. . I f, 2.2,? ...,2 v J Jrk : . i :1. ”fit. , 1 . O O O p 1 . . . y ... “$3951. 4 .. v Q. 25...“? 5.31.; . v — .v I n v I." I v 3. ’ you c 0 ”£1“, .. n)" '3’ a c ’Wflvdhrtw V .0 In! 3.53.3v «n , 113.. . .(rr/ . .1! :1. . Yrvfrrv-l: ~.T........v. .4154 \L. ‘ 4 12.: .'X.. 4 ‘ 1.34.1.174 fir ..rt.v.‘vvrv..1!-ir \ vr.u:l.1.1:1vvlvv n. y tit-2...}. if 4v.v . 1 I. 'auvi ft \ \fvcr-l ,(vn.‘r Pvrrvvfr 111 4vv. ..7..4.¢ .1!v..ir.l!’ a s , 9.2 .11.. vv‘y.)l...l. leu.v 7.51:! ,L... ...3 3.. 5 L 2,2,...3...‘ z lllllllllllllglllllllLllllllllfllllllllljlllllfll 1 LIB RA R Y Michigan State Unit" ty -ri This is to certify that the thesis entitled An Information Theory of Communication Functions: A Focus on Informational Aging presented by James A. Danowski has been accepted towards fulfillment of the requirements for Ph.D. . Communication __ degree in _______._ 4196/3017 Véwz 46 / Major professor .‘lmx v ' i! n £41 ' December 9, 1974 Date__—__ 0-7 639 ABSTRACT AN INFORMATION THEORY OF COMMUNICATION FUNCTIONS: A FOCUS ON INFORMATIONAL AGING By James A. Danowski As the demographic structure of the population shifts toward becoming increasingly older, greater research attention to aging is called for in understanding the implications of these processes. In this work a functional view of aging is taken which focuses attention on communication behaviors, both mediated and interpersonal. A major thrust is the explication of a set of concepts and interrelationships labeled informational aging. Key features of this perspective are that informational aging: l) is a curvilinear process occurring over the life—span, 2) involves changes in the complexity of interpersonal information space, mediated information space, and the individual information processing system, and 3) has as a major determinant the perceived individual control over the environment. In developing the notion of informational aging, an information theory of communication functions is explicated. At the core of this model is a principle of optimal information, which suggests that individuals have a preferred range of uncertainty and that deviations above or below this range create forces to alter the amount and entropy James A. Danowski of communication with the environment through both media and inter— personal sources. Four functions of communication are proposed. Two uncertainty reduction functions are stress release and forgetting, while two uncertainty expansion functions are stimulation and learning. Control over the environment is treated as an important determi- nant of whether more change attempts will be directed toward the medi- ated or interpersonal sphere, as well as the amount of uncertainty discrepancy that is likely to be sustained by the individual. Also explained by this cybernetic model are the attributions individuals make regarding the functions of their communication behaviors. Control over environment is proposed as a major predictor of the extent of attribution and the relative evaluation of mediated and interpersonal communication. In testing components of the model, the primary research involves a field survey of a retirement community in the Mid—Michigan area. The research setting is composed of l98 residents living in a set of one— story apartment buildings on a 58 acre site. Nearly all residents are self—sufficient in meeting basic needs. The average age is 72 and 8l% are women. The population of the retirement community was censused to enable mapping and structural quantification of the interpersonal connmnica— tion network operating within the residential community, in addition to the measurement of other variables. Personal interviews were conducted by a team of twelve trained interviewers. James A. Danowski Hypothesized relationships were tested with partial correlations, and the major findings of the study are summarized as follows. The individual with high control over the environment talks to a relatively large number of people and uses little television. Overall, the high control individual attributes less uncertainty balancing functions to his/her communication behaviors, and slightly more to interpersonal communication than to television use. Considering all the individual's communication behavior, we can predict more by knowing the magnitude of the person's perceived control over environment, than by knowing the person's age. 0n the other hand, the low control individual consumes greater amounts of television messages and talks less to people in a face—to— face mode, while attributing more uncertainty balancing functions to both mediated and interpersonal communication. Evidence also suggests that regardless of level of control, persons experiencing uncertainty overload consume greater amounts of mass communication. Those who are experiencing uncertainty underload have interpersonal networks of low integration, suggesting a high interpersonal content entropy to balance the underload. Also, persons experiencing uncertainty overload attribute communication behaviors to uncertainty reduction functions. AN INFORMATION THEORY OF COMMUNICATION FUNCTIONS: A FOCUS ON INFORMATIONAL AGING By . \(.,‘ x; Z- 0 T , James A. Danowski A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Communication T975 Accepted by the faculty of the Department of Communication, College of Communication Arts, Michigan State University, in partial fulfillment of the requirements for the Doctor of Philosophy degree. W MW. Director of Thesis Guidance Corrmittee: W, 1/ W , Chairman IJQA‘4JAAJ 1’ I ii ACKNOWLEDGMENTS This work was accomplished under the guidance of Richard V. Farace, to whom I owe many thanks and appreciation for his multi—faceted con- tributions. Bradley S. Greenberg, Joseph Noelfel, and Michael Moore were very helpful Committee members. A number of friends freely gave their valuable time and energy during various phases of the research and without their assistance the project could not have been conducted. My grateful thanks go particularly to Donna Danowski, and also to Betty Farace, Michael Pacanowsky, Sue Rose, Jennifer Shelby, James Taylor, Judy Taylor and the others whose names are too numerous to list. I also appreciatively acknowledge the cooperation of the management and resi— dents of the Grange Acres and Cedar Place retirement communities, who opened their lives to scientific scrutiny. TABLE OF CONTENTS CHAPTER Page PART I THE CONTEXT FOR THEORY AND RESEARCH Foreword ................................................ l I. THE PROBLEM ............................................. 2 Overview .......................................... 2 The Validity of Communication Theory .............. 2 Functional Analysis ............................... 7 Summary ........................................... 9 II. COMMUNICATION AND AGING ................................. ll Overview .......................................... ll Theories of Aging and Communication ............... l2 Communication implications ..................... l3 Interpersonal Communication ....................... l4 Predictors of interpersonal communication ...... l4 Interpersonal communication as predictor ....... l6 Mass Communication ................................ l8 Portrayal of old people ........................ 18 Television exposure ............................ 20 Television content preferences ................. 22 Other media .................................... 22 Media relationships to other activities ........ 24 Summary ........................................... 28 III. PREVIOUS VIEWS OF COMMUNICATION FUNCTIONS ............... 29 Overview .......................................... 29 General Functions of Communication ................ 29 Functions of Mass Communication ................... 3l Brief history .................................. 3l Theoretic perspectives ......................... 32 Methodological approaches ...................... 33 Summary ........................................... 36 iv TABLE OF CONTENTS--continued CHAPTER Page PART II AN INFORMATION THEORY OF COMMUNICATION FUNCTIONS Foreword ................................................ 37 IV. INFORMATION SPACE: DEFINING THE BASIC CONTEXT OF THE THEORY .................................................. 38 Overview .......................................... 38 Information Space ................................. 4l Defining Information .............................. 43 Traditional definition ......................... 43 Information as form ............................ 43 Information, Content, and Context ................. 44 Distinguishing information and uncertainty ..... 45 Creating Uncertainty and Information .............. 46 Creating Chunks ................................... 46 Linking Chunks Together ........................... 48 Individual and Social Factors ..................... 49 Summary ........................................... 49 V. MEDIATED AND INTERPERSONAL COMPONENTS OF INFORMATION SPACE ................................................... 50 Overview .......................................... 50 Mediated and Interpersonal Components ............. 50 Interpersonal Component ........................... Sl Mediated Component ................................ 52 Communication Structure ........................... 53 Measurement .................................... 55 Locus of definition ............................ 56 Channel entropy ................................ 56 Content entropy ................................ 57 Form entropy ................................... 57 Examples of Interpersonal Component Entropy ....... 57 Examples of Mediated Component Entropy ............ 60 Summary ........................................... 62 VI. DEFINING BOUNDARIES: ENVIRONMENTAL, EGO, AND COGNITIVE SPACE ................................................... 63 Overview .......................................... 63 Boundaries ........................................ 63 Ego space ...................................... 64 Boundary as zone ............................... 65 TABLE OF CONTENTS—-continued CHAPTER VII. VIII. Cognitive Space .................................... Cognitive Uncertainty .............................. Cognitive Complexity ............................... Uncertainty Preference ............................. Summary ............................................ THE THEORETIC CORE: OPTIMAL INFORMATION, FUNCTIONS OF COMMUNICATION, CONTROL OVER ENVIRONMENT, AND ATTRIBUTIONS OF COMMUNICATION FUNCTIONS ............................... Overview ........................................... Balance Model ...................................... Functions of Communication ......................... Uncertainty Discrepancy and Communication Behaviors Control Over Environment ........................... Effort .......................................... Balance utility ................................. Relative use .................................... Generality of optimal information ............... Attributions of Communication Functions ............ Attribution Propositions ........................... Summary ............................................ INFORMATIONAL AGING ...................................... Overview ........................................... Process View ....................................... Types of Change .................................... Factors in Informational Aging ..................... The U—Curves ....................................... Control and Informational Aging .................... Trends in Mediated and Interpersonal Communication. Ego Space and Informational Aging .................. Uncertainty Preference and Informational Aging ..... Functional and Chronological Aging ................. Summary ............................................ PART III THE RESEARCH Foreword ................................................. vi lO7 TABLE OF CONTENTS—~continued CHAPTER Page IX. RESEARCH PROCEDURES AND OPERATIONAL DEFINITIONS .......... 108 Overview ........................................... lO8 A. Research Procedures ................................... lO8 Design Selection ................................... lO8 Sample Selection ................................... lO9 The Primary Research Setting: Grange Acres ........ llO Contact with the Community ......................... llZ Pilot Tests ........................................ ll3 Interviewer Recruitment and Training ............... ll6 Response Rate ...................................... ll7 Demographic Characteristics of Grange Acres ........ ll8 Generalizability ................................... ll9 Propositions to be Tested .......................... l20 B. Operational Definitions ............................... l21 Uncertainty Scale .................................. l2l Control Over Environment ........................... l26 Network Analysis ................................... l29 The variables ................................... 132 Media Exposure Items ............................... l34 Media entropy measures .......................... l36 Functions of interpersonal communication ........ l37 Functions of television ......................... l4O Socio-economic Status .............................. l44 Age ................................................ l45 Summary ............................................ l45 X. RESULTS .................................................. l46 Overview ........................................... l46 Analysis Procedures ................................ l46 Descriptive Statistics and Inter-item Correlations. l47 Tests of Hypotheses ................................ l49 Summary of Results ................................. l57 XI. INTERPRETATION OF RESULTS AND IMPLICATIONS ............... l60 Overview ........................................... l6O A. The Results in Light of Methodological Considerations. l6l The Findings ....................................... l6l B. Future Research Suggestions ........................... l66 Improved Operations ................................ l66 Age Distributions .................................. 168 Time ............................................... l69 Elaboration of the Research Program ................ l7O Summary ............................................ l7O vii TABLE OF CONTENTS--Continued Page LIST OF REFERENCES .............................................. T72 APPENDIX--INFORMATION LETTER AND QUESTIONNAIRE .................. T82 viii LIST OF TABLES TABLE _‘ T0. TT. T2. T3. T4. T5. T6. . Descriptive Statistics on Uncertainty Items ................ . Inter-item Correlations for Uncertainty Items .............. . Descriptive Statistics for Control Over Environment Items.. . Inter-item Correlations Among Control Over Environment Items ...................................................... . Factor Analysis of Control Items ........................... . Descriptive Statistics for Intra- and Extra-community Items . Media Exposure Descriptive Statistics (hours per day) ...... . Descriptive Statistics for Functions of Interpersonal Com- munication Items ........................................... . Correlations Among Functions of Interpersonal Communication Items ...................................................... Factor Analysis of Functions of Interpersonal Communication Items ...................................................... Descriptive Statistics for Functions of Television Items... Inter-item Correlations Among Functions of Television Items Factor Analysis of Functions of Television Items ........... Descriptive Statistics for Indices ......................... Correlations Among Indices ................................. Comparison of Control Over Environment with Age in Predic- tive Power ................................................. ix Page l24 T25 T27 T28 T28 T33 T35 T38 T39 T40 T4l T42 T43 T48 T50 T56 LIST OF FIGURES FIGURE Page l. Methodological approaches to functions of communication.. 36 2. Basic components of the model ............................ 4O 3. Levels of information space .............................. 4T 4. Three levels of communication structure .................. 54 5. Information and liking ................................... 7O 6. Four functions of communication .......................... 79 7. A cybernetic model of optimal information ................ 8l 8. Summary of hypothesis testing ............................ l58 PART I THE CONTEXT FOR THEORY AND RESEARCH FOREWORD Part I sets the foundation for the development of an information theory of communication functions in Part II, and the report of research designed to test components of the theory in Part III. Part I begins with a discussion of aging as a focus of communica— tion research (Chapter I). This chapter then evolves into a discussion of prior research in communication and aging (Chapter II). Part I closes with a review of prior conceptualization and research in the general area of communication functions (Chapter III). CHAPTER I THE PROBLEM Overview This chapter offers two thrusts in defining the context for this work. One thrust is the role of aging in improving the external and internal validity of communication research. The other is the function- al approach to communication theory construction. We conclude this chapter with an overview of the explication of theory and research aimed at the development of an information theory of communication functions, featuring informational aging. The Validity of Communication Theory In the context of social research, two important dimensions are frequently discussed: internal validity and external validity. Because of the deep intimacy of theory and methods, these two dimensions may also be applied to theory. Can communication theory be made more internally and externally valid? We seek to answer this question by transforming a wide range of communication variables into a theory cast in the context of aging. First, the external validity of communication theory and research will be discussed. Two age groups appear to have received most attention in communi— cation research--college students and children. The drawing of these kinds of samples in communication research has limited its generaliza— bility. Less attention has been given to researching the adult; and moreover, old people have been given negligible research attention. The great bulk of research, primarily experimental, has been conducted with college underclassmen, and in recent years considerable attention has also been given to children, particularly in the area of mass com- munication and the effects of televised violence. Interestingly, changes in the demographic structure of the popu- lation are at variance with the demographic structures of the samples generally drawn in communication research. The average life span of males and females has increased from 49.2 years in l900 to 70.5 in l969 (United States Bureau of Census, T973). Correspondingly, the percent— age of persons over the age of 65 has increased from 5.4% in T930 to 9.9% in T970. In addition to longer expected life spans, an additional demographic factor impinging on the changing age distribution is the declining birth rate in the United States. Demographic projections reflecting the coupling of these two phenomena show that the proportion of older people will continue to increase to l3.l% in 2020 (Cutler, Bengtson, and Dowd, T974). Thus, if communication theory and research are to be more generalizable, these demographic factors alone call for greater attention to aging. Considering now the internal validity of communication theory and research, research attains ultimate internal validity when the logic of research design is isomorphic with the logic of theory. Nherein, then, lies the internal validity of theory? Theory attains ultimate 45.....-- —- $5.. _. internal validity when the logic of theory is isomorphic with the logic of human information processing. And the latter may be viewed as the logic of human experience. This matching of the logic of theory and the logic of human information processing and experience is this cri- terion of validity that we seek in the present theory and research. The important link between the two comes through the widespread view that human experience is process. This is seen both in the physi- cal sphere and in the social sphere. The work of Einstein (l920) has highlighted a process view of experiencing the physical world, and in the social sphere, communication theory has been oriented to a process view through the work of Berlo (T960, T969, T970). The methodology of communication research has also contributed to a process view through the increased development and use of multivariate statistical procedures. The more humanistic areas of social science, reflecting a blending of threads of Gestalt psychology and Eastern philosophies and practices, have facilitated the growth of a process view in both communication theory and methodology. In modern models of communication, process views are typically manifested in the inclusion of feedback as a necessary component. The concern with feedback as a concept is typically expressed at a micro— level of analysis, the dyadic level communication system. However, more macroscopic foci such as mass communication, small group communication, and organizational communication have also built feedback notions into contemporary theory and research. Regardless of the level of analysis, when feedback is recognized conceptually, tim§_is implicitly prominent, for feedback occurs typically as a response to some pattern that is received and decoded by an information processor. Both of course cannot occur simultaneously. Therefore, in a process view of communication theory and research feedback and time are essential. When the concern for explaining the communication behaviors of individuals is coupled with a process view, manifested in the concept— ualization and measurement of variables from a time perspective, age_ and maturation become important. The same clocks we use to measure time in testing theories are the same clocks we use to measure a person's life. Thus, it is difficult to separate aging from communication be- haviors if individuals are studied over relatively long periods of time. This is not to suggest, however, that time and aging are synonom- ous. While chronological age is measured solely on the dimension of time, agjpg_is more than just time. Unlike chronological age, which is linear and unidimensional, aging is multidimensional and non-linear. Key factors in the non-linearity of aging are birth and death. These periods involve massive energy reorganizations in the person, with birth an important marker in the initial development of consciousness and death a marker in the disintegration of consciousness. This energy and information reorganization is of such intensity that energy first becomes so highly modulated that it develops the capacity to feedback on itself, developing information processing patterns, awareness and the ability to decide how to pattern itself. Then in later portions of the life span these processes decline, and eventually the unique multi— dimensional non-linearity of experience loses dimensionality and becomes invariant. --—~_-..._..._— - These intense changes in information organization, involving the inter-play of functional relationships among a set of variables, gives rise to a functional view of aging, and a functional approach to explain— ing communication in aging. A process view suggests that age becomes both a criterion and a predictor variable, with the reciprocal effects of aging and other communication variables becoming important. ”How can age function as a dependent variable in relationships with aspects of human behavior?" A functional approach to aging suggests that while physiological, psychological, social and communication processes vary with time, time itself is not their determinant. Time merely tracks their interplay. However, at a macroscopic level there is such con- sistency in the slopes of form—by—time relationships across individuals that it may appear that time itself is the essential cause of change. Hence, chronological age has been a primary scientific variable in aging research. In communication research, age has had low status as a variable. Typically, its effects are "removed” from some more important relation- ships, provided, of course, that it is used in empirical analysis at all. However, when the theoretic concern for explaining communication behaviors from a process view is linked with functional notions of time varying processes, an approach to theory emerges that has a distinct developmental perspective. We label this approach informational agjpg. In remaining chapters we will report a theoretic and empirical explora— tion of informational aging. Control over environment will be a key construct that yields functional transformations in informational aging over the developmental span. In the research testing the theory, there are constraints prevent- ing the examination of a complete developmental range of subjects. However, the segment of the life-span about which least is empirically known will be focused upon. The average age of our respondents is 72 years. This provides balance to the non-representativeness of samples that have been limited greatly to college students and 'children in a major proportion of communication research. In the next section, since the theory and research will take a functional perspective in its basic operation, the major characteristics of functional analysis will be briefly explicated. Functional Analysis Functional or teleological explanations may be traced back to the work of Aristotle, specifying that objects move in physical space to their natural resting places. Thus, their movement is goal-directed. More recent explications of functional explanations in social sciences have arisen in the work of Merton (T967) in sociology and Radcliffe- Brown (l952) and Malinowski (T945) in anthropology. Essentially, these perspectives focus on system adaptation and adjustment, and do not operate on a deductive, deterministic model. Monge (T973) in discussing functional explanation cites a definition of function by Radcliffe—Brown (l952) that is a useful one: The function of any recurrent activity such as the punishment of a crime, or a funeral ceremony, is the part it plays in the social life as a whole and therefore, the contribution it makes to the maintenance of the structural continuity. The concept of function as here defined thus involves the notion of a structure consisting of a set of relations amongst unit entities, the continuity of the structure being maintained by a lifeeprocess made up of the activ- ities of the constituent units. (p. T80) -‘—.