred C., Pomeroy, Wardell B. and Hartin, Clyde E. "Social Level and Sexual Outlet," in Class, Status and Power, ed. by Reinhard Bendix and Seymour hartin Lipset, Illinois: The Free Press, 1953, 300-308. Knupfer, Genevieve "Portrait of the Underdog,” Public Opinion Quarterly, Spring, 1947, 103-114. Lazarsfeld, Paul F., Berelson, Bernard and Gaudet, Hazel The People's Choice, New YOrk: Columbia University Press, 1948. 14. 15. 19. 20. 21. 22. 23. 24. 25. 78 Lionberger, Herbert F. "Some Characteristics of Farm Operators Sought as Sources of Farm Information in a kissouri Community," Rural Sociolop , 1953, 18, 327-338. Lionberger, Herbert F. and Coughenour, Milton C. Social Structure and Diffusion of Farm Information, Columbia, missouri Agricultural EXperiment Research Bulletin, 631. Lynd, R.S. and H.I. Hiddletown, new Yorl: Harcourt Mather, William G. "Income and Social Participation," American Sociological Review, June, 1941, o, 3BO-3d3. Po e Liston "Relicion and the Class Structure " , C) 9 The Annals of the American Academy of Political and Social Science, march, 1943, 54-91. Reissman, Leonard "Class, Leisure and Social Partici- pation," American Sociological Review, 1954, 19, 70-64. Rogers, Everett Diffusion of Innovations, New York: The Free Press of Glencoe, 1952. Schramm, Wilbur and White, David "Age, Education and Economic Status as Factors in Newspaper Reading" in Schramm, Wilbur fiass Communications, Illinois: University of Illinois Press, 1949. "The Different Distribution of Sykes, Gresham K. Social Forces, Hay, 1951, 29, Community Knowledge," 376-382. Troldahl, V.C. Yediated Communication and Personal Influence: A Field Experiment. Doctoral dissertation, University of Linnesota, 1963. (2 vols.) Useem, J. and others "Stratification in a Prairie Town," American Sociological Review, 1942, 7, 331-342. Walker, Helen K. and Lev, Joseph Statistical Inference, flew York: Holt, Rinehart and Winston, 1953. ‘Warner, w. Lloyd and others Democracy in Jonesville, new York: Harper, 1949, Chapter 12. wright, Charles and Hyman, Herbert ”Voluntary Association liemberships of American Adults: Evidence from National Sample Surveys," American Sociological Review, June, 1958, 23, 284-294. APPEHDIX A OCCUPATIONAL-PRESTIGE SCALE INSTRUCTIONS We want to know how much prestige different people have. One way to estimate a person's prestige is by obtaining a description of his occupation. This infor- mation is useful because, for most people, certain jobs carry more prestige than other jobs do. For example, say that you meet two people for the first time. You find out that they have different types of jobs. Even if you know nothing about them but their jobs, you often get a feeling that one of these persons has more pres- tige than the other one. Another way to put it is that people tend to show more reSpect for persons in some jobs than for persons in other jobs. We want you to read some descriptions of the occupations held by several persons we studied recently. We need your estimate of how much prestige each occupa- tion has. To assist you, we have already had some people rank several occupations from high prestige to low pres- tige. From this, we have prepared an "Occupational Prestige" scale, which varies from 0 (very low prestige) to 12 (very high prestige). (See page 3) At each point on the scale, an occupation is shown in capital letters to tell you one type of job which has that amount of prestige. For additional information, a second example is shown in parentheses at every point on the scale. An extra category is provided at the t0p and bottom of the scale in case you find some occupation that has either higher prestige or lower prestige than any of the occupations shown on the scale. Before you begin your coding of the occupational descriptions I give you, read over the master Occupa- tional Scale very carefully. Notice that a LAWYER has more prestige than an ASSISTANT PUBLIC SCHOOL SUPERIN~ TENDENT, who has more prestige than a CONSULTING CHEMICAL ENGINEER, who has more prestige than a SALES ENGINEER FOR .AK ELECTRONICS PARTS COMPANY, and so forth down the scale. Coding each occupational description we give you will take a little time. Make each decision carefully, because your'decisions will have an important influence on our study. Code each occupation in this way: Say that the first occupational description you are asked to code is .a GLASS INSTRUMENT MAKER. Starting at the bottom of the scale, say to yourself that a glass instrument maker is ‘higher in prestige than a berry picker, higher than a 'waitress, hi her than a boiler Operator, higher than....., 'until you find that you don't agree with the statement you.just made. For example, you may believe that a