THROUGH TELEVTSTBN A STUDY OF souRcE EFFECTIVENESS .g: ThESIS for NE Degree of Ph D MICHIGAN STATE UNIVERSITY HELEN ELIZABETH PARRIs { 1971, ~ ‘ EAcILIIIIIINE INDINs FAMILY PLANNING PROGRAM-ff _.., I III III INNNIIIII II II IIIIIII IN N I; II {I‘r‘b .. This is to certify that the thesis entitled Facilitating India's Family Planning Program Through Television: A Study of Source Effectiveness presented by Helen Elizabeth Parris has been accepted towards fulfillment flgfl requirements for f ' " Jegree in [76644. 12 Z(//)c Lj/ {1/1” , - Major professor 2"" I W» Date , / ,1 / 0-7639 ABSTRACT FACILITATING INDIA'S FAMILY PLANNING PROGRAM THROUGH TELEVISION: A STUDY OF SOURCE EFFECTIVENESS BY Helen Elizabeth Parris The present study investigated the relative effec- tiveness of two types of sources used on television to impart a family planning message to adults in the villages of Delhi State, India, provided with television sets. The sources were a village opinion leader and an urban medical expert. Television effectiveness was measured in terms of changes in (1) knowledge, (2) attitudes, and (3) motivation to seek additional family planning information (information- seeking behavior). These television effects were measured immediately before and after exposure to television and again four weeks after exposure. Methodology Two experimental television treatments (programs) essentially equated on all factors other than the source variable were professionally produced for television and were recorded on half-inch video-tape for playback in the Helen Elizabeth Parris experimental villages. Twenty experimental villages were randomly selected from among the Delhi State villages receiving television and those villages were randomly assigned to one of the two comparable television treatments. Ten adult males were selected randomly as experimental sub- jects from each sample village. The degree of each subject's modernity was measured prior to treatment. The dimensions of modernity measured are (1) education, (2) mass media exposure, (3) cosmopolita- ness, (4) political knowledgeability, and (5) community participation. The degree of each subject's perceived homophily with the source (similarity in certain attributes) and perceived source credibility were measured immediately after treatment. Based on relevant theory and research, it was predicted that low modern subjects would perceive the village source as more homophilous with themselves and more credible than the urban source. Low modern subjects would, therefore, tend to obtain higher scores on measures of the television effects as a result of exposure to the village source. It was likewise predicted that high modern subjects would perceive the village source as more homophilous with themselves than the urban source. However, finding urban values more salient than rural, high modern subjects would perceive the urban source as more credible and tend to obtain higher scores on measures of the Helen Elizabeth Parris television effects as a result of exposure to the urban source. A two by four experimental design was followed to collect the data. Summary of Major Findings As predicted, both high and low modern subjects perceived the village source as decidedly more homophilous with themselves than the urban source. Also as predicted, subject's degree of perceived homOphily with the source to whom he was exposed was related to his level of modernity: the more modern the subject, the more homophilous his relationship with the source tended to be. Contrary to prediction, all subjects tended to perceive both sources as highly credible relative to the measures used. As predicted, low modern subjects exposed to the village source gained significantly more knowledge of ' family planning than their counterparts exposed to the urban source. Contrary to prediction, high modern subjects exposed to the village source also gained significantly more knowledge than their counterparts exposed to the urban source. In other words, all subjects tended to gain significantly more knowledge from the village source than from the urban source. Contrary to prediction, differences in the effective- ness of the alternate treatments were not related to levels of modernity on any of the television effects measured. Helen Elizabeth Parris The present experiment demonstrated the efficacy of experimental research in a village setting as a viable means of determining those approaches to television pro- gramming which are effective in accomplishing specified develOpmental objectives. FACILITATING INDIA'S FAMILY PLANNING PROGRAM THROUGH TELEVISION: A STUDY OF SOURCE EFFECTIVENESS BY Helen Elizabeth Parris A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY College of Education C Copyright by HELEN ELIZABETH PARRIS 1971 ACKNOWLEDGEMENTS Regrettably, the following pages of scientific report- ing do not reveal the human story behind my research col- laboration in India. It is a gratifying story of urban/rural, Indian/American cooperation in a world where intersystem cooperation is a highly valued yet often elusive phenomenon. I am particularly indebted to India's Central Family Planning Institute (CFPI) for serving as my official col- laborating agency; to All India Radio-Television for con- tributing personnel services and technical facilities for the professional production of two experimental television programs; to the Development Department of the Delhi State Administration for coordinating the activities of our investigation team in the experimental villages; and to the Electrical Engineering Department of the Indian Insti- tute of Technology, Kanpur, for the loan of portable video- tape equipment, without which a television laboratory experiment could not have been conducted in the field. I am most grateful to four scholars for their academic guidance: Dr. Charles Schuller, my academic advisor, fostered and perceptively directed my inter- disciplinary approach to the study of rural developmental television from the first days of our association. By ii fortunate coincidence, Dr. Everett Rogers, my official thesis advisor, arrived in India while I was drafting the thesis chapters. His rich experience as a communication researcher was brought to bear on all subsequent chapter revisions and his communication perspective is mirrored throughout the work. For more than 15 months while I was in India, Dr. S. D. Kapoor, Psychologist and Senior Research Officer of the CFPI expertly guided the application of theoretical aspects of the study and procedures for its execution to the realities of village India. Dr. D. C. Dubey, Sociologist and Senior Research Officer of the CFPI also contributed his broad knowledge of rural Indian research to the study. The United States Agency for International Develop- ment (USAID/New Delhi) funded the study and the Ford Foundation made an initial contribution toward its imple- mentation. I am indebted to both of these organizations and, in particular, to Mr. Michael Jordan, Population Development Officer of USAID, who served as project admin- istrator. Mr. Jordan's deep personal interest in the project was an important factor in its successful execu- tion. My deepest appreciation is extended to the two "sources" (they shall remain anonymous) who willingly made a contribution to scientific knowledge by appearing in the experimental television programs. Also most deserving of my appreciation are those village respondents who kindly iii offered their cooperation to our research investigators in the midst of their wheat harvesting season. iv TABLE OF CONTENTS ACKNOWLEDGEMENTS . . . . . . . . . . . LIST OF TABLES. . . . . . . . . . . . LIST OF FIGURES . . . . . . . . . . . Chapter I. THE NATURE AND SCOPE OF THE STUDY. . . II. The Background and Purpose of the Study. x The Need for Studying Television Source Effectiveness . . . . . An Overview of the Research Question. . The Study in Perspective. . . . . The Research Method . . . . . The Substantive Content for Research. The Variables for Study and Their Relationships . . . . . . . Independent Variable A: The Treatment Modes. . . . . . Independent Variable B: Degree of Modernity of Subjects. . . . Perceived Source Credibility as an Intervening Variable . . . . . pQThe Effects Variables Measuring Cogni- tive, Effective, and Motivational Change. . . . . . . . . . THEORETICAL FOUNDATIONS FOR THE DEVELOP- MENT OF EMPIRICAL TESTS . . . . . . Overview of the Research Question. . Measures of Individual Modernity . . Development of an Empirical Test to Measure Individual Modernity . Page ii ix xi cnua +4 Id 12 l7 19 19 22 28 3O 37 37 38 40 Chapter Page Measures of Homophily . . . . . . . 44 The Present Study Compared with Prior Homophily Studies . . . . 47 Development of the Homophily Instrument. . . . . . . . . 50 Measures of Perceived Source Credibility. 56 Credibility as a Multidimensional Construct . . . . . . . . . 59 Operationalizing Perceived Source Credibility . . . . . . . . 62 Measures of the Television Effects Variables . . . . . . . . . . 66 Knowledge-gain . . . . . . . . 66 Knowledge Retention. . . . . . . 68 Attitude Change . . . . . . . . 68 Persistence of Attitude Change-. . . 70 Information-seeking Behavior Change . 70 Persistence of Information-seeking Behavior Change . . . . . . . 72 Review of the Hypotheses for Study. . . 72 III. THE RESEARCH METHODOLOGY . . . . . . . 75 Design of the Stimulus Treatments . . . 75 Design and Production Procedures . . 77 Execution of the Research Design . . . 86 Locale of the Study. . . . . . . 86' Selection of Experimental Villages. . 88 Selection of Respondents . . . . . 90 Assignment of Villages to Treatments . 93 The Pilot Test . . . . . . . . 97 Conducting the Experiment. . . . . 102 Administering the Post Posttest. . . 105 Analysis of Data. . . . . . . . . 106 Methods of Analysis. . . . . . . 106 Statistical Tests of Significance . . 113 vi Chapter I Page IV. THE RESEARCH FINDINGS. . . . . . . . 123 Significance Tests of Perceived __ HomOphily. . . . . . .. . . . 123 }\Significance Tests of the Television Effects . . . . . . . . . . 126 Knowledge-gain . . . . . . . . 127 Knowledge Retention . . . . . . 131 Attitude Change. . . . . . . . 135 Persistence of Attitude Change. . . 136 Information-seeking Behavior Change and Persistence of Information- seeking Behavior Change . . . . 138 V. CONCLUSIONS, IMPLICATIONS, AND RECOMMENDATIONS. . . . . . . . . . 141 The Research Summary . . . . . . . 141 Summary of Major Findings . . . . . 142 The Perceived Homophily Findings . . . 144 Interpretation of Perceived HomOphily Findings. . . . . . 145 Perceived Source Credibility as an Intervening Variable . . . . . . 146 The Knowledge-Gain Effect . . . . . 147 Interpretation of the Knowledge- Gain Effect . . . . . . . . 148 The Optimal HeterOphily Assertion. . . 149 The Extent to which the Research Findings can be Generalized. . . . 151 Implications of the Present Experiment . 155 Relevance of the Research Findings . 155 Relevance of the Experimental Method of Research. . . . . . 157 Recommendations for Future Research . . 159 The Study of Television Source Effectiveness . . . . . . . 159 Future Application of the Experi- mental Method of Research . . . 162 Indian Television Research in Perspective . . . . . . . . 166 vii Chapter BIBLIOGRAPHY APPENDICES . APPENDIX A: APPENDIX B: APPENDIX C: Television Research Questionnaire (English Version) Television Program Script. Experimental Villages: to Treatment viii Assignment Page 168 174 175 196 206 Table 3.1. 3.2. 4.1. LIST OF TABLES Caste Affiliation of Respondents . . . . Occupations from which Respondents Earned More than Half their Income . . . . . Degree of Intravillage Heterogeneity and Intervillage Homogeneity Indicated by Computing the Means and Range of Sources Obtained from Measurement of Three Dimensions of Modernity . . . . . . A Hypothetical Example of Mean Effects Obtained in an After-Only and a Before/ After Research Design . . . . . . . Replications for Two-Way Analysis of Variance Tests of Hypotheses I, II, and III. 0 O O I I O O O O O 0 Respondents' Perceived Source Credibility Scores Obtained in the Alternate Treatment Modes . . . . . . . . . Summary Data from Two-Way Analysis of Variance of the Homophily Effect Showing Main Treatment and Modernity Effects and their Interaction . . . . . . . . Mean HomOphily Scores and Standard Devia- tions in the Alternate Treatment Modes . Mean Perceived Homophily Distance Scores Across Treatment Modes and Levels of Modernity . . . . . . . . . . . Summary Data from Two-Way Analysis of Variance of the Knowledge-Gain Effect Showing Main Treatment and Modernity Effects and their Interaction . . . . Mean Knowledge—Gain Scores Across Treatment Modes and Levels of Modernity . . . . ix Page 94 95 108 114 116 119 124 124 126 129 129 Table 4.6. 4.10. 4.12. 4.13. Mean Knowledge Retention Scores.Across Treatment Modes and Levels of Modernity Summary Data from Two-Way Analysis of Variance of the Knowledge Retention Effect Showing Main Treatment and Modernity Effects and their Interaction Mean Gain Scores on the Television Effects Measured at the Time of Treatment and Four Weeks Later. . . . . . . Summary Data from Two-Way Analysis of Variance of the Attitude Change Effect Showing Main Treatment and Modernity Effects and their Interaction . . Summary Data from Two-Way Analysis of Variance of the Persistence of Attitude Change Effect Showing Main Treatment and Modernity Effects and their Inter- action . . . . . . . . . . Mean Persistence of Attitude Change Scores Across Treatment Modes and Levels of Modernity . . . . . . . . . Summary Data from Two-Way Analysis of Variance of the Information-Seeking Behavior Change Effect Showing Main Treatment and Modernity Effects and their Interaction . . . . . . Summary Data from Two-Way Analysis of Variance of the Persistence of Informa- tion Seeking Behavior Change Effect Showing Main Treatment and Modernity Effects and their Interaction . . Page 132 132 134 136 137 138 139 139 Figure 1. 3. 4. LIST OF FIGURES Theoretically predicted relationships be- The The Smoothed frequency polygons showing distribu- tween high and low levels of subjects' modernity and degree of perceived homophily with sources in the alternate treatment modes. . . . . . . . . Optimal heterOphily assertion showing the relationships between high and low levels of subjects' modernity and degree of perceived homophily and perceived source credibility. . . . . . . . cptimal heterophily assertion showing the relationship between high and low levels of subjects' modernity and degree of homOphily and source credibility, two perceptual variables intervening between the treatment modes and the television effects . . . . . . . . . . . tions of modernity scores of Mode I, Mode II, and total reSpondents . . . xi Page 26 29 32 110 CHAPTER I THE NATURE AND SCOPE OF THE STUDY The Background and Purpose of the Study The Government of India is in the process of develop- ing a television communication system for the purpose of accelerating India's rate of educational, social, and economic growth. The use of television for rural develop- ment will be emphasized. In addition to the existing television facility in the capital city of Delhi, during the Fourth Five-Year Plan (1969-1974), television stations will be set up in Bombay/ Poona, Srinagar, Kanpur/Lucknow, Madras, and Calcutta. The Indian Government will provide television receivers for community viewing to approximately 400 villages in the proximity of each television station. Also India will soon receive broadcasts transmitted by satellite. An agreement between the Department of Atomic Energy (DAE), the Indian agency for space affairs, and the National Aeronautical and Space Administration (NASA) of the United States, provides for a synchronous satellite to be stationed over the Indian Ocean in 1974 for a period of one year. NASA will be responsible for the space segment of the satellite and the DAB will be responsible for the supply and maintenance of all ground receiving equipment. Approximately 2,800 villages in seven lesser devel- oped and widely scattered regions of India will receive television directly from the satellite on specially equipped receivers. Although these regions are represented by six different language groups, linguistic barriers to communica- tion can be surmounted to some extent, since the satellite will have the capability of transmitting two audio channels simultaneously with the video channel. Three separate rural programs, each presented in two languages, will be offered by satellite on a time-sharing arrangement. ‘ The television stations at Delhi and Bombay and a television facility at the headquarters of the Indian Space Research Organization (ISRO)l at Ahmedabad will be equipped to generate programs to the satellite for national distribu- tion. All of the television stations previously mentioned will be capable of receiving and redistributing satellite broadcasts in addition to broadcasting locally produced programs to their respective audiences. At the conclusion of the one-year DAB-NASA experiment, ISRO plans to launch its own satellite and continue to expand the country's satellite broadcasting service. 1ISRO is the agency which administers the Indian space program and serves a comparable function to that of NASA in the United States. I-J' Presently, the television facility in Delhi is the only television broadcast operation in India. In 1967, eight years after the Delhi television station commenced broadcast service, approximately 80 villages of Delhi State were provided with community television receivers by the Indian Government. At that time, a bi-weekly agricultural program was included in the station's evening program schedule. The rural Delhi television program, a continuing "experiment" in rural adult education, offers timely information to farmers concerning agricultural innovations. The Delhi television signal is now being strengthened to cover all of Delhi State and parts of five neighboring states within a radius of 40 miles from the transmitter. Additional villages within the expanded coverage radius are soon to be provided with television receivers. India's commitment to television technology provokes serious consideration of the kinds of programs that will be transmitted. Sarabhai (1969, p. 2) Chairman of the DAB, stated that in a country with limited economic resources, such as India, it is necessary "to gain insights on the manner in which television can be used as a direct instru- ment for promoting the development tasks of Government, so that the medium can be regarded as an inveStment rather than an overhead." Specifically, Sarabhai and other Indian communication officials want to know what can be done to optimize the effectiveness of television, particularly in rural develOpment. (Approximately 80 per cent of India's total pOpulation is rural.) They are interested in pro- gramming innovations which, when systematically applied to program design, can capture the interest of villagers and at the same time move villagers incrementally toward self help and village development. And they realize that the best time to discover improved procedures for program design is now, before the expansion of television tech- nology. There are few studies that offer mass media practi- tioners insights into the practical tasks of program develOpment. Theoretical propositions about persuasive messages, mostly of Western origin, have rarely been tested on Eastern soil. Referring to family planning communica- tion research, Rogers (1970, p. 15) commented, "Especially ignored in past research have been message variables. . . . We know precious little about the effectiveness of one- sided versus two-sided message treatment (e.g., is it better to admit that a family planning innovation has some disadvantages along with its advantages?), the use of fear appeals, and the credibility of various message Sources . . . " Several studies, reviewed later in this chapter, have been conducted in India to determine whether television can be effectively used as an educational medium in rural develoPment. Based on these studies there is a good deal of evidence to suggest that television can be an effective teaching tool. But, given that television is and will be used increasingly in rural development, it is important to discover those message strategies which when systematically incorporated in television programming can lead to the accomplishment of specified educational objectives. The present study is the first television study conducted in India to investigate the relative effective- ness of comparable television programming treatments. Specifically, the purpose of the study was to discover whether a village opinion leader2 and an urban expert3 _are differentially effective in deliveringdevelopmental information and advice through television. The content used in preparing television materials for evaluation concerned family planning. The locus of the study was those villages of Delhi State that have been receiving television for more than three years. The Need for Studying Television Source Effectiveness "The effectiveness of a communication is commonly assumed to depend to a considerable extent upon who delivers it" (Hovland and others, 1965, p. 19). Since Indian 2An opinion leader is an individual who has the ability to influence other individuals' attitudes in a desired way and with relatively high frequency (Rogers, 1969a, p. 88). 