~_ While this definition is developed in the context of a macro functional- ism perspective, the basic structure of the definition is applicable to a more micro level. In the theory to be explicated in this work a func- tional perspective will be taken on the behavior of individuals. Little change in the above definition is necessary to make the transition: The function of any recurrent communication behavior is the part it plays in the individual information space as a whole and there- fore, the contribution it makes to the maintenance of the struc- tural continuity of information space. The concept of function involves the notion of a structure consisting of a set of rela- tions among unit entities (components of individual information space), the continuity of the structure being maintained by a life-process made up of the activities of the constituent units. Monge (l973) in drawing on the work of Nagel (l957) and Merton (T967) developes the following logical requirements for functional analysis: (l) identification of the system, i.e., a set of interrelated parts are identified that may be viewed as a whole. It is not necessary that all the parts be specified so long as (a) the system as a whole is identified, and (b) those parts that are necessary for the analysis are identified. (2) specification of the environment in which the system operates. This generally means the specification of all those factors that are not a part of the system but which may affect the particular behavior of the system being studied. (3) determination of some trait, attribute, or property of a sys- tem which is considered essential for the continuation of the system. (4) specification of the range, i.e., the different values that the trait mgy_assume as well as the range within which it must stay if the system is to remain in operation. (Every trait is assumed to be a variable, even if only dichotomized into present-not present.) (5) a detailed account of how the parts collectively operate to keep the value of the trait within the limits required for the existence of the system despite other changes in the system or impinging influences from the environment. Such a system is said to be functional, self-maintaining, direc- tively organized, or goal-directed with respect to the trait. Empirical requirements are also specified for functional analysis: (l) Some indication must be made of the system operation. This implies (a) a set of criteria for deciding when the system ceases to exist or operate gga_system, as well as (b) measure- ment that can reveal the current state of the system. (2) The range which the system trait covers must be measured. Linkage must be demonstrated between the required range (a subset of the possible trait range) and the continued operation of the system. (3) The several mechanisms identified in the logical analysis must be measured and their relationship with the trait must be identified. The mechanisms are considered conceptually independent but may be empirically interdependent. (4) If any environmental factor affects the system trait, it must also be measured and its relationship with the trait and system mechanisms (if there is a relationship) indicated. (p. 6l-63) In the following chapters the explication of the logic of the information theory of communication functions, and the empirical research testing this theory, will meet the above nine criteria for functional analyses. The core of the functional theory will be made more precise through the development of a cybernetic model, specifying the functional interrelationships of the system components in more rigorous terms. Summary In this chapter two general topics have been discussed. First, the role of aging in improving the external and internal validity of communication theory and research was focused upon. With respect to external validity, it was posited that theory and research will benefit from increased attention to the older segments of the population, both because they have been undersampled in the bulk of previous communica- tion research, and because the age distribution of the population is changing in the direction of becoming ”older,” making it even more justified to study older people with the aim of improving the generali- zability of research. With respect to internal validity, it was pro— posed that a process view of communication implies that changes in com- munication behaviors over the life—span are important to explain at a theoretical and empirical level, if theory and research is to more adequately correspond to human experience. Second, background on the general theoretic perspective to be taken in the remaining chapters was developed. A functional approach was delineated, specifying conditions of logical and empirical adequacy for functional analyses. The remaining theoretic and empirical chapters will be devoted to meeting these conditions. CHAPTER II COMMUNICATION AND AGING Overview Overall, in the history of communication research a great deal of attention has been given to samples of young persons, particularly in recent research of televised violence effects on children. Communica— tion behaviors in aging, however, have received little research atten— tion, and thus no cohesive body of literature can be identified. In this section this rather limited research literature will be surveyed. The following sources were searched: Dissertation Abstracts, Psycho- lpgical Abstracts, Sociological Abstracts, Journalism Quarterly, Public Opinion Quarterly, Journal of Communication, American Sociological Review, American Journal of Sociology, Journal of Gerontology, Gerono- tologist, Aging and Human Devel0pment, and the card catalog listings of the Gerontology Center Library at the University of Southern California. In organizing the discussion, three components of communication research in aging will be identified. While these levels are specified here in the context of aging, they are general to studies of all human communi- cation behavior. The three components are: l) interpersonal communi- cation, 2) mediated communication, and 3) relationships of interper- sonal to mediated communication. Before reviewing this literature, TT T2 general theories of aging will be reviewed in light of their communica- tion implications. Theories of Aging and Communication Social gerontology has seen the development of some major theo- retic perspectives on the aging process. It is characteristic that these theoretic views are normative in nature. Rather than attempting merely to explain the behaviors of persons as they age, they seek to provide guidelines for how people ought to successfully age. Because of this emphasis, these developments are not theories in the sense that they present a systematic set of relationships among variables for empirical testing, but are merely perspectives on behavioral phenomena. The two prominent theoretic views on social aspects of aging are dig; engagement theory and activity theory. Disengagement theory (Cumming and Henry, l96l) proposes that decreasing contact with the environment in aging is a natural process. Thus, those who decrease contacts with the environment will successfully age. It is postulated that disengage— ment is primarily initiated by the individual, but that both society and the individual benefit from reduced contact. The individual is viewed as released from roles that strain his capacity under the decreas— ing physical strength of increasing age, while the society receives new opportunities for younger people. Activity theory takes an opposing view on successful aging. It strongly reflects the dominant cultural values in this society revolving around the work ethic (Kleemeier, T964). This perspective proposes that for an individual to be happy in aging, he or she must be activee—active T3 both in personal ways and active socially. In general, when both activity theory and disengagement theory are empirically tested, mixed results emerge. To this date there appears to be no clear-cut resolu- tion of this basic controversy. Communication implications. The implications of disengagement theory for communication behaviors suggest that interpersonal communica- tion decreases with aging, with less amount of communication, less variety in kinds of people communicated with, and less variety in com- munication content. Although less directly implied, the perspective would seem to suggest that mass communication use should also decrease with aging, since there is a general decrease in concern for the outer world and increased attention to the self and inner world. Examining the implications of activity theory for communication behaviors, it seems likely that the theory would call for maintenance or increases of interpersonal communication with aging. Number of con— tacts with others should increase, as well as the number of different people communicated with, and the variety of content in communication. Considering mass communication, it is difficult to unambiguously determine the implications drawn from activity theory for mass communi- cation behaviors. Although because the focus is on activity, this perspective perhaps suggests that people maintain or increase their use Of mass communication in aging. However, an increase in mass communi- cation behaviors as a replacement for decreased interpersonal communi- Cation would probably be viewed unfavorably from an activity theory Petspective. The author's evaluation of the general orientation of the tfneory suggests that under these conditions attention would be placed T4 on increasing the decremented interpersonal communication, and the dis- placement into mass communication use would be viewed as insufficient to maintain appropriate levels of activity. In the remainder of this chapter, we will review empirical research regarding interpersonal and mass communication behaviors in aging. The research, while restricted largely to communication behaviors, will illustrate the mixed results that are received when these two theories are directly tested with broader ranges of behaviors. Interpersonal Communication Research directly concerned with interpersonal comnunication be- haviors in aging is nearly non-existent. However, while the author could identify no research deriving from the discipline of communication theory and research, there are some studies in social gerontology that are concerned with social interaction in aging, particularly in the context of friendship patterns. This work may be viewed as closely re- lated to interpersonal communication behaviors, both theoretically and operationally. First, research will be reviewed that examines inter- personal communication as a dependent variable. Then, research will be considered that examines interpersonal communication as a predictor of other variables. Predictors of interpersonal communication. Rosow (T967) conducted research in a large number of apartment buildings to determine the rela- tionships between the older age density (old age homogeniety) of resi- dential areas and the number of friends held by older persons living Tr! these varying areas. The basic results of the research show that as T5 old age density increases, number of friends nominated by older resi- dents increases. This suggests that age similarity or homophily leads to increased likelihood of communication. Also, social class was found to have an impact on the density and friendship relationship. Middle class respondents tended to have friendship patterns less subject to density influences; much of their communication was with others outside of their immediate neighborhood. Working class were much more local in focus and more subject to influences of density on friendship patterns. Rosenberg (T968) explores the relationships between consonance and dissonance between the individual's social characteristics and those of the local residential community, and friendship patterns. Blue-collar workers or retired blue-collar workers, were interviewed (n=l596). Those aged 65 or older numbered 668. Results indicated that old men (over age 65) living in neighborhoods where their wealth, occupation, or race differs from other local residents, tend to be isolated from friends. This relationship was strongest for ”solvent” rather than "poor" working class males. Bultena (T968) tests the hypothesis that with increasing age, there is an increasing ratio of communication with older people to com- munication with younger people. An area probability sample of six Wisconsin communities, ranging from small rural to large urban, resulted in a sample of 434 retired men. Social interaction was measured by the frequency of contact each respondent had with spouse, children, grand- children, siblings, relatives, and friends, at least weekly. Social interaction within organizations or formal groups was also measured. T6 Results showed that elderly males had a substantially greater amount of face-to-face contact with age-mates than with younger persons. However, as age increased, there was greater contact across a wider age range; respondents aged 80 and older showed greater interaction with younger people than older people, and an absolute increase in contact with younger peOpTe with increasing age. Advancing age was also associated with decreased participation in formal groups. Interpersonal communication as predictor. Tobin and Neugarten (l96l) report a panel study of a stratified random sample of Kansas City residents between the ages of 50 and 70, with an additional single point in time sample ages 70-80. The concern was measurement of the relation— ship between social interaction and life satisfaction. Four social interaction variables were used: the gmpppt_of interaction, the number of different people that the respondent interacts with in a month, the perceived current interaction rate relative to age 45, and the number of different roles in which the individual interacts. The first three variables may be considered communication variables, if interaction is viewed as comprised of verbal and non-verbal communication. The life satisfaction index included measures of pleasure in activities, meaning— fulness of life, success in goal attainment, positive self—image, and being in a happy and optimistic mood. Results of the research indicate that as social interaction increases, life satisfaction increases, and this relationship becomes stronger with increasing age. Lowenthal (T964) studied the relationship between social isolation and admission to a mental hOSpital. Two samples were used—-the popula— tion of 534 persons admitted to a mental hospital, and a stratified T7 random sample of 600 community persons in the same residential areas. Results indicate that extreme life-long isolation does not strongly associate with the development of mental disorders in old age, while marginal life-Tong social adjustment is associated with mental disorder in old age. Late developing isolation is linked with mental disorder, but this isolation may be the consequence of mental disorder, not neces— sarily the antecedent. However, the cross-sectional data does not enable the separation of the relative strengths of the bidirectional components of the relationship between isolation and mental disorder. Palmore (T968) reports results of a four-wave longitudinal study (n=l27) of relationships of aging to activities and attitudes. Among the findings are that frequency of contact with family and friends shows no decrease over a ten year period. Amount of total activity shows a positive relationship to satisfaction, while interaction with family and friends does not. Pastorello (T972) reports analysis of two-point in time (T968, T97l) national panel data in testing the relative effects of change in qualitative and quantitative social interaction on change in life satis— faction and other variables. Results of a path analysis show that quantitative interaction has no direct impact on life satisfaction, but that it is associated with negative age—related self-evaluations. Furthermore, increasing quantitative interaction is positively associated with normative integration, i.e., decreasing anomie, however, the in— creased normative integration leads to lower life satisfaction. This suggests that to be integrated into the predominant cultural normative T8 system that holds negative views of aging is detrimental to life satis- faction. The qualitative measure of interaction--whether the respondent has a “confidant"-—shows a strong positive association with changes in life satisfaction. However, this relationship holds only for a confidant serving as a significant other. In summary, the potpourri of research reported here that can be indirectly linked to interpersonal communication behaviors is indicative of the need for further research in this area. Greater consistency in the conceptualization and measurement of variables related to communica- tion, and more intensive follow-up research, extending homogeneous lines of research would be valuable. Mass Communication As is the case with research on interpersonal communication, research on mass communication and aging is very limited. There are few studies directly concerned with aging and mass communication and most of these are unpublished Ph.D. dissertations. Slightly more studies are concerned with the general population, yet report data for older people. In the following review of literature, three basic groups of research will be examined: a) portrayal of old people on television and in print media, b) mass communication exposure and content prefer- ences of old people, and c) relationships between mediated and inter- personal communication in aging. Portrayal of old_people . A literature search revealed only two studies concerned with content analysis of television messages, and several studies concerned with less ”mass” media--children's books and T9 magazine short stories. First, the latter studies will be briefly reviewed. Martel (T968) content analyzed four popular American maga— zines, sampling 40 to 85 stories from several years between T890 and T955. The results generally suggest a declining status for older people over this period, with the prime of life increasingly characterized as young adulthood rather than later life. Earlier prestigious references to age, such as wisdom, experience, and seasoning, shift to references to being past the prime and out of touch with the social mainstream. Earlier indications that old people have close friendships with young people substantially decline. Seltzer and Atchley (l97l) content analyzed children's books, selecting samples at three 30 year intervals from T870. Content analy- sis was performed with semantic differentials, and word counts for references to old people or things. Results showed a slight decline in evaluations of aging over the time-span, but the difference was sig- nificant only for the first two comparison points. There was also a small decrease in reference to old people and things. Robin (T974) analyzed three sets of children's reading textbooks for content about death and dying. Little direct death content was found in the primary grade materials, although some symbolic references to death were found. References to death were more frequent in upper elementary grades, with casual treatment in a nonaffective manner. In general the process of death was infrequently mentioned. Analyzing a random sample of 30 half—hour prime-time segments of programming available in Portland, Oregon, Peterson (l973) tested two hypotheses: a) old people are less likely to be proportionately visible 20 than people who are not old, and b) old women are less likely to be proportionately visible than old men. In addition, it was expected that old people would be portrayed in an unfavorable image, given their status in the society. The investigator indicated surprise at the results that there were greater proportions of old people on television (T3%) than there are in the general population (9%), although women were greatly underrepresented (9l% were male). With regard to image, old people were portrayed unfavorably T8% of the time, neutrally 23% of the time, and favorably 59% of the time. Aronoff (T974) in a very poorly reported study analyzed data per- taining to age in the ”Cultural Indicators“ television archives at the University of Pennsylvania. The analysis was based on 2,74T characters sampled from prime-time television network drama from T969 and l97T. Results show that roughly 5% of all male characters and 5% of all female characters were classified as elderly. With aging there is a trend indicating increased portrayals of "bad guys” and decreased portrayals of "good guys," also a large decrease in portrayals of ”success” is accompanied by a large increase in portrayals of ”failure.” Television exposure. As in the area of interpersonal communica- tion, there is a scarcity of research concerned with mass communication behaviors. Most related research, however, is concerned with television, and we will briefly review this work before considering use of other media. Television is generally the most important leisure activity for the older person. A number of studies report this finding. Cowgill and Baulch (T962) in a study of the leisure activities of a sample of 2T Wichita persons aged 60 and over, report that television was the most frequent leisure activity by a decisive margin for both men and women. Zborowski (T962) reports that television is perceived by a majority of 204 members of the Age Center of New England as a primary recreational activity. Beyer and Woods (T963) reporting the results of a survey of 5,000 social security beneficiaries regarding daily activities, indi- cated that 70% of the respondents watched television on the previous day. Bracey (T966) in a comparative study of British and American pensioners, states that although fewer British pensioners owned tele- vision sets, those that did used it as the primary leisure activity. Interestingly, he states: "In both countries we met the exceptional individuals who kept television in its place as a supplement to ordinary living and not as a way of life in itself....” (emphasis added, p. T43). Considering viewing behaviors, Beyer and Woods (T963) in the study of 5,000 social security beneficiaries report a median viewing time of three hours per day. More recent data from a national sample (Bower, T973) suggests that the average person over age 50 spends about four hours per day watching television, slightly higher than those under 50. Davis (T972) in a study of 289 urban and T74 suburban residents of Los Angeles, reports a median viewing time of roughly 4 hours. Steiner (T963) finds that the number of programs viewed per week increases among persons 55 and over. In the study of social security beneficiaries mentioned above, it was also found that people report viewing more television as they get older. Thirty-nine percent of people (Iver 60 report an increase, while only T6% report a decrease (Opinion 22 Research Corporation, T957). In a cohort analysis, Bower (l973) found that all of seven cohorts showed an increase in viewing over a ten year period, with the youngest cohort showing the least change. Television content_preferences. A number of studies on content preferences of older television viewers show consensus that older people most prefer news and public affairs programming (Steiner, l964; Glick and Levy, T962; Schalinske, T968; Davis, T972). While there is less consensus on the remaining preferences, there is a trend toward comedy and variety as second preference, with less convergence on further con- tent preferences, since researchers define content categories in differ- ent ways. There is also some research on general attitudes toward television. Bower's (l973) comparison of results of Steiner's T960 study and a later T970 study suggests that young and old are most favorable toward tele- vision, with cohort analyses suggesting that this is due to a life-stage factor and not a generational one. Other media. Reading is a frequent activity for old people, particularly of newspapers. Research by Hoar (T960), a descriptive study of mass communication behaviors of 200 old people in Oxford, Mississippi, supports this assertion. Also, Beyers and Woods (T963) found that 60% of their sample reported reading on the previous day for an average of one hour. Furthermore, it appears that reading activity may increase with advancing age. The same researchers report that read- ing is particularly high among their sample over 80. However, it should be cautioned that making inferences of the effects of aging from cross- Sectional data is dangerous, since results may be attributable to 23 generational or cohort factors, or to life-stage or developmental factors, and measurement at one point in time cannot separate these com- peting explanations. This criticism applies as well to most of the studies reviewed in this chapter. Meyershon (T96l) reports a T96l study conducted by the Television Bureau of Advertising of 3,000 households showing that persons over 50 read newspapers more than persons under age 50. With respect to newspaper content, Schramm and White (T949) report from cross-sectional data that with aging, there appears to be a trend toward reading newspapers less for entertainment and more for informa— tion and serious viewpoints on public affairs. There is a correspond— ing increase across age groups of the percentage of “news" read in the paper. More recent readership studies also support this view. Reading of books and magazines tends to decline with age from the teens, but indicates an upturn around age 60 (Opinion Research Corpora- tion, T957). However, earlier research by Lazarsfeld and Kendall (T948) shows that age—related declines in book reading occur only for college educated, while less educated old people show no substantial difference in book reading from that of younger people. Radio use has received little attention in the context of aging. Parker and Paisley (T966) in a study of l,9OO adults in two California communities report that radio use declines rather substantially from early'to late adulthood. However, with increasing age there is greater selection of infbrmational radio content than entertainment content. Movie use generally shows very large declines with advancing age. Meyerisohn (l96l) presents data from three years——T945, T947, l957-- 24 computed from national samples of Lazarsfeld (T948) and Opinion Research Corporation (l957), showing that roughly 50% of the youngest age groups (2l-29) report attending a movie at least once a week, while only about l5% of persons over 50 report attending at least once a week. In T957 this figure was at a low of 8% for those over 50. In another analysis, only T% of persons over 60 report attending a movie on the previous day, while roughly T5% of young people aged l5—l9 do. Media relationships to other activities. While the existing re- search on communication activities in aging has fallen largely into two discrete groups--interpersonal and mass communication--there have been a small number of studies that bridge both of these areas. An important question in communication theory, at a more general level than the particular context of aging, is "what relationships are there between interpersonal and mass communication?” Some have approached the ques- tion from the perspective that the primary relationship is a positive one, while others suggest that it is primarily negative. The former view is emphasized by Atkin (T972) in his explication of the communica- pory utility of mass communication--people use media because they pro- vide the basis for t0pics of interpersonal communication, hence media use is based on the anticipation of its social utility. The latter view--that media are a substitute for interpersonal communication, or a functional displacement-~15 held by other scholars. Horton and Wohl (T956) and Horton and Strauss (l957) view television use as a form of ”para-social interaction.” Robinson (T972) hints at the substitution relationship in his suggestion that television involves a time—trade-off with other social activities, implying that similar needs .._.