glass 1 2 instrument maker is higher in prestige than an auto mech- anic, but not be sure he is higher than a rate clerk for a transportation company. As soon as you get this un- sure feeling, you are nearing the prestige level of the occupation you are coding. When you get this unsure feeling, assume for the moment that you have found the prestige level of a glass instrument maker. Before you make a final decision, however, check whether you believe a glass instrument maker has less prestig§_than the occupation mentioned in the prestige category above the one you are about to place him in. If this is not so, you may want to make your rating of the occupation higher than the prestige level you first considered as appro- priate. When more than one level seems to fit, pick the one you think fits best. When you have made your final decision, place the number of the prestige level, from O to 12, on the recording sheet provided. High Prestige 12 11 10 0 Low Prestige 3 Occupational Prestige Scale MORE PRESTIGE THAN A LAWYER LAWYER (Sociology professor) ASSISTANT PUBLIC SCHOOL SUPERINTENDENT (Aircraft instructor for air force) CONSULTING CHEMICAL ENGINEER (Grade school teacher) SALES ENGINEER FOR ELECTRONICS PARTS COMPANY (Purchasing agent for electric company) SOIL CONSERVATION WORKER (Automotive cost accountant manager) CREDIT MANAGER FOR ADVERTISING SERVICE (Cost estimator for a glass company) RATE CLERK FOR TRANSPORTATION COMPANY (Offset photographer) AUTO MECHANIC (Salesman for thread corporation) BOILER OPERATOR FOR PICKLE FACTORY (Service station operator) WAITRESS (Newsstand operator) BEAN ND BERRY PICKER (Garbage collector) LESS PRESTIGE THAN A BEAN AND BERRY PICKER APPENDIX B INTERVIEWER-SCHEDULE INSTRUCTIONS FRONT PAGE: The "reSpondent number" given in the blank following C6-C7-C8 is very important information to us. If your original questionnaire for a respondent is Spoiled be sure to put this reSpondent number, and the respondent's address on the Spare questionnaire you use. Also put the reSpondnet number on the green and yellow pages attached at the end of the questionnaire. QUESTIONS 9-18: Several of these questions refer to "our‘ own home town,""local problems," "our community," etc. For most of reSpondents, Detroit will be their "home town." In the suburbs, some reSpondents may ask you whether you mean their suburb or Detroit as a whole. Tell them that we mean "the Detroit area as a whole." QUESTION 19: Let the respondent name as many magazines as he can without being probed. Then probe with "Any others?" If he names one or more magazines after this probe, probe once more with "Any others?" Then quite. Magazines named after a probe should be placed in the blank following that probe. QUESTION 22-23: Here we are interested only in news broad- casts on radio and television. If the reSpondent wonders ‘whether he should include Special news background programs and news panel shows like Meet the Press, tell him "No." Consider only regularly scheduled news broadcasts. QUESTION 24: On this question, write down whatever topics the reSpondent gives you. Let him decide if it is a topic "that has been getting a lot of attention in the news lately." If the respondent wonders, we mean news covered in news- papers, magazines, television, and radio. We want a maximum of four topics from each respondent. Probe with "Any others?" until you get four topics from the respondent, or until you don't feel you should press him any more. DON'T GIVE HIM .ANY HINTS! That would ruin the purpose of the answers we are trying to get. QUESTIONS 25-40: At question 25 you begin a long series of questions. These questions will be based on the topics the respondent has mentioned to you in question 24. In general, we want two types of information here. First, has the re- Spondent asked anyone for opinions on these topics. If so, 'we'd like to know what the person he asked is like. WE DON'T CARE HOW CASUAL THE SITUATION WAS IN WHICH HE ASKED SOMEONE FOR AN OPINION. WRITE DOWN WHATEVER THE RESPONDENT SAYS IN RESPONSE TO YOUR QUESTIONNING. Second, we'd like to know whether other people have asked the respondent for any opinions 1 2 recently. If so, we want to know what the other person was like. In both cases, if the reSpondent did not ask anyone, and nobody asked him, you will skip the questions pertaining to the description of the other person. Watch the instructions for when to skip and where to skp to. QUESTION 26 and 34: In the past, we've found that most people are very willing to give you the name and address of the other person. If they say that they do not want to, DO NOT TRY TO PERSUADE THEM TO. Just say "okay," and go on. QUESTIONS 27-28 and 35-36: We are doing a Special