3An ex ert is an individual who has special skill or knowledge in a subject. television will be called upon to play a vital role in rural family planning education, the question of who should convey family planning messages on television is of con- siderable importance. The question of appropriate message sources is of immediate concern to personnel who design the rural Delhi television program. In 1970, more than three years after its inception, the rural Delhi television program was expanded from a 20 minute bi-weekly program to a 30 minute program offered three times a week. In addition to regularly scheduled agricultural features, the program now offers features encouraging family planning, improved nutrition, sanita- tion, and the like. Traditionally, the rural program has relied heavily on the use of urban experts to convey agricultural information. Occasionally, agricultural Opinion leaders from the Delhi‘ State villages are invited to appear on the program to discuss agricultural innovations. But, for the most part, agricultural experts from the Indian Ministry of Agriculture and the extension division of a local agricultural university have been featured. The decision to use experts to impart technical information may be a valid one. Many of the Delhi State farmers have witnessed high yields of grain produced on university demonstration plots and they recognize the: desirable consequences of agricultural innovation--a1though farmers differ in the extent to which they adopt these innova- ‘*-\, tions. When the intent of a television communicator is to impart technical information about an innovation, and when there is evidence to suggest that, in general, members of the viewing audience are not adversely disposed toward the innovation, but, rather, need information related to its application, then the use of an expert source on television may have merit. However, when the intent of astelevision~communicator is to encourage social change, it may be found that, among some types of villagers, a village opinion 1eader_appearing on television is a more effective message source than an urban expert. Messages intended to evoke changes in tradi- tiohal patterns of behavior usually convey value judgments and may be controversial. Such messages may be more effec- tively communicated by a television source whom the viewer perceives to hold values similar to his own. The message incorporated in the present investigation pertained to the benefits to be derived from adoption of the small family norm4 and was categorized as serving a social change func- tion. It was in the context of such a message that tele- vision source effectiveness was studied. 4The small family norm is a prescribed social be- havior concerning the desirability of having a family of few children, specifically two or three, rather than many children. The present assessment of the message source variable as a determinant of desired television viewer response is of practical as well as theoretical importance. If an urban expert and a village opinion leader should be found to have a significantly different impact on rural Delhi television Viewers, then such information can be taken under advisement by television personnel who plan and design the rural Delhi television program. As a result of a number of experiments similar to the present one, testing the source proposition in various contexts in Delhi State and other rural regions of India slated for television, it may be possible to supply program personnel with widely applicable guide- lines concerning appropriate source selection. System- atically applied to program design, these guidelines could help to increase the goodness-of-fit of family planning and other developmental messages to specified target audiences of rural Indian adults. An Overview of the Research_guestion The present investigation pertained to the relative effectiveness of two types of television sources used to impart a family planning message to adults in the villages of Delhi State which are provided with community television receivers. The sources were a village opinion leader in matters pertaining to family welfare and an urban medical -.__._” expert. The research question was to determine whether “. ~——.. one of these two types of sources is more effective than the other in producing desired viewer response. Effectivee ness was measured interms of changes in_(l)know1edge of \r . .—*A —~-. ‘M ......r-._,._ ._ _ _. 7 content, (2) attitudes toward family planning, and (3) motivation to seek additional family planning information (information-seeking behavior). These variables constituted w...____ the "television effects." A television effect is a response H .1- .n‘ u... of a receiver to‘a television message. The differential effectiveness of the sources was assessed through viewer response measured at the time of exposure to television and again four weeks after exposure. In cooperation with All India Radio-Television (AIR-TV),5 two experimental television programs (experi- mental treatments) were designed. The programs were es- sentially equated on all factors other than the experimental factor: the sources used to impart the family planning message. Both programs were 14 minutes in length; both contained the same concepts and the same number and sequence of concepts. Audiovisual relationships and the quality of production were held constant in both. Only the sources presenting the message were systematically varied in the alternate experimental programs. The programs were recorded on two-inch video-tape and were subsequently re-recorded on one-half inch video-tape. 5AIR-TV is the television branch of the Ministry of Information and Broadcasting. 10 The re-recorded programs were shown on television in the experimental villages by means of portable video-tape equipment. The research of Rogers and others (1969a) supported ... the contention that a village Opinion leader appearing on television would be perceived by low modern viewers as mere like themselves in certain attributes than an urban expert. Sharing a life-style relatively similar to_these '\i viewers, the village opinion leader would present the message in terms of the viewers' experience. —Themyillage opinionleader would, therefore, be perceived by low modern viewers as more credible than the urban expert. Credibility is the degree to which a communicator is perceived as a competent and trustworthy source of informa- 2.1.9.11- Low modern viewers would, thus, tend to score higher on measures of the television effects as a result of exposure to thefiyillage source rather than the urban source. High modern viewers, although perceiving the village leader as more like themselves in certain attributes than the urban source, would, nevertheless, perceive the urban source as more credible. That is,high modernviewers, Often influenced by communication coming to theirvillage ”fromrqutside sources, and holding the norms of their village lessrsalient than other villagers, would respect the judgment 5. :‘i' ".5". L1 ’ '-.",/ “ "TJ 11 5.5 of an urban expert_in matterspertaining,to family planning. High mpdern viewers would, therefore, tend to score higher Non_measures of the television effects as,asresult_of exposure , to the urban source rather than the village source. The two sources used in the alternate television programs were as nearly equated as possible 12,999r communi- cation.ability, and personal dynamism. Both sources were committed to family planning and themselves raised small families. However, the sources differed radically in terms of years of schooling, place of residence, and occupation. Subjects' perceptions of these differences in the sources' cuiturel Orientation was of primary concern in the research investigation. ' 7 1 Each subject in the sample was exposed to one and only one experimental television treatment. By using video- tape equipment to present the television treatments, it was possible not only to randomly select subjects for treatment in each experimental village, but also to randomly select the villages for treatment and to randomly assign villages to the experimental treatments. By employing such randomiza- tion procedures,6 the sampling equivalence of the treatment villages was established and measures obtained across treatments could be compared. 6Randomization is a selection procedure which gives each unit—in the pOpulation an equal chance of being included in the sample. It is a way to insure that treatment groups do not differ except by chance. 12 Had the treatments been broadcast, control over television exposure could not have been assured. Thus, only by means of portable video-tape equipment was it possible to focus on the question for study. Explanation of the variables for study will be pro- vided following a discussion of the research method employed in the present research and the substantive content used in develOping the experimental television treatments. The Study in Perspective The Research Method The present study was a field experiment.7 It differs from field experiments previously undertaken in India to assess the effectiveness of television in one fundamental way: the present experiment incorporated what Campbell and Stanley (1963, p. 13) refer to as a true experimental design8 as compared with quasi-experimental designs9 customarily used. 7A field experiment is an eXperiment conducted in a setting approximating a natural environment rather than an artificially created environment. 8A true experimental design is a research design in which units in the population are randomly assigned to treatments, or to treatment and control groups, and in which all factors other than the experimental factor are system- atically controlled. True experimental designs attain a high degree of internal validity. Internal validipy is the degree to which extraneous factors that might produce effects confounded with effects of the experimental stimulus are controlled. 9A quasi-experimental design is a research design in which units in the population are systematically assigned to 13 Television (and radio) studies conducted in India to date have assessed programs that are broadcast. Since community television receivers have been placed in approxi- mately 80 villages of Delhi State by the Indian Government, it is not possible for a researcher to control television viewing by randomly assigning television villages to a control group and then refusing viewing privileges to persons in those villages for the sake of experimentation. Therefore, in studies planned around regularly scheduled broadcasts, control villages are systematically selected from among those villages that have n23 been provided with community receivers. Experimental and control villages are Often matched in size, accessibility to transportation facilities, and other factors which are believed to be related to the television effects. Although matching helps to equate comparison groups on a few relevant factors, the groups are never equated on all_relevant factors. Whenever villages are systematically rather than raRESHEE‘géiécséa, the sampling equivalence of treatment and control groups cannot be established and the possibility is always present that selection factors might produce effects confounded with the effect of the experimental treatment. Thus in assessing regularly scheduled treatments, or to treatment and control groups. Researchers use quasi-experimental designs when full control over the scheduling of the experimental stimulus is not possible. 14 broadcasts, researchers are faced with a situation in which a relatively high degree of experimental control is lacking. A number of field experiments using quasi-experimental designs have been conducted in India tgyassess the effective- ness of radio or television in combination with group dis- cussion. Neurath (1960) conducted the first major field experiment designed to assess the effectiveness of radio rural forums.lo Responses from forum members in 20 villages were compared with responses from individuals in 20 villages without forums. Control villages were of two_basic types: thgee with radio but no forums, and those with neither mum radians? forum. I w I ~ -— Using gain scores from pre- and post-broadcast measures, Neurath found impressive gains in knowledge of innovations in forum villages as opposed to small gains in both types of control villages. He also found that gains in knowledge were not restricted to the literate members of the forums; the illiterates started at lower levels of information but actually gained more. As part of a cross-cultural study, Roy and others (1969) compared radio forum and non-forum villages. Taking observations over a period of 12 months, Roy found that knowledge changes and positive changes in adoption of 10Radio rural forums are village organizations which give their members a chance to discuss innovations following regularly scheduled broadcasts. 15 agricultural and health practices were significantly greater among forum participants (and among the rest of the heads of household in forum villages) than among respondents in the control villages. In 1960, shortly after the Delhi television station initiated broadcast service, 71 tele-clubs (the television counterpart of the radio rural forum) organized in the lower-middle income neighborhoods of Delhi were evaluated (UNESCO, 1963). Findings revealed that television combine 9":- ‘ I . . . ,4 "~ I,” 7. '1'1‘! with group discussion is an excellent means Of imparting’ knowledge and stimulating plans for action concerning the responsibilities of citizenship. These and numerous other channel combination studies conducted in India reveal that radio forums and tele-clubs can be a catalyst for social and technological change. -LE- hasmalso been found that television broadcasts, unsupported byggnterpersonal communication, can be an effective develop- mentalfaid. In 1967, shortly after tele-clubs were organized in the villages ovaelhi State, it became apparent that many of the farmers who watched the rural Delhi program were interested in learning how to apply agricultural innovations but were not interested in discussing the consequences_of innovations which they mostly favored. Thus researchers decided to investigate the effectiveness of the rural broadcasts alone. 16 Mishra (1967) evaluated the impact on respondents of five agricultural programs in five television villages of Delhi State. Using a before/after design, knowledge- gain and knowledge retention were calculated. Mishra found both effects to be significant in all five programs across all villages sampled, so the programs had a measurable effect. The National Council of Educational Research and Training (NCERT, 1968) studied knowledge of agricultural innovations, willingness to try innovations, and actual adoption as a result of exposure to seven rural Delhi television broadcasts. In an after-only design, NCERT researchers compared results Obtained from a random sample of 100 farmers in the experimental television villages with results Obtained from an equal number of farmers in non- television villages. Significantly higher scores were obtained on all the effects variables among respondents who were exposed to the agricultural television programs. All of the previously mentioned field experiments, and many others conducted in India to date, have addressed the question: "Are the broadcast media, taken alone or in combination with interpersonal activity, an effective means ef bringing about social change?" Infithese_studies—te1e- V1510“ PEQSEams are evaluated_as a composite stimulus. Thus there is no way of knowing the contribution to tele- vision effectiveness of any given element Operating within 17 a program. Studies investigating the relative effective- ness of television, such as the present one, assess the impact of individual program elements for the purpose of verifying program improvements. The Substantive Content for Research Thirteen million persons are added to the population in India each year (Patel, 1969, p. 1). Because production gains are largely off-set by population gains in the country, official priority has been given to a high-powered mass education program in family planning launched throughout India (Patel, 1969, p. 4). Television, as it expands, will be called upon to make a contribution to the educational effort. In a study of family planning communication conducted by the Social Research Division of the Central Family Planning 1 it was found that few Institute (CFPI) Of India (1969),1 persons sampled listened to radio to know about government policy on developmental programs. The CFPI study supports the position of Mitra (1967), who maintained that radio programs which stress the concern Of India's social and political elite with the national population problem are <3f little interest to the village dweller. That the mass Inedia in developing countries tend to stress national IPrOblems rather than the personal needs Of individuals ‘ . 11The CFPI is an autonomous research agency under the Mlnistry of Health and Family Planning of India. 18 who comprise the mass public is a well known fact. Review- ing the literature on mass media research in the developing countries, Rogers (1969c, p. 3) commented, "The media feature messages about development plans and projects . . . but there is little content about the specifics of new ideas in agriculture, health, or family planning." Dubey (1969b, p. 18), commenting on the need for new approaches to mass media use in family planning in India, stated that, although the media have been generally successful in creating an awareness of family planning, their motivation-creating, action-stimulating potential has not been sufficiently exploited. The gap between awareness and practice, between desired and actual family size, remains wide.. The gap may be due to weak motivation or a high degree Of ambivalence for family planning (Dubey, 1969a, p. 40). It would appear, therefore, that the most important role television can play in the family planning program is to stimulate interest in family planning and to encourage action. Rogers (1969c, p. 3) stated that "Mass media EEXposure is able to create a generally favorable mental Sem.toward change, but it is seldom able to form or change Specific attitudes toward new ideas." Television's challenge is, thus, not one of trying to change individuals' deeply- sHeated negative attitudes toward family planning, for this objective is beyond the scope of the medium. Rather, the l9 challenge lies in eroding the inertia of tradition that inhibits an individual from learning that, through know- ledge, he can exercise control over his environment. By deepening knowledge of the economic, health, social, and marital benefits to be derived from adopting the small family norm, an individual, over time, acquires new experience-~new frames-Of-reference upon which to base his judgments about family planning. In designing the two experimental family planning programs for the present research, every effort was made to apply theoretical knowledge about effective communica- tion. The general program objective was to encourage adoption of the small family norm. Family security now, and in the future, was the theme. The programs stressed the health, educational, and economic benefits to be derived from adopting the small family norm. An explana- tion of procedures used for program design and production is furnished in Chapter III. The Variables for Study and Their Relationships Independent Variable A: The Egeatment Modes Treatment modes are programs designed for experi- lEEntal purposes in which all variables are held constant Efifigmm for one variable which is manipulated. 20 In the present research, the source delivering the message was the manipulated variable. Treatment Mode I featured a village opinion leader in matters pertaining to family welfare. Treatment Mode II featured an urban 12 medical expert. The television source as a variant.--Systematic selection of a village source and an urban source for alternate television treatments is valid only if the term "source" is established as a variant. That is, it must be possible to reliably discriminate between the sources based on decidedly different amounts of some common characteristics they possess. The question of establishing the present sources as variants was handled by the research- er's objective and categorical assignment of sources to treatments based on the sources' (1) level and type of education, (2) occupation, and (3) place of residence. The village source completed eight years of schooling in a Hindi medium school, was headmaster of his village school, and resides in a typical village in Delhi State. The urban Source, a medical doctor, graduated from an accredited medical college, holds a prestigeous position in his pro- fession, and resides in urban Delhi. These observable differences in the cultural orientation of the sources ‘ 12Hereafter the village opinion.leader and the urban medical expert will be referred to as the "village source" and the "urban source," respectively. 21 gave the researcher reason to believe that the experimental modes are distinctly different treatments. Subject's perceived homophily with source.--Systematic selection of sources for the alternate treatment modes based on Objective assessment Offers no assurance that subjects will similarly discriminate between the sources. Unless subjects can make such a discrimination, it is impossible to explain the outcomes of the research in terms of the variable manipulated in the treatment modes. Reviewing the literature on relational concepts of communication, Rogers (1969b, p. 4) stated, "The transfer of ideas most frequently occurs between a source and a receiver who are alike." Since the effectiveness of a communication is highly related to the degree to which a receiver perceives himself as similar to a source, it was important in the present research to investigate the similarity in the relationship existing between source and subject as perceived by the subject. Specifically, it was necessary to determine whether homophily was dif- ferentially perceived between the urban source and the village source relative to the subject's assessment. Perceived homophily is the degree to which pairs of individuals who interact perceive themselves as similar in certain attributee. If subjects could discriminate between the two sources on the basis of homophily, then 22 the alternate treatment modes could be considered a reliable variable for eXperimental manipulation. The attributes of perceived homophily which were measured are (1) education, (2) socio-economic status, and (3) modern value orientation. Explanation of these attri- butes is furnished in Chapter II. The first research question was to ascertain the degree to which subjects perceived source differences on the basis of homophily. All hypotheses in the present 13 study are stated in the null form. Hypothesis I is offered in its null form as follows: Hypothesis I: Subjects' degree of perceived homophily ' with source does not differ across the treatment modes. Independent Variable B: Degree of Modernity of Subjects Modernization is a process by which individuals anticipate and/or adjust to change which is directed toward a technologically complex style of life. Modernity_is the state of an individual on the continuum of modernization. Some individuals, by virtue of a more modern orientation toward life than their fellow villagers, are l3A hypothesis can never be proved with absolute certainty, since there is always a probability, howsoever small, that statistical differences between sample means occur simply by chance. However, a null hypothesis, stating that there is no significant difference in sample means, can be rejected with certainty if statistical differences fall in the region of rejection. The region of rejection consists of a set of possible values which are so extreme that when the null hypothesis is true, the probability is very small that the sample actually observed will yield a value which is among them (Siegel, 1966, p. 13). 