~_ 25 are being fulfilled. A more direct explication is provided by Rosengren and Windahl (T972). They suggest that consumption of mass communication is a functional alternative to direct interaction, developing the con- struct of interaction potential as a predictor of the degree of involve- ment with media, the amount of media consumption, and the degree of reality proximity of media content. Interestingly, none of these researchers consider aging as an advantageous means of testing these propositions. Because important changes in activities typically occur in the sixth decade, the behavior of older people may yield valuable insights into these possible trade- offs or facilitative relationships between mediated and interpersonal communication. There have been a small number of researchers in the area of aging that have directly attended to the relationship of mediated and inter- personal communication. However, little of this research is published in widely accessible sources. Cassata (T967) appears to be the first researcher to directly focus on the interrelationship of mediated and interpersonal communication. In an unpublished dissertation, she reports research in the Age Center of New England (n=l77) showing that as the degree of retirement and the amount of free time increase, con- sumption of mass media generally increases. Also, self-perceived decreases in personal conversations and letter writing at age 45 are associated with increases in media consumption. She concludes that there is some support for a substitution hypothesis. Indirect evidence for substitution comes from research reported by Beyer and Woods (T963) finding that widowed and single older people, 26 particularly men, show higher television viewing rates than those living with others. However, research by Davis (T972) shows a slight negative relationship. Schramm (T969) proposes that mass communication provides the elderly with an important link to the environment in a period of life when more direct contact with others tends to decrease. However, he reports no research directly supporting this assertion. Meyersohn (l96l) also without direct data, suggests that tele- vision supplies the "illusion of contact” (p. 270) for old people, substituting for their exclusion from social experiences. At the same time he advances the facilitation function, similar to Atkin's views, that television also provides a stimulant for conversation and rapport between members of different generations who might otherwise find it difficult to communicate. Along this facilitation Tine, Davis (T972) finds that communication with children and with friends is positively related, although weakly, with amount of television viewing. Graney and Graney (T974) report results of a longitudinal panel study of a group of elderly persons (n=46) designed to test the propo- sition that decreases in social activities will be related to corree sponding increases in mediated communication. Their results show that the substitution relationship is most strong with the following pairs of activities: radio listening and frequency of participation in organized religious activity, frequency of visiting friends and rela- tives in the community and extent of radio listening, and frequency of telephone use and frequency of visiting friends and relatives in the community. Parallel, not inverse, relationships were found with 27 television viewing and organizational participation, and radio listening and organizational participation. Although these mixed results were obtained and correlates were weak, the authors conclude that there is support for a substitution hypothesis, at least within specific subsets of activities. Considering the evidence presented in this section, there is sup- port, although rather weak, for the substitution hypothesis, and there is also evidence for weak support of the facilitation hypothesis. In Part II we seek to clarify this relationship through explicating a general theory of the relationships between interpersonal and mediated communication use, and individual information processing characteristics. Part III presents research testing the theory. In general the theory may be viewed as taking a form of substitution approach in explaining the relationship of mediated to interpersonal communication use. First, the information theory of communication functions will be developed at a general level, applicable to all stages of the life-span. Then, the implications of the theory for aging will be drawn from the general level as the construct of informational aging is developed. Although the general theory must be explicated before notions of informational aging can be fully discussed, at this point some important characteristics of informational aging will be highlighted to provide an aging context for the theory to follow. Some key features are: T) informational aging is a curvilinear process occurring over the life—span, 2) informational aging involves changes in the complexity of inter- personal information space, mediated information space, and the individual internal information processing system, 28 3) a key determinant of the processes of informational aging is the perceived control of the individual over environment. Summary In this chapter we have surveyed prominent theories of aging in light of their implications for interpersonal and mediated communication behaviors. Then, research in sociology that can be recast in a communi- cation perspective was reviewed, dealing primarily with friendship patterns in aging. Following this, research in mass communication and aging was examined. Major topics were: content analyses of mediated communication, media exposure and content preferences of older people, exposure to other media such as books, newspapers, radio, and movies, and finally the interrelation of interpersonal and mass communication. Related research in this latter category was reviewed in light of two alternative approaches to their Tinkage—-facilitation and substitution. In the next chapter more general views of communication functions will be considered, both at a general level and in the area of mass communi- cation. CHAPTER III PREVIOUS VIEWS OF COMMUNICATION FUNCTIONS Overview This chapter will provide the historical context of the concern for the functions of communication. First, general explications of communication functions will be briefly reviewed, then mass communica- tion functions will be reviewed. This platform will provide the basis for the development of the information theory of communication func- tions in Part II. General Functions of Communication While the model developed here is general to both interpersonal and mass communication behaviors, most conceptualizations of the func- tions of communication focus exclusively on mass communication. There are only a relatively small number of authors who develop perspectives on communication functions that explicitly apply to interpersonal com- munication variables, and either explicitly or implicitly to mass com- munication variables. One approach to specifying functions of communication is to simply inventory them, in effect developing a "shopping list” of communication functions. Often these approaches are found in the less scientific conceptions of communication. For example, Hance, Ralph and Wiksell 29 3O (T969) refer to the purposes of communication as informing, persuading, and entertaining. The major problem with approaches that merely inventory communication functions is the lack of theoretic perspective. Hence, the contribution of these approaches to theory building is limited, and perhaps they serve to inhibit theory development. Berlo (T970) has proposed three general functions of communica- tion applicable to systems at all levels of analysis. Production com- munication involves the exchange of messages directed toward the achieve- ment of instrumental objectives. Maintenance communication involves the exchange of messages that are concerned with the intercomponent relationships in a system. Innovation communication involves the exchange of messages concerned with new alternatives for the behavior of system components. While an important feature of this view is its generalizability across system Tevels--from individuals through large organizations, and upward--it suffers from a lack of theoretic principles that bind these functions into a network of relationships with other intra— and inter—system variables. Mortensen (T972) takes Katz's (T970) conceptualization of the functions of attitudes and relates them to communication. Four functions are dealt with: adjustment, ego—defensive, value-expression, and knowl— edge. A major problem with this development is that the conditions under which these four functions operate are not systematically expli- cated, and the pattern of interrelationships among the functions are nOt specified. In relatively more comprehensive and potentially useful attempts, Mi ller (T966) and Miller and Steinberg (T974) suggest that peOple 3T communicate to control the environment. Along related lines, but from a different perspective Villard and Wipple (T974) propose that people communicate to negotiate and support self-identities. The utility of these models is contingent on future research. Their important value at this point is that they offer potential for theory building, unlike the "shopping list” approaches. Functions of Mass Communication Brief history. Attention to the functions of mass communication for users has been in the stream of social research for a considerable period. Among the earliest to approach media use from a "uses and gratifications" perspective were Cantril and Allport (T935). In the Psychology of Radio they viewed the medium as providing the listener with a link to the outside world, also conceptualizing all media in terms of the degree of social participation each provides. Among a larger body of later research, the following are significant. Herzog (T944) researched the uses and gratifications of radio “soap operas” for listeners. Berelson (T949) developed notions about the functions of newspapers by examining what happens during a newspaper strike. Riley and Riley (l95l) conclude that boys listen to action— adventure radio shows to give them ideas for group activities. Schramm (T949) and others after him, have used a four-fold typology of media functions that distinguishes “immediate” and ”deferred gratifications“ and “informational” and “entertainment” motives. Horton and Whol (T956) view television, particularly game and audience-participation shows, as the basis for ”para-social interaction,‘ supposedly providing the same 32 benefits as the ”real thing." Katz and Foulkes (T962) discuss in detail television as an escape mechanism. Many other studies from radio in the thirties through television in the seventies, while not directly concerned with the uses and grati- fications of mass communication, offer evidence that may be recast in this light. However, a more direct concerted effort at furthering knowledge about uses and gratifications of mediated communication has taken shape in this decade. While earlier research in this vein was strictly American in origin, this more recent work has been advanced with a major thrust from European researchers. Useful reviews and critiques may be found in Katz, Blumler, and Gurevitch (T974) and Rosengren (T972). Therefore, we will briefly review selected works within the context of general theoretic and methodological approaches, attempting to note where appropriate recent research not covered in these reviews. First, theoretic perspectives and implications will be discussed. Theoretic perspectives. Earlier work on media functions typically attempted to deal with a single medium, such as radio, newspapers, or even television. More recent work examines the functions of media within the larger life space of the individual. Thus, we see a concern for: l) explaining the use of all media, rather than a limited subset. 2) the general psychological functioning of the individual, par— ticularly in the area of basic human needs (Katz, Gurevitch, and Haas, T972), 3) social contexts in which individuals are imbedded, and the social phenomena that impinge on media uses and gratifications (Rosengren and Windahl, T972; Atkin, T972). 33 Perhaps as a result of this increased scope, an increase in the scope of mass communication theory will more rapidly develop. Implicitly or explicitly, most approaches are concerned with the gratifications of pe§g§_that the individual obtains through using media. Hence, man is viewed as a teleological system, engaging in goal-seeking behaviors. Earlier work focused primarily on delineating alternative gratifications, while more recent developments attempt to determine the underlying needs within which gratifications are rooted. Other work (Graney and Graney,l974) assumes both a given type of gratification and need performed by one kind of behavior, and looks for a systematic relationship with another type of behavior that would suggest substitu- tion or functional equivalence. While gratifications and needs are still essential to this approach, the focus is on relationships among sets of behaviors, and not their underlying causes. Within the general needs approach to the functions of mass com- munication, a most critical question somewhere down the line must be “how do these needs develop?” This is a particularly sticky ghestion since as Nordenstreng (T970) discusses and Katz et_al, (T974) echo, the media themselves may have an important role in structuring the needs that they supposedly satisfy. This kind of perspective would re-orient this line of theory and research somewhere between the view that media act on people, and the counter-balancing view that it is the audience that rather independently acts on media. Methodological gpproaches. In the functions and gratifications earea there appear to be two distinct strategies, and one hybrid strategy, for'empirically answering the question of why people use mediated 34 information. One strategy is to ask people directly why they use media and categorize these responses. These categories then become the theoretic functions of communication used in further research. Blumler, Brown, and McQuail (T970) use this approach in a study of British media users in an attempt to link television gratifications with the larger social environment. Greenberg (l974) also used this respondent-generated approach in determining the functions that television serves for British children. The derived functions are then correlated with other variables such as media exposure, perceived reality, attitudes toward violence, and so on. Another approach is to avoid asking people to report directly why they use media, but to develop some indirect means of assessing these reasons. For example, Graney and Graney (l974) attempted to determine if elderly used media as a functional displacement for interpersonal communication through a longitudinal study of the relationships between media use and wide range of other social activities. Respondents were pp§_asked if media did indeed serve as a substitute; the relationship was indirectly assessed by examining the change in the relationships of the behavioral variables over time. Olsen (T960) examined relationships between residential mobility and movie attendance without directly interviewing respondents, but through the analysis of archival data. Another variant on this theme is provided by Robinson (T972) who draws evidence for three functions of television use through a secondary analysis of existing survey data. Both approaches have limitations. The approach that asks people lNhy media are used, and stops there, suffers from lack of the capability 35 of developing abstract theory. This approach offers no underlying explanatory principle, but is merely a descriptive inventory of uses as perceived by media consumers. Theory emerging from this approach can only be as sophisticated as the cognitive system of the median individual in the sample, unless the category systems are modified, or by chance, more abstract theorizing happens to be congruent with respondent gener— ated gratifications. The second approach may be criticized for the inability of the design to allow people being studied to consciously validate or invalidate the researcher's theoretic formulations, and also for the inability to provide hueristic input into the scientist's in— formation system, perhaps Teading to new theoretic insights. An approach falling between these two uses elements of both. The scientist develops a framework relatively independent of the cognitions of those being studied, but then translates this framework into vari- ables that then allow those being studied to respond directly to the derived functions, evaluating how well they apply to them. While it is possible that this intermediate strategy may suffer from an aggregation of the problems of both of the other two, it i§_likely that greater benefits will accrue. Lyle and Hoffman (l97l) take this approach in researching the uses that children make of television. Katz, Gurevitch and Haas (l973), in their study of the relative need satisfaction of Israelis by the various media, surveyed previous literature, used their own insights into the Israeli media environment, and phen_pilot tested in developing 35 needs statements that were subsequently related to media use. 36 This is also the general approach used in the research to be reported in Part III. A theoretic rationale for functions of communi- cation is developed independent of direct consultation with those being researched. Then, these functions are translated into linguistic statements that are intended to represent them at a lower level of abstraction, more within the everyday standardized usage of the popula— tion under consideration. Finally, respondents evaluate how well each statement then applies to them. Figure T pictorially represents these three methodological approaches. 'Functions E Functions ! Respondent Researcher g Generated Generated ' .Only TOnly l \ / A it 4,? ' Functions ; Researcher Generated‘ and Respondent Validated Figure l. Methodological approaches to functions of communication. Summary Having reviewed previous developments in the uses and gratifica- tions of communication at the general level, and also in the context of mass communication, a foundation now exists for the deveTOpment of a comprehensive information theory of the functions of both mediated and interpersonal communication. Part II will concentrate on this development. PART II AN INFORMATION THEORY OF COMMUNICATION FUNCTIONS FOREWORD Having established the context for the information theory of communication functions in Part I with a discussion of two general themes-—aging and communication, and general prior conceptualizations of communication functions——in Part II the theory is fully developed. Chapter IV develops a definitional context for the model and provides a general discussion of information space. Chapter V explicates two key components of information space, mediated and interpersonal communica— tion systems, focusing on the entropy of the respective components. Chapter V1 is concerned with distinguishing subsets of information space from the perspective of the individual, defining boundaries that separate environment, ego, and cognitive space. The key operating principle of the model, the principle of optimal information, deriving from a balance logic, is developed in Chapter VII. This chapter also explicates the corresponding four functions of com- munication, discusses the construct of control over environment as an important variable in the model, and accounts for the attributions the individual makes about his own behavior with regard to interpersonal and mediated communication. Chapter VIII draws from the previous more general theoretic developments in explicating the construct of informa- tional aging. 37 CHAPTER IV INFORMATION SPACE: DEFINING THE BASIC CONTEXT OF THE THEORY Overview The general objective of the theory developed in Chapters I through VIII is to account for the manner in which individuals process information within information space. An information processing per- spective is an eclectic one. It draws upon theory and research in psychology, social psychology, sociology, and the physical sciences, particularly in the area of communication technology. Out of these varied areas a thread of common conceptualizing may be observed. The units of analysis, whatever they may be, are viewed as systems inter- acting within an environment. These interaction processes take form and content through the processing of information. Information input is received from the environment, transduced or changed in some way, and may then be stored by the system or provide the basis for some output to the environment. This focus on information offers the potential for increased integration of scientific work in these areas into a more coherent overall frame- work. As an information processing perspective is an eclectic one, so too is the model of the functions of communication that will be developed. 38 39 It will draw on work in the areas of mass communication, cognitive psychology, social psychology, and communication engineering, among others. As such, the model will provide a theoretic rationale to explain a relatively large range of human communication behaviors. The major groups of variables and relationships to be integrated in the model appear in Figure 2, on the following page. They are individual information processing characteristics, control over environment, mediated communication, interpersonal communication, and attributions of communication functions to mediated and interpersonal communications. As the theory is developed it will be useful to overlay another organizing framework on information space, related to the above, but not completely overlapping with it. Their relationships will become clear in the discussions to follow. To provide an overview at this point, the three levels of information space are presented in Figure 3, on page 4T. The overall set of spheres is the global individual informa- tional space, and each of the three smaller spheres is a subset of the larger one containing it. The boundaries that distinguish environment, ego, and cognitive space will be explicated within the development of Chapter VI. Before moving into the operation of the components of the model, attention will be given in this chapter to the development of a defini— tional context for the concept of information space. Following a general orientation to this notion, the constructs of information and uncertainty will be defined. Next is a discussion of how the individual creates lnl:Ormation and uncertainty in information space through the chunking of exPEE‘rience, and the linking of these chunks into patterns of interrela- tion ships. 40 INDIVIDUAL INFORMATION PROCESSING Uncertainty Preference (UP) l r—a- Input (1)] ——a> I : UP I . l : if I < UP if I > UP T if I = UP I ‘ l ' . I . . I Create Maintain Create | .