experiment on the verbal descriptions you get from respondents regarding people's occupations. It is eSpecially important, therefore, that you probe for a very Specific description of the occupation. How- ever, do not ask them the C MPANY the person works for; this is sometimes threatening to respondents. QUESTION 31: Probe for detail here by saying "Could you explain that a little more?" QUESTION 32 and 39: If a reSpondent says "I just don't know” to this question, just record that comment in the margin to the right of the answers given. QUESTION 42: Let the respondent answer this question by giving you a number. When he has done so, code his answer into one of the three categories shown. QUESTIONS 43 and 44: YOu will have to use your judgment in coding reSponses to these two questions. It Should not be difficult, however, as we want you to use these rules as guidelines: (1) If the respondent answers either "yes" or "no" immediately after you ask yes or no for question 44. (2) If the person hesitates for a few moments before answering, put him in the appropriate "probably? category. (3) If the person cannot decide how to answer the question, code him as a "don't know." QUESTION 45: You mention only two of the three categories to the respondent in this question. Use the "about the same; don't know" category only if the reSpondent is not willing to choose between the "more" and "less" categories. QUESTION 49: Some people may say that they get together with friends and neighbors quite often, but not with people they work with. We want the total number of times a week they get together with friends and neighbors and work mates. 3 PROCEDURE FOR HAIDIKG OUT YELLOW AND GREEN PAGES: When you hand the reSpondents the yellow or green page, be sure to supply him with a pencil and with something to write upon. While he is responding to the questions, do not make him anxious by watching what he is doing or by getting fidgety while he is reSponding. DO NOT RUSH THE RESPONDENT! If he says he does not know one or more answers, tell him: "Just guess, we'd like you to pick one answer for each question." When the reSpond- ent is done, check whether he answered every item. If he did not, be sure he didn't overlook the item. Do not force him to answer every one if he doesn't want to, but note on the blank questions that he "didn't know." QUESTION 50-53: This is a three-part question. It is hard for reSpondents to recall many organizations if they belong to many. Also, they may not know which organiza- tions we are interested in. After they have named one or more organizations, probe for more by repeating the examples given in the question. To begin with, you are simply trying to get them to list the organizations they belong to. For your second set of questions, you will ask "Are you an officer in...." for each of the organizations they have named. For your third set of questions, you will ask "How many meetings of the .... have you attended out of the last four?" Do this for each organization. This question may take a fair amount of time for some people in the sample. QUESTION 54: If respondents answer "yes," we would like a fairly Specific description of the topic. However, be careful not to "help" the reSpondent. Ask only non- directive probes such as: "Could you describe it a little more?" QUESTIONS 55 to 68: Notice that you must get two responses to each statement. Code the first reSponse as "just don't know" only if he is unwilling to pick "agree" or "disagree." If he says "just don't know,” you do not ask the second part of the question. His reSponse would have to be "indifferent." QUESTION 10 (page 12): Code a person's race as "other" if he is oriental or indian, or if you don't know for sure if he is either white or negro. QUESTION 13 (page 13): If a person has 12 years of school, but did not get his high school diploma, we assume he did not complete 12 years. Code him 9-11 years. Likewise with college; if no bachelor's degree, code him 1-3 years of college. This question is intended to measure only academic education. If a person is a practical nurse, or a regis- 4 tered nurse that graduated from a hOSpital program rather than a university nursing program, those years of education do not count here. Likewise, with persons who attended “commercial college” or a ”business college" that is not part of a regular university. Trade schools do not count either. If you are not sure how to code this question, simply write down any information you have that will help us decide how we want it coded. QUESTION 15-16 (page 13): Again, we want you to give Special attention to getting very Specific descriptions of the occupation of the main wage earner. We will be using the answers you