23 more receptive to new ideas and practices (Inkeles, 1969, p. 210). Television personnel may wish to tailor their programs to these relatively modern villagers in an attempt to speed up the adoption process, or they may have reason to tailor their programs to the more traditional villagers in order to establish a motivational base for future action programs in the field. If findings in the present research show that the two types of television sources are differ- entially effective in imparting knowledge, changing atti- tudes, and motivating subjects to seek further family planning information, then these findings can be of value to television personnel in selecting the appropriate type of source for reaching subgroups of viewers in their rural audience. Under Hypothesis II, subjects' degree of perceived homophily with source across levels of modernity (high, above average, below average, and low) was tested. To obtain modernity scores, the following dimensions of modernity were measured. 1. Education 2. Mass media exposure 3. Cosmopoliteness 4. Political knowledgeability 5. Community participation 24 Reasons for selecting these dimensions are explained in Chapter II. Hypothesis II is formally stated in null form as follows: Hypothesis II: Subjects' degree of perceived homophily with source does not differ across levels of modernity. The next research Objective was to find out whether subjects at the various levels of modernity differentially perceived homophilic relationships with the village and urban sources. HomOphily scores were compared across treat- ments in order to determine whether possible differences in the degree of perceived homophily in the alternate treat- ments are attributable to levels of modernity. According to theoretical prediction, subjects at all levels of modernity would perceive themselves as more homophilous with the village source than the urban source, although to different degrees. The high modern subjects would perceive themselves relatively more homophilous with the village source than the low modern subjects. Since by definition a village Opinion leader is only slightly more modern than the average villager, it was predicted that high modern subjects would perceive themselves as having a somewhat better education, higher socio-economic status, and more modern values (the attributes of perceived homophily measured) than the village source. However, in comparing homOphily relationships across treatment modes, the high 25 modern subjects would consider themselves relatively more homophilous with the village source than the urban source. Figure l, visually illustrates the theoretically predicted relationships between high and low levels of modernity and perceived homOphily in the alternate treat- ment modes. To compare the relationships between levels of modernity and perceived homophily across treatment modes, the following null hypothesis was tested. Hypothesis III: Subjects' degree of perceived homophily with source is not a function of treat- ment mode/modernity interaction. Optimal heterophily as related to modernityge- Although a high degree of similarity between individuals facilitates communication, it does not usually facilitate the flow of new ideas (Rogers, 1969b, p. 20). In a change- oriented situation in which the objective of a change agent14 is to influence his client,15 the change agent's orientation should be somewhat more modern and expert than that of his client. If both parties are very much alike, change cannot occur, since the client is as modern as the change agent (Rogers, 1969b, p. 13). On the other hand, if the degree of dissimilarity, or heterOphily (the polar opposite of — 14A change agent is a person whose purpose is to influence innovation decisions in a direction deemed desira- ble by a change agency (Rogers, 1969a, p. 169). 15A client is a person whom a change agent seeks to influence. _ ” 26 High" Urban Expert Treatment Mode II ’5 s c -d 4.) c o 0 >1 p '2 3 High Modern Subjects c o 2 Village Leader Treatment Mode I Low —~ Low Modern Subjects Figure l.--Theoretically predicted relationships between high and low levels of subjects' modernity and degree of perceived homophily with sources in the alternate treat- ment modes.* *The shorter the arrow, the greater is the degree of perceived homOphily with the source. 27 homophily), between the parties is too great, meanings are not shared and messages become distorted. Rogers theorized that there is an Optimal distance between two parties in a social change situation that can Rogers termed this distance lead to desired results. The present "optimal heterophily" (Rogers, 1969b, p. 11). researcher further theorized that the heterophily distance that is optimal differs, depending on the degree of Imodernity of the client--the more modern the clienty the lgreater the distance than can be epanned before reaching ;§he level of optimal heterophily. Applying these theoretical assertions about Optimal kneterophily to the present research, it was reasoned that time more modern subjects in the sample (clients), because (hf? greater receptivity to new ideas, would be influenced 133' a relatively heterOphilous and urban-oriented source (cflnange agent). Prone to make independent evaluations, these subjects would tend to respect the judgment of the Utrioan source appearing on television. They would perceive TliJn a more credible source of information than the village Source in matters pertaining to family welfare. On the other hand, less modern villagers are usually irlfluenced by those individuals whose attitudes, values, 'aIlCi ibeliefs conform to village norms. Opinion leaders in a. "fiLllage serve as role models for traditional villagers alléi often legitimize appeals coming to the village from 28 outside sources. For the less modern subjects in the sample, the village source appearing on television would be perceived as more credible than the urban source. The researcher planned to test the Optimal heterophily assertion by measuring the differential credibility of the alternate sources as perceived by subjects differing in degree of modernity and perceived homophily with the source. Optimal heterophily is Visually presented in Figure 2. ‘Eerceived Source Credibility as Igp Intervening Variable Perceived source credibility is the degree to which figcommunicator ispperceived as a competent and trustworthy Enource of information. The rationale for using competence and trustworthiness as dimensions of perceived source credibility and definitions of these terms are provided 111 Chapter II. To test subject's degree of perceived source credi- ijrlity as a function of the relationship between his degree (Di? modernity and his degree of perceived homophily with SOurce, the following null hypotheses are presented: Subject's degree of perceived compe- tence in the source is not a function of the relationship between his degree of perceived homophily with the source and his degree of modernity. Hypothesis IV: Subject's degree of perceived trust in the source is not a function of the relationship between his degree of perceived homophily with the source and his degree of modernity. Hypothesis V: 29 ,.NM«HAAHOOHO mounom Oo>woouom cam maflrmofion OO>HMOHOQ mo moummo cam mpwcumpofi .muooflndm mo mao>ma 30H one Ema: cmospon mmHEmcoflumHmH may mcflsonm coauuommm mafinmomouon Hmfiflumo on911.m musmfim 30A I >HH£QOHODO£ Hmcwmumz.ll..lI mavennsm :HOOOS BOA saardoumpme Hmeflumo H moo: pcoEpmouB \ Hocmon mmcaafl> \. w 0 z . w I/ \ I muommbsm sumac: roam . w. 1. rA 3 spHHHbaemuo o oousom AUAVQX W Oo>flooyom m. n HH opoz usoEuonB unomxm menu: K as 30 The Effects Variables Measuring Cognitive, Effective, and Motivational Change The television effects in the present study are (l) knowledge-gain, (2) knowledge retention, (3) attitude change, (4) persistence of attitude change, (5) information- seeking behavior change, and (6) persistence of information- seeking behavior change. Effects 2, 4, and 6 were measured four weeks after exposure to treatment. If differential effectiveness between the treatment Inodes exists, then these differences should be reflected :in terms of the television effects. Although it was theoreti- <:ally predicted that the village source would be perceived 13y subjects at all levels of modernity as more homophilous 'than the urban source, high modern subjects should tend to sscore higher on measures of the television effects as a aresult of exposure to the urban source. The rationale for 'this prediction is that high modern subjects should tend 1:0 perceive the urban source as more credible than the \rillage source. Low modern subjects, perceiving the village Source as both more homophilous and more credible than the Taliban source, should tend to score higher on measures of the television effects as a result of exposure to the Vi llage source . Confirmation of the homophily and credibility hl’Pcheses would indicate that homophily and credibility, Ipsalttzeptual variables intervening between the experimental 31 stimulus and the television effects, had a bearing on the effects obtained. The optimal heterophily assertion is disgramed in its expanded version in Figure 3. The six television effects were compared across treatments to find out whether they are attributable to (l) the treatment modes, (2) levels of modernity, or (3) an interaction between the treatment modes and levels of modernity. Knowledge—gain.--Knowledge-gain is defined as any Ipreeposttest change in a person's cognitive learning koehavior resulting from a specific learning experience. A knowledge test was administered to respondents :immediately preceding and following exposure to television ‘treatment. Since the same facts and concepts were presented :in both treatment modes, knowledge-gain scores could be <:ompared across treatments. The knowledge-gain hypotheses are as follows: Hypothesis VIa: Knowledge-gain does not differ across the treatment modes. Hypothesis VIb: Knowledge-gain does not differ across levels of modernity. Hypothesis VIc: Knowledge-gain is not a function of treatment mode/modernity interaction. Knowledge retention.--Knowledge retention is the SZSESLree to which a person is able to remember stimulus inputs a‘C3.$ired as a result of exposure to a specific learning SiiSIaeerience. 32 . muhuwmmm GOHmHK/Qflwu. OLD. one mmcoE psoEumouu oswlaooBuon msflco>HOch moanmaum> Hospmoouom 03p .huflaflnfloouu OOHDOm ochNwHAMOEOS mo monmoc com wpflcuocofi .mDoonQSm Mo mHo>oa 30H cam amen coo3pon marmGOHumaoH on» mcfl3onm cofluuommm >HH£QOHOpo£ HmEflumo oseII.m musmflm maflrmouopos Hmcflmumz III-III maflnmououo: Hmeflpmo IIII. muomflndm :Hoooz BOA .30q omcmno Hofl>mcon mcflxoom Icofiumauowcfi mo momcumwmumm .OI H 0602 l. pcoaumona omcmno Hofl>mnon Hopped ommHHfl> mcfixoomIcoHumEHOMcH .mI \ m own 5m /, . omcmno cwocdm rmamr, \ opsuflpum mo oosoumwmnom .vl . h a a H OH mango 3339a .mI pght o tI mousom 08 GOHDGODOH omOOHBOQM .ml .Oo>fimouom sawmIomooa30cx .HL muoommo coflmfl>ofio¢ ore II HH moo: unofiummne uuomxm swan: a .s mnnurauoo Karuxapow 33 Four weeks after experimentation the knowledge test was again administered in order to ascertain the differential amount of knowledge subjects remembered from the television treatments. By measuring knowledge reten- tion, the stability of the television effects was estab- lished over time. Hypothesis VIIa: Knowledge retention does not differ across the treatment modes. Hypothesis VIIb: Knowledge retention does not differ across levels of modernity. Hypothesis VIIc: Knowledge retention is not a function of treatment mode/modernity inter- action. Attitude change.-—An attitude is defined as a set <>f interrelated beliefs or predispositions focused on an (object in a situation which comprise an agenda for action (Rokeach, 1968, p. 120). Attitude change is a shift in ean attitude from a previous position. While it was not anticipated that either treatment nubde would alter strongly held negative attitudes toward fénmily planning, minor positive shifts in attitude might Cuzcur. Also, the firming of attitudes in a positive direction might occur among individuals who are ambivalent toward fami ly planning . Hypothesis VIIIa: Attitude change does not differ across the treatment modes. Hypothesis VIIIb: Attitude change does not differ across levels of modernity. Hypothesis VIIIc: 34_ Attitude change is not a function of treatment mode/modernity inter- action. Persistence of attitude change.--Persistence of attitude change is the degree to which a respondent's attitude remains stable over time. If attitude change in a positive direction is evidenced as a result of exposure to a more homophilous and credible source, then this change may persist four weeks after exposure. The post posttest of attitudes toward family planning concepts presented in the television treat- Inents offered further indication of the stability of attitudes over Hypothesis Hypothesis Hypothesis time. IXa: IXb: IXc: Persistence of attitude change does not differ across the treatment modes. Persistence of attitude change does not differ across levels of modernity. Persistence of attitude change is not a function of treatment mode/ modernity interaction. Information-seeking behavior change.--Information- jséaeking behavior is the manifestation of a desire to know {HCDIe about an idea or_practice. Information-seeking fisflbavior change is an increase or decrease in a desire JEEL_1earn about an idea or practice. Short-term exposure to mass media rarely results £111 loehavioral change (Klapper, 1961). Although television Vi@Wers are not likely to adopt family planning methods 35 as a result of exposure to one television program, they might be motivated to seek additional family planning information from interpersonal change agents. Thus, it is important to find out whether television has the potential for creating in individuals a desire to explore new ideas, to enlarge their capacity for new experience. Hypothesis Xa: Hypothesis Xb: Hypothesis Xc: Information-seeking behavior change does not differ across the treatment modes. Information-seeking behavior change does not differ across levels of modernity. Information-seeking behavior change is not a function of treatment mode/ modernity interaction. Persistence of information-seeking behavior change.-- IPersistence of information-seeking behavior change is the (degree to which a desire to seek information remains stable over time . There is no research evidence to indicate that ITIOtivation to seek information will persist over time as a :result of exposure to an influential source. The following hyrpotheses are Hypothesis Hypothesis Hypothesis Offered for their heuristic value: XIa: XIb: XIC: Persistence of information-seeking behavior change does not differ across the treatment modes. Persistence of information-seeking behavior change does not differ across levels of modernity. Persistence of information-seeking behavior change is not a function of treatment mode/modernity inter- action. 36 The value of the present study will depend not only on its usefulness to television practitioners in Delhi State, but also on its contribution to a body of knowledge about effective develOpmental communication. Studies unrelated to theory offer little more than isolated bits of information. Findings obtained from such studies can neither predict regularities in human behavior nor be generalized beyond the limits of the immediate experiment. In the present experiment, the variables for study are related to theoretical propositions from which they were deduced. These relationships are discussed in Chapter II. CHAPTER II THEORETICAL FOUNDATIONS FOR THE DEVELOPMENT OF EMPIRICAL TESTS Overview of the Research Question The purpose of the present study was to investigate the differential effectiveness of two types of television sources used to impart a family planning message on televi- sion. The sources selected were a village opinion leader and an urban medical expert. The television effects, knowledge-gain, attitude change, and information-seeking behavior, were measured at the time of exposure to tele- vision treatment and again four weeks after exposure. It was theoretically predicted that the low modern subjects sampled would perceive the village source as more homophilous with themselves and more credible than the urban source. These subjects would tend to obtain higher scores on measures of the television effects as a result of exposure to the village source rather than the urban source. High modern subjects would also perceive the village source as more homophilous with themselves than the urban source. But influenced by urban values and norms, these 37 38 subjects would perceive the urban source as more credible than the village source and would tend to obtain higher scores on measures of the television effects as a result of exposure to the urban source. The purpose of this chapter is to eXplicate the variables for empirical investigation and the theoretical propositions from which they were deduced. Working defini- tions for each variable studied will also be provided. Measures of Individual Modernity Modernization is agprocess by which individuals anticipate and/or adjust to change which is directed toward a technologically complex style of life. Modernityyis the state of an individual on the continuum of modernization. The process of becoming modern is not merely one of adjusting to conditions as they are. Rather it is one of adjustment under changing conditions, of predicting the need for change. Because one of man's basic drives is to increase order, and thus exercise control over his environ- ment, modernization is a dynamic process. As men progress toward desired goals, technological advancement kicks the society into new gear in anticipation of new needs to be met, new goals to be achieved. In Rapoport's words (1967, p. 191), movement toward higher order implies "a struggle . . . against the chaotic tendency of nature, a struggle of creating order out of chaos, knowledge out of ignorance, insight out of illusion, freedom out of compulsion." 39 In the present study, the individual, rather than his social group, was the unit for investigation. The following indicators of individual modernity were taken from a largerlist offered by Rudolph and Rudolph (1967, p. 3): 'Modernity' assumes that local ties and parochial’ perspective give way to universal commitments and cosmopolitan attitudes; that the truths of utility, calculation, and science take precedence over those of the emotions, the sacred, and the non-rational; that the individual rather than the group be the primary unit of society and politics. While there appears to be a fair amount of agreement among scholars, both Western and non-Western, as to the individual attributes which constitute modernity, assessment of modernity needs to be related to the cultural context in which a study is conducted. Instruments based solely on a Western yardstick of modernity would fall short in detecting many valid indices of change taking place in the villages of India. What may appear to be a radical departure from societal norms from a non-Western point-of—view may, by Western standards, be judged as an insignificant event. As Srinivas (1964, p. 127) observed, "Change is much more serious and pervasive in small and stable societies where the same people are invOlved with each other in a number of relationships, than in hugh industrial societies . . . where relationships between individuals are more specialized and disparate." 40 In an attempt to measure individual modernity as objectively as possible, the following four models of modernization developed by Sen and Roy (1966, p. 51) were considered in the present research. (1) (2) (3) (4) Westernization: Westernization is the process in which an individual or a group internalizes the values and behavioral patterns of the Western industrial societies. This aspect of moderniza- tion can be more readily observed in large cities of India where individuals are often exposed to the values of Western societies. Urbanization: Urbanization is a process in which an individual or a group of individuals increas- ingly looks to the urban industrial society as its reference group. Renaissance: Renaissance . . . is a process of synthesis in which an individual or a group is highly aware of other systems, selects new ideas and practices emanating from these systems, internalizes and adopts them, but at the same time maintains strong links with tradition. Sanskritization: Sanskritization is a sub- process of renaissance in which an individual or a group goes through an initial stage of traditionalism to obtain status. By utilizing these models in concert, individual mani- festations of social and technological change may be identi- fied. Although heavy reliance was placed on the urbanization model in the present study, some of the items used to measure modernity relate to the other three models as well. DevelOpment of an Empirical Test to Measure Individual Modernity Modernization is a multidimensional construct; that is, many components, taken together, comprise an individual's 41 degree of modernity. By pooling the items used to measure the components, a modernity score was obtained for each respondent. The components selected to measure modernity in the present study are (1) education, (2) mass media exposure, (3) cosmopoliteness, (4) political knowledgeability, and (5) community participation.1 Education.--Since the type of socialization an individual undergoes determines in large measure how he will react to his environment, education emerges as a primary and, perhaps, antecedent variable in explaining modernity (Inkeles, 1969). The school serves in two fundamental ways to facilitate changes in the socializa- tion process: It broadens knowledge through training in literacy and substantive content, and it molds values and beliefs in preparation for adult responsibilities. Because of this dual role of the school, items measuring literacy and years of schooling were accorded equal numerical value in the modernity test. Mass Media Exposure.-—Education and literacy (generally facilitate mass media exposure (Rogers, 1969a, p. 328). It should be emphasized that degree of media exposure is not necessarily an indicator of modernity. {Hue degree of exposure to radio offerings which reinforce ‘ 1The interview schedule for empirical measurement of allvariables is furnished in Appendix A. 42 local norms may be directly related to a localite orienta- tion. However, the act of listening to the radio may reflect a certain desire for contact with outside sources and for this reason, degree of eXposure was measured. Type of exposure, however, is certainly an important factor for consideration, for, as Inkeles (1969, p. 210) indicated, "Persons who manifest modernization character- istics strive to keep up with news of national and inter- national import in preference to items dealing with sports, religion, or purely local affairs." In the modernity instrument, listening to radio programs which provide a modernizing influence was thus given twice as much weight as programs which likely reinforce the status quo. Both degree and type of eXposure to radio and films were measured. Cosmopoliteness.