__ greater _____ entropy less .3 entropy entropy I T T T T ' l s T T I I . L, I J I Attribute l Attribute I Unc. Exp. I Unc. Red. Function I Function 4 /\ \\ I /’ T \\ 1 / CONTROL OVER ENVIRONMENT /' \ if Low/ \,jf High MEDIATED COMMUNICATION Figure 2. Basic components of the model. <%——-Referent function ~e—————— Comparator function Environmental change function INTERPERSONAL COMMUNICATION 4T Environmental space .y—Ego space Cognitive space f A/ / V Figure 3. Levels of information space. Information Space Throughout the development of the theory of communication func- tions the concept of information space will enter into the explication. What is the concept of information space? Information space is the space that we as people live in. It is anchored not in some separate world "out there,” but in human consciousness. Thus, it is a space defined relative to the awareness of the individual, although it does not exist independent of social processes. Information space takes form and content from the structure and functioning of the individual information processing system, forming experience out of energy and matter. The projection of this basic logic 42 of experiencing into an environment creates information space, with which the individual intimately interacts. In a similar sense Berlo (T967) explicates this notion for the scientific experience, suggesting that theory is a relativistic projection of created relationships onto nature. Why use the concept of §pgge? For thousands of years information was thought to have spatial characteristics. The Vedas of India con- ceptualized thought in this manner roughly 2000 B.C. (Coster, T934). In current thinking the notion of spatial relationships becomes clear when information is defined as pattern among a set of elements, for to have elements and to have pattern there must be a space within which they are located. In the physical sphere, objects provide a reference system of their relative location to one another, creating a space. That is, the objects or physical elements create the space of which they are a part. Analogously, information creates the space of which it is a part. It forms a reference system for the relative location of pat- tern densities to one another. These areas of higher density may be defined as elements in a discrete sense, although it may be more appro— priate to view them more continuously as areas where the ratio of within pattern strength to between pattern strength is high. Similarly, although we treat physical objects as discrete entities, theory in physics suggests that they are concentrations of energy in which simul- taneous and infinite wave functions interact in such a way as to give the appearance to us of discrete ”objectness”. In the context of the present theoretic development, information space is the sum total of an individual's creation of pattern among 43 elements. This notion is more fully elaborated in the next section a the construct of information is explicated. Defining_Information Traditional definition. In the tradition of Western thought before the development of communication engineering, the concept of information has been treated synonomously with "knowledge." As such, information has been intimately related to substantive content. Beca of this content-bound nature of information the concept traditionally has had little theoretic utility in communication science. If a conc cannot be effectively abstracted beyond the particular spatial and temporal characteristics of the contexts in which it appears, then it is not possible to develop theory using the construct that itself is transcendent of particular contexts. A content—bound definition of information inhibits the abstraction of the concept beyond the conten that sets the context for its definition. Information as form. Another way to conceptualize information to treat it as a form—bound rather than content—bound concept. Becac form lies within the structure of content--the relationships among cc tent elements-—and is separable from specific content, it is a consic ably more abstract and generalizable concept. Hence, this characteri tion of information ought to have more potential utility in the pro- cesses of theory building. A form—bound view of information derives from the conceptualize tions of Shannon and Weaver (T949). In this development information defined as the ”reduction of uncertainty.” What then is uncertaintyf 44 Uncertainty is closely related to the concept of entropy in phys Uncertainty is lack of order, lack of predictability, or lack of ity. Given some possible set of alternatives for the occurrence event, uncertainty exists to the extent that which alternative w occur cannot be predicted in advance beyond chance probabilities some of this uncertainty is reduced and the probability of predii the occurrence of an alternative is greater than chance, then in' tion exists in this system. There is less than randomness, less maximal disorder in the system. In the general case, information will be defined as the pe of pattern among a set of elements. When there is pattern among ments, there is some degree of order, predictability, and structl less than maximal uncertainty. Without pattern, that is without information, a set of elements is in complete disorder and randOI This concern with pattern is then a concern with the form and st' of human experience, with content playing a relatively minor rol. the overall perspective. Information, Content, and Context The notions of form and content are very often difficult 6 separate. Within a given context, however, this difficulty does arise. The elements that are related to one another constitute content aspects of the context; the particular relational pattem the elements is form. However, when a particular element is exm a Systemic view suggests that it is itself composed of sub-eleme linked together in a pattern of interrelationships. Therefore, 45 another level each element has both form and content aspects, and might go up and down levels of abstraction shifting the context a ing phenomena in this way. This does not suggest, however, that tion is content-bound, for the nature of the structure of the set elements transcends the particular elemental context. Structure measured independent of the characteristics of the elements thems This is not the case with the traditional view of information, wh there is pnly_content boundedness and no notion of an underlying within content. In attempting to usefully measure aspects of content and fc relating them to other concepts, the scientist must be careful to the elemental context at an appropriate level of abstraction. Th particularly critical when content-free metrics of form are used. Along with the greater Operational freedom provided, comes the in of increased attention to defining precisely the levels at which measures will be applied. This operational flexibility necessita greater conceptual precision. Distinguishing information and uncertainty. The concepts c certainty and information have produced considerable confusion ir literature. At times the concepts are not distinguished from eac and are used interchangeably, and at other times one of the conce emphasized while the other is nearly ignored. The original expl' by Shannon and Weaver defined information as the reduction of unc and nothing more. Uncertainty may then be considered potential tion. Thus, the distinguishing characteristic is time. Given a éalternatives for the occurrence of some “event,“ the only distin< 46 between potential information or uncertainty and actual informati the point in time that one alternative in the set is selected, an transmitted or occurs. Information is then processed and uncerta reduced. Creating Uncertainty and Information The original formulation of information theory was rooted i context of communication hardware systems. For this technology t develop it was necessary to make the assumption of a closed syste logics. This is exemplified in the notion of conservation of ene If, for example, we assume that energy is never created or destrc then all the alternative energy states of components in the syste known once the input energy level is known and the structure of t ponents is known. Working within this context, when Shannon and undertook the development of a logical and mathematical system tc account for the transmission of pattern, it was natural to assume completely known set of patterns which could be potentially trans Perhaps as a result, there is little concern in the original deve for explicating the processes by which uncertainty comes about. In the remaining sections of this chapter the creation of l tainty and its reduction through the establishing of patterns wi' discussed. First, the chunking of experience will be examined, ' the linking of these chunks into larger wholes. Creating Chunks Through socialization processes individuals are taught how UP awareness into units or categories of experience (Bruner, T95 47 These units, categories, or chunks of experience then become bas elements of experiencing the environment. Initially in a child' development, there is likely to be considerable error in the man which awareness is chunked, if error is defined relative to the behaviors of those who have already learned the rules. As the i further develops,the boundaries of these chunks become more clea defined. The use of particular chunks is reinforced by others, uses of other chunks are not, and as such the information proces system begins to take on structure. By the time the individual has matured, the units of exper are often so heavily engrained that the environment is experienc an objective independent phenomenon. The person observes others similar frameworks in experiencing, and while the individual may others to have different responses to experiential units, he/she assumes that the stimuli are constant for all. In fact we might late that in addition to these indirect factors Western culture socializes individuals to view the world as reified and indepenc observer, with its emphasis on discoveryrather than creation ir development of science and technology. While the greatest portion of the chunking of experience ‘ units is likely to be the result of learning to experience thr0l others first defining the categories of experience, there also 2 to be physiological limitations that set the boundary parameters eXperience. The structures of the sensing organs and the nervo: are relatively constant across people, and while there are dlStl (TF these characteristics, we may view the central tendencies of 48 distributions as the approximate boundaries to the chunking proces an aggregate social level. Physiological structures then provide level of commonality in the chunking of experience, but within the boundaries there is great latitude for the creation of chunkings 0 wide variety. In this sense, the information space may be conside finite, but boundless. Linking Chunks Together The chunks of experience may be considered the substance of experience, while the manner in which the chunks become arranged i patterns of interrelationships may be viewed as the structure or s of human experience. The process of linking chunks together is th process of creating this structure. Elements of experience are co into larger wholes through the establishing of relationships among experiential chunks. Through linking chunks of experience togethe particular ways, the individual creates order and systematization ceiving. These linking patterns become more predictable as they d through use, and the individual develops generalizable rules for T chunks. This process of integrating and synthesizing experience i more meaningful wholes is the processing of information. A critical component of information processing is formed by attributions the individual makes about the causal relationships 5 experiental chunks. These attributions form the basis for immedie response, and for the long range planning of information processir Within a context. In latter chapters detailed attention will be 9 t0 attributions the individual makes about communication behaviors 49 Also, the concepts of cognitive complexity and subjective and uncertainty are intimately related to this notion of linking c These constructs will be discussed in Chapter VI. Individual and Social Factors In sum, the creation of experiential chunks and the link these chunks together are determined both by individual and sc factors. Consensus and rules (Cushman and Whiting, T972) amor of individuals set the overall parameters for experience. Inc differences come into play as variation around these rules. T experience is neither completely common at a social level, nor idiosyncratic at an individual level. Human experience involv factors in dynamic interplay. Over time some of these once it Operations become linked into networks of individuals, forming rules with new variances. Summary This chapter has set the global definitional context for information theory of communication functions by defining infr space, information and uncertainty, and the processes by whicl created. The next chapter differentiates information space ll and interpersonal components, discussing means of describing ‘ turing of these components. CHAPTER V MEDIATED AND INTERPERSONAL COMPONENTS OF INFORMATION SPACE Overview This chapter is concerned with explicating two primary components of information space: mediated and interpersonal communication. A third component of information space--the direct experience of environment, without the exchange of messages with others either verbal or non-verbal or either directly or through media--is of little importance in communi— cation theory and research and will not be treated in this development. Following a discussion of the two primary communication components, a framework for conceptualizing the structure of these components of information space will be developed. An information theory perspective is taken in explicating channel, content, and form entropy within the two respective components. Mediated and Interpersonal Components In the previous discussions of information space, it has been treated as relatively undifferentiated. One division is developed here in terms of mediated and interpersonal communication components of information space. Later, information space will be differentiated into environmental, ego, and cognitive space. 50 5T Interpersonal Component Communication research overwhelmingly shows that interpersonal communication exchanges have great importance in influencing individual behaviors and cognitions. Correspondingly, an important tradition in social research has been to examine the interpersonal communication net- works in which individuals are embedded (Festinger, Schacter, and Back, T950; Laumann, T973; see Collins and Raven, l969 for a review of group network research; see Richards and Lindsey, T974 for a review of net— work analysis). Even in the area of aging, in which social science research has been rather limited, there are studies concerned with the interpersonal networks of old people (referenced earlier). This social network perspective on individual human communication may be termed a structural approach. Most of this research, however, has been structural only in the broadest and weakest sense of the term. Rather than examining the precise structural configurations of networks in which people are embedded, a great deal of research has been concerned with structure only in an aggregate manner. That is, communication "network" variables have been operationalized in ways that preclude the possibilities of examining patterns of interrelationships. For example, many researchers have operationalized "network” variables by procedures such as counting the number of people a given individual talks with about a particular t0pic, or cumulating the total amount of time a given individual communicates with others, regardless of structure. Even more imprecise, the researcher may simply ask whether or not a person communi- cated with other people about a particular topic, as studies of news diffusion often have done. It is unlikely that these gross measures of 52 interpersonal communication structures alone will enable the development of advanced theories of human communication at an individual level. In brief, it is proposed that a network approach to interpersonal communication provides a useful framework within which to organize the study of this component of human communication. At a macro level, the precise structural configurations in which individuals are embedded may be examined. At an intermediate structural level, the content of inter- personal communication may be examined. And at a micro level, the form of particular dyadic communication relationships may be examined. Mediated Component Mediated information is another major source of information for the individual. However, when pitted against interpersonal communica- tion in the measurement of relative impact, it has generally played a secondary role for most kinds of samples (for a direct example see Woelfel, McPhail, and Gillam, in press). Nevertheless, research has shown mediated communication to be an important influence on cognition and behavior (a current example comes from the research on televised violence and aggression reported in Rubenstein et_gl,, T972). This is particularly the case when there are high degrees of mediated message repetition, or when the persons measured have little direct experience in an area where impact is being assessed, but have a relatively high amount of mediated experience (Greenberg, l97l). In the next section a general framework for examining the struc- ture of both mediated and interpersonal communication at three levels of abstraction is developed. The basic notions for the conceptualizations 53 derive from an information theory perspective, hence the entropy within these three levels is of importance. Communication Structure Communication relationships may be viewed from three levels of structure: channel, content, and form. Channels refer to the relative- ly discrete linkages that bind systems into communication relationships. In mass communication,channels refer to the standard media of radio, television, newspapers, etc. In interpersonal communication, channels refer to the overall relationships that link pairs of individuals to- gether through the processes of communicating. Thus, channels or links form the basis for the communication networks that link elements of social systems together. Content refers to the messages that are exchanged within mediated and interpersonal channels. As such, content has topical or semantic characteristics. In other words, content refers to what it is that is being communicated about. Form within content deals with the syn— tactical characteristics of particular messages: how the message is organized, or how the subcomponents of the message relate to one another in some pattern or structure. As Figure 4 (on the following page) illustrates, all of these three levels are intimately intertwined. Channels are defined by the particu— lar ways that messages or content flows among a set of elements. Channels might be viewed as the “trace” of the repeated flow of specific messages among systems over some period of time. Likewise, messages are defined by their form characteristics. While content may be held Content Entropy Form Entropy LL Figure 4. Three levels of communication structure. constant as form is allowed to vary, thus still not recognizably chang- ing content, there are limits to how much form can be changed before the manifest content of communication changes. Form may then be viewed as aggregations of micro-forms, with varying levels of abstraction of form within the overall message structure. In sum, channels are aggre- gations of message or content effects, content is aggregated form, form is aggregated micro—form, down to the lower limit of human pattern recognition. A fourth group of communication variables is somewhat orthogonal to the other three groups. These variables index amount of communication, 55 measuring how much communication is occurring at various levels. Amount measures may index the frequency with which particular channels are used, the frequency with which types of messages are exchanged within a particular channel, the frequency with which form characteristics Of messages are processed, etc. Or, these variables may index the duration Of channel use, message type exchange, form type exchange, etc. Aside from their utility as measures that might stand alone in relationships to other kinds of variables, amount variables may be used in conjunc- tion with the other three groups of levels variables in the formation Of hybrid variables that quantify the variance, predictability, or structure of communication relationships. In the next section these kind of meta-variables, or "variance Of variables,” will be generally explicated. At each of the three levels Of structure a corresponding entropy variable will be discussed. First however, measurement char- acteristics generalizable across the three variables will be considered. Measurement. The definitions of the three constructs are closely linked to the Operations of basic information theory. Common to each is a notion Of alternatives within a particular level of analysis and relative probability of the occurrence of the alternatives. Thus, each may be operationalized by a basic information theory formula: : ‘2 pi '092 pi r log2 N where the relative entropy in the set Of elements is represented by Hr; pi is the relative probability of the occurrence of an alternative; N is the number of alternatives; and log2 is the basic function calculat- ing the number Of bits, where one bit is the reduction Of uncertainty by 56 one half. This kind of measure determines the amount of entropy in the set relative to the amount of entropy that is maximally possible, i.e., if each alternative were equally likely. Thus, it is a ratio of Observed entropy to potential entropy. Locus of definition. Cutting across the three measures of channel, content, and form entropy is the locus of definition of these measures. In the general case these measures will be divided into "objective” and ”subjective" measures. In an earlier paper the author discusses a sub- set of these, source and receiver defined entropy, within the context of mass communication but not interpersonal communication (Danowski, l974a). Objective and subjective methods of measuring entropy may be briefly defined as follows: objective entropy is defined through data gathered by an Observer of someone's communication behavior, or by measuring the entropy of messages in the environment of a person and assuming this is what is perceived by the individual. Subjective methods Of measuring entropy involve asking the individual to directly report on perceptions of patterns in information space. _In the follow- ing discussion of entropy measures, general reference to locus of definition will be made. Channel entrgpy. Channel entropy varies as a function Of the number of channels used by a person and the relative probability of their usage within some standard unit of time. As the number of channels used increases, and/or the relative probability of their usage becomes more equally distributed, the greater the channel entropy. Sources of data necessary for calculating this measure may be either observing the behaviors of the person, or asking the individual, or others familiar 57 with his or her behavior to report on them. Content entropy. One type of content entropy varies as a function Of the number of alternative message types used in communicating, and/or the relative probability of their usage. As either or both increases, content entropy increases. This kind of data may also be gathered through observation or self— or other-reports. Another type of content entropy is a function of the individual's ability to predict sequences Of message types, characteristics, or sub- message components that will occur in some later portion of a communica- tion experience, or in some distinct future communication experience. Since this kind of content entropy is a subjective one, data may be gathered only by directly asking the respondent for a judgment. Form entropy. Form entropy varies as a function Of the number of alternative message substructures, and the relative probability of their occurrence. As the number of substructure alternatives increase, and as their occurrence becomes more equiprobable, form entropy increases. Form entropy measures are typically below the level of awareness of the communicator. Therefore, they cannot be assessed by directly asking the respondent for a judgment. Currently, they must be indexed through systematized ”content" analysis techniques, by trained coders or pro- grammed coding machines. Examples of Interpersonal Component Entropy In discussing measures of interpersonal communication, the con- struct of zone will be used. This concept has previously been expli- cated by Barnes (T972) in a somewhat different manner. It has been 58 adapted for use here because it is consistent with spatial notions developed in earlier sections. A zone is defined through the structural arrangements of communi- cation channels, and is constituted by the systems that are linked through message exchange to a particular focal system. A first-order zone is defined by the set of systems that are directly linked to a focal system, and the communication channels linking these systems to the particular system are one step links. Correspondingly, a second- order zone is formed by those systems that are linked to the one-step systems, but not directly linked to the focal system. Thus, information must travel through two channels to flow between the focal system and systems in the second-order zone. To simplify the discussion in the following section, first order zones are implied when the unqualified term "zone“ is being used, and although the variables are discussed in the context of individuals, the logic is general to other levels Of analysis. Zone size is a way to measure channel entropy. It increases as the number Of people linked to a focal person increases. Zone size might be discussed in the general case with reference to all persons communicating with an individual, or it might be discussed with refer- ence to more narrowly defined, discrete subsets of people, such as those in the occupational setting, or the family setting, etc. Interpersonal channel entropy increases as zone size increases, since there are more alternative channels through which information can flow. Zone integration indexes the extent that the persons to whom a particular individual is linked are linked to each other. The greater 59 the zone integration, the greater the extent that the first order zone is interconnected among the component individuals. Zone integration provides an indirect means of indexing content entropy, which increases as the number Of alternative topic categories becomes larger and as the relative probability of the use of these alternatives becomes more equal. This relationship between integration and content entropy is likely to exist because the greater the integration Of the individual's zone, the greater the likelihood that homogeneous messages will be exchanged among the members. Also, fewer alternative message types will be exchanged, and the individual will be more likely to predict the sequencing and content of messages over time. This proposition is sup- ported by research Of Schacter (l95l) and Back (l95l) in work concerning group uniformity. Form entropy in the interpersonal context is linked to the dyadic relational patterns that emerge through communication transactions. The works of Mark (l97l), Ericson (l973), Millar (T973), and Rogers and Farace (in press) suggest that a highly important aspect of relational patterns is the control dimension. In this context form entropy is a function of the variability' in relational control configurations. The greater the number of alternative relational control configurations used in an interaction sequence, and the more equal the probabilities that these patterns are used, the greater the form entropy of interpersonal communication relationships. Form entropy is not a direct zone variable. It operates at a lower order of abstraction, defined within specific dyadic subcomponents 60 of the zone. Nevertheless, an aggregate form measure at the zone level could be developed by computing the average form entropy across all zone subcomponents. The most common measure of interpersonal communication network characteristics is gmpgpr_of communication. Interestingly, this parallels media exposure measures--the most widely used variables in that area of research. Amount Of communication is not a direct structural variable, since it aggregates message exchange rates without regard for the speci- fic flow patterns. Amount may be viewed as a measure of system Openness (Katzman, T970). Examples of Mediated Component Entropy Simple exposure measures have been the mainstay Of mass communica- tion research. More recently, content measures in relationship to the behaviors of media consumers have become important, as witnessed in the television violence and effects area. And in addition, the uses and gratifications area has given attention to the functions that different kinds of content serves. The area of conceptualization and operationali- zation in mass communication that has received the least attention has been the specific structure of media use—-patterns and variances in exposure and use of mediated information. Although overall the specific attention to detailed structures Of media use has been minimal, in recent years applications of information theory to mass communication have constituted the major attention to mediated communication structure. These applications will be briefly reviewed with attention to the use of information theory and related theoretic metrics. 