get in the Special experiment. GETTING RESPONDENT'S NAME. If the reSpondent prefers not to give his name, that's okay. If possible, however, we would like the name, in case we have to call back for additional information. APPENDIX .c DETROIT OPINION AND INF 11:mesz STUDY C1 C3 518 Project Number C . C4 C5 02 Phase Number C6 C7 Respondent Number C8 Hello. . .I'm from Michigan State University. We're doing some research on the opinions people have about a variety of issues. One of the persons chosen for this study was the Nan Woman. . .of your household It's very im ortant that we find out the opinions of everg person selec ed for the study. (Arrange With respondent or interview.) To begin with. . .I'm going to read some statements people have made as their 0 inions on several topics. Ybu may agree with some of t ese statements. . .and disagree With thers. After I have read each statement. . .please tell me how you personally feel about the statement I have read. 9. Here's the first statement.'. .Before we try to solve all of the world's problems, we should take care of those in our own home town. . .Which of the following answers best fits how you feel about this statement? 0 strongly agree 1 agree 2 don't know (or ref.) 3 disagree 4 strongly disagree 10. The next statement. . .I'm.more interested in the problems of our state than in local problems. 4 strongly agree agree 2 don't know (or ref.) 1 disagree 0 strongly disagree 2 11. You should get to know as many people as you can. 0 _strongly agree 1 _agree 2 :don' t know (or ref.) 3 :disagree 4_ wstrongly disagree 12. Only people who have grown up in our community can really understand our local problems. 0 _strongly agree 1 agree 2- _don't know (or ref.) 3 :disagree 4_ _strongly disagree 13. It is more important to know several people in one particular line of work than to know people in many types of work. 4___strongly agree 3____agree 2___don't know (or ref.) l___disagree O___strongly disagree 14. There are only a limited number of people in this community with whom I have a lot in common. 4___strongly agree 3___agree 2___don't know (or ref.) 1___disagree O___strongly disagree 15. Young people who go off to college should come back to their ometown to live when they finish their education. 0 “strongly agree “agree _don't know (or ref.) :disagree _strongly disagree 1 2 3 4"” 16. It not howm people you know that is important... but the e of people you know. 4 3 2 l 0 _strongly agree _agree _don't know (or ref.) :disagree _strongly disagree q .3 17. Community leaders should be people who were born and raised in the community. Q___strongly agree ___agree 2___don't know (or ref.) 3___disagree 4;__strongly disagree 18. National issues have a bearing on local problems. 4___strongly agree 3___agree 2___don't know (or ref.) l___disagree O___strongly disagree 19. Now I'm going to ask you some questions about your use of the mass media...First...what magazines do you read regulagly...that is, at least three out of every four issues See 5 mma ee d' responses) Any others? Any others? 20. What about newspapers...on an average day...how much time do you Spend reading your daily new3papers? minutes (See summary sheet) 21. About how much of that time do you spend reading the main news stories of the day? minutes (See summary sheet) 22. About how frequently do you listen to news broadcasts on the radio...would it e several times a day...once or twice a day...every other day...about once a week...or less 0 ten? 4 49%several times a day 3:32 once or twice a day 2 1 every other day l_75 about once a week 0::E less often 4 23. How about television..how often do ou watch news broadcasts on television . . . would it e more than once a day . . .about once a day . . .every other day . . . about once a week . . .or less often? 4 42% more than once a day 3-53 about once a day 2::5 every other day l 5 once a week 0: less often 24. Now something slightly different. . .Can you . . . off the top of our head . . .think of three or four topics or issues that ave been getting a lot of attention in the news lately? I. 2. 3. 4. 25. Have you asked anyone for his or her opinion on any of these topics during the past week or two? 0 To 2 Yes If NO, skip to If YES, go to Question 33 next question 26. Which of these topics did you ask this person about? Now a few questions about the person you talked to . . Could I get the person's name? Do you know where this person lives? (Address) 27-28. Do you know (his) (her) occupation? 5 29. Is this person a member of your family, a neighbor, a relative, someone you work with, or someone else? 0 family 1 neighbor 2 relative 3 works with him 4 someone else (Specifyz) 11f FAMILY, skip to question 31 30. How well do you know this person. . .would you say he's one of your closest friends. . .a fairly close friend . . .a casual acquaintance. . .or someone you had not met before? 