--Cosmopoliteness is an individual's degree of orientation toward environmental stimuli external to his own social system. Attitudinal statements represent- ing either a high or low degree of cosmopoliteness were presented to the reSpondents. The statements called for dichotomous answers such as "agree" or "disagree." Attitudinal rather than behavioral measures of cosmopoliteness were used since the latter measures may not be equally valid indices of modernity among farmers and village dwellers such as fruit sellers, for example, who make daily trips to Delhi to sell their produce. These latter individuals tend to interact in the city with other 43 commuters of similar status; their contacts with modern urban persons are likely to be impersonal. On the other hand, farmers make fewer trips to the city than fruit sellers. But in negotiating farm loans and securing consumer markets and farm equipment, farmers are involved in closer and more personal relationships with urban persons of the middle class who are moving upward on the modernization continuum. Since the nature of urban contact may be directly related to modernity, the present researcher decided to avoid behavioral measures of cosmopoliteness such as "degree of interaction with outside sources," and "number of trips to the city." Attitudinal items selected for measurement in the present research were those indicat- ing preference for ideas and behaviors originating outside the village social system. These measures were considered to be valid indices of modernity for farmers and village commuters alike. Political knowledgeabilipy.--Political knowledge- aabilipy is the degree to which an individual comprehends gfiacts essential to his functioning as an active and eeffective citizen (Rogers, 1969a, p. 56). According to Rogers (1969a, p. 328), education, Inass media exposure, and cosmopoliteness are likely to JJead to a higher degree of political knowledgeability lDeacause an individual possessing these qualities has 44 access to more and varied information. Political knowledge- ability is thus a consequent of the three primary components of modernity. Community_participation,——Communipy participation is the degree to which an individual takes an active part in or is a member of formal community organizations. As previously mentioned, the move from dependence solely on familiar ties to a village frame-of—reference represents a significant break with the past for some individuals. Affiliation with village organizations which are traditional in nature, such as religious organ- izations, for example, received only half the weight of affiliation with organizations directed toward the achieve- ment of modern objectives, such as farm cooperatives or village improvement societies. The total score Obtained from measurement of all five components of modernity represented a composite moder- 2nity score for each respondent. By pooling the components (of modernity, it was possible to calculate each individual's <flegree of modernity on a continuum ranging from high to low . Measures of Homophily Homophily is defined as the degree to which pairs 51f individuals who interact are similar in certain attri- 1DIItes. 45 Typically, researchers have measured homophily through the use of objective tests. That is, researchers usually Observe certain characteristics shared by individu- als who interact. When characteristics such as age, race, social status, values, etc. are positively correlated between two individuals, or among members of a group, the researcher considers the result to be a measure of likeness or homophily. Measurement of perceived homophily differs from that of objective homophily in that the perceptions of the indi- viduals who interact are taken into account. When individuals are asked if they perceive the existence of similarities, the procedure is referred to as a measure of perceived homophily. Review of the research literature on homophily reveals that, to date, little attention has been paid to perceived homophily. Lazarsfeld and Merton (1954), for example, used objective measures to evaluate homophily among friends and friendship groups. They found a high correlation between friendship and the observed tendency toward similarity in values and social status. Triandis (1968, pp. 393-340), investigating simi- larities in the manner in which pairs of individuals cognitively process environmental events, found a direct relationship between cognitive similarity of individuals (as objectively measured by the researcher) and their ability to communicate effectively. 46 In developing countries, homophily is Often measured in change oriented situations in which the purpose of one party, the change agent, is to influence the other party, the client. Studies of homophily in change oriented situa- tions differ from other homophily studies in that the element of "intent" to influence the client is present. Furthermore, individuals who seek advice tend to interact with change agents slightly more modern than themselves. Almost all of the studies of homophilic relationships conducted in change oriented situations, to date, have been in the field of agricultural innovation diffusion. Most of these studies have used an objective measurement of homophily. The Allahabad Agricultural Institute of India (1957) noted that village-level change agents with only an elementary education were more effective in inducing adoption of agricultural innovations among village persons than change agents with a high school or college education. The Allahabad study indicates that degree of similarity in formal education is directly related to effective client/ change agent interaction. In a study of Colombian peasants, Ramos (1966) found that the more social distance Colombian peasants "perceived" between themselves and the extension workers, the less favorable attitudes they held toward the change 47 agents, the less credibility they placed in them, and the less interpersonal communication they had with them. Ramos thus studied homophily as a perceived phenomenon. Rogers (1969a, p. 237), in studying heterophily (the polar opposite of homophily) in three modern Colombian villages, revealed that opinion seekers tend to obtain information and advice from opinion leaders who are slightly more competent than themselves in technical know— ledge and innovativeness. Rogers found that 30 per cent of the opinion seekers sought advice from individuals in their own adopter category, whereas 51 per cent of the Opinion seekers went to slightly higher adoptor categories for advice. Only 19 per cent went to sources of informa- tion lower on the adoption ladder than themselves. However, the degree to which Opinion seekers interact with higher adoptor categories differed among the more modern and traditional villagers. Rogers concluded that the more modern the village, the more likely it is that opinion seekers will seek advice from individuals higher on the adoption ladder than themselves. The Present Study Compared with Prior Homophily_Studies The present study is similar in some respects to those agricultural diffusion studies conducted in a social change oriented situation. The objective of the television source was to provide the viewer with information about 48 the benefits of adopting the small family norm. Although the element of interpersonal communication was absent, it was asserted that the television medium, interposed between source and receiver, would not alter basic homophily/ heterophily relationships with the effects. Viewers would be able to perceive similarities or differences between themselves and the television source, since tele- vision, with its elements of motion, timeliness, and spontaneity—-and its appeal to both auditory and Visual senses--comes closer to stimulating direct experience than any other mass medium. The present study differs from other social change oriented studies of homophily in two fundamental ways, how- ever. In prior research on homophily, direct interaction between client and change agent has been possible. In the present instance, the communication flow was one way, since one party functioned only as receiver and the other as sender of the message. While the viewer could psychologically interact with the television source--he could make covert evaluations of the message or express his reactions to other members of the audience--the source could not modify his own behavior in response to the viewer. Because of the lack of feedback,2 homophily was analyzed in terms of the relationship the viewer perceived as existing between 2Feedback is a response to a communication by which the sender of a message is informed by the receiver con- cerning the former's success in accomplishing his communica- tion objective. Hy-t 49 himself and the source. Thus, communication was a non- reciprocal phenomenon in the present case. Second, client-change agent interaction Often continues over a period of months or years. The parties come to know each other well, and, therefore, the researcher has reason to believe that homophily and increased inter- action are causally related, or, to paraphase Heider (1958, p. 185), "that personality similarity induced liking, and liking induced the assumption of similarity." In the present situation, the subject was exposed to the source for the duration of the television program only, and there is no way of knowing what the interaction pattern would have been in a voluntary face-to-face en- counter. However, to compensate for short-term formal exposure to the television source, the subjects were provided with a good deal of background information about the source to enable them to make homophily evaluations. Immediately preceding the experimental treatment, the research project supervisor displayed a large photograph of the source before the subjects and simultaneously inform- ed them about the place of residence, formal education, and occupation of the source "who will tell you a story on tele- vision tonight about real families who live in a village much like your own." Also in the introduction of the program, a television "compere" (referred to as an announcer 50 in the United States) made similar remarks about the source. Furthermore, during the 14 minute program, sub- jects were able to assess the source's physical character- istics, speech, mannerisms, dress, and general demeanour. As a result of these cues, it was believed that subjects would be able to form distinct impressions about the source in order to make appropriate evaluations. These cues formed the basis for empirical measurement of homo- phily which immediately followed the experimental treat- ment. Development of the Homophily Instrument HomOphily, like modernity, is a multidimensional construct. That is, homophily evaluations are based on any number of different attributes, such as age, sex, social status, etc., the attributes selected for evaluation depend- ing on the context in which a communication is given and the nature of the topic discussed. In measuring homophily in the present instance, it was important to select only those attributes that relate directly to the credibility of the source, i.e., attributes thought to have a direct bearing on the subject's evaluation of the source's compe- tence and trustworthiness. Tests measuring both objective and-perceived homophily were administered to subjects at the termina- tion of the experimental program. The objective test served as a validity check on the perceptual test. 51 In the objective test the attributes that were selected are (1) education, and (2) occupation. Subjects were reminded of the source's education and occupation and were then asked to furnish the same information about themselves. It was thought that this technique of forcing comparisons would bring up to the conscious level, evalua- tions on the specific dimensions of homophily subsequently investigated in the perceptual test. Since objective measurement preceded subjective measurement, the former might serve to prompt the appropriate evaluations called for in the latter. The three attributes of perceived homophily measured are (1) educational status, (2) socio-economic status, and (3) modern value orientation. Each of these attributes will be disucssed in turn. Educational status homophily,-—Educational homophily was measured as the degree to which a viewer perceives correspondence between himself and the television source in amount of formal education. The amount of schooling a subject perceives a source as having may be directly related to the subject's assessment of the source's competence. A village person who has had no formal schooling may consider a university trained and professionally experienced medical expert to be too theoretical in his analysis of village life and rural family problems. A village person may perceive a 52 village opinion leader, with only a few years of formal schooling in the Hindi medium, as a more qualified source to offer advice. As Mathur (1969, p. 6), Joint Director of Family Planning for All India Radio, pointed out, "The local leaders of opinion, persons respected in the villages . . . and even trained 'dias' (midwives) who speak the same idiom as an uneducated village person, carry far greater conviction when they speak about limiting families." The more educated village person who is motivated to seek new information from outside sources may, on the other hand, look to the medical expert as a competent and trustworthy source. Educational status homophily was measured in terms of the amount of perceived education of the source relative to the subject's education. Socio-economic status homOphily.-—Socio-economic homophily was measured as the degree to which a viewer perceives correspondence between his social and economic position and that of the television source. For the low modern viewer, the village opinion leader may instill more confidence because he is a role model representing an acceptable way of life, yet perceived to be of sufficiently higher status that the viewer himself. On the other hand, a high modern viewer who attaches value to urban norms may perceive the medical expert as a person whose judgment he can respect and trust. 53 Items relating to effective income and occupational role status were used to measure degree of perceived socio- economic homophily. Modern value homophily,——Modern value homophily was measured as the degree to which a viewer perceives correspondence between his values related to modernization and those of the television source. In a transitional society, where social change communication is continually impinging on the village from outside sources, a villager may tend to compare his outlook toward social change with that he perceives the change agent to hold. This comparison may be made in situations in which the visual media are used as well as in inter- personal situations. Perceived similarity in values toward social change may influence the viewer's evaluation of the source's trustworthiness in a family planning context. The high modern viewer may perceive the urban source to be a person who is sincerely interested in helping village people to plan ahead; who shares his desire for social change; who, like himself, is striving to improve his condition and provide opportunities for his family. The traditional viewer may characterize the urban source as an urban agent of persuasion and, thereby view his motives with suspicion. 54 Modern values selected for measurement are (1) innovativeness, (2) achievement motivation, and (3) cosmo- politeness. A definition of each of these components is provided: (1) Innovativeness is the degree to which an indi- vidual is earlier than others in his social system to adopt new ideas (Rogers, 1969a, p. 56). The less innovative viewer may be influenced by the village source, whom he perceives to be only slightly more innovative than himself. The more innovative viewer may perceive the urban source as sharing his interest in new ideas and, as a result, confer more credence upon him. (2) Achievement motivation is a social value that emphasizes a desire for excellence in order for an individual to attain a sense of1personal accomplishment (Rogers, 1969a, p. 54). The experimental programs emphasized the fact that couples who plan for their families are better able to provide opportunities for their childrens' education and for self-improvement. A villager may interpret this com- munication in one frame-of-reference when it is imparted by a respected village leader and quite another when it is imparted by an urban expert. As evidenced by Asch (1952, p. 420), "The meaning of any communication is partially determined by the reputation of the person who makes the statement." The traditional viewer may suspect that the source reputed to be an eXpert is advocating Opportunities that estrange an individual from his village way of life. 55 The same message delivered by a village leader may be inter- preted as offering opportunities that help a rural person to better adjust to his indigenous environment. (3) Cosmopoliteness is an individual's degree of orientation toward environmental stimuli external to his own social system. If a source is considered an outsider, his sugges- tions about village improvement may hold less relevance. The traditional viewer may, therefore, place more credence in a source perceived as more like himself in values per- taining to system orientation. In measuring perceived homophily, a modified version of the Cantril ladder was used.3 Respondents were presented with a picture of an ll-step stairway such as they might have in their houses, and were questioned accord- ing to instructions developed by Kapoor.4 The instructions given to investigators asthey relate to perceived educa- tional homophily are as follows: (To respondent) Think of a person you know who is highly educated. (Pause) Have you thought of him? What is his name? This is a staircase such as you have in your house. (Show) There are many steps on this staircase. (Point them out) 3This subjective rating technique was developed by Cantril (1965) to determine level of individual aspirations. By considering the best and least satisfactory life imagina— ble, an individual estimates his present situation relative to self-imposed limits. 4Dr. S. D. Kapoor, Psychologist, is Senior Research Officer for the Central Family Planning Institute, New Delhi. 56 I will place the educated person you mentioned at the top of the staircase. (Be sure that respondent is Observing) ' Now, please think of a person who has no education. (Pause) Have you thought of him? What is his name? I will place this person with no education at the bottom of the staircase. (Be sure that respondent is observing) Now between these two persons, on which step will you place (name of source)? Give your answer by putting your finger on one of the steps. Now tell me, between these two persons, on which step will you place yourself? Please put your finger on the step. By subject's positioning of himself in relation to the source on the stairs, degree of perceived homophily with the source was established. Differential degrees of perceived homophily with source could thus be compared across the treatment modes. Furthermore, the relationships between perceived homophily, modernity, and perceived source credibility could be compared across the treatment modes. Measures of Perceived Source Credibility Perceived source credibilipy is defined as the degree to which a communicator is perceived as a competent and trustworthyysource of information. I Many of the source credibility studies conducted in developing countries might more appropriately be called 57 "channel"5 credibility studies. That is, the credibility of such channels of communication as interpersonal, film, radio, neWSpaper, etc. are compared in order to determine the relative credibility of each channel. Ramos (1966) in Colombia, conducted a channel credibility study using the paired comparison method. Six "sources," or suppliers of agricultural information, were rank-ordered by village respondents according to the degree to which they were preceived as credible. Sources were presented to respondents in all possible paired combinations and respondents were asked to select the more credible source in each pair. In descending order of credibility, source rankings were extension agent, school teacher, radio, neighbors, salesmen, and neWSpapers. The paired comparison method was used by Shankariah (1969) in India to compare the perceived credibility of seven suppliers of agricultural information in a progressive and a non-progressive village. In the progressive village, personal contact with agricultural experts was perceived as most credible, followed by radio, progressive farmers, block extension agency, demonstration, bulletins, and newspapers. In the non-progressive village, demonstration received the highest ranking with agricultural experts ranked third and radio ranked fifth among the respective 5A channel is the vehicle through which a message is transmitted from a sender to a receiver. 