6T Channel entropy was measured in research by the author in relation- ships to interpersonal communication network integration (Danowski, T974a). The measure was formed by considering each mass communication medium as an alternative channel (television, radio, newspapers, maga- zines, etc.), then calculating the ratio of amount of exposure to each medium to total media exposure, and subsequently entering these data into the general relative entropy measure noted earlier. Content entropy was measured by Hanneman (T970) in an experiment designed to examine relationships among television message uncertainty and physiological and self-report measures of arousal and aggression. Content uncertainty was not measured by a formal information theoretic measure, but by deleting scenesfrom prime-time programs, thus creating uncertainty about the direction the plot would take. Judges then tried to predict what would occur under different deletion conditions, and subjects rated message uncertainty as a manipulation check. The author used another type of content entropy measure in sec- ondary analysis of data concerning television exposure and aggression for adolescents (Danowski, l974b). Alternative categories of television content types were the basis for the measure, and the relative exposure to each type was determined. The values then entered the relative entropy formula described earlier. Following the development of the original researchers (McLeod pt_gl:, T972), the content alternatives of crime, westerns, adventure—drama, comedy-variety, situation comedy, and game shows were used. Form entropy has been developed by Watt and Krull (T975) who created six measures of television message form which they have 62 subsequently used in relationships with aggression Of adolescents: set time entropy, set incidence entropy, verbal time entropy, verbal incidence entropy, set constraint entropy, and non—verbal dependence entropy. A factor analysis of these measures based on a sample of l58 prime time television shows from July, l97l yielded two factors: ”unfamiliarity", composed of the set constraint and non-verbal dependence entropy measures and ”dynamics”, composed of verbal time, verbal inci- dence, and set time entropy. Wartella and Ettema (T975) use the same measures of Watt and Krull to quantify the entropy of television commercials. A factor analysis performed on the coding Of 40 commercials resulted in the emergence of two factors. The first factor is loaded upon heavily by the visual indicators——"visual uncertainty"—-while the verbal auditory indicators load heavily on the second factor-—“auditory uncertainty.” Non-verbal dependence entropy loads evenly on the two factors. Summary In this chapter we have separated information space into two components--mediated and interpersonal-~and have explicated the entropy of channel, content, and form within each component. In the following chapter another division of information space is developed, separating it into environmental, ego, and cognitive space. CHAPTER VI DEFINING BOUNDARIES: ENVIRONMENTAL, EGO, AND COGNITIVE SPACE Overview In previous chapters information space has been carved into medi- ated and interpersonal components. In this chapter another division is proposed that is relatively orthogonal to this structure. Environ- mental, ego, and cognitive space within information space are dis— tinguished. Control over environment separates ego space from environmental space; freedom from time and location constraints separates cognitive space from ego space. In focusing on cognitive space, two types of uncertainty are distinguished: subjective cognitive uncertainty, and attributed cogni- tive uncertainty. Both are related to the cognitive complexity of the individual. The chapter closes with a discussion of uncertainty preference. Boundaries The concept Of information space is not tied to a particular level of analysis. As the level Of abstraction Of the system being examined shifts, so too shifts the level of abstraction Of information space. For example, a cultural, organizational, small group, dyadic, or individual information space may be defined. Since in this explication 63 64 the individual level of analysis is used, the individual information space is focused upon. However, it is important to differentiate aspects Of information space, and not treat it as a unitary entity. The previous chapter proposed a major differentiation into mediated and interpersonal components. As will be apparent later, in order to develop theoretic relationships about the ways individuals operate in information space, in addition the individual ego must be separated from the larger space, and the cognitive space separated from the ego space. Ego space. The distinction to be drawn here will separate ego from environment, the boundaries of which are the boundaries of overall information space. A boundary zone will be distinguished defining the ego as a subsystem of larger information space, and the dimension used to distinguish ego from the larger space will be control (Rotter, T966; Lefcourt, T966; Smith, T970; Mischell, Zeiss and Zeiss, T974). The concern will be not with some "Objective“ measure of control, but with control pprceived py the individual. A boundary of information space distinguishing environmental from ego space will be defined as a point at which the individual perceives himself to have greater than 50% control over components in the informa- tion space. Those components within this area over which the individual has greater than 50% control will constitute the ego space. Those areas over which the individual perceives less than 50% control will be the environment within information space. Since individuals will vary in their perceptions of control over components Of information space, 65 people will distribute according to how large their ego space is rela- tive to the size of information space. Boundary as zone. Boundary has been discussed to this point as discrete. It is perhaps more appropriate to view boundary as gppe, rather than a discrete line. A basis for determining boundary zone is the variance in the individual's perception of control over components. Over time, a person is not likely to have a completely stable assessment of control. For example, at one point in time the individual may per- ceive his control over some component of information space as 53%, at another time 45%, at another time as 5l%, etc. Over time, then, the line of greater than 50% control will shift; the region in which this line varies will be the boundary_zone. It should be noted that although we have used very precise per- centages in this illustration, it is not suggested that individuals actually conceptualize raw percentages in their feelings of control over the environment. We do suggest, however, that individuals at least make dichotomous discriminations: "this is under my control ... that is not under my control." Others may perceive degrees of control with varying precision. Nevertheless, whether people actually think in terms of ratios is not important, for the value of using this metric in the explication is to foster clarity and potential theoretic precision. Most individuals asked to make these kindsof ratio judgments will prob- ably do it quite validly. 66 Cognitive Space To this point, two subcomponents of information space have been discussed: environmental space and ego space. It will be remembered that ego space has been distinguished from the larger information space, leading to the discrimination of environmental space. In this section ego space will be further subdivided. The concept of cognitive space will be distinguished from ego space. While environmental space and ego space are defined through imme- diate relationships of the individual to other systems, hence being largely time and space bound, cognitive space is relatively time inde- pendent and immediately unbounded by spaces shared with others. In short, cognitive_space is defined by the processes of thought, that mpy occur partially or nearly completelyrwithout reference to the nature of immediate contact with the environment. This information/uncertainty may be retrieved from “storage“ or memory and simply reminisced, or it may be reformated into new elements and linking relationships, pro- jected into the past, present, or future. In the remaining sections of this chapter the focus will be upon explicating aspects Of this cognitive space. The constructs of subjec- tive and attributive cognitive uncertainty will be linked to the con- struct of cognitive complexity, and the correspondence between ego space and cognitive space will be discussed. Cognitive Uncertainty The concept Of cognitive uncertainty is difficult to precisely explicate. It has seen usage in everyday language, which perhaps itself 67 impedes the development of theoretic explication. It has seen a variety of uses in social science theory and research. However, its development in communication theory has been rather limited. Atkin (T972) proposes that information is selected from mediated sources tO reduce cognitive uncertainty. However, no conceptual or operational definition of cognitive uncertainty is Offered. This is indicative of the difficulty that is faced in explication. In this section some progress will be made toward these ends. In everyday usage uncertainty typically refers to not being ”sure” about something, not ”knowing,' or having incomplete knowledge or having rival or competing explanations for something in a particular context and not knowing “which is right." This general kind of uncertainty, viewed from the phenomenology of the individual will be labelled subjec- tive (S) cognitive uncertainty. §_uncertainty is the extent to which the individual is confident that a particular configuration of alterna- tive matter, energy, or information states is appropriate in some particular context. §_uncertainty will increase as the number of altern- ative configurations under consideration increases, and as their prob- abilities Of occurrence become more equal. These configurations may occur in three general areas: l) the environment within information space, 2) the ego space within information space, and 3) the relation- ship between components of ego space and components of the environment. Time is another dimension for organizing the construct of §_cog- nitive uncertainty. Uncertainty may be rooted primarily in the past, in the present, or the future. With respect to the past, the individual may be uncertain whether behaviors in a past situation were appropriate, 68 or whether the characterization or abstraction of the context in memory is accurate. Or, in the general sense there might be past uncertainty about aspects of the three major components specified earlier. Present uncertainty arises in the individual's attempt to deter- mine appropriate behavioral configurations, the value structure that one ought to hold, or the manner in which one ought to interpret events in the three primary components of information space. Future uncer- tainty arises in predicting alternative future configurations of matter, energy, or information in the three primary components. Cognitive Complexigy The above view of cognitive uncertainty is from the perspective Of the individual. Cognitive uncertainty may also be defined from the perspective of someone Observing the encoded representations of another person's cognitive processes or inferring an individual's cognitions from his behaviors. This is the prime focus of attribution psychology (Kelley, T972). In this sense, uncertainty involves an Observer's ability to predict the particular cognitive activities Of another over time. Given awareness of some behavioral or perceptual contexts, the extent to which the cognitive reactions of another can be predicted is the basis for this second type Of uncertainty, which will be labelled attributive (A) cognitive uncertainty. An individual's cognitive uncertainty is attributed based on the inferences an observer makes about that person's behaviors. The key to the predictability of a person's cognitive responses is in the cognitive complexity of that person. The work of Schroder, 69 Streufert, and Driver (T967) has been concerned with the explication validation of a set of constructs labelled cognitive complexity. The view cognitive complexity as composed of three basic parts: l) differentiation: the number of different dimensions used in processing information. 2) discrimination: the fineness of discrimination of informati along a dimension. 3) integration: the extent to which relationships are formed among dimensions in processing information. The greater the cognitive complexity of an individual, the grea the uncertainty in predicting the precise cognitive responses of an individual to information from environmental space and from ego space If the individual uses more alternative dimensions in processing info tion, then it is more difficult to predict which particular dimension will be used at a given point in time as information is processed. A the more finely the individual discriminates in processing informatio along a given dimension, the more uncertainty there is in predicting location of a unit of information on a cognitive dimension. In addit the more the individual integrates different dimensions, and the more these integrations deviate from probability values for the population or how "creative," or ”new,” these integrations are, the more uncerta one is about the particular pattern of integration that will occur. short, the more cognitively complex the individual, the greater the A cognitive uncertainty. It is likely that attributive and subjective cognitive uncertai eare correlated. That is, individuals who are more cognitively comple vvill .also tend to have more §_cognitive uncertainty. Since the 70 cognitively complex individual views the world in a complex way, ing a more precise multidimensional organizing structure on expe he/she is likely to perceive greater uncertainty about: l) envi mental space, 2) ego space, and 3) relationships of ego space to ment. The individual will perceive greater numbers of alternati' relevant to a particular context within one Of these three areas in addition is likely to perceive them as more equally likely to appropriate, given the individual's more relativistic logic. Uncertaintnyreference Experimental research shows that when individuals are pres. with stimuli of varying complexity or entropy, they distribute w' respect to preference for level of uncertainty or complexity (Ecl T963; Munsinger and Kessen, T964; Dorfman and McKenna, T966; Rum] McNeil and Rule, T970; Swartz and Herbik, l97l). The relationsh' between uncertainty/complexity and liking or preference is an in' U-shaped one. Figure 5 illustrates this relationship. LL— Preference \ Liking Uncertainty Figure 5. Information and liking. 7T This suggests that each individual has a preference for uncg that operates as a "referent" in a homeostatic system. While most the evidence comes from experimental studies that measure liking i stimuli that vary in uncertainty, there have also been some attemt develop general scales to measure such a construct. A recent atte was made by Cohen (l974) with a related construct he labeled ”c0pi with uncertainty.” In earlier work, Adorno 23.91: (T950) and othe have developed the construct of ”tolerance for ambiguity“ to acc0L such variables as authoritarianism and ethnic prejudice. Rokeach explicated the construct Of dogmatism, that has seen a great deal use in interpersonal communication research. Jones and Gerard (TS discuss the construct Of "information dependence“ in a related mar Summar In this chapter information space has been differentiated ir three major components: environmental space, ego space, and cogni Space. The boundary separating ego space from environmental space the zone in which the individual perceives greater than 50% contrc over components of the space. Within this region defined by the c boundary is the volume of an individual's ego space. Cognitive s; a subset of ego Space, is unbounded by innediate contact with the environment. Cognitive uncertainty and complexity are characteris of cognitive space which determine preference for uncertainty witl ego and environmental space. The nature of this preference is suc that deviations above and below a preferred range produce an unfa\ response by an individual, and this relationship between the exper 72 uncertainty of ego and environmental space and uncertainty preference is therefore an inverted U-shaped curve. The following chapter is concerned with specifying the role of this preferred range of uncertainty in the operation of the principle of Optimal information. A homeostatic principle is proposed through which the individual attempts to maintain relationships within environ— mental and ego space that optimize the correspondence between preference and actual experience. CHAPTER VII THE THEORETIC CORE: OPTIMAL INFORMATION, FUNCTIONS OF COMMUNICATION, CONTROL OVER ENVIRONMENT, AND ATTRIBUTIONS OF COMMUNICATION FUNCTIONS Overview This chapter explicates core constructs and principles in the information theory of communication functions. The chapter begins with an explication Of the basic operating principle of the model being developed in this work: the principle of optimal information. Follow— ing this is a discussion of the functions of communication: uncertainty expansion and uncertainty reduction, with two subtypes within each. Subsequently, a cybernetic model of the relationship between uncertainty discrepancy and the entropy Of ego space is developed. The variable control over environment is then discussed in the context of the relative use Of mediated and interpersonal communication, leading to the development of an explanatory function linking relative use to the effort required to use each component of information space, and the construct of the balance utility Of the respective components. Finally, in the last section the bridge is made from the abstract development of communication functions to the individual's own attribu— tions of functions to communication behaviors, specifying propositions that enable the validation or invalidation of the model through direct responses of individuals. 73 74 Balance Model In this section the focus is upon how the individual functions in information space, how the structure of information space is changed, and why it is changed. The basic explanation for these processes is a principle of balance, in line with a tradition of balance models in other areas Of cognitive and interpersonal behavior. The work of Heider (T958) in the area of cognitive balance in interpersonal perception is often recognized as the key initiator of a series of balance model explications. Newcomb's balance model (T953) often receives attention as an important contribution to social science. The A-B-X model of interpersonal relationships forms the basis for the later development and elaboration of the "coorientation model" Of Chaffee and McLeod (T967). The balance model of Festinger (l957) in the area of cognitive psychology has had a very large impact on social science. In this development the individual essentially operates in the cognitive domain to reduce the incidence of cognitive dissonance, or lack Of balance among cognitive elements. Woelfel's (l973) Linear Force Aggrega- tion Theory has the notion of balance as a key principle underlying the Operation of the model. An individual's behaviors and cognitions are a weighted average of proposed behaviors and cognitions received in messages from the environment. In the present development, the primary principle that underlies the model Of communication functions is the principle of optimal informa— rjpp, This balance principle suggests that people operate as homeo- static systems, seeking a level of information input that will maintain uncertainty within parameters that are necessary for the individual to 75 function effectively in information space. Deviations above or bel this range of uncertainty will cause the individual to alter inforn input to bring uncertainty back within preferred parameters. When input deviates to the extent that generalized cognitive uncertainty no longer within preferred levels, arousal of the individual will increase to the extent that it becomes dysfunctional for the indivi performance. This excess arousal, occurring as a result of both de tions above and below preferred levels, will cause the person to ch information input to bring the experienced and preferred uncertaint back into correspondence. Research evidence supporting the general principle Of optimal information can be gathered. An important feature Of this research evidence is that information theoretic measures of the stimuli are in experimentation. Thus, it integrates well with the model we are developing here. Jones and McGill (T967) performed two experiments regarding t effects Of deprivation and satiation on subsequent information-seek behaviors. In the deprivation experiment, subjects were allowed to select different rates Of information in high and low tones, rangin from 0.0, 0.33, 0.67, to T.00 bits, broadcast at l second intervals Results showed that more entropic tonal patterns received the most selections. In the satiation experiment, individuals heard complet unpredictable tones (T.OO bits). One group listened for one hour before being allowed to select the preferred level of information, other'group for 5 hours. For the remainder of a ten hour session, Jectns were allowed to select their own information levels. 76 Results of the experiments showed that after deprivation, more entropic patterns were selected, while after satiation, more negenti patterns were selected. Other research by Thornton and Jones (T965 and Thornton (T966) also shows that after deprivation treatments, 3: jects select more entropic patterns. Thus, research indicates that principle of optimal information is a viable mechanism for explainn the operation of the individual in information space. In essence then, there is a proposed U-shaped relationship fO' effective functioning of the individual. When uncertainty falls abi or below preferred levels, excessive arousal will occur and effecti' functioning will be low, while within preferred levels, arousal and functioning will be optimum. Evidence for this U-shaped information input and arousal rela‘ ship was demonstrated in research by the author (Danowski, l974b). relationship between television content entropy and self-reported measures of aggression for a sample of adolescents was found to be : nificantly U—shaped. However, the relationship was not strong. Furthermore, there is controversy regarding the arousal and aggreSS' relationship, suggesting caution in applying these results in suppO‘ of the proposition. Hanneman (l970) found that greater uncertainty the plot Of violent television programs created greater physiologica arousal, but did not examine the relationship in a way that would e: determination of a curvilinear relationship. Now that the principh Optimal information has been explicated, in the following section i VWFTT be linked to two general functions of communication. 