3 one of closest friends 2 fairly close friend 1 casual acquaintance 0 had not met him before 31. Why did you happen to ask this person for his Opinion? 32. As a result of this conversation. . .would you say that you formed any new Opinions. . .changed any of your gig Opinions. . .or did you come away from the discussion with the opinions you had before? 2 formed new opinions l changed old opinions 0 came away with same as before 33. During the past week or two. . .has anyone asked you for your opinions on any of these topics in the news? 0 79?. 1x70 2.2.11 Yes [If YES, ask: If NO, skip to question 41 34. Which of the topics did this person ask you about? Could I get the name of this person? 6 Where does this person live? 35-36. Do you know (his) (her) occupation? 37. Is this person a member of your family, a neigh- bor, a relative, someone you work with, or someone else? O___family l neighbor 2 relative 3 works with him 4 someone else (Specify:) I If F ‘JIILY, skip to question 39] 38. How well do you know this person. . . would you say he's one of your closest friends . . . a fairly close friend . . . a casual acquaintance . . . or someone you had not met before. 3 one of closest friends 2 fairly close friend 1 casual acquaintance 0 had not met him before 39. Do you think that you influenced this person to form any new Opinions . . . to change any of his old 0 inions . . . or do you think his opinions remained a out the same? 2 formed new opinions I changed old opinions O remained about the same 40. Without going through this whole series of questions again . . .I'd just like to know whether anyone else has asked you for your opinions on any of these topics in the news during the past week or two? 0 93%No or was not asked [if YES, ask; I Which of the - tOpics did 2_mZ_Yes he ask you about? 7 41. About how often would you say people ask you for your Opinions on topics which get a lot of attention in the news. . .would it be several times a week. . .about once a week. . .once or twice a month. . .or less than once a month? 3 30% several times a week 2 28 about once a week 1 17 once or twice a month 0 25 less than once a month 42. About how many people you know look to you for Opinions on major tOpics in the news? 0 54% No one 2 35 l to 3 persons 4 If 4 or more persons 43. If someone you know. . .said that he depended a great deal on your judgment regarding major news topics... would you believe him? 416%T3__surelg1 y wproba 2 ll don' t know ‘ 1 15 probably not INTERVIEWER OM44-'definitely not [JUDGE CODE 44. Would you like to be thought of as a person who others depend upon in making up their minds about major issues in the news? 4 21% unqualified "yes" 3 12 "I Guess so" 2 5 don?t know 1 :1 "probably not" 0 40 unqualified no INTERVIEWER JUDGE CODE 45. Compared with your circle Of friends. . .are you more likely. . .or less likely. . .to be asked for opinions on topics in the news? 4 35” more likely 2 about the same, D.K. 0 41 less likely 46. When you and your friends discuss topics in the news, what part do you play?. . .do you mainly listen. . .or do you try to convince them of your ideas? 8158], mainly listen 28 both, don't know try to convince them 1f" 47. How important is it to you to be considered a person whose Opinions on topics in the news are well- founded?. . . Is it very important. . .fairly important. . .n9£.very important. . .or not at all_important? 3 _1873 very important 2 47 fairly important 1 27 not very important 0 8 -not at all important 8. About how Often. . .on the average. . .do you Oet together with your relatives. . .would it be severa times a week. . .once or twice a week. . .once or tw1ce a month . . .or less Often? 3 20% several times a week 2 34 once or twice a week 1 23 once or tw1ce a month 0 23 less Often 49. About how Often do ou get together socially with friends and neighbors. . .or wit people you work with?. . .would it be several times a week. . .once or twice a week. . .once or twice a month. . .or less Often? 3 20% several times a week 2 4 once or twice a week 1 0 once or twice a month 0 12 less Often Now I'm going to give ou a sheet of paper with some uestions on it. dlike you to read each uestion care- ully. . .then placed an "X"y in the blank in ront of the answer you consider most apprOpriate. Please choose only one answer for each item. . .If you don't know. . .go ahead and guess. Here' 5 the sheet. . . HAND RESP Ii DENT THE YELLOW PAGE I [Take back YELLOW page. See that all items are answered. 9 50-53. Now I'd like to know what organizations you are active in. . .that is. . .organizations such as civic groups, clubs or lodges, PTA, church groups, veterans' organizations, and the like? See summary sheet) 52. How many meetings 51. Are you an have you attended 50. Officer in... out of the last four _. Yes No 0 l 2 3 4 Yes NO 0 1 2 3 4 __ Yes N o O 1 2 3 4 Yes NO 0 1 2 3 4 Yes NO 0 1 2 3 4 Yes NO 0 1 2 3 4 54. Now a somewhat different topic. . .During the past week or two, have you read anything about the dangers of a nuclear war. . .or how you might protect yourself from a nuclear explosion? O No Yes [If YES, askzl Do you remember what ~4— - the particular tOpic was? (See summary sheet) Now I'm going to read you several statements people have made on this topic. After I read each statement please tell me whether. . .in general. ..you agree or disagree with the statement. Then tell me how strongly you feel about your Opinion. 55. Here's the first statement. . .Community fallout shelters may not save us, but they are the only chance we have to survive. . .do you agree or disagree? 2 69% agree 0 25 disagree 1 6 just don't know 10 56. How strongly do you feel about your answer? 3 41% very strongly 3 32 strongly l 2] moderately O 5 indifferent 57. The next statement. . .There is really no protection against the effects of radioactive fallout. . .agree or disagree? 0 29% agree 2 53 disagree l [3 just don't know 58. How strongly do you feel about your answer? 3 25% very strongly 2 3; strongly l moderately O 13 indifferent 59. The building of community fallout shelters is wrong because it increases the "war scare." O 13% agree 2 23 disagree 1 9 just don't know 60. How strongly do you feel about your answer? 3 31% very strongly 2 E9 strongly 1 5 moderately O 5 indifferent 61. If we had a nuclear attack, I would go to a community fallout shelter. 2 74% agree 0 ]3 disagree 1 8 just don't know 62. How strongly do you feel about your answer? 3 48% very strongly 2 39 Strongly 1 § moderately O 5 indifferent 11 63. Community fallout shelters would not be practical in my community. 0 19% agree 2_§§_.disagree 1 15 just don't know 64. low strongly do you feel about your answer? 3 40% very strongly 2 3% strongly l l moderately O 9 indifferent 65. The drive to build community fallout shelters is merely a money-making scheme. 0 135 agree 2 22 disagree 1 10 just don't know 66. How strongly do you feel about your answer? 3 37” very strongly 2 1 strongly l 16 moderately O 6 indifferent 67. Our community officials should begin plans now to provide fallout protection for our entire community. 2 77% agree 0 15 disagree 1 2 just don't know 68. How strongly do you feel about your answer? 3 92% very strongly 2_;;_ strongly l 16 moderately O 4 indifferent Now I'm going to give you another sheet of paper with questions on it. Again I'd like you to read each question carefully. . . then place an "X” in the blank in front of the answer you consider most apprOpriate. Remember. . .please choose only one answer for each item. If you don't know. . .go ahead and guess. _Here's the sheet. . . [hAND RESPONDENT THE GREEN PAGE] [dhke back GREEN page 12 CARD TWO C1 C2“ 518Project Number C C4 C5 02_Phase Number C6 C7 “ReSpondent Number C8 9. FIELD CODE: S X: l 48¢Male 0 Female 10. FIELD CODE: RACE: 2 721White 1 22 Negro O ] Other 11. Now I'd like to finish with just a few questions about yourself. . .First. . .what is your marital status? Are you single, married, separated, widowed, or divorced? l 5“ Single 2 8Q“; Married and living with spouse Separated 4 6 TJidowed Divorced 6 ] Other (Specify:) 12. How many children under 18 years of age do you have living at home? 0 agog None 1 15 One 14 Two 3 8 Three 4 5 Four 5 g Five 6 3 Six or more 7 i Single What is the name of the last school or college you attended? 13. What was the last grade you completed in school or college? 13 0 w o to 4 years 1 18 5 to 8 years 2 2% 9 to 11 ears 3 .__12 yearsy(HS diploma) 4’12 1 to 3 years of college 5 5 4 years of college (degree) 6 3 More than 4 years of college 7 - Refused 8 ] Don't know 14. And what is your age? 0 1% Under 20 1 20-24 2 25-29 3 30-34 2* 2323.2 .18. ' 6 45-49 7 50-54 8 55-59 9 22 60 and over 15-16. What kind of work does the main wage earner in your household do? (See summary sheet) [ PROBE FOR DETAIL; 17. And what was the last grade in school or college com- pleted by the main wage earner in your household? 3 5% g to 4 years 3% to 8 ears 2 9 to llyyears 3 25 12 years (HS diploma) 4 13 l to 3 years of college 5 5 4 years of college (degree) ? 5 More than 4 years of college - Refused 8 ] Don't know Finally. . .Could I get your name? THANK YOU VERY'NUCH for your c00peration. Your Opinions are very important to us. 14 Summary Sheet Question Number 19. 20. 21. 51. 52. 54. Number of news magazines read (Time, Newsweek, and U.S. News & World Report). 0 85% 2 2% 1 l3 3 0 Time Spent reading the neWSpapEr. 0-20 minutes 21% 55-65 minutes 32% 25-50 minutes 32 70 or more 15 Time Spent reading the "main news Stories in newspapers. 0-5 minutes 1 % 30-40 minutes 22% 10-15 minutes 24 45-50 minutes 17 20-25 minutes 10 55 or more 12 Actual number of offices held in organizations: 0 83% 2 3% 1 13 3 1 Sum of meetings attended for all organizations. 0 52% 5 4% 1 6 6 4 2 6 7 1 3 5 8 5 4 ll 9 or more 6% Awareness of the mailed message: 0 61% Hasn't read anything about nuclear war. 1 _2_ Has read something about nuclear war, but no indication of Specific message awareness. Indication of awareness of message content. 0 H is GREEN PAGE 18. At the present time, what is the status of the fallout shelter program in Detroit: the city has decided not to set up any community fallout shelters. the city has not set up any community shelters, but will next year. the city has set up many community shelters, but as not stocked any of them. >( the city has set up many shelters and has stocked many of them with food. Correct responses 3 % Incorrect responses 66% 19. “Fallout" from a nuclear explosion is composed mainly of radioactive: _L__fragments of the bomb itself. 3< pieces of dirt stirred up by the explosion. water vapor Eroduced by the explosion. Smoke partic es caused by fire after the blast. Correct reSponses 20% Incorrect responses 80% 20. In this part of the United States, radioactive fallout giuld travel primarily in which direction from the nuclear ast site: ____north south )3 east west Correct responses 31% Incorrect reSponses 69% 21. A fallout shelter would offer protection from nuclear explosion only if: }{ it keeps radioactive particles from entering the Shelter. the walls of the shelter are airtight. outside light is kept from the shelter. the walls of the shelter are given a Special insulated coating. Correct responses 62% Incorrect responses 38% 22. After a nuclear explosion, one should stay in a fallout shelter (except for short durations) for about: a day. a week. 2; two weeks. a month. Correct responses 42% Incorrect responses 58% 23. Protection from radiation fallout: would require building large concrete chambers underground. :1 would require only slight revision of many existing buildings. would require a massive program of building family shelters. is impossible; you can't really protect yourself. Correct reSponses 2 % Incorrect responses 71% YELLOi-J 35g; 24. Astronaut John Glenn recently announced that he was entering the race for: U.S. Representative from Ohio. President of the United States Governor of Ohio. X U.S. Senator from Ohio. Correct responses 64% Incorrect reSponses 36” r 1'0 25. One reason city officials give for the violence found in Detroit schools is that: there are no policemen patrolling Detroit schools. the law doesn't allow judges to give teenagers jail sentences. ‘ v A few youths are sentenced because the training school is crowded. Detroit teachers don't want any teenagers put in jail. Correct responses 55% Incorrect reSponses 45% 26. The 24th amendment added to the Constitution of the United States this month: ‘ A forbids charging anyone a poll tax to vote in federal elections. makes it a federal crime to allow segregation in schools. places a death penalty on the crime of putting bombs in airlines. makes it legal to say prayers in public schools. Correct responses 30% Incorrect reSponses 70% 27. A star witness for the prosecution in the Hoffa jury- tampering trial was: a woman secretary that worked for Hoffa. ){ an officer of a Teamster Union Local in Louisiana. a president of a trucking company. Robert Kennedy, attorney general of the U.S. Correct reSponses 54% Incorrect reSponseS 46% 28. A Detroit ordinance that would give property owners complete freedom to choose who they will sell or rent their property to: was passed by the Common Council last week. .4L.was declared unconstitutional and banned from the August ballot. produced a civil rights demontration at the state capital last week. did not get enough signatures backing it to qualify for a vote. Correct reSponses 40% Incorrect responses 60% 29 . Last week, Cuba Shut off the water supply to the United States marine base at Guantanamo, Cuba, because: the United States cut off diplomatic relations with Cuba. Arnerican businessmen are buying Egyptian tobacco instead of Cuban. :3 the United States coast guard seized some Cuban fishing boats. the American Red Cross never gave Cuba the tractors it promised them. Correct responses 74% Incorrect reSponses 26% room USE 0m wees“ J1! mt. xv” HICHIGRN STRTE UNIV. LIBRQRIES 31293104106012