58 sources of information. Shankariah reasoned that the difference in credibility ratings between the two villages may have been due to the differences in value orientation of respOndents in the respective villages. To date, perceived source credibility as an effect of experimentation has received little attention among researchers in Western countries. Traditionally, sources are objectively assigned to treatment and, thus, credibility is handled as an experimental variable. During the 1950's, Hovland and others conducted numerous experiments using credibility as an experimental variable. In one of his classic experiments on source credibility, Hovland (1951, pp. 635-650) rated sources as of high or low credibility and systematically assigned them to treatments. The message, a persuasive appeal, was held constant across treatments and the effects studied were the recipient's evaluation of the presentation and acceptance of the opinions advocated by the communicator. Hovland found that opinions of the recipients were changed immediately after communication in the direction advocated by a credible (trustworthy) source. In contrast to the previously mentioned study, when source credibility is treated as an effect of com- munication, the variable must be measured in terms of the recipient's perceptions of the source. The term, "per- ceived," implies that the source is evaluated on the 59 recipient's criteria of credibility, and not by criteria objectively assigned by the researcher. Thus, when source credibility is treated as an outcome of research, the dimensions recipients use to evaluate sources must be established and Operationalized. Credibility as a Multidimensional Construct Perceived credibility, like perceived homophily, is a multidimensional construct measured on several dimen- sions; the dimensions used for evaluation of a source depending on the context in which the communication is presented and the topic discussed. Hovland (1965, p. 21) was aware of the existence of more than one dimension of credibility when he theorized: A recipient may believe that a communicator is capable of transmitting valid statements, but still be inclined to reject the communication if he suspects the com- municator is motivated to make nonvalid assertions. It seems necessary, therefore, to make a distinction between (1) the extent to which a communicator is perceived to be a source of valid assertions (his 'experiences') and (2) the degree of confidence in the communicator's intent to communicate the assertions he considers most valid (his 'trustworthiness'). In any case, the weight given a communicator's assertions by his audience will depend upon both of these factors, and this resultant value can be referred to as the 'credibility' of the communicator. Other researchers have considered criteria other than expertness and trustworthiness as bases for source evaluation. Krech and Crutchfield (1962, p. 231), for example, asserted that, among other characteristics, 6O "attractiveness" and "group affiliations of the communicator" are important in determining communicator effectiveness. Such characteristics as "sincerity," "prestige," "attractive- ness," "awe," "affection," etc. have been similarly con- sidered by researchers over the past several years. Although it is now generally accepted that credi- bility is a multidimensional construct, until recently little attempt was made to disentangle the various factors contri- buting to differential evaluations of credibility. Berlo (1969, p. 564) realized the need to empirically determine the dimensions of credibility when, in reference to Hovland's conceptualization of expertness and trustworthiness, he stated, . . . NO evidence has been obtained of the stability and independence of either perceived trustworthiness or expertness, nor of the independence of either from such variables as sincerity, affection, admiration, prestige, and the like. If the latter variables are in fact inherent in trustworthiness, they should be so considered. If they are independent, there does not appear to be any logical or theoretical reason to exclude them from the set of perceptions which the receiver has of a message source; i.e., to reject them as evaluative criteria which affect the influ- ence of the source as a transmitter of information. In factor analyzing semantic differential scales to determine the evaluative criteria of subjects in his study, Berlo (1969) found two psychologically independent dimensions of credibility. Together, these dimensions explained a large portion of the over-all variance. These dimensions Berlo termed "qualification," and "safety." A 61 third dimension, "dynamism," is statistically independent but may not be psychologically independent. According to Berlo (1969, p. 575), "The dynamism dimension can be con- ceived of as an intensifier. In other words, given an evaluation of a source as safe-unsafe or qualified— unqualified, the polarity or intensity of these evalua- tions of the source is intensified through perceptions of high dynamism." Some of the evaluative adjectives in the semantic differential scales most closely associated with the qualification dimension are "trained," "experienced," "authoritative," "informed," "educated," and "expert." The safety dimension include evaluative words such as "just," "honest," "friendly," and "safe." In comparing Berlo's empirically derived dimensions of source credibility with those conceptualized by Hovland, a considerable degree of isomorphism in meaning can be evidenced. The evaluative words generated by Berlo's subjects which come under the heading of qualification are apparently closely associated with expertness as defined by Hovland. This similarity was particularly striking, according to Berlo, when specific sources were evaluated in a relevant context, such as, for example, Jawahar Lal Nehru discussing India's neutralism. Used in a specific context, the term "qualification" can be considered approximately the semantic equivalent of eXpertness. Berlo found the safety dimension, however, to be more general than trustworthiness. Trustworthiness was conceptualized by Hovland as the perceived intent of the communicator, while safety included, along with intent, an affiliative relationship between source and receiver as perceived by the receiver. It would appear, then, that whether or not a communicator's motives are perceived as sincere and honest, depends on the reference group to which the communicator belongs. Operationalizing Perceived Source Credibility TF7 TTT‘ Ideally, an analysis of credibility such as that undertaken by Berlo in the United States should have been conducted in connectiOn with the present study. However, such an extensive undertaking in preparation for the present research investigation was not feasible. Since there was no reason to suspect that the dimensions used by Hovland would not hold in the present research context, and since it is important for media practitioners in India to know how members of their rural audience evaluate sources based on expertness and trustworthiness, these dimensions were selected for investigation with only one minor alteration. In the present research the term "competence" was substituted for expertness. Evaluation of a source's competence applies equally well to a village Opinion 63 leader as it does to an urban expert. A viewer may perceive a village leader as competent in matters per— taining to rural family life but not expert, since the term "expert" usually denotes specialized training. Competence was measured as the degree to which a viewer perceives a television.source as informed on the tOpic under discuesion. As previously mentioned, Berlo‘s safety dimension included affiliative relationship5.as well as intent. In the present study, however, affiliative relationship was considered an antecedent to the evaluation of source ‘trustworthiness. If, for example, the reference group to which the source belongs is perceived by a viewer as relatively homOphilous with his own, the source may be judged as a person likely to make statements based on motives related to the viewer's best interests. Thus, trustworthiness.in the present research was restricted to Hovland's narrower interpretation of the term as meaning the intent of the communicator. Trustworthiness was measured as thefdegree to which a viewer perceives the motives of\e1te1evision source as sincere and directed towardfihis beet interests. The dynamism of a communicator--his enthusiasm, and energy level--is also considered by Berlo to be a 5An affiliative relationship is an association between a pair offindividuais based on their belonging to a common reference group or groups. 64 possible criterion influencing perceived communicator credibility. In the present research, attempt was made to hold the dynamism factor constant.. Since there was no theoretical reason to believe that dynamism is neces- sarily associated.with a particular type of source, and since, according to Berlo (1969), dynamism may not be psychologically independent from competence and trust— worthiness, selection of both television sources was made on the basis of their having personality attributes which qualified them to be called "dynamic." These attributes of the sources are discussed in Chapter III. Both structured and unstructured questions were used to measure the dimensions of source credibility. preceding each unstructured question,.respondents were asked to consider whether or not the source was competent/ trustworthy. If the answers were affirmative, subjects were then asked,."Why do you think so?" Three coding categories were devised by the researcher to establish the degree to which respondents perceived the sources as competent and trustworthy. The categories were high, above average, and average. In the structured questions, respondents were presented with a series of statements made by the source in the program and were asked to evaluate the source on a three-point scale based on each statement. The state- ments served as stimulus cues to help the respondents 65 recall initial impressions formed of the source. Respond- ents' replies were forced into three categories by the interviewers. In evaluating source competence, for example, each stimulus statement was followed by, "Do you think (the source) knows enough about these things to make such a statement--by that I mean, does he know much, some, or very little about these things?" Degree of perceived credibility was thus measured in both structured and unstructured questions with the latter type Of question serving as a validity check on the former. Measures of the Television Effects Variables The six television effects variables measured are (l) knowledge-gain, (2) knowledge retention, (3) attitude change, (4) persistence of attitude change, (5) informatiOn- seeking behavior change, and (6) persistence of information seeking behavior change. Measures of each of these variables will be discussed in turn. Knowledge-gain Knowledge-gain is defined as any preeposttest changes in a person's cognitive learning behavior resulting from a specific learning experience. The knowledge test measured learning based on three of the lower levels of cognitive learning outlined by Bloom (1956). In ascending order of complexity, they are (1) knowledge, (2) comprehension, and (3) application. 66 Knowledge, according to Bloom (1956, p. 62), includes those behaviors and test situations which emphasize rememberinge-either by recognition or recall--of ideas, materials, or phenomena. A number of two-choice items in the knowledge test measured recognition of factual material included in the television programs. Referring to recall, Bloom (1956, p. 201) stated, For measurement purposes, the recall situation involves little more than bringing to mind the appropriate material. . . . The knowledge objectives emphasize most the psychological process of remembering. Test items relating to recall measured the respond- ent's ability to specify correctly certain fundamental information furnished in the program, such as, for example, "A school-going child needs such things as books, school funds, stationary, and uniforms. About how much does it cost per year to give these things to a boy in the 7th class?" According to Bloom (1956, p. 204), comprehension represents the lowest level of understanding . . . such that the individual knows what is being communicated and can make use of the materials being communicated without necessarily relating it to other materials or seeing its fullest implications. Questions testing comprehension were similar to the knowledge questions, except that they required of respondents the ability to interpret in their own words the information provided in the program. An 67 example of a comprehension question is "Can you name the possible reasons children fall sick?" Bloom (1956, p. 205) defined application as the use of abstractions in particular and concrete situations. In testing for application, respondents were asked to select appropriate answers from two-choice questions such as, "Who would be more help to you on the farm, an educated son or an uneducated son?" These questions were followed by, "Why do you think so?" Correct answers were those which conveyed the general meaning intended in the program, but credit was also given to all logical answers to the questions. Logical answers were determined by two Indian coders who are graduate students in sociology. It was assumed that respondents would be likely to give more and more logical answers after exposure to treatment. This assumption tended to be true across both treatment modes. The knowledge test was developed prior to the production of the experimental programs. In fact, program materials were systematically designed around the specific educational objectives that were to be tested. The knowledge test was administered immediately preceding and following exposure to treatment. Thus the differential amount of knowledge acquired as a result of exposure to the alternate treatment modes was established. 68 Knowledge Retention Knowledge retention is the degree to which a person is able to remember stimulus inputs acquired as a result of exposure to a specific learning experience. The knowledge test was again administered four weeks after the experiment in order to determine the dif- a I ferential amount of knowledge retained as a result of exposure to the alternate treatment modes. f.” Attitude Change é} An attitude is a set of interrelated beliefs or predispOsitions focused on an object in a situation which comprise an agenda for action (Rokeach, 1968, p. 120). Attitude change is a shift in an attitude from a previous position. Rokeach theorized that an attitude is comprised of beliefs that are states of expectancy about the environ- ment. The statement, "I believe that . . . " is the foundation from which an individual's judgments are made. Since beliefs are functioning in the evaluation of any given stimulus object, it is beliefs which direct an individual toward a particular response. In the present research, beliefs were focused on the object, which is the small family norm. The situation in which the small family norm was evaluated relates to the concepts presented in the experimental programs, i.e., 69 the health, educational, and economic benefits derived from ad0ption of family planning. Both object and situa- tion were combined in most of the items used to measure attitudes. The following statement taken from the attitude test furnishes a case in point: "It is desirable to limit the number of children you have [object] so all children 3...: can get proper education [situation]." The scale used to measure attitudes contained statements relating to many of the major concepts stressed ;j in the programs. The statements were either highly favor- ; able or unfavorable toward family planning. Respondents were asked to state whether the statements were true or false. Correct answers (favoring family planning) were given a numerical score of one. Incorrect answers were given a score of zero. The summation of item scores in the scale represented the respondent's attitudinal score. By definition, attitudes are asserted to be agendas for action. According to Rokeach (1968, p. 120), "attitudes have a behavioral component, because all beliefs comprizing them represent predispositions, which when activated, will lead to a response." While adoption of family planning methods could not be expected to result from exposure to one television program, it was considered possible for sub- jects ambivalent toward family planning, or those who had yet to form attitudes on the topic, to demonstrate some evidence of positive evaluation of the small family norm. 70 The attitude scale was administered immediately preceding and following stimulus treatment in order to assess differential change in attitudes as a result of exposure to the alternate treatment modes. Persistence of Attitude Change Persistence of attitude change is the degree to which a respondent's attitude remains stable over time. Hovland and Weiss (1951) demonstrated that agree- ment with an opinion recommended by a trustworthy source may be fairly high immediately after exposure, but will gradually decline with the passage of time. Conversely, there is a decreased tendency over time to reject the material presented by an untrustworthy source. The researchers reasoned that, over time, the source ceases to be a mediating cue to influence evaluation of the message. The researchers termed the result of this disassociation of source from content the "sleeper effect." In the present research, it was important to deter- mine the degree to which possible attitude change would be differentially retained over time as a result of exposure to the alternate treatment modes. By administering the attitude test four weeks after experimentation, the sta- bility of attitudes was assessed. Information-seeking Behavior Change Information-seeking behavior is the manifestation of a desire to know more about an idea or practice. 71 Information-seeking behavior change is an increase or decrease in a desire to learn about an idea or practice. In family planning communication, the gap between cognitive awareness of the small family norm and adoption of family planning methods is wide (Dubey, 1969a, p. 40). However, if individuals, as a result of exposure to an optimally heterOphilous television source, are motivated to seek additional information from family planning field representatives who visit their villages, then the tele- vision medium may serve as a viable supplement to inter- personal development efforts. The information-seeking behavior variable was included in the present research to discover whether television might serve as a motivation- creating link between knowledge and action. Information-seeking behavior was measured by administering pairs of unstructured and structured ques- tions. In the structured questions, re5pondents were asked if they wanted to know more about the health, educa- tional, and economic benefits derived from planning for a family. If respondents replied negatively, it was assumed that they were not motivated to seek additional information. If they replied affirmatively, they were then asked to state all possible reasons HEX they wanted to know more. The information-seeking scale was administered before and after exposure to treatment. Ability to give more and more valid reasons for wanting additional information after '1' 2 exposure suggested that respondents had been motivated to think about the program content and might be open to further exploration of family planning. Persistence of Information-seeking BehaviOr Change Persistence of information-seeking behavior change a is the degree to which a desire to seek information remains stable over time. The information-seeking behavior scale was adminis- J tered along with the knowledge test and attitude scale four weeks after experimentation to test the differential motivational effects over time resulting from exposure to the alternate treatment modes. Review of the Hypotheses for Study The hypotheses for study are listed in sequential order as follows: Hypothesis I: Subjects' degree of perceived homophily with source does not differ across the treatment modes. Hypothesis II: Subjects' degree of perceived homophily with source does not differ across levels of modernity. Hypothesis III: Subjects' degree of perceived homophily with source is not a function of treat- ment mode/modernity interaction. Hypothesis IV: Subject's degree of perceived competence in the source is not a function of the relationship between his degree of perceived homophily with the source and his degree of modernity. HypothesiS‘V: Hypothesis VIa: Hypothesis VIb: Hypothesis VIc: Hypothesis VIIa:' Hypothesis VIIb: Hypothesis VIIc: Hypothesis VIIIa: Hypothesis VIIIb: Hypothesis VIIIc: Hypothesis IXa: Hypothesis IXb: Hypothesis IXc: Hypothesis Xa: HypothesiSEXb: 73 Subject's degree of perceived trust in the source is not a function of the relationship between his degree of perceived homophily with the source and his degree of modernity. Knowledge-gain does not differ across the treatment modes. Knowledge-gain does not differ across levels of modernity. Knowledge-gain is not a function of treatment mode/modernity interaction. Knowledge retention does not differ across the treatment modes. Knowledge retention does not differ across levels of modernity. Knowledge retention is not a function of treatment mode/modernity interaction. Attitude change does not differ across the treatment modes. Attitude change does not differ across levels of modernity. Attitude change is not a function of treatment mode/modernity interaction. Persistence of attitude change does not differ across the treatment modes. Persistence of attitude change does not differ across levels of modernity. Persistence of attitude change is not a function of treatment mode/modernity interaction. Information-seeking behavior change does not differ across the treatment modes. Information-seeking behavior change does not differ across levels of modernity. Hyppthesis Xc: Hypothesis XIa: Hypothesis XIb: Hypothesis XIc: 74 Information-seeking behavior change is not a function of treatment mode/ modernity interaction. Persistence of information-seeking behavior change does not differ across the treatment modes. Persistence of information-seeking behavior change does not differ across levels of modernity. Persistence of information-seeking behavior change is not a function of treatment mode/modernity interaction. _.I I‘m CHAPTER III THE RESEARCH METHODOLOGY . . . . air-1 The degree to which the findings obtained from an experiment correspond with reality depends, in part, upon the extent to which systematic error is controlled. As W4 Schramm (1967, p. 3) commented, a comparative experiment is extremely difficult to design and conduct satisfactorily “because of the problems that arise in controlling the variables so that one knows the comparison is truly measur- ing clear and defined alternatives." In the present research, it is important to interpret findings in the light of factors which imposed limitations on the achievement of Optimal control. Since these factors may be inherent in the design of the stimulus treatments as well as in procedures employed in conducting the experiment, both aspects of the research methodology will be discussed in this chapter as well as data analysis procedures. Design of the Stimulus Treatments In designing the present stimulus treatments, two major factors, which had a bearing on the results obtained, were identified and systematically treated. One crucial factor previously mentioned is the selection of television 75 76 sources to appear in the alternate treatment modes; the selection criterion being the distinctly different type of cultural orientation of the two sources. Possible failure to produce significant differences on the effects variables, if these differences truly exist, could be attributed to the fact that the characteristics which qualify the sources to be termed "urban expert" and "village opinion leader," overlap when they should be mutually exclusive. The other factor concerns the degree to which it is possible to control all variables operating in the stimulus treatments other than the variable for experimental manipula- tion. If not controlled, any one of a number of variables, such as the length of treatment, the number, nature, and sequence of concepts presented, the audiovisual relation- ships, and the quality of production of the treatments, could interact with treatment effects and thereby obscure the presence of true effects. A third factor, unrelated to experimental control, is the design of experimental treatments which correspond as closely as possible with programs regularly broadcast to the rural Delhi audience. Findings obtained from experimental treatments which All India Radio-TV (AIR-TV) programing officials consider appropriate for their rural audience are more likely to be taken under advisement by these officials than findings obtained from treatments employing some unorthodox approach. They also have greater generalizability. Care was, therefore, taken to develop treatments which fulfilled the research requirements and, at the same time, were as representative as possible of operational broadcasts. Design and Production Procedures 3 The prOducer of AIR—TV's rural Delhi program cole thH laborated with the present researcher in the design and , production of the experimental treatments.1 The first és'" task was to select an appropriate format, or form of pre- L. sentation, for the programs, one with which the rural Delhi audience is familiar. The decision was made to use the "interview format," a format in which discussion is led by a professional tele- vision performer, referred to in India as a "compere." Using this format, the sources presented in the respective programs were engaged in a discussion with the compere, and visual materials such as films and still pictures relating to the issues under discussion were inserted in the programs at appropriate intervals. The responsibility of the compere was to equate the content across both pro- grams by asking the sources structured questions designed to evoke responses directly related to the information selected for inclusion in the programs. Since no novice 1In discussing design procedures, the treatments will be referred to as "programs." 78 to television could perform this difficult task, the services of an experienced compere--the compere for the regular rural Delhi programs-were acquired. The next procedural step in program design was to determine the program objective based on assessment of the family planning informational needs of rural Delhi citizens. Officials in the Ministry of Health and Family Planning and the Central Family Planning Institute in India were con- sulted. They advised the author that a large percentage of adults in the rural Delhi area have a general awareness of the need for family planning but they are nevertheless slow to arrive at personal commitment and adoption of family planning practices. These officials maintained that, while the move from awareness of family planning to adoption can only be consummated over time through the com- bined use of mass media and interpersonal channels of com- munication, television can, nevertheless, deepen under- standing of family planning concepts and establish a climate for commitment and later adoption. The purpose of the experimental programs was thus to move subjects beyond superficial knowledge to a fuller understanding of family planning concepts. The producer and researcher next determined a message strategy for implementing the program objective. It was decided that subjects should be encouraged to make rational decisions concerning their individual need for 79 adopting the small family norm. Decision-making is a process which, when applied to problem solving, negates fatalism. If individuals can be encouraged to weigh alternatives in order to arrive at a rational decision, they can move from passive acceptance of what they consider to be their fate a to purposeful action. To encourage decision-making, both arguments favoring having a large and small number of children were presented in the programs. The sources suggested that subjects weigh the good and bad sides of arguments presented t] in "deciding the number of children it is best for you to have." Two arguments supporting the benefits derived from having a large family were selected for brief presentation in the opening portion of the programs and were followed by three arguments supporting the benefits of a small family. The latter arguments were (1) better health through proper diet, (2) educational opportunities for the children, and (3) opportunities for economic advancement for the family. All information presented in support of each argument related to "family security now, and in the future." Family security was thus the theme of the programs. Specific information relating to each argument was next determined and sequenced. An information outline (see Appendix B) contained the educational objectives2 of the programs and 2An educational opjective is a specific content item to be learnedfiwhich is specified by the designer of instruc- tional material before the material is formally prepared. 80 served as a guide for the compere in interviewing the respective sources. The next step was to select appropriate examples to illustrate each content item. The decision was made to have the television sources support their arguments through reference to actual village families. Specifically, a l the sources, guided by the compere, would compare the life- style of two village families, one with three children3 and one with many. The families would not appear on television j but reference would be made to them through still pictures 5 and short silent film sequences shot in the village. Families, whose life-style met the qualifications specified in the information outline, were found in a Delhi State village which received television; that village was deleted from the population of villages from which the research sample was randomly drawn. Two brothers had married sisters, each of the two families had approximately the same number of acres under cultivation, yet striking comparisons between the families, one with three children and one with eight, were apparent on all items specified in the information outline. Still pictures and films of the families were made in their homes and fields. 3From a villager's standpoint, there may be valid reasons for having several children (see traditional argu- ments in the information outline). Using a model family with two rather than three children might have thus reduced acceptance of the arguments in favor of family planning. 81 A search for a village opinion leader to appear in the Mode I program was next conducted. Selection of sources was based on objective assessment of sources' (1) level and type of education, (2) occupation, and (3) place of resi- dence. Not only was it necessary to ascertain that the sources differed radically on these characteristics, but it was also necessary to insure that the village source was considered an opinion leader in his village in matters pertaining to family welfare. The characteristics of a village opinion leader offered by Rogers (1969a, p. 227) served as a guide to source selection. Rogers indicated that, when compared with their followers, village opinion leaders are characterized by such qualities as (l) more formal education, (2) higher levels of functional literacy, (3) more innovativeness, (4) higher social status, (5) more cosmopoliteness, and (6) higher empathy.4 Taken by themselves, however, these character- istics (the indices of modernity) carry no guarantee of leadership acceptance. If an individual rates Ego high on any one of them, he may be considered an outsider or as different in his own village and have little influence over the beliefs and behaviors of other villages. The task in the present instance was, therefore, to find a source who conformed to village norms and was highly influential in his village in matters pertaining to family welfare. 4Em ath , according to Rogers (1969b, p. 13), is the ability of a person to project himself into the role of others. 82 Whether or not the source was a satisfied user of a particular contraceptive method was not considered germane to the research, since the program content did not relate to contraceptive practices. However, the source would need to be committed to the idea of family planning and himself have raised no more than three children. ‘Since' emphasis in a- the programs was placed on the benefits derived from having a small family, it was important that the source's family K1":- _' , _ IF-fT size be consistent with his beliefs on the topic of family planning. After considerable investigation, an ex-schoolmaster living in a Delhi State village receiving television was selected; that village was also deleted from the population of experimental villages for sampling. The schoolmaster, referred to in his village as "Panditji," or respected teacher, was 70 years of age, had received eight years of schooling in a Hindi medium village school, was still vigor- ous and deeply involved in village social and educational affairs, and had raised only two children. In his manner of living, speech, dress, and general demeanor, the village leader appeared only slightly more modern than the average villager, yet the high degree of respect he enjoyed in his village qualified him as an opinion leader in family matters. The village leader agreed to serve in the research assign- ment. 83 A search for an apprOpriate medical doctor living in Delhi was then conducted. It was important to equate the two sources on communication skills, age, and personal dynamism.S For example, if one source should be much younger than the other, differences obtained on measures of the television effects might be credited to differences in the age of the sources as well as to differences in their cultural orientation. The doctor selected, an ex-Commissioner of Health of Delhi State, was approximately the same age as the village leader and had raised three children. Like the 'village leader, he was highly articulate in communicating his ideas to others--although his communication ability reflected his years of training in English medium urban schools and in medical college. In his mannerisms, speech, and personal appearance, the doctor was a prototype of an urban professional man. He was poised, self-confident, and enjoyed an excellent reputation amOng members of the medical profession in Delhi. In the judgment of the television producer and researcher, both sources were considered dynamic, although to a somewhat different degree: the village leader was slightly more emphatic in his speech delivery and gestures than the doctor. That is, he was observed to project 5D namism is the amount of energy and emphasis a source prOJects into a communication act. 84 himself into his presentation somewhat more actively than the doctor. Therefore, a possible limitation in the present research is the fact that differences in source dynamism might have had a bearing on the results obtained as well as differences in the cultural orientation of the sources. After source selection, the sources were individually exposed to the village families whose life-style they would refer to in their performance on television. It was, of course, important that the sources become thoroughly familiar with the living conditions of the families on all relevant counts. And it was equally important to discover the sources' perceptions of the realities concerning the families presented to them. If the sources' perceptions were incongruent with each other, or with those of the producer and researcher, then the possibility that the sources would address the issues outlined and would take the same position on those issues would be jeopardized. If, on the other hand, the sources were prompted by the producer as to what to say and when to say it, they would be inhibited in speaking in their own idiom and the programs would be equated on the variable for experimental manipulation. But by asking the sources to recount their perceptions of the families im- mediately after exposure to them, it was possible to estimate how close to the targeted content they might come. Any incongruity would then have to be corrected through prompting. 85 Interviews with the sources immediately after expo- sure to the families revealed that their perceptions were, by-and-large, on target. The doctor required no prompting; the village leader was prompted on one issue. Next, a television script in the Hindi dialect of the Delhi villages was developed for the benefit of the program production staff and the compere. The compere became familiar with the script but was requested to use it only as a means of coordinating his performance with that of the production staff.6 The compere did, however, memorize the information outline in order that he could equate his questions put to the sources and thereby control source response. He was also thoroughly briefed concerning the research objectives and the need for controlling source response. Only when the researcher was certain that the compere understood the vital importance of his role, were the dates for program production scheduled with AIR-TV officials. The two programs were produced in the studios of AIR-TV on different occasions and thus neither source auditioned the other's performance. The village opinion leader, who appeared first, followed the information outline unfaultingly. The doctor did likewise several days later. Both programs were 14 minutes in length and the production quality of both programs was comparable. 6An English version of the television script is furnished in Appendix B. 86 The programs were recorded on two-inch video-tape and were subsequently re-recorded on one-half inch video- tape for playback in the experimental villages. The two- inch master tape was preserved for the duration of the eXperiment. This precautionary measure was taken so that new recordings could be made in the event that the one- half inch tapes were accidentally erased or mutilated. Execution of the Research Design Locale of the Study The rural area of Delhi State is inhabited by approximately 300,000 people living in more than 275 villages which surround the Capital City of Delhi.7 Approximately one-third of the rural Delhi citizens live in villages with a population range of 2,000 to 5,000 while a little less than one-third live in villages with a population range of 1,000 to 2,000. There are no villages in Delhi State with a population of more than 5,000 inhabitants. Delhi state is divided into.five development blocks and is bounded on the East by the State of Uttar Pradesh and on the other three sides by Haryana State. The basic occupation of the people of rural Delhi State is agriculture. These people tend to speak a common dialect of Hindi, which 7These figures are taken from the Census of India, 1961, Vol. XIX, Part II, Delhi. The population of Delhi State has increased since the 1961 Census was taken probably to about 300 villages with a total population of 365,000. 87 probably reflects the dominant influence of urban Delhi as Opposed to that of neighboring states. , Generally speaking, the influence of modern Delhi has had its greatest impact on villages in closest prox- imity to the city limits. Less land per capita is under cultivation in these villages, and consequently these villages are developing social patterns that are character- ized by a trend away from the self-contained village unit and toward individualism and specialization. Relatively more individuals living in these villages commute daily to offices, factories, and unskilled jobs in the city. However, all villages in Delhi State are under the influ- ence of the city and in all there are manifestations of social and technological change. Presently, all of the Delhi State villages are electrified and most are accessible by all-weather roads. Adoption of agricultural innovations has raised the standard of living of a large percentage of the Delhi State farmers. Village schools are better attended today than they were even three years ago, commuter transporta- tion services to the Capital City are continually expanding, and small industries are developing in many Delhi State villages.8 8Facts concerning the rapid development of the Delhi State villages were furnished by Dr. K. N. Singh, and Dr. S. K. Sharma, Professors and Senior Extension Officers, Division of Agricultural Extension, Indian Agricultural Research Institute, New Delhi. OO 03 In spite of improvements in their standard of living, many rural Delhi citizens adhere to social and cultural traditions that have remained relatively unchanged through the centuries. Living.under the joint-family system, ascription through caste affiliation, conformity to village norms, observance of age-old customs and habits, all reflect a deep sense of continuity with the past. There is no doubt that modern technology has had an impact on the peOple of rural Delhi State. But to an appreciable extent, the new technologies have been assimi- lated into a traditional village way of life and the result is a synthesis of old and new, and indication of the capacity of the Indian agricultural village to absorb and transform outside influences into indigenously rural Indian patterns. Selection of Experimental Villages In 1967, the Department of Atomic Energy (DAE) of India provided 80 Delhi State villages with television receivers for community viewing. Television villages were selected by the DAB on the basis of their being (1) electri- fied, (2) accessible by all-weather roads, (3) agriculturally innovative, (4) progressive in character, and (5) able to provide proper facilities for community viewing.9 The number of villages receiving television sets in each develop- 'ment block was proportionate to the number of villages in 9Although almost all of the Delhi State villages can meet most of these selection criteria today, such was not the case in 1967. 89 that block. All villages selected were within the 23 mile range of the Delhi television transmitter. At the time the random sample of experimental villages was drawn for the present research, plans for the placement of television receivers in several additional Delhi State villages were being executed. The new tele- vision villages, however, were not included in the popula- tion of villages for sampling. The novelty of initial exposure to television, referred to as "the novelty effect," tends to bias viewer response since individuals usually react to the uniqueness of the experience rather than to the presentation itself. Mixing old and new television villages could, therefore, produce differential perceptions of the viewing experience, a factor that might increase error variance as well as interact with the type of exposure. To further safeguard against novelty effects, only those television villages among the original 80 in which the sets had been working for a period of two months prior to sampling were included in the population of villages. Although 61 television villages met the criteria for inclusion, nine of those were eliminated for various reasons. All-weather roads leading to two villages were temporarily in poor condition. Three villages were excluded because community television viewing habits were somewhat altered due to the fact that the sets were located in the 90 headquarters of local government agencies. Two villages had been utilized in the production of the television programs, and two villages were set aside for pilot test- ing. The remaining 52 villages constituted the population from which a random sample of 20 villages was drawn. These villages (listed according to their block designation in Appendix C) can be said to be representative of the type of villages which will continue to be served by television originating from the Delhi television station. Selection of Resppndents The original plan was to randomly sample 20 couples from each treatment village, the population of couples for sampling to include only those couples in which the wives were within the reproductive age range. However, the female portion of the sample was dropped for two reasons: Preliminary investigation revealed that many village men . were opposed to the idea of mixed male-female television viewing and that a request for mixed viewing might dis- courage the male subjects from agreeing to serve in the experiment. Second, from a theoretical standpoint, the inclusion of women is of questionable value, since the degree of homophily perceived by a female with a male television source would be low. While it is important to investigate the reactions of women to family planning television programs, the idea was dropped in the present research in favor of a male-only design. 91 Also, the number of male respondents per experi- mental village was reduced from 20 to 11. It was originally planned that each inVestigator should conduct two pre- and posttest interviews per experimental session. The pilot test revealed that such a procedure takes more than five hours to accomplish and is taxing on both respondents and investigators. Also, the time lapse between treatment and posttest for the second group interviewed could decrease the response validity of the design, since individuals in the second group could talk among themselves while waiting to be interviewed. It was decided that 11 males would be randomly selected in each treatment village. Respondents who for some reason were not available for the post posttest interview would be dropped from the sample in the analysis of data. The experimental treatment was, therefore, administered to 11 respondents per village in the hope that at least 10 respondents per village could be summoned for the post posttest. The resulting population of respondents in each village from which a simple random sample was drawn was composed of all males (1) whose wives were within the reproductive age range, and (2) who were known to be rela- tively consistent viewers of television. Selection of males based on their viewing habits was made in order to reduce the novelty effect. The criterion concerning the 92 reproductive age range of wives was necessary since the television programs were designed to encourage decision- making regarding adoption of the small family norm. To execute sampling procedures based on these two criteria, lists of all couples of reproductive age in the experimental villages were secured from the Primary Health Centers10 in each of the five development blocks of Delhi State. Village leaders in each experimental village were than asked to select from their village list 40 males whom they considered to be the most frequent viewers of tele- vision. From each village list of 40 males, a sample random sample of 11 males was drawn for inclusion in the experiment.11 Selection from 40 names made it possible to estab- lish a wide range of response on measures of the modernity variable. Had village leaders been asked to select only 20 names from the reproductive age lists, they might have presumably selected village persons similar to themselves in degree of modernitye-persons with whom they tend to interact most frequently. But it is not likely that as 10A primary health center is a unit of state govern- ment responsible for administering health and family planning action programs at the village level. 11A simple random sample is one in which each person in the population has an equal chance to be included. It differs from a stratified random sample in that the latter divides the population into homogeneous subparts--such as men and women--and takes a random sample of each subpart or strata (Backstrom and Hursh, 1963, p. 26). 93 many as 40 persons in any given village are at the same approximate level of modernity. Thus 40 males per village selected_from the reproductive age list was the population from which simple random selection was made. Descriptive data concerning the caste affiliation and occupation of the respondents in the sample are fur- nished in Tables 3.1 and 3.2. The average age of respond- ents was 35. Assignment of Villages to Treatments Certain factors inherent in the research design made systematic random assignment12 of villages to treat- ments necessary. Experimental sessions in all 20 villages could not be held simultaneously but rather were held on successive nights. When sessions are not held simultane- 13 may obscure differences on ously, inter—session history the dependent variables, if such differences truly exist. Inter-session history is attenuated, to some degree, when a large number of sessions are held over time and the probability of the occurrence of unique events is thereby distributed across treatments. However, one type of 12Systematic sampling is a procedure of sampling from a list of cases by randomly selecting the first case and then taking every kth case thereafter (Selltiz, 1959, p. 523). "‘ “* 13Inter-session history is an event (such as a festival or unusual news) affecting daily routines which can contribute to a rival hypothesis and obscure possible differences between treatment means on measures of the effects variables. 94 OOH OOH ucou mom mH OH Amummmon .mumnumn .mwcc>uom .muomoosmv mummo omHsoorom .m mm HO Amnoucomnmo one mcommE .mnouuom .muo>mm3v woummu HmsoH .v H m AcoEmmochse pom mnooouuv whcmHm> .m ON om “whoomoH OOMHHH> HOEHOH .mumfinmwv Hyposx .uMO .N O m AmuoomoH mOOHmHHOH .mnozomoav cHfismnm .H AOOH N zv AOOH N zv moHnomoumU mummo HH moo: H moo: Anamo ham :HO .mpsmocommom mo GOHHMHHHmm< mumMUII.H.m mqm cH mcoHHMMOOOO ocMH mo mwnom once HO HsOm mchso Hoanmm OOMH mo mouom HOOM corp mmOH mchso moanmm weapon panama HOHOQOH who omoHHH> cH onsuHOOHumm IOOH u zv HH moo: AOOH H H 7.: QUOZ mucoocommom mo chHummoooo Aucmo umm :Hv .mEoosH “Hon“ MHmr one» who: oocnmm mucoosoomom £OH£3 Scum mcoHumosOOOII.m.m mqmda 96 historical event had to be avoided if possible: If news about the experiment should happen to diffuse from an already-treated village to one scheduled for treatment, responses obtained from the subsequently scheduled village might be biased. Experimental villages in close proximity to each other were thus treated on consecutive nights to reduce the possibility of intervillage communication about the experiment. Had the need for proximity in selection of villages been the only assignment limitation, experimental villages could still have been randomly assigned to treatments. But in this event, one treatment mode might appear on several successive occasions. Without backup in the form of auxiliary video-tape equipment—-and knowing that the equipment was not guaranteed to perform effectively in such high temperatures as were experienced during the experimental period--it was decided to alternate treat- ment modes, rather than to gamble on simple random assign- ment at the risk of having to abandon the experiment with only a few sessions in one mode and a number of sessions in the other. As it turned out, the video-tape equipment worked successfully on all occasions, but on a number of occasions, equipment failure immediately preceded or followed an experi- ment and led the researcher to believe that settling for systematic random assignment of villages to treatment was a wise decision. 97 The Pilot Test In the pilot test not only was it necessary to try-out the questionnaire in order that necessary revision might be made before full-scale experimentatiOn, but also to judge the effectiveness of interview procedures, test the workability of the video—tape equipment under field conditions, and assess the effectiveness of procedures used to prepare villages for the experiment. To avoid a time lapse between exposure to treatment_ and posttest interviewing, it was necessary to transport to the field in one conveyance an investigation team con- sisting of 11 investigators, the project supervisor, an electronics technician,14 and a general assistant. It was also necessary to install the television equipment in the home of a village leader who had agreed to host the experi- ment. Community television sets provided to villages by the Indian Government are usually located in convenient gathering places such as schools. In an experiment it is important to control for extraneous activity that might 14It turned out that the video-tape equipment was not compatible with television sets and so it was necessary to improvise a television monitor (similar in appearance to a television set but electronically different). The improvised monitor required modulating equipment which was beyond the capability of the research team to operate. Also, in each village, power outlets differed and electrical. power fluctuated widely. Thus the electronics technician served an important function in assuring consistent opera- tion of the viewing equipment on the experimental occasions. 98 bias response and create differences in the application of the treatment. Thus the community television viewing location would not be conducive to crowd control and had to be avoided in favor of the home of a village leader, prefer- ably a home at the outskirts of the village with doors that could be locked to insure privacy. In the present experiment, an experimental labora- tory was transported to the field--a procedure not required in field experiments in which the effectiveness of regularly scheduled broadcasts viewed on community receivers are evaluated. Pilot testing revealed that the novelty of setting up a laboratory in the field can create a great deal of excitement and general commotion in a village. Although the village host of the pilot test assured the researcher that the experiment would be conducted in the privacy of his home, he was unable to control the curiosity- seekers that gathered there. Lack of control for crowds 15 of the could seriously jeopardize the internal validity experiment, and means had to be found to conduct the experi- ment under less obtrusive conditions. It took the better part of the twoeweek period between the initial pilot test and the first eXperimental session to work out a plausible solution to the problem. The experimental schedule had lSInternal validity is the degree to which all variables ii an experimental design are controlled except for the experimental variable. For a detailed explanation of internal validity, see Campbell and Stanley (1963, pp. 5-16). 99 been set with village leaders well in advance of the pilot test and could not be moved forward. Therefore, the idea of conducting two pilot tests, one in each treatment mode, had to be abandoned in order that strategies could be altered and more systematic procedures developed for con- ducting the experiment. The Assistant Development CommisSioner of the Delhi State Administration took the problem encountered in the pilot test village under advisement and made the Block Development Officers (BDO's) responsible for keeping the curiosity-seekers in the experimental villages from diSturbing the experiment.16 BDO's are highly respected by village persons who turn to them for many types of assistance. Thus the BDOs' requests are not likely to be disregarded. The BDO's, assisted by Village Level Workers who are responsible for executing administrative programs at the village level, contacted village leaders and asked them to cooperate on the experimental occasions. On each experimental occasion a BDO, or his appointed 16The Development Department of the Delhi State Administration is composed of officials who are appointed to carry out the policies of the political units of govern- ment at the block and village level. As a political unit, Delhi State is actually one district composed of five development blocks. The Development Commission has jurisdiction over the BDO's who are responsible for administering rural social and economic development pro- grams at the block level. BDO's have authority over VLW's who act as the agents of the BDO's at the village level. Each VLW in Delhi State is responsible for implementing development programs in 10 villages. 100 representative, accompanied the research team. As soon as the research team arrived in the village, village leaders designated by the BDO gathered in front of the host's home where they remained throughout the experiment for the purpose of controlling crowds. At times when the village leaders failed to perform their function satisfactorily, the BDO gently reminded the leaders of the responsibility they had agreed to assume. Without the presence of a BDO in the pilot test village, village leaders treated their responsibility lightly. While the BDO's and VLW's were making preparations to execute their responsibilities in the approaching 20 day experiment, other problems encountered in the pilot test had to be corrected. The pilot test revealed that a man-machine systems approach, encompassing interviewers, respondents, village leaders, BDO's, the electronic tech- nician, back-up assistants, and the viewing equipment, had to be developed. Any approach less efficient than a highly integrated and coordinated system could result in undue harassment of the interviewers by the crowd, unnecessary delays in expediting the experiment, restlessness on the part of respondents, and general failure to achieve the objectives of the experiment. Although the investigators, graduate students in various universities of Delhi, had had previous experience 101 in conducting surveys,l7 none of them had had training in experimental investigation. Immediately preceding the pilot test the investigators had received three days of training for their experimental assignment. However, the pilOt test revealed that extra training would be necessary in order to cope with a number of unique circumstances encountered in conducting a laboratory experiment in the field. For example, it was necessary to charge several investigators with the responsibility of assisting the electronics technician in setting up the viewing equipment while other investigators scouted the village with village leaders in order to locate tardy respondents. It was also necessary for investigators to learn how to pace their interviews in order that all could finish their interviews at approximately the same time, and thereby keep the respondents from talking among themselves. ; The pilot test took more than five hours to com- plete, retaining the respondents far beyond the limits of their attentiveness. Thus procedures for setting up the equipment, pairing each investigator with his respondent, 17Some of the investigators were working on a second master's degree. All were receiving training in the social sciences. Three investigators had had con- siderable experience in conducting rural investigations and were of invaluable assistance in training investigators who had not conducted rural interviews. All of the investigators had had considerable experience in urban interviews. JU2 rapport building, pretesting, administering the stimulus, posttesting, and closing the session had to be highly synchronized to increase efficiency and reduce confusion and the time taken in conducting the experiment. Measuring instruments were also revised following the pilot test. As anticipated, the interview schedule was too lengthy and items which produced little variability, or appeared to the majority of respondents as irrelevant, were deleted. The items retained in the schedule were analyzed independently by three investigators experienced in conducting interviews in the villages of Delhi State. After each investigator had considered the possible meanings respondents might attach to each item, their evaluations were compared in the presence of the researcher. In the course of four sessions, all items of doubtful validity were rephrased. A final validity check on all the items was made by the Senior Research Officers of the Central Family Planning Institute. Also, all itemswere checked to make sure that the language used was that spoken in the experimental villages. A check was made on the knowledge test to ascertain that the questions corresponded with content presented in the television programs. Conducting the Experiment Systematic procedures developed as a result of the pilot test experience paid rich dividends on each experimental 103 occasion. There were exceptions to the smooth flow of events, however, and in no village were all man-machine functions perfectly coordinated. On five out of twenty occasions, a few of the randomly selected subjects (ranging from two in one village to five in another) were not present and it became neces- sary for the village leader hosting the experiment to select his neighbors or other village substitutes who were available and willing to serve. Selective sampling procedures on these occasions probably increased systematic error and to some extent affected the internal validity of the design. Fortunately, however, the treatment modes were altered so that systematic error occurred in both of the modes and the number of substitutes in the two modes was approximately equal. Some lowering of morale among the investigators occurred about halfway through the experiment. The experi- ment began in mid-April and ran into May, 1970, a very hot and physically enervating time of year in Delhi State. The heat and dust in the villages, the exacting nature of the task, the monotony of the daily routine, all were factors contributing to the drop in investigator morale. The project supervisor and this investigator kept close track on the performance of each investigator and every possible consideration was given to them to facilitate htheir work. On about the sixteenth night of the experi- ment, with the end of the arduous task in sight, initial 104 enthusiasm for the work returned and was retained through- out. The morale problem likely contributed to differences in application of the treatments and to systematic error. However, since the treatment modes were alternated, such differences occurred about an equal number of times in each treatment. The BDO's handled their assignment effectively and the curiosity-seekers were fairly well disciplined by the village leaders. An electrical power failure in one village caused a 15 minute delay in administering the treatment and some conversation among respondents was unavoidable. The fidelity of the television picture left somethingto be desired on almost all of the experimental occasions. Snow often appeared on the screen and defini- tion of the picture was never sharp, although at all times it was possible to perceive the video image. On the whole, the administration of the experiment was fairly consistent across all 20 occasions. The effi- ciency procedures developed for conducting the experiment trimmed approximately one and one-half hours off the time spent in the pilot test village; the average time spent in each'experimental village being approximately three and one-half hours.) The need for expediting procedures in the present research cannot be over-emphasized. Un- necessary delays and confusion are detrimental to an experiment in that they allow spurious factors to confound the effects of the stimulus treatment. But in the present instance, it was of utmost importance that delays be avoided. The experiment took place during "rabi" harvest, a time of year when the farmers of Delhi State are harvest- ing their winter wheat crop. During this periOd, the farmers, tired from their daytime labors in the fields, are not anxious to take part in extraneous activity. That they agreed to cooperate in the present experiment, demonstrated a large amount of consideration On their part. It was important to show appreciation for their cooperation by expediting the experiment so that they would not be re- tained past their usual bedtime. Although the pressures of synchronizing all pro- cedures was a strain for the investigators, the respondents did not appear conscious of this fact and the experiment appeared to be conducted in an informal, relaxed manner. Much of the obtrusiveness experienced in the pilot test was overcome and a number of spurious factors that could have seriously jeopardized the experimental results were essentially controlled. Administering the Post Posttest Four weeks after the experiment, three investigators returned to the field to administer the post posttest. Approximately 65 per cent of the subjects were interviewed on the first callback. Two subsequent callbacks were 106 sufficient to obtain interviews from 100 respondents (ten per village) in each treatment mode, or a total of 200 respondents. The 20 respondents absent from their village on all callbacks were dropped from the sample and observa- tions from 200 respondents, 100 in each mode, were finally used in the analysis of data. Post posttest interviews were not scheduled at any set time of day and investigators conducted private interviews with respondents in their fields and homes. .Accurate records were kept of each return visit, and of the degree of cooperation of all resource persons in the vil- lages who assisted in facilitating the follow-up test. Analysis of Data Methods of Analysis Unit of analysis.--Systematic random assignment of 'villages to treatments in the present study permitted the researcher to assume that the villages appearing in the alternate treatment modes were essentially equivalent. frherefore, it could be assumed that villagers were compa- rable across treatment modes as well, since the assignment <3f subjects to treatments was automatically determined by their village assignment. Ideally, to meet the assumptions for an inferential test of significance, the unit of analysis should be the Village. Analysis of variance, the parametric test used 107 to test most of the present hypotheses, assumes that the villagers in the sample are randomly drawn from a common pool which has homogeneous variance. In such a case, variance can be partitioned on the basis of individual villagers and the condition of a normally distributed population can be assumed. Because individuals rather FA_J than villages were used as the unit of analysis in the present research, the assumption of independence was not 1.. m LN- -"A I I. fully met. Liberties are sometimes taken in using villagers rather than villages as the unit of analysis, providing ’ the assumption of independence does not seem grossly violated. Support for the decision to use villagers as the unit of analysis in the present experiment is evidenced by the large amount of heterogeneity which existed within each village sampled and the homogeneity that existed across villages. Table 3.3 shows the range of variance within each village included in the sample, using three dimensions of modernity for illustration purposes. Two practical considerations influenced the researcher's decision to use the individual subject as the unit of analysis: First, the 20 villages sampled would have furnished few replications in the treatment cells for two-way analysis of variance. By using indi- vidual scores it was possible to increase the number of observations per cell and thus to increase statistical 108 mNIo . . O.mH O IN h.v O IO b.w ON O IN O.m ONIO m.OH OHIO 0.0 OH O IO 0.0 OOIO m.hH OHIO m.b OH O IH m.v ONIO m.OH OHIO m.v OH O IN w.¢ HNIO w.mH OHIH m.> OH s IH O.v ONIO O.vH OHIO O.v OH O Im O.m HmIm N.OH OHIO O.h OH O IN v.m ONImH O.hH OHIO N.n OH O In O.m NNIO m.OH O IO H.O NH O Iv n.m OOIO H.OH O IO v.m HH HH moo: O IN m.v mNIHH m.mH OHIO 0.0 OH O IN v.v ONIO 0.0H O IO H.O m O IN m.v ONIHH N.>H OHIO 0.0 O O IN m.¢ ONIO O.mH OHIO h.m n O Iv H.O ONIOH v.OH OHIO m.m O O IH m.¢ ONIH >.mH OHIO v.5 O O Im O.m ONIO N.nH OHIO N.m v ONIm m.m ONIO H.OH O IO O.m O O Iv N.m ONIO m.OH OHIO m.m N O IO O.m ONIO N.mH OHIO m.v H H moo: omcmm . COOS omcmm COOS omcmm com: mmocouHHomOEmoo whomomxm OHooz mmmz coHumooom mmmHHH> auchoooz mo msoncoEHO mo .02 OOOU .muHchoooz mo mconcoEHa means no HGOEOH5mmoz Scum oochuno moouaom mo omcmm can memos on» mcHusmfioo an coumOHocH huHocomoeom mmmHHH>HOHCH one huHocmmonmumm OOOHHH>MHHGH mo mmnmoall.m.m MHmda 109 precision. Second, the Delhi State villages, under the modernizing influence of urban Delhi, are relatively less autonomous and self-contained than villages located in the more remote and less accessible rural areas of India. Close contact with the city has diffused the locus of social and political power in the Delhi State villages and, therefore, a relatively wide spectrum of individual differences exists among individuals within a single village. Analysis of modernity scores also revealed that the assumption of equal variance of the population sampled had been essentially met: the sampling equivalence of the ‘treatment groups was essentially achieved and values of 'the combined modernity scores had approximately a normal (distribution with only a slight negative skewl in each tzreatment mode. That equivalence was essentially achieved :is demonstrated in Figure 4. Treatment exposure.--Because randomization procedures L‘lsed in the present experiment theoretically established tine sampling equivalence of the treatment groups, subjects VVEEre exposed to one and only one television treatment and \ 18Skewness is the piling up of cases at one or the c>‘tlher side of the distribution of values of a variable. A Ilcxrnal distribution of values forms a bell-shaped curve. sskewed distribution is asymmetrical. In the present case, ‘3}HOOIHouw< so OH oochuno muOOMOMIOOOz mo OHQmem HOOHuonuoomr «II.v.m mHm<8 115 values than the source. To facilitate computer processing of the data,19 negative scores were transformed into posi- tive scores by adding a value of plus 10 to all scores. Once the scores has been transformed, the homophily hypotheses were tested for statistical significance by calculating F obtained in the sample and comparing it with the appropriate critical value of alpha. Analysis of variance disclosed the relationship of the two independent variables to the dependent variable, homophily. The experimental design was a 2x4 factoral design with two treatment modes and four levels of modernity. Observations from 200 subjects, 100 in each treatment mode, formed eight treatment cells consisting of 25 observations each as shown in Table 3.5. Table 3.5 indicates that each level of modernity occurred with each treatment mode and 'that the combinations occurred an equal number of times. Hypothesis I: Subjects' degree of perceived homophily with source does not differ across the treatment modes. Under Hypothesis I, mean homophily distance scores were compared across treatments. The test of Hypothesis I was performed to determine whether differences in perceived hOmophily can be attributed to the treatments, irrespective of. levels of modernity. More specifically, the purpose of ‘ 19Data processing was handled by the Computer Center, Prflgram Evaluation Organization, Planning Commission, Vernment of India, New Delhi. ' 116 TABLE 3.5.--Replications for Two-Way Analysis of Variance Tests of Hypotheses I, II, and III. Treatment Modes Levels of Modernity Mode I Mode II High 2' 3(- (N 25) (N 25) Above average 3' i. (N 25) (N 25) Below average Y' i. (N 25) (N 25) Low Yb if (N 25) (N 25) the test was to determine whether subjects could discrimi- 2nate homophily/heterophily differences between the village (opinion leader appearing in Treatment Mode I and the urban Inedical doctor appearing in Treatment Mode II. It was theoretically predicted that the mean distance ascore of Mode I would be significantly smaller than the mean (distance score of Mode II. (To conform to statistical con- ‘vention, the null hypothesis was formally stated.) Rejec- 1:ion of the null hypothesis and the hypothesized ordering <>f the means would indicate that the village source was IPreceived by subjects, on the average, as significantly fluore homophilous with themselves than the urban source. Tune critical F value was 5.15 for significance with l and J~92 degrees of freedom. 117 Hypothesis II: Subjects' degree of perceived homophily with source does not differ across levels of modernity. The test of Hypothesis II was performed to determine whether differences in perceived homophily can be attributed to levels of modernity, irrespective of the treatment modes. The levels of modernity were high, above average, below average, and low. Cut-off scores of the modernity test were values corresponding to the quartiles of the test. These values were 46/60, 36/60, and 26/60. It was theoretically predicted that mean homophily 20 distance would be negatively related to modernity. That is, high modern subjects would perceive a more homophilous relationship with the source to whom they were exposed 'than would low modern subjects. Rejection of the null Ihypothesis and the hypothesized ordering of means would .support the theoretical prediction. The critical P value vmas 3.23 for significance with 3 and 192 degrees of freedom. Hypothesis III: Subjects' degree of perceived homo- phily with source is not a function of treatment mode/modernity inter- action. There was theoretical reason to believe that there “Nbuld be no treatment mode/modernity interaction and that tile null hypothesis would not be rejected. Subjects at k 20A low numerical homophily score indicated high Perceived homophily with the source. 118 all levels of modernity would perceive a more homophilous relationship with the village source than the urban source. The critical F value was 3.23 for significance with 3 and 192 degrees of freedom. The credibility hypotheses.--The credibility hypotheses were presented at the conclusion of Chapter II. No attempt was made to perform statistical tests related to these hypotheses. Table 3.6 shows that there was a lack of variability of response on measures of per- ceived credibility. Across all subjects sampled, both sources were perceived as highly competent and trustworthy relative to the measures obtained.' Since there was no possibility that degree of source credibility could have been related to degree of homophily or modernity, it was pointless to further test these hypotheses. Pilot testing of the credibility instrument revealed what the researcher considered to be adequate variability of response. However, as mentioned previously, exhorbitant and unchecked excitement reigned on the pilot test occasion and interfered with the orderly execution of the experiment. IData obtained from interviews conducted in such an environ- Jnent were quite possibly inaccurate. One might conclude 'that the pilot test was an excellent pretest of the experié Imental method, a pretest which led to reexamination and nMDdification of procedures for efficiently conducting an e):periment under field conditions. But it was not an afiiequate pretest of the interview schedule. In retrospect, 119 ooa. ooa, ,. cod . , ooa ucmo nmm o N a H 302M u.coo m H m H 3oq h m 5 Ha snacmz Hm mm mm hm amnm flood u zv nooa n 29 . iced n 21. need u 21 HH moo: H moo: HH moo: H moo: mmflnomwumo .cmuo mmmcflcuuo3#deB. . .pmuu mocmuomaou. mcoo Aucmo mom cHV .mmooz pcmaumoua mumcnmuafl map ca Umcflmuno mmuoom huflawnflomuu condom wo>wmoumm .mucmccommmMIl.w.m mamfls 120 the sizeable amount of variability produced from instru- ments other than perceived source credibility may be con- sidered the result of careful preplanning mixed with good luck rather than the result of adequate pretesting. It is possible that credibility differences between the sources were perceived, but that the credibility instru- ment failed to uncover them. Appreciable credibility dif- ferences might have been detected if the subjects had been exposed to both sources and were asked to make differential evaluations of them. Respondents may be prone to make stimulus judgments based on what they think the researcher ‘would like to know. Unless the purpose of the research can be disguised, the researcher is likely to obtain inaccurate data. Of course, in an experiment, exposure to both sources for the sake of discriminating credibility differences should be avoided, since exposure to more than one treatment can also bias response. The fact remains, however, that a comparable measurement technique used in the present experiment might have prompted differential credibility response. The television effects hypotheses.--The six tele- vision effects measured are (l) knowledge-gain, (2) know- ledge retention, (3) attitude change, (4) persistence of attitude change, (5) information-seeking behavior change, and (6) persistence of information-seeking behavior change. 121 Hypotheses related to these effects are listed in sequential order at the conclusion of Chapter II. As in the case of the homophily hypotheses, two-way analysis of variance, fixed effects model, was the pro- cedure used to test the possible significance of differences in means obtained on measures of the television effects. The critical F value for tests of the treatment modes as a main effect was 5.15 for significance at the .025 level with l and 192 degrees of freedom. The critical F value for tests of the modernity and interaction effects was 3.23 for significance at the .025 level with 3 and 192 degrees of freedom. Analysis of variance disclosed the relationship of the two treatment modes and four levels of modernity (a 2x4 factoral design) to each of the television effects. The eight treatment cells in the analysis of television effects are those previously shown in Table 3.5. The purpose of testing the television effects hypotheses was to determine whether the effects can be attributed to (l) the treatment modes, irrespective of modernity, (2) levels of modernity, irrespective of the treatment modes, or (3) treatment mode/modernity inter- action. It was theoretically predicted that high modern subjects would tend to obtain higher scores on measures of the television effects as a result of exposure to the urban source. Low modern subjects would tend to obtain 122 higher scores on measures of the television effects as a result of exposure to the village source. Thus interaction effects were predicted in tests of the television effects hypotheses. In testing Hypotheses VI, VIII, and X (Knowledge- gain, attitude change, and information-seeking behavior change), pre- to posttest gain scores were used as the dependent variable. In testing Hypotheses VII, IX, and XI (knowledge retention, persistence of attitude change, and persistence of information-seeking behavior change), pre- to post posttest gain scores were used as the dependent variable. Significant gain from pre- to post posttest would indicate that differences between the treatment modes on measures of the television effects remained stable over time. CHAPTER IV THE RESEARCH FINDINGS Significance Tests of Perceived Homophily Hypothesis I: Subjects' degree of perceived homophily ' does not differ across the treatment modes. As theoretically predicted, subjects, on the average, perceived the village opinion leader appearing in Treatment Mode I as significantly more homophilous with themselves than the urban medical expert appearing in Treatment Mode II. Table 4.1 shows that the F value for the difference in treatment means is 45.10, which is greater than the 5.15 required for significance at the .025 level. Therefore, the null hypothesis, stating that subjects' degree of perceived homophily with the source does not differ across the treatment modes, is rejected. The alternate hypothesis, stating that a difference in treatment means exists, is accepted. Means and standard deviations obtained are reported .in.Table 4.2. The smaller homophily score for subjects in Bmode I than for subjects in Mode II signifies the existence <>f a greater degree of perceived homophily with the source in this mode . 123 124 TABLE 4.1.--Summary Data from Two-Way Analysis of Variance of the Homophily Effect Showing Main Treatment and Modernity Effects and their Interaction. Source of Variance SS* df* MS* F* A Treatment modes' 6996.25 1 6996.25 45.10 B Modernity 11949.59 3 3983.19 25.68 AxB Interaction 99.43 3 33.14 0.21 Within groups 28539.70 192 155.10 Total 47584.99 199 * SS = Sum of squares df = Degrees of freedom MS = Mean squares F = Sample F TABLE 4.2.--Mean Homophily Scores and Standard Deviations in the Alternate Treatment Modes. Mean H Standard Treatment N Scores Deviations Mode I Village 100 72.95 16.01 source Mode II Urban 100 85.03 12.81 source Difference 12.08 125 Hyppthesis II: Subjects' degree of perceived homophily with source does not differ across levels of modernity. As theoretically predicted, mean homophily distance is negatively related to modernity,l irrespective of treat- ment mode. The F value for the difference in homophily means across levels of modernity is 25.68, which is greater than the 3.23 required for significance at the .025 level. The null hypothesis, stating that subjects' degree of perceived homophily with source does not differ across levels of modernity, is rejected. The alternate hypothesis, stating that subjects' degree of perceived homophily with source differs across levels of modernity, is accepted. It can be seen from the row means in Table 4.3 that, irrespective of treatment exposure, the more modern the subject, the more homophilous the subject's perceived relationship with the source tended to be. Since it is obvious that homophily distance is related to modernity, no attempt was made to evaluate the statistical signifi- cance of any given pair of homophily means. , Hypothesis III: Subjects' degree of perceived homophily with source is not a function of treat- ment mode/modernity interaction. As predicted, the interaction effect between treat- ment modes and modernity is not significant. Table 4.1 k 1A low numerical homOphily score indicates high Perceived homophily with the source. 126 TABLE 4.3.--Mean Perceived Homophily Distance Scores Across Treatment Modes and Levels of Modernity. Levels of 1 Combined Modernity N Mode I N Mode II N Modes (row means) High 25 63.33 25 76.33 50 69.83 Above average 25 67.12 25 80.04 50 73.58 Below average 25 76.04 25 88.83 50 82.43 Low 25 85.33 25 94.91 50 90.12 100 100 200 (column means) 72.95 85.03 shows that the F value for interaction is .21, which is smaller than the 3.23 required for significance at the .025 level. Null Hypothesis III, stating that subjects' degree of perceived homophily with source is not a function of treatment mode/modernity interaction, is not rejected. Table 4.3 shows that additivity is present in the homophily relationships reported. Mean homophily difference attri- buted to the alternate treatments did not depend on a particular level of modernity. Significance Tests of the Television Effects Major emphasis in the present study is focused on the question of source effectiveness measured in terms of (l) knowledge-gain, (2) knowledge retention, (3) attitude Change, (4) persistence of attitude change, (5) information- Seeking behavior change, and (6) persistence of 127 information-seeking behavior change. Since laboratory control procedures were employed in the present research, thereby reducing the possibility of rival hypotheses, significant findings obtained on the television effects variables can be attributed to the causal factor, subject's degree of perceived homophily with the source. Perceptual factors other than homophily may have contributed to the television effects (perceived source credibility may be one). But these factors were not uncovered in the research. All of the television effects hypotheses were sub- mitted to F tests in two-way analysis of variance, fixed effects model, to determine (1) treatment effects, (2) modernity effects, and (3) the effects of treatment mode/ modernity interaction. These F tests furnish answers to three major questions posed in the examination of effects data: Is one treatment significantly more effective than the other? Do individuals at one level of modernity tend to obtain significantly higher scores than those at another level? Is a significant difference in performance within a treatment qualified by the level of modernity of the Subjects in that treatment? [Knowledge-Gain Hypothesis VIa: Knowledge-gain does not differ across the treatment modes. 128 Subjects exposed to Mode I, the village source, gained significantly more knowledge, on the average, than subjects exposed to Mode II, the urban source. The F value for the difference in treatment means is 5.88, which is greater than the 5.15 required for significance at the .025 level. Null Hypothesis VIa is, therefore, rejected and the alternate hypothesis, stating that knowledge-gain differs across the treatment modes, is accepted. A summary of data obtained from statistical analysis of the knowledge-gain hypotheses is furnished in Table 4.4. It may be seen from the column means reported in Table 4.5 that subjects exposed to Treatment Mode I gained more knowledge, on the average, than subjects exposed to Treatment Mode II. Hypothesis VIb: Knowledge-gain does not differ across levels of modernity. Modernity as a main effect in the test of Hypothesis VIb is statistically significant. The F value for the dif- ference in knowledge-gain means across levels of modernity is 5.03, which is greater than the 3.23 required for signi- ficance at the .025 level. The null hypothesis is rejected and the alternate hypothesis, stating that knowledge-gain differs across levels of modernity, is accepted. Row means Shown in Table 4.5 indicate that the less modern the sub- ject, the larger his gain in knowledge tended to be, irrespective of treatment exposure. 129 TABLE 4.4.—-Summary Data from Two-Way Analysis of Variance of the Knowledge-Gain Effect Showing Main Treatment and Modernity Effects and their Interaction. Source of Variance SS df MS F A Treatment modes 133.33 1 133.33 5.88 B Modernity 341.85 3 113.95 5.03 AxB Interaction .64.45 3 21.48 0.94 Within groups 4165.66 192 22.68 Total 4705.31 199 TABLE 4.5.--Mean Knowledge—Gain Scores Across Treatment Modes and Levels of Modernity.* Levels of Combined Modernity N Mode I N Mode II N Modes ' (row means) High 25 5.00 25 2.41 50 3.70 Above average 25 5.79 25 4.91 50 5.35 Below average 25 6.29 25 6.08 50 6.18 Low 25 8.87 25 5.87 50 7.37 100 105 200 (column means) 6.84 4.82 * Pre- to posttest gain scores were used as the dependent variable. 130 A good deal of caution needs to be exercised in interpretation of the significance of the modernity effect. As Campbell and Stanley (1963, p. 15) indicated, a regres- sion effect2 is almost always present in experimental designs in which extreme groups are sampled. Extreme low scorers on a pretest are aided by the probability of improvement on a second test; since they rarely perform less well on a posttest, their scores tend to regress toward the total posttest group mean. The same principle tends to operate in reverse for high pretest scorers. That is, high pretest scores tend to regress downward toward the total posttest group mean. Preliminary analysis of data revealed that the low modern subjects in the present sample, less educated and less exposed to external communication, tended to be low scorers on the pretest of knowledge. Because a regres- sion effect was likely present in the statistical analysis of Hypothesis VIb data, modernity should not be considered as a meaningful main effect. Hypothesis VIc: Knowledge-gain is not a function of treatment mode/modernity interaction. 2A regression effect is an artifact of the statisti- cal analysis of data which supplements the gain scores for the below-mean pretest scorers, and tends to cancel it for the high pretest scorers (Campbell and Stanley, 1963, p. 15). 131 Contrary to theoretical prediction, the treatment mode/modernity interaction effect did not produce a signi- ficant F and the null hypothesis is not rejected. The F value for the interaction is .94 which is smaller than the 3.23 required for significance at the .025 level. There is no significant difference in the amount of knowledge gained across levels of modernity as a result of exposure to the alternate treatment modes. For interaction to have taken place as theoretically predicted, high modern subjects would have gained more knowledge from the urban source in Mode II than from the village source. As it turned out, subjects at all levels of modernity tended to obtain higher knowledge-gain scores as a result of exposure to the village source rather than the urban source. Knowledge Retention The relationships of the two treatment modes and four levels of modernity to the dependent variable, knowledge retention, are reported in Table 4.6. Hypothesis VIIa: Knowledge retention does not differ ' across the treatment modes. A treatment effect in the test of Hypothesis VIIa did not obtain. The F value for the difference in treatment means is 3.23, which is smaller than the 5.15 required for significance at the .025 level. The null hypothesis, stating that knowledge retention does not differ across 132 TABLE 4.6.--Mean Knowledge Retention Scores Across Treatment Modes and Levels of Modernity.* Levels of Combined Modernity N Mode I N Mode II N Modes (row means) High 25 . 3.62 25 1.37 50 2.50 Above average 25 3.08 25 3.41 50 3.25 Below average 25 7.04 25 6.87 50 6.95 Low , ‘25 9.62 .. 25 4.62 ...50 7.12 100 100 200 (column means) 5.84 4.07? * Pre- to post posttest gain scores were used as the dependent variable. TABLE 4.7.--Summary Data from Two-Way Analysis of Variance of the Knowledge Retention Effect Showing Main Treatment and Modernity Effects and their Interaction. Source of Variance SS df MS F A Treatment modes 150.52 1 150.52 3.23 B Modernity 847.50 3 282.50 6.07 AxB Interaction 211.89 3 70.63 1.51 Within groups 8561.75 192 46.53 Total 9771.66 199 133 the treatment modes, is not rejected. Summary data obtained from statistical analysis of the knowledge retention hypothe- ses are furnished in Table 4.7. Table 4.8 shows that pre- to posttest gains in know- ledge resulting from exposure to the village source rather than the urban source did not remain highly stable over time. However, Table 4.8 reveals an interesting trend in the differential effectiveness of the treatment modes among low modern subjects over time. On the average, low modern subjects exposed to the village source obtained higher scores on the post posttest of knowledge than they did on the posttest administered immediately after exposure to treatment. Conversely, on the post posttest of knowledge, low modern subjects exposed to the urban source tended to revert to their pre-televisiOn position.. Hypothesis VIIb: Knowledge retention does not differ across levels of modernity. Modernity as a main effect in the test of Hypothesis VIIb is significant. The F value for the difference in knowledge retention means across levels of modernity is 6.07, which is greater than the 3.23 required for signifi- cance at the .025 level. The null hypothesis is rejected and the alternate hypothesis, stating that knowledge reten- tion differs across levels of modernity is accepted. Row means shown in Table 4.6 indicate that the less modern the 134 .cmmmn mum mamms Ham £0fl£3 no mHmEMm mnu mo muflm on» mgmoflocw mmmmnucmumm ca mumnfisz i Ammv Ammv Ammv Ammv Ammv Ammv Ammv .«Ammv . . . . . . . . mammamc mm H mo m mo I we I ma H mm I am me I CD can H0H>mcmm . . . . . . . . ummnumoa manxmmm oo we mm mm mm mv mm vm on mum IGOHumEHowcH . . . . . . . . pmhwamp om om mm mm I cm H mo wH 0H I Op mum . . . . . . . . umopumom mm Hm om ma mm as 00 mm on mum mmU5pfluu¢ . . . . . . . . omsmamo No v mm 0 av M pm H mm m we 5 mo m mm m ow mum . . . . . . . . pmmuumom mm m mo m am a av m hm m mm m an m co m. on mum omcmH3OQM 30a mmmum>< mmMHm><.I£mflm 30A. mmmum>¢ mmmum>¢ nmfim Bonm m>on¢ 30Hmm m>on€ mmuoom,v muommmm HH moo: H moo: came conH>mHmB .Hmumq wxmwz Hsom can ucmfiummua «0 mafia may um monummmz muommmm :oflmw>mama 059 so mmHoom meow cmmEII.m.v mqm