77 Functions of Communication The principle of optimal information suggests that when info tion input falls above or below the preferred level, the individua alter the environment to establish balance. When input falls belo preference, uncertainty expansion is necessary, while when input r above preference, uncertainty reduction is necessary. In line wit view are two general types of communication functions: uncertaint expansion and uncertainry_redugtion. The concepts of information overload and information underlo. useful in referring to the direction of uncertainty discrepancy. T these two load concepts, overload has received considerably more a' tion, probably because it is generally recognized that post-industl society is experiencing an "information explosion." Miller (T964,' has given considerable attention to information overload with respc its effects on mental health, and on the methods of coping that sy: undertake that appear generalizable across systems levels. Underload receives relatively little attention in social sci: Probably the line of research that best applies to the concept of I load is sensory deprivation (for a review see Zubek, l969). HOWEVT this research remains rather esoteric, with little attempt to deve' links to a wider non-laboratory world. Nevertheless, it is probab‘ ppt_the case that the experience of underload rarely occurs for im viduals in the population. The evolution of production efficiency industrial settings has created likely conditions of underload for workems. It is commonly expressed in everyday life that factory w higtily predictable and boring. 78 Underload also appears to be experienced at a higher rate than overload by another social group-—old people. In everyday conversatic and in applied social gerontology it is often heard that many Old peop experience a lack of activity and an excess boredom. Many programs aimed at improving the quality Of life for old people attempt to in— crease the information input to the individual through increasing acti ity levels. To summarize at this point, underload, or negative uncertainty discrepancy, will lead to the use of communication for uncertainty expansion functions; overload, or ppsitive uncertainty discrepancy, wi lead to the use of communication for uncertainry reduction. Another dimension of uncertainty discrepancy will be introduced that allows the differentiation of each type of discrepancy and the corresponding type of communication function into two subtypes. This dimension is the level of specificity of uncertainty. Specificity wil be discussed at two levels--diffuse and specific. Diffuse uncertainty is more affective in nature, while specific uncertainty is more cognii Highly specific uncertaintyarises in narrowly defined contexts. Diffuse uncertainty is generally topically relevant. For example, an individual may be uncertain about what has happened in national politi affairs on a given day, or uncertain about whether what was said to a particular person was appropriate, or uncertain about how the economic climate will affect purchasing power. Diffuse uncertainty_is more generalized than specific uncertaint ancl extends more fully throughout the individual's life space. It is less; likely to be topically relevant uncertainty, and more likely to 79 manifest in the individual's emotional and autonomic states. Diffuse uncertainty may be the result of large accumulations of specific uncer- tainty discrepancies that are unresolved. Examples of diffuse uncer- tainty might be an individual's uncertainty about mental functioning, life style, self-worth, and so on. When level of specificity is linked with communication functions, two uncertainty expanding and two uncertainty reducing functions result. The diffuse uncertainty expander function is labelled stimulation, while the diffuse uncertainty reducer function is labelled stress release. The specific uncertainty expander function is labelled learning, while the specific uncertainty reducer is labelled forgetting. Figure 6 graphically displays these functions of communication. UNCERTAINTY UNCERTAINTY REDUCERS EXPANDERS l "Forgetting" l - "Stimulation" .L__, Figure 6. Four functions of communication. Uncertaintnyiscrppancy and Communication Behaviors It is now possible to specify theoretic relationships between uncertainty discrepancy and amount of both interpersonal and mediated communication, as well as relationships between uncertainty discrepancy 80 and the structure of communication relationships. Both amount and struc- ture of communication produce changes in the entropy Of relationships in the mediated and interpersonal spheres, although one measure of entropy is more precise than the other. By identifying changes in the struc- ture of communication, changes in uncertainty are more clearly determin- able, than is the case with identifying changes in amount of communica— tion. On one hand, if amount of communication increases, there is greater likelihood that more alternative patterns are exchanged, and hence there is greater uncertainty. However, it is possible that in- creased amount does ppt_change the range of alternative patterns or their relative probability but indexes more of the same alternatives in the same distribution pattern used earlier. Thus, amount is likely to be positively associated with uncertainty, but with a potentially high level of measurement error in some cases. On the other hand, changes in the entropy of channel, content, or form are linked to more direct changes in uncertainty, since here there is a correspondence of struc- tural measures that does not occur between amount and uncertainty. From another perspective, with changes in uncertainty discrepancy come changes in the functionality of communication, and as functionality increases, the amount of use of communication will increase. This coupled with the above distinctions provides a concise rationale for the following propositions: p]: The greater the uncertainty discrepancy (overload or under- load) at a point in time, the greater the amount of subse- quent communication (mediated and interpersonal). 8T With respect to entropy of channel, content, and form, the follow- ing relationships are hypothesized with uncertainty discrepancies: p2: The greater the uncertainty overload at a point in time, the less the entropy of subsequent communication relation- ships (mediated and interpersonal). P33 ships (mediated and interpersonal). The greater the uncertainty underload at a point in time, the greater the entropy Of subsequent communication relation— Figure 7 summarizes these relationships and their rationale by illustrating a cybernetic model of the operation of the principle of optimal information and its corresponding impact on the structuring of I:UP I : ifIUP a I l lfI=UP l . T l . I Create Maintain Create . L» greater entropy less I entropy entropy : l I T l I I ' L ______ I ______ __ J ______ J Environmental change «9—— function Figure 7. A cybernetic model of optimal information. 82 Control Over Environment In previous portions of this chapter, it has been suggested that uncertainty discrepancy occurs for individuals. Why does this discrep— ancy occur? This is the case because the environmental space, and portions of the ego space are not under the complete control of a per- son. If they in fact were, then it would be unlikely that uncertainty discrepancies would occur at all, for the individual would be likely to structure the outer information space in patterns congruent with prefer— ences. Control over environment is, then, of critical importance in this model. The following proposition specifies this relationship: The greater the control over environment, the less the uncertainty discrepancy. P4: Not only will control over environment have an impact on uncer- tainty discrepancy, but it will have a major role in determining the relative use of mediated and interpersonal communication, and the rela— tive attribution of functions to each of these modes. In this section relationships between control and use of mediated and interpersonal communication will be discussed; relationships regarding attributions follow subsequent to this discussion. Effprp, In explaining the relative use of mediated and inter- personal communication as a function of control over environment, the concept of gffprp_is introduced. Depending on the amount Of control perceived by the individual, different levels of effort are perceived necessary to change aspects of outer information space. The person with high perceived control over environment will perceive changing the interpersonal communication component as requiring less effort than a 83 person with lpy_perceived control. However, control is likely to have less Of an impact on perceptions of effort required to change the mediated component than the interpersonal component, because media require relatively low effort to use. It is generally recognized that media are widely available and distributed in the population. Marketing strategies of media-related organizations have generally succeeded in creating media environments that are easy to use and require low expenditure of effort, since the effectiveness of advertising expenditures increases through an increas- ingly positive relationship between exposure and frequency of resultant purchasing behaviors. In addition, media designs have traditionally been "one-way" information flows. Thus, the increased involvement required by "two-way" interactive feedback systems in most other environ- mental components (especially social relationships) may be avoided by media users. Anecdotally, it would seem reasonable to assume that turn— ing on the television or picking up the newspaper, for example, would require considerably less effort or cost than changing jobs, getting a divorce, moving to another geographic area, etc. Overall, considering the characteristics of media it seems more appropriate to assume a largely constant perceived effort for mediated components across levels of control. Balance utiliry, Mediated and interpersonal communication compon- ents are likely to have differing reward value for the individual in restoring balance in the information processing system. This reward value will be conceptualized as balance utility. Mediated communication, particularly of the "mass” type is likely to have lower balance utility 84 than interpersonal communication, when the amount of communication in each component is held constant. While in the interpersonal mode, the individual has considerable freedom to structure the situation by the use of feedback in the message exchange process, in mediated communica- tion this feedback capability is substantially reduced. Once the individual selects a particular chunk Of messages, the "objective” con- tent cannot be changed--the only way to substantially alter the communi— cation situation is to selectively process messages or to discontinue their use. Media themselves vary in how much constraint is imposed from the source end on receiver selectivity. As one example, while print media are not time bound, electronic media typically are (save for the intervention of tape recording). In short, interpersonal communication will generally have higher balance utility than mediated communication. Interpersonal modes of communicating Offer greater specificity and efficiency in restoring balance, not only because of greater feedback potentials, but perhaps because uncertainty discrepancies are more subject to causes arising more directly out of interpersonal situations. Relative use. Having discussed control, effort, and balance util- ity, these variables can now be integrated with the relative use of mediated and interpersonal communication. The individual's relative use Of mediated and interpersonal communication may be characterized by the following formulae: - Eli (l) SCmode _ k E (2) If SCmm > SClp + MM (3) If SCmm < SCip + IP 85 where SC is the selection coefficient for a mode, mm and MM are mass media, ip and IP are interpersonal communication, BU is balance utility, E is effort, and k is a constant associated with the mode. Thus, the component with the highest ratio of balance utility to effort expenditure will be selected to reduce uncertainty discrepancy. Control determines the value of the effort variable. Generality of pptimal information. We have specified that although mediated and interpersonal communication have differential balance utility, uncertainty discrepancy can be reduced by either. However, it may be argued that the principle of optimal information is not a general one, and that imbalance may be restored only by changing the component of the environment that is most responsible for the dis- equilibrium. For example, this argument would suggest that imbalance experienced in the occupational setting must be reduced through a change in the occupational setting itself, and not through watching television. Research does pp§_support this argument. Thornton and Jones (T965) conducted an experiment in which two groups of subjects underwent deprivation for an 8 hour period. Both groups were deprived Of visual information, while at the same time one of the groups was given maximally entropic auditory tones and the other was given maximally negentrOpic tones. After six hours of these condi- tions, subjects were allowed to select their own level of entrOpy in a series of yisggl_patterns, which they processed for 2 hours. Results showed that deprivation and satiation of the auditory mode were respectively counterbalanced in the visual mode in the remaining two hours. Thus, evidence is provided for the generality Of the principle 86 of optimal information. This suggests that the operation of the prin- ciple of Optimal information is mode and content free. The following propositions summarize the preceding discussions regarding the relationships between control and mode of communication: p5: The greater the control over environment, the greater the interpersonal communication. p6: The greater the control over environment, the less the media exposure. The following section focuses on the individual's attributions of func— tions to communication behaviors. Attributions of Communication Functions To this point the discussions of the information theory of communi- cation functions has been in “objective“ terms, i.e., from the perspec— tive of the external scientific observer. This section serves an important role in linking the model to the functioning Of the individual from his own perspective. The concern is with attributions the indi- vidual makes about his own communication behaviors that are in line with the development of the model. Given different levels of uncertainty discrepancy and control over environment, does the individual make attributions about communication behaviors that are consistent with the theoretic development Of the functions of communication? The area of attribution psychology, largely evolving through the work of Kelley (T972) and his associates, attempts to understand the manner in which the individual orders experience. A major assumption of this line of social theory and research is that the person attributes causes to relationships among variables in a manner very similar to 87 the processes in which the scientist engages in the construction of theory. Some of the key assumptions of this paradigm are: l) an effect is attributed to one of its possible causes with which over time it varies, 2) interpretation of a given covariance relationship often depends upon a person's prior causal attributions, 3) the role of a possible cause in producing a given effect is discounted if other plausible causes are also present. In a classic study, Schachter and Singer (l962) have demonstrated that attribution of causes to one's prp_behavior is a function of both changes in autonomic arousal and salient environmental information. In this light, in the context of attributing functions to communication behavior: l) the individual will be cognizant of changes in awareness Of cognitive and affective internal processes that restore balance in the information processing system, and 2) the individual will be cog- nizant of a covariation of communication behaviors--either talking to people, or using media--and the restoration of balance. On the other hand, the individual will also perceive a covariance between communicating with others and through media, and deviation amplifying processes that create imbalance. However, when the indi- vidual is asked why he/she communicates, it is unlikely that communica- tion behavior will be attributed to the maximization Of deviation amplifying processes. A simple explanation for this is that it is rare in our culture (or at least relative within subcultures) that high value is placed on the maximization of dysfunctionality, particularly at a psychological level. While there are obviously individuals who do Operate in this manner, they are likely to be defined as deviants and 88 subject to sanctions through incarceration, hospitalization, or the initiation of some form of therapy whose expressed purpose is to produce deviation reducing processes of behavior. Attribution Propositions In this section the theoretic propositions regarding the attribu— tions of communication functions will be explicated. The logic for these propositions has been developed in earlier sections of this chapter, in which the concern was with predicting communication behaviors as a func- tion of control and uncertainty discrepancy from the perspective of the independent Observer. In predicting the individual's own attributions, the rationale for the propositions derives from coupling this earlier logic with the notions developed in the preceding section. p7: The greater the control over environment, the less the attribution of communication functions to interpersonal communication and mediated communication. The greater the control over environment, the greater the relative attribution of communication functions to inter- personal than tO mediated communication. P81 p9: The greater the overload at a point in time, the greater the subsequent attribution of communication functions to uncertainty reduction. p10: The greater the underload at a point in time, the greater the subsequent attribution of communication functions to uncertainty expansion. Summary To summarize this chapter, we have explicated the core theoretic underpinnings of the information theory of communication functions. The principle of Optimal information has been developed and linked to 89 two general functions Of communication. A cybernetic model was then developed to account for changes in uncertainty discrepancy and changes in the entropy of ego space. Control over environment was discussed in light of the relative use of interpersonal and mediated components of information space, serving an integral explanatory function; also in- cluded were the constructs Of balance utility of each component, and effort required in using each component. In the last section, the link was made betweenthe “objective" accounting of the individual's opera- tion in information space, and the individual's own attribution of functions to his/her communication behaviors. This chapter and the preceding theoretical chapters provide the basis for application of these notions to aging. In the following chapter the construct informational aging is developed. The concern is with explaining the Operation of the individual in information space over the life-span. CHAPTER VIII INFORMATIONAL AGING Overview This chapter is an explication of the construct of informational aging. This introduces a developmental perspective to the information theory of communication functions. Changes in the manner in which indi- viduals Operate within information space over time are focused upon. The chapter begins with a general discussion Of two major types of change that may be Observed in individual information processing: cohort and developmental. Subsequently, informational aging is dis- cussed in terms Of three basic factors: change in overall information space, change in size and complexity of ego space, change in relative mediated and interpersonal communication use, and change in uncertainty preference. Finally, this functional view of aging is contrasted with a chronological view. Process View To this point the explication of the information theory of communi- cation functions has taken a process view. The concern has been with changes in the use of interpersonal and mediated modes of communicating as a function Of uncertainty discrepancy and control over environment. This concern with process, however, has been at a micro level; relatively 90 9T short-range changes have been focused upon. Nevertheless, the process view of communication functions may be implemented at a more macro level. While we have assumed a relatively stable overall information space, and have dealt with the individual's operation within it, we will now examine major shifts in the overall information space that occur over the life-span of individuals. In earlier development of the construct control over environment and uncertainty preference, constant values for an individual were assumed. In contrast, this seCtion focuses upon the life-span development Of control and uncertainty preference, and rela- tionships to other portions of information space. Types of Change Two basic types of change are important to consider in the exami- nation Of changes in information space over time; changes may be attributable to generational or gphprp_factors, or changes may be attributable to maturational or developmental factors (for a discussion see Riley, T973). Cohort changes refer to changes occurring in the information space of a group of individuals who have experienced similar historical events; events in the macro system structure impact on the bulk of individuals in a particular period Of time. For example, the birth cohort Of the early T9505 experienced the widespread diffusion and use of television from their initial entry into the social system, while the birth cohort of the post World War I period have no comparable experience with television in their lives. Cohort effects assume that a given cohort, a birth cohort for example, will respond as an aggregate to environmental factors in a relatively homogeneous way. A major 92 assumption of a cohort approach is that there is greater variance between cohorts than within cohorts in effects processes. Developmental changes refer to changes that are generalizable across cohorts, and are the result of basic changes in physiology, psychology, and micro social relationships that are cohort independent. These developmental changes may be roughly summarized as changes attributable to the effects of ggjpg, Thus, cohort changes are macro system linked, while developmental changes are micro system linked. However, when taking a systems view Of human activity, both the macro and micro aspects Of information space are in dynamic interrelationship. In other words, both types of change are mutually reciprocal. Events in the macro system structure arise, in part, out of the networks Of interrelationships linking individual information spaces together. These events in turn impact the micro system structures. In short, the two types of change are not discrete and clearly conceptually segmentable. The critical distinction is the degree of homogeneity of change across aggregates Of individuals. Cohort changes will make relatively homogeneous impact on particular cohorts, while developmental changes will be localized to much smaller and widely dispersed sub—groups of the social system that happen to experience similar combinations of micro variables in time sequence. Since there are more variables Operating at the micro level with respect to individual information space, with more plausible permutations of relational patterns than at the mcaro level, and since biological changes have more specific impact, it is likely that there will be greater variance in developmental changes than cohort changes in individual 93 information space. In the remainder of this chapter, the focus will be upon these developmental changes in individual information space. We will refer to these processes as informational gging, Factors in Informational Agipg_ In keeping with the model developed in earlier chapters, three major components of informational aging will be discussed in the follow- ing portions of this chapter: T) changes in perceptions of control over environment, 2) changes in the structure of ego space, and 3) changes in preference for uncertainty. Each of these are interlinked in a pattern of relationships, and are likely to be mutually reciprocal. However, we will begin with a discussion of control as a focal variable. Before moving to this discussion, the general notion of U-shaped relationships will be briefly discussed. The U—Curves In reference to the notions of process, one sometimes sees refer- ence to the words of Thomas Wolfe, "You can't go home again, or to those of Heraclites, ”You can't step in the same river twice.” Perhaps in the basic phenomenological or ethnomethodological context this is an appropriate assertion, for an earlier raw experience can never be directly recreated. However, such an assertion is counter to the basic assumptions of science, for science is concerned with time-independent abstractions. Thus a variable taking on a value at one point in time may have taken on that same value at an earlier point in time, or may take on that value at a future point in time. With some forms Of cycli— cal processes, through abstract representations Of behavior the 94 individual does 'go home again' and does 'step in the same river twice.‘ The basic graphic representation of these phenomena is a U—shaped or inverted U—shaped curve. Correspondingly, when information processing is studied, the system often returns to a value of some state or criterion where it has been located before. One example was discussed in Chapter VII regarding the relationship between information input and arousal. At high and low levels of input the system is highly aroused, while at intermediate levels the system exhibits low arousal. In the current discussion informational aging will be characterized as a set Of U-shaped relation— ships, tracking changes over time in the individual's structure of outer information space and uncertainty preference. In brief, in informational aging the individual returns to an earlier set of system states, abstractly retracing its steps from begin- ning to middle, and from middle to end. These beginnings and ends might be characterized as the entrances and exits to another form of energy organization. Except for time, the entrances and exits are one. Control and Informational Aging In earlier explication, control over environment has been chosen as a variable to serve as a major focal point in the information theory of communication functions. Correspondingly, control over environment will be used as a focal point in explicating the construct of informa— tional aging. Let us now consider the development of control over the life—span. 95 When the human organism is very young, it perceives itself as having so little control over environment that it does not distinguish itself from it; it has no concept of actor. Gradually, the person begins to distinguish itself from environment and develops perceptions that by engaging in particular behaviors, regular effects result. Thus the person becomes an individual, forming perceptions of increased ability to control the environment. The individual increases in control through childhood and adoles- cence; typically in this period, however, levels of control are not achieved that the average adult exhibits. The socialization system places restrictions on the opportunities the individual has for develop- ing high levels of control. For example, the child must generally accept the definition of experience that the parent imposes; also, the child must attend school, cannot legally use drugs like alcohol and tobacco, drive, vote, and so on. Adulthood brings with it the opportunities to maximize perceptions of control over environment. The individual is granted maximal social PY‘TViTeges and can to a greater extent structure activities in line With preferences. While there is tremendous variance within an adult poPli'laition on this dimension, when a single individual is examined over time, it is most likely that the period of early to middle adulthood WT 1 1 be that person's time of greatest opportunity for developing maximal peT‘Ceptions of control over environment. With advancing age through the middle adult period, the later adUTt period, to the old age period, for most people the macro system StY‘UCture is likely to increasingly limit these opportunities for the 96 exercise of control. Occupational settings are typically geared for the "young" in order to maximize the return on expended training and development resources, and the middle to later adult may often perceive a withdrawal of earlier organizational levels of support. After retire- ment, Often mandatory, the limitations on developing increased control are strong, and greater declines in control are likely to be observed. The individual is loosed from the occupational role, Often the marital role, generally receives substantially reduced financial income, and is Often perceived by others as less able to engage in previous activi- ties, whether or not this is actually the case. These macro system factors are not likely to be the only ones contributing to decreased control over environment with advancing age. Some Of the reasons are perhaps bio-physiological in origin. Restric- tions in physical capacity may somewhat directly contribute to decreas— ing perceived environmental control. However, wide-ranging evidence suggests that psychological and social factors have a large impact on health, indicating that these factors may perhaps be most important (Moss, T973). A review of the literature on perceptions of Old people (McTavish, l97l) shows that they are perceived as generally ill, tired, not sexually interested, mentally slower, forgetful, less able to learn, grouchy, withdrawn, feel sorry for themselves, less likely to partici- pate in activities, isolated, unproductive, and defensive. Running through these perceptual elements appears to be a strong underlying dimension of less perceived control over environment. At the micro social level it is likely that these perceptions will be communicated 97 to the old person either directly or indirectly through the manner in vvhich others behave toward the old person. Communication theories would suggest that messages containing these kinds of elements will have a large impact on the self—conceptions of old people. The work of Woelfel (l973) suggests an explanation for the effects of these kinds of messages on the self-conception. The individual's conception of self and behaviors are a direct function of information received from significant others. Self-conception and behavior may be represented as the simple linear aggregate of information received from these sources. Thus, given that these perceptions shown TX) be held by research are communicated to the old person, the theory suggests that this information will have a large impact on the old person%;perceptions of loss of control over environment. In sum, considering these macro and micro factors over the life- span, an inverted U-shaped curve for perceptions of control over environ— ment is likely to be observed. Some research evidence supports this view. One of the subcomponents of the control scales developed by Rotter (l966) is control over the political system, or political efficacy. In fact items of this type appear in standard operationalizations of the construct. Research on political efficacy shows a commonality between the young and the old. Research by Martin, Bengtson, and Acock (T973) and Miller, Brown, and Raine (l973)--a cohort analysis--has shown that both young and old are low in political efficacy, while those in the middle of the chronological age distribution are highest. However, a rather limited cohort analysis by Cutler and Bengtson (l974) does not give these clear—cut results. 98 In an earlier section it was proposed that control is related to communication patterns. In particular, as control decreases, the use of interpersonal communication decreases and the use of mediated com- munication increases. If this proposed relationship is valid, then through Observing changes in the use of television, changes in control can be inferred (although cautiously). Research shows that both the young and the Old are the highest consumers of television programming, and the young and the old likewise hold the most favorable attitudes toward the medium (Bower, T973). Evidence from cohort analysis shows that these are not simply cohort effects, but are in large part develop- mental in nature. In short, it appears to be a reasonable proposition that changes in control over environment over the life-span may be represented effectively by an inverted U—shaped curve. Trends in Mediated and Interpersonal Communication Linking the earlier explication Of the relationship between con- trol and the relative use of mediated and interpersonal communication with this developmental view, it will be the case that the use Of mediated and interpersonal communication will generally vary inversely over the life-span. In the early years a relative high mediated and low interpersonal use will be observed, in middle life a relative high interpersonal and low mediated, and in late life a relative high mediated and low interpersonal. As was mentioned in Chapters 11 and 111, research and theory on the relative displacement of mediated by interpersonal communication is 99 beginning to receive attention, although in absolute terms little has ‘yet been done. Work by Graney and Graney (l974) cited earlier suggests that in a sample of elderly persons, as some activities decreased, use Of mediated communication increased. Also the work of Cassata (l969) gives some evidence for this kind Of relationship through the use of self-reported past and present communication behavior. Let us emphasize that chronological age is not the primary pre- dictor of these relative relationships of mediated and interpersonal connnnfication, but control over environment is. Thus, persons scattered at spots all over the age distribution will exhibit the predicted rela- tive use as a function of control. Persons experiencing temporary losses in control perceptions, because Of major disturbances in the family, with friends, at work, etc., will exhibit the predicted shifts in relative use of mediated and interpersonal communication. This high- lights the importance Of a functional view of aging--in this case informa- tional aging——rather than a chronological view. These issues will be addressed more fully later in this chapter. Ego Space and Informational Aging In linking control over environment with ego space, as the indi— vidual acts less directly upon environment the structure of ego space will change. The individual with lower control is likely to reduce the amount of communication with the interpersonal component, and thus will exchange information with less heterogeneous and more interconnected network nodes. Hence, the entropy Of channels, content, and form of interpersonal relationships will decrease, and this component of T00 information space will become less complex and more predictable. 'Therefore, there will be a tendency for individuals at both ends of the chronological age distribution to have a less entropic interpersonal information space. At the more extreme ends Of the chronological age distribution, the entropy Of mediated information space is also likely to be less entropic with respect to channel, content, and form. The person with lower control is more likely to exhibit less selectivity in processing mediated information, and less actively seek varied mediated information input. For example, the young child of 2-3 years is likely not to read time newspaper, magazines, listen to radio, and attend movies, but is highly likely to watch television. Within the television sphere, the young child is likely to have highest interest in a very limited range Of program types, perhaps primarily children's programming such as Sesame Street and cartoons. The young child is also likely to prefer less entropic form within television messages. Research by Wartella and Ettema (T975) shows that in children's viewing Of television commercials, controlling for form entropy, the younger the child, the less the atten- tion to the television commercial. Along similar lines, the very old person is likely to decrease in reading newspapers and magazines, listening to radio, and attending rmivies, but is much more likely to increase in watching television, thus showing a decrease in channel entropy. The Old person is also more likely to have restricted content preferences indicating decreased con- tent entropy. Research shows that old people show an unusually high perference for news and public affairs programming (Meyersohn, l96l; TOT Steiner, T963; Davis, T972). Preference for program form is also likely to show decrements in entropy, with the old person tending to exhibit preference for more predictable structures. Preliminary research shows that preferences for form entropy are represented by an inverted U-shaped relationship, with young and Old preferring more predictable program form, and middle age persons preferring less predictable form. The peak appears to occur at about 25 years of age (Watt, T974, personal communication). Uncertainty Preference and Informational Aging It is unlikely that an individual's preferences for uncertainty ranain constant across the life-span. Changes in uncertainty prefer- ence will be intimately linked to changes in control and information space entropy. The following formula represents the development of uncertainty preference over time: UP = UPtn_ + a It . Uncertainty tn n—l preference at a given point in time is equal to the uncertainty prefer« ence at the preceding point in time, plus or minus a quantity that is a function of the input entropy at the preceding point in time, multiplied by a fractional constant that will be labelled the adeptability coefficient. In short, this suggests that uncertainty preference changes as the individual experiences an input entropy level that is discrepant from the original preference. In other words, an individual experiencing a repeated positive uncertainty discrepancy will gradually increase in uncertainty preference, while an individual experiencing a repeated negative uncertainty discrepancy will gradually decrease in uncertainty T02 preference. The acceleration of these changes will be a function of the adaptability of the individual. Persons high in adaptability will change more rapidly than persons with low adaptability. This line of development may appear contradictory to the earlier explication of the changes in entropy of information space associated with changes in uncertainty discrepancy. There it was proposed that the experience of a discrepancy between preferred and actual uncertainty would lead to a change in the actual uncertainty experienced rather than to a change in preference. However, this assumption was made for short periods of time, but over long time frames preference will be changed as a function of what is experienced, as the immediately preceding discussion suggests. At a conceptual level, Ashby (T968) has dealt with this relation- ship between environment and internal system structure in very broad terms, applicable to a wide range Of systems and levels Of analysis. He proposes a "law Of requisite variety.” Essentially, this law posits that in order for variety or complexity to be reduced in a system com- prising the environment Of another system, the reducer system must have a level Of variety or complexity in its internal processes that corre- sponds to the complexity or variety within the environmental system. Interestingly, research not directly aimed at validating this notion for human systems, has provided strong evidence for the validity of this relationship. Schroder, Driver and Streufert (T967) report and review research that shows when individuals or groups are placed in complex environments, their cognitive complexities increase. On the other hand, when groups or individuals are placed in simple environments, T03 complex cognitions do not develop. In more extreme research contexts, it has been shown that under conditions of sensory deprivation, the complexity and structuration of cognitive systems deteriorate rather significantly. Subjects report that they cannot think clearly, that uncontrolled thoughts rapidly Occur, that they are unable to concen- trate (Zubek, T969). Principles Of educational methodology have long recognized this relationship, for much of the role of the educator is the structuring of complex environments for the student, so that a complex process of ondering the environment will be imprinted upon the student's cognitive pnocesses. Principles of child-rearing also adhere to this notion, and in recent years the concern for creating complex visual and verbal environments for the developing infant has reached widespread diffusion through the population. From a developmental perspective, the indi- vidual is then socialized into a level of cognitive complexity and preference for complexity and uncertainty in the environment. When the characterization of the process Of changes in uncertainty preference represented by the formula presented earlier in this section is coupled with notions developed earlier that changes in control over environment and correspondingly that changes in the entropy of ego information space are U-shaped, the implication is that preference for uncertainty is U-shaped across the life—span. At both extremes of the chronological age distribution, uncertainty preference is likely to be low, while in the middle high. Another factor that may contribute to the curvilinear pattern in uncertainty preference is the adaptability coefficient itself. While it T04 is likely that mean adaptability varies eprpss_individuals, adaptability is also likely to vary yiphip_individuals over the life-span. It is clearly and commonly recognized that adaptability of the very young child is high, then decreases through adulthood. It is much less clear what happens to adaptability from middle adulthood to old age. Common wisdom suggests that adaptability decreases with Old age, for I'you can't teach an old dog new tricks." We might speculate, however, that adapt- ability increases with advancing age, but the individual may adapt more rapidly in a direction of decreased uncertainty preference. In the con- text of decreasing control and decreasing entropy of ego space, the individual may more rapidly adapt, yielding an acceleration in decreas- ing uncertainty preference. To summarize this section, the introduction Of this characterization of changes in uncertainty preference with notions of adaptability playing an important role has introduced a learning function into the model. Functional and Chronological Aging The thrust of this chapter has been to explicate a functional view of aging, information aging. We have attempted to show that age per_se is not the important determinant of communication behaviors, but that more abstract and generalizable variables are. Some of these variables are associated with age to be sure, but not because of something directly attached to the passage Of years, but to the social context the person operates within over the life—span. The variables we have chosen to explicate in the information theory of communication functions will operate with a great deal of T05 variance across the life-span. There appears to be a pattern to the variance, however. Macro and micro system factors contribute to the development of greater homogeneity in the variance in these variables and relationships in the extreme early and late periods of life, although variance in the end segments still remains rather high. Research in many areas Of gerontology has found this to be the case; while many think of old people as a homogeneous minority, there is great variation within. However, because of these macro and micro system factors, some positive correlation between functional and chronological age should occur. This functional view of aging is generally a healthy one for social science, for it avoids the development of a science that researches isolated subgroups of the population with little attention to the development of general theories. The notions of functional aging in the context of communication remove the focus from studying Old people per_se_toward a study of informational development over the life span. The implications of what has been explicated in this chapter suggest that the use of the functional vs. the chronological perspective on aging will yield differing theoretic outcomes. The thrust of the chapter can be represented in the following proposition: Control environment is a stronger predictor Of changes p : ll in communication behaviors than chronological aging. Summary The conceptual treatment of the information theory of communica— tion functions is now complete. We have progressed from the basic explication of information space, to the mediated and interpersonal T06 components of information space, the boundary demarcation, the theoretic core Of the principle of optimal information, functions of communication, control over environment, attributions of communication functions, and finally we have progressed to the integration of the theory into the context of developmental informational aging. In Part 111 this theory will be subject to operationalization, empirical test, and discussions of theory and research implications. PART III THE RESEARCH FOREWORD Having now explicated the context for the theory, and having expli« cated the information theory of communication functions, we will now present the research component of the theory construction. Chapter IX discusses the research procedures, and Operational definition of con— structs. Chapter X presents the results of hypothesis testing. And finally, Chapter XI discusses potential factors impinging on the results obtained, and develops strategies for improving the quality Of the theoretic test in future research. T07 CHAPTER IX RESEARCH PROCEDURES AND OPERATIONAL DEFINITIONS Overview In this chapter discussions will focus on the operational pro- cedures of the present research. Part A reviews research procedures, while Part B develops Operational definitions of constructs. In Part A, with regard to the sample studied, we will discuss sample selection, generalizability, the research setting, contact with the research site, and pilot testing, Further, interviewer recruitment and training will be reviewed, and then response rates in the paimary target sample speci- fied, followed by presentation Of demographic data. Part B explicates the operational propositions to be tested, and operational definition of constructs, most of which are defined through a multi-stage empirical process. A. Research Procedures Design Selection A field study is the most appropriate type of general design for testing components Of the theory of communication functions. Not only is it desirable to conduct the research in a naturalistic setting to maximize external validity, but particularly in these early phases of T08 T09 testing parts of the model and in dealing with older people as data sources it is best to gather data from them in their most regular environments. In this way the researchers are able to learn the most detail about the behaviors of this population segment. In addition, precise interpersonal communication network data are required to opera— tionalize variables in the model, and this necessitates that a whole system of individuals be censused, which is feasible only with field interviews. Given these constraints it would still be desirable to conduct field experimentation. However, manipulation of variables in the naturalistic environment is not logistically feasible, and because resources for the study are very limited, a cross-sectional survey design is chosen for the initial test of the model. Sample Selection There are two main problems influencing sample selection pro- cedures. First is the very limited number of older people concentrations in the Lansing, Michigan area. Since the project is not funded, and the time involved in screening respondents for a large area sample is great, it is desirable to locate a residential area having a high con— centration of older people, thus reducing interviewer travel difficulty. A second factor is the desire to gather network analysis data suitable for the NEGOPY network analysis program (Richards, Farace, and Danowski, T973; Richards, T974). This requires a complete enumeration of the elements of a system. Hence, a community sample is necessary that has a great deal of intra—community interaction. TTO Two communities with more than one hundred residents are located in the Lansing area. However, one Of these has only recently begun Operations and initial background interviews reveal that although resi— dents talk considerably in the apartment building, many have not learned full names or room numbers of those with whom they talk. These data are essential for network analysis. Thus, this particular site is used for three pilot tests of the instrumentation. The other community, Grange Acres, is the primary research site. The Primarngesearch Setting; Grange Acres All respondents live in a retirement community located in Meridian Township, a residential area in Greater Lansing. The retirement commun- ity is located in a somewhat open area of the township on a 58 acre site. It is relatively isolated both geographically and visually from homes in the surrounding area. There is a band of trees around the perimeter which sets off the community from the locale around it. The community was established by a non-profit corporation with national membership, with the assistance of Federal Housing and Urban Development monies. A husband and wife currently manage the facility and are responsible to a Board of Directors. The community contains 205 living units. Most are either one- bedroom or efficiency units, with a small number of two-bedroom units. The two-bedroom units are located in two apartment buildings, each Of which has two floors. All other units are one-story and each building contains approximately 6 units, which are contiguous. Rents range from $95 to $2l2 per month. Twenty-six units are vacant. TTT One hundred units were constructed in T968 and an additional l05 units were constructed in T973. The newer units are more aesthetically pleasing than the Older ones and are at the upper end of the rent scale. Each of the efficiency and one-bedroom units has a patio door Opening to the outside. Many residents engage in ornamental landscaping in the area surrounding their patio. A number of services are provided for residents Of the community. There is a two-story community building that houses the management offices, recreational facilities, work shops, and meeting and lounging rooms. There is a more heavily wooded area on the grounds designated as a "park." In the middle of the complex are shuffleboard courts. A small bus/van fondly referred to by some residents as the "Green Hornet” is used for weekly shopping trips to surrounding business areas; residents sign up for weekly transportation services in advance. There are no permanent health care facilities located on the grounds, although a visiting nurse holds open clinic hours once weekly for those desiring medical attention. The Older units are serviced by a master television antenna. However, many residents expressed strong dissatisfaction with signal reception quality, indicating that it appeared that the wiring was dis- connected. This emerged as a primary complaint against the quality of life in the community. The new units are served by a decentralized antenna system, with each group of 6 units having an individual antenna. No complaints were voiced about reception quality in these areas. The community has an established social organization composed of residents. There are a number of standing committees that coordinate TT2 various social activities in the community on a regular basis: Arts and Crafts, Visitation, Pot Luck, Bingo, and Flowers. There are no formal procedures for the selection of committee members--staffing takes place through volunteering and solicitation. Contact with the Community The initial contact with Grange Acres was made through a telephone call to the management, indicating interest in meeting with them in person. The initial visit was concerned with describing the nature and scope of the study and gathering background information about the retire— ment community. At this meeting, management agreed to participate in the research. A second meeting was held to develop a more detailed description of the research project and to gather more background infor- mation on the residents and the community. At both meetings management suggested interviewers attend a "pot luck” dinner, held each month, to become acquainted with the residents and establish rapport with them. It was indicated to us that this had been a successful approach in another survey conducted a year earlier by another group of researchers. Eight members of the research team attended the July pot luck, bringing dishes of food to be shared with the residents. At this meet— ing the author was introduced as the project director, and he in turn briefly described the survey and introduced the other study team members. Throughout the meeting, the behaviors of the residents suggested that the research group was warmly received, and that there was interest in participating in the survey. Many individuals implored that particular research team members come to interview them in their apartments. TT3 Second contact with the residents was made through attendance Of a monthly "bingo nite" by some of the research team members. As at the earlier event, research team members participated fully in the activi- ties, bringing prizes, playing bingo, and also calling bingo numbers. Following these two activities, a letter was distributed to each resident's mailbox by management, which briefly described the research and informed residents that interviewers would be calling upon them in the next two weeks for a personal interview (see Appendix). Shortly after the distribution of the letters, interviewing began. After all interviews were completed, the project director and another team member attended the August pot luck to debrief the resi- dents about the study and to thank them for their participation. This activity was cut short by the resignation of Richard Nixon as President, expected that evening, which interested a large number of residents in shortening the length of the pot luck dinner. However, the residents showed a very favorable response to our thanks for their participation in the survey. Pilot Tests Simultaneously with the initiation of contact with Grange Acres, the three pilot tests were conducted in another retirement community in the Lansing area, ”Cedar Place.” In contrast tO the primary research site, this community is housed in a building with 6 floors. There were approximately T80 residents at the time Of the pilot testing. All of these residents are relatively ”new" since the building has only been open for about 4 months. TT4 Although no systematic data were gathered to enable a comparison Of demographic characteristics with Grange Acres, the residents of Cedar Place appear to be of same or slightly lower sociO-economic status (SES) as residents of Grange Acres. This is an advantage rather than a dis— advantage since the lower SES respondents are likely to have most dif- ficulty with the interview questions and format. Hence, pilot testing with this sample is more likely to yield a questionnaire that would be usable with a wider range of respondents, since the ”lowest common denominator" is designed into the instrumentation. On other character- istics, the samples also appear to be relatively comparable. The manager's estimate of the average age in the pilot test building is 72, and in the primary survey the measured average age is 72. Both communi- ties seem to have a corresponding large percentage of women; 8T% of Grange Acres residents are women. Management of Cedar Place was cooperative in allowing ”sign-up” sheets to be placed Tn heavily trafficked areas of the building to enlist pilot test respondents. The researchers were also allowed to approach potential respondents in the building and request their participation in an interview. The sign-up sheets did not yield an adequate number Of respondents, so many individuals were approached in the public areas of the building and interviewed in the various lounging and recreation areas on the first floor. Three pilot testing phases were completed. First, open-ended focused interviews were conducted with T3 respondents. Some predeter- mined questions were asked regarding activities and media use, but most of the interviews progressed according to the particular situation. TT5 There were two basic objectives of this phase of interviewing. One Objective was to develop experience in how the Older person views the world, and to determine what the major components of the older person's life are, for use in constructing the life uncertainties scale. A second objective is more methodological in nature. We were interested in determining how the older person responds to the interview setting, and to specific question formats and scaling techniques. Each of the interviewers participating in this phase who were the core of the latter interviewing team completed detailed attitude and behavior sketches Of each person interviewed. These were reproduced and distributed, and interviewers then abstracted out the major life components of the older person as they interpreted these data. The author then compiled the interpretations of the other interviewers and used them as inputs into formulation of the scales for life uncertainties. Following the first pilot test, a draft interview protocol was drawn up consisting of potential closed-ended questions to be used in the final data gathering. Eight respondents were involved in this phase of testing. This interview format required approximately three times as long to complete as the final interview format. Unfortunately, the protocol contained many questions measuring constructs that later had to be reduced or eliminated completely. This phase allowed the testing of question wordings, sequencing, scaling techniques, and efficiency. It was determined from the second pilot testing phase that the interview protocol required far tOO much time to complete. After rough- ly half an hour, many respondents verbally and non—verbally expressed TT6 fatigue and boredom with the survey. Therefore, many of the multiple items used to operationalize constructs were dropped, and the complexity of the scaling techniques was reduced. This was particularly the case for the life uncertainties scale, which presented the greatest diffi- culty for respondents. A third pilot testing phase was initiated. The primary objective was to make final adjustments in questionnaire length and scaling techniques and to determine how to alter the wordings of specific ques- tions to achieve reasonably normal distributions. Sixteen residents of Cedar Place participated in this phase. Following this third pre- test, a final version of the questionnaire was developed and then used in the primary research site. A copy of this questionnaire appears in the Appendix. Interviewer Recruitment and Training Twelve interviewers were recruited from within the University community, and were not remunerated for their services since the study was conducted without funding. All interviewers were either under- graduates in the Department of Communication, masters degree candidates, doctoral students, or spouses of department members. Most of the stu- dents participated for credit through an independent study program. Half of the interviewing team had previous field experience in inter- viewing, and nearly all had contact with interviewing in a classroom setting, either learning how to conduct interviews or teaching others how to conduct them. TT7 A training session was held in which interviewing principles were reviewed, the specific interview instrumentation was examined in detail, and mock interviews were held. A packet of materials was distributed to interviewers including general materials concerning interviewing skills and procedures, and specific materials dealing with interviewing the elderly. In the field, each interviewer returned after the first interview and discussed it at length with the project director or field supervisor. This procedure was selected over more extensive preinter- view training because other elderly respondents were not readily avail- able for pilot testing, and the small sample size of Grange Acres re- quired that none Of these residents be used for practice interviews. Response Rate The response rate for the survey is particularly important to this research, since in order to accurately measure the interpersonal com- munication network, all respondents in the social system must ideally be interviewed. In general, the response rate for this survey is very good. Of the 205 units in the community 26 (l3%) were vacant, and T9 (9%) had double occupancy. Thus, there were l98 possible respondents for the survey. Of these possible respondents, interviews were com- pleted for T62 persons (82%), constituting the sample for the analysis. The remaining 36 persons (T8%) who were not interviewed distribute into the following groups of reasons. Three residents (2%) were hospitalized and 8 residents (4%) were too ill to be interviewed, although they were currently living in the community. The management asked that 3 persons TT8 (2%) not be interviewed because of emotional difficulties. It was determined by talking with neighbors that 5 residents (3%) were definitely on vacation, while an additional 5 residents (3%) were not at home after seven call backs. The two managers were not interviewed, and the remaining 9 individuals (5%) refused to be interviewed. A number of these refusers indicated to the interviewers that they had heard from others that the interview questions were not appropriate to the problems that Older people have. From information relayed from the interviewers and by tracing the communication network patterns through the unreciprocated links, it was determined that the bulk of these refusals were attributable to a single person who completed the interview and provided a communication contact list. This person expressed dissatisfaction with the interview, and was later seen express- ing this to a neighbor who subsequently refused to participate. Demogrephic Characteristics of Grange Acres The sex distribution in Grange Acres reflects a general demographic trend in the larger social system for persons of a comparable age range. There are 8T% women and l9% men living in the retirement community. The average age of residents is 72 years with a standard deviation of 8.3 and a range Of 50 to 94. The median education Of residents was re- ported to be l0.8 years with a range Of 3 to T9 years. The mean length of residence in the community is 23 months. With respect to transporta- tion, 54% of the residents own a car. TT9 Generalizability There are a number of problems with generalizing the results of this research to Older people in general. Most obvious is the type Of residential living experienced by the sample. Most older people in this country do not live in areas with very high older people concentra- tions, and where they do, most of these areas are multi—story buildings, rather than aggregates of single story units. Since a listing of the population of similar retirement communities is not feasible, there is no evidence that this community is representative. Another significant factor is the selection process by which indi— viduals become residents of the community. First, individuals decide to locate in the community, based on their own perceptions and atti- tudes. There may be factors operating that attract only a specific type of resident. A second consideration is that management also exer- cises selectivity in screening interested persons for residence. It was stated to the researchers that people must be able to care for them— selves and be mentally alert to be allowed to reside in the community. It was also noted that there were no minority residents, a phenomenon that was not Observed in the pilot testing site, which contained Blacks and Spanish—speaking persons. Thus, management exercises its own selec- tion criteria to at least the stated degree in determining who will reside in the community. In short, considering these factors the generalizability of the research is limited. However, since the interest is not in describing the characteristics of a population, but in testing for the relationships T20 of a number Of variables at an individual level of analysis, generaliz— ability is Tess critical, particularly since this is a new area of research requiring a great deal of future development. Propositions to be Tested Before moving to an explication of the operational definitions of variables in the theory, it will be useful to review the propositions to be tested: H]: The greater the uncertainty discrepancy at a point in time, the greater the amount of subsequent communication (mediated and interpersonal). The greater the uncertainty overload at a point in time, the less the entropy of subsequent communication relationships (mediated and interpersonal). The greater the uncertainty underload at a point in time, the greater the entropy of subsequent communication relation- ships (mediated and interpersonal). The greater the control over environment, the less the uncertainty discrepancy. The greater the control over environment, the greater the interpersonal communication. The greater the control over environment, the less the media exposure. The greater the control over environment, the less the attribution of communication functions to interpersonal com- munication and television. The greater the control over environment, the greater the relative attribution of communication functions to inter- personal communication than to television. The greater the overload at a point in time, the greater the subsequent attribution of communication functions to uncer- tainty reduction. T2T 10: The greater the underload at a point in time, the greater the subsequent attribution of communication functions to uncertainty expansion. H1]: Control over environment is a stronger predictor of changes in communication behaviors than chronological aging. It will be noted that there are minor adjustments in hypotheses 7 and 8: 'mediated' is substituted by 'television.' This is done because both time constraints and respondent interest could not sustain the asking of the same set of functions questions for four or five different media. Since television has been shown to be the most important leisure activity for Old people (see Chapter II), there is strong justification for the selection of this particular medium. B. Operational Definitions In the following sections the Operational definitions of variables used in the analysis will be presented. Most of the variables are con— structed out of sets of component items, therefore we report descriptive statistics, inter—item correlations, and factor analyses as they are appropriate to the Operational definitions of the constructs. Uncertainty Scale In the theoretic section the construct of uncertainty discrepaney was discussed. Uncertainty discrepancy is a function of the difference between the preferred level of uncertainty and the experienced level of uncertainty. Experienced uncertainty may be within the preferred range, yielding no discrepancy, it may be pelop_the preferred range, or it may be above the preferred range. T22 In operationalizing this construct an attempt was made to create a close correspondence between the theoretic definitions and the Opera— tional procedures. A scaling technique was developed that asked respond- ents how certain they were about a particular dimension of the life experience, and once this judgment was made the respondent was asked to indicate how certain he would like_to be on this dimension. For example, it might be asked: How certain are you about how independent you are? How certain would you like to be about how independent you are? In the pilot tests this kind of scaling created a great deal of difficulty and confusion for the respondent; Often the latter question would produce noticeable stress in the respondent, who would Often be un- able to answer the question or would not understand it. After testing a number of variations Of these kinds Of questions, it was decided to scale the construct in a simpler but less precise way, having the respondent reply to an assertion such as "I would like to be more sure I ani an independent person." The responses were scaled with seven point Likert scales of the agree-disagree type. This type of scaling was easier for the respondent to cope with, although some of the items remained difficult for respondents. The primary disadvantage of this technique is that it less directly measures uncertainty discrepancy. Discrepancy must be inferred from the response to a single item rather than two. The respondent who agrees with the item above is assumed to exhibit a positive uncertainty discrepancy, varying with the magnitude of the agreement. Those disagreeing with the item are assumed to exhibit a negative discrepancy, increasing with the l23 magnitude of the disagreement. Neutral responses are taken to indicate a balanced state, the respondent judging that the current state of affairs is preferred. However, in actual interviewing there was a tendency for respond- ents to avoid selecting the neutral position on the scale. Thus, measured values may to some extent be a function of the scale and not the indexed subjective state. Also, "neither agree or disagree" did not appear to be a common reflector of the manner in which the respondents cognized their attitudinal states. Judging from responses it appeared that when individuals were satisfied with the current state, they tended to slightly disagree with the statement. While this introduces some measurement error, it is unlikely that the technique is completely in- valid. Furthermore, the correlational measurement of relationships with this variable is likely to reduce the magnitude Of problems with the non-alignment because of the relative nature of the techniques. In sum, while the technique has somewhat questionable reliability, it is proposed that it has acceptable, although not strong, face validity. The dimensions used to construct the uncertainty scale were developed through the exploratory pilot testing; a set of key life dimensions was abstracted. Since this determination involved consider- able subjective judgment by the author, to eliminate some possible bias each pilot interviewer abstracted his/her own list of key life dimensions. These were then compiled by the author and formed a basis for determining the overall set of generalized dimensions to measure uncertainties. These generalized dimensions are represented in the following items used in the research: T24 . I would like my daily activities to be less usual ... less routine. . I would like to be much more sure that I will visit with my family often. . I would like to be much less sure about what lies ahead in the near future. . I would like to be very much more certain that I will have transportation when I need it. . I would like to talk more to people different from me. . I would like to be much more sure I could depend on others I know. . I would like to be much less sure that my health will be the same in six months as it is now. The means and standard deviations for these items appear in Table l, indicating that the distributions are reasonably normal. Table l. Descriptive Statistics on Uncertainty Items Variable Mean Standard Deviation (T) activity 4.24 T.67 (2) family 3.57 l.7l (3) future 4.27 T.67 (4) transportation 3.45 2.T7 (5) people 3.29 T.6l (6) independence 4.07 l.83 (7) health 4.28 l.89 Regarding the relationships among the items, it is not theoretical- Ty expected tnat these items inter-correlate highly, since they are measuring uncertainty discrepancy in a widely varying set Of areas. T25 If there are in fact high correlations this would indicate that uncer- tainty discrepancy is a highly situationally independent phenomenon, suggesting that it is a "personality variable." However, it is proposed here that it is a variable arising out of direct on-going interactions between individual and environment. Since environments are typically highly differentiated and ”compartmentalized," the overlap among the items should be low. As expected, the inter-item correlations in Table 2 show an overall pattern of relatively low interrelationships among items; the largest inter—item correlation is 0.37. Table 2. Inter-item Correlations for Uncertainty Items (l) activity T.00 (2) family 0.27 T.00 (3) future 0.33 0.2T T.00 (4) transportation 0.T5 0.32 0.T8 l.00 (5) people 0.T6 0.T3 0.0T 0.2T T.OO (6) independence 0.23 0.37 0.T8 0.37 0.22 l.00 (7) health 0.22 0.T2 0.T5 0.T2 0.07 0.22 T.00 T 2 3 4 5 6 7 This consideration provides evidence for the theoretic assumption that the source of uncertainty discrepancy does not influence the nature of the relationships among uncertainty discrepancy and the other variables in the model. Therefore, the uncertainty discrepancy indices are formed from the linear combination Of the seven items. However, there are in effect two variables contained within each Of the component scales. As was discussed previously, an agreement T26 response to a question indicating a preference for more uncertainty is an indicator Of underload, while a disagreement response is an indicator of overload. The strength of the agree-disagree response is a measure of the magnitude of the discrepancy. Since in the analysis we want to represent both the underload and overload conditions without the con- founding of one by the other, two separate indices are formed: an over— load and an underload index. After appropriately reversing some of the scales, the overload index was created by assigning a value of zero to a neutral through very strongly agree response to an item indicating that more uncertainty was preferred; the three levels of disagreement were coded as one through three, yielding an overload variable taking on four possible values: 0, T, 2, 3. Each respondent's value on each of the seven items was summed to form the overload index. In a parallel fashion, the uncertainty under- load variable was formed by assigning a value of zero to all responses ranging from neutral to strongly disagree, and values of one through three to the corresponding levels of agreement. The scores resulting from this recoding for each item were summed together to yield the uncertainty underload index. As the value of each of the over- and underload variables increases, the magnitude of the discrepancy between preferred and experienced uncertainty increases. Control Over Environment Below are the items designed to measure the construct of control over environment. They are the remaining items from the pilot testing Of twelve items designed to operationalize this construct. T27 The author adapted some items from Rotter's (T966) scales, modifying the wordings, and also wrote original items. They are also seven point Likert scales: —l . I usually plan my daily activities in advance. 2. When I make plans ... I am almost certain I can carry them out. 3. It is not easy to change my life ... too many things stand in the way. 4. Most things that happen to me are just a matter of luck. In Table 3 the descriptive statistics for these items appear. The variables appear to be relatively normally distributed. Table 3. Descriptive Statistics for Control Over Environment Items Variable Mean Standard Deviation (l) Plans 4.05 l.94 (2) Carry out plans 2.75 T.6l (3) Change 3.79 l.8l (4) Luck 4.50 l.72 Table 4 (on the following page) shows the inter-item correlations among the control items. The correlations range from low to moderate, generally indicating rather low reliability. A Factor Analysis (principal components with varimax rotation) was performed on these items, shown in Table 5, resulting in a factor structure that is moderately clean in distinguishing the planning items and the change and luck items. It was assumed in the construction of T28 Table 4. Inter-item Correlations Among Control Over Environmental Items (T) Plan T.OO (2) Carry out 0.23 l.00 (3) Change -.02 0.T0 l.00 (4) Luck -.05 0.T0 0.33 T.00 T 2 3 4 Table 5. Factor Analysis of Control Items Variable Factor l Factor 2 Plan 0.09 0.48 Carry out 0.T5 0.49 Change 0.56 0.04 Luck 0.58 0.0l these scales that the planning items were indicators of control over environment, but less central ones than the change and luck items. During pretesting and final interviewing, it became apparent from observ- ing respondents' reactions to the questions that the planning items elicited unexpected responses. Judging intuitively and by responses to the change and luck items, it appeared that some individuals with high control over environment responded in agreement with the planning items, while a roughly corresponding proportion responded oppositely. It appeared that in many cases, respondents with judged high control perceived planning to be a burden to a spontaneous and high quality Of life. These considerations appear to be reflected in the factor T29 structure. Thus, control over environment is operationalized by the factor scores for the first factor, which has highest loadings for the change and luck items. Network Analysis The operationalization of the interpersonal communication network variables was a multi-stage process. Two general types of interpersonal communication data were gathered: intraecommunity communication, and communication with persons residing outside the community. First, the intra-community procedures for operationalizing variables will be dis— cussed. Initially the overall process will be summarized, then the component activities will be reviewed in greater detail. The primary data used to enable the calculation of network vari- ables was specific dyadic data on "who talks to whom." This information was subjected to a computerized network analysis procedure, that in— corporates the calculation of network variables in the analysis process. These variables were taken directly from computer output. Respondents were asked to name individuals in Grange Acres with whom they communicate through telephone or face—to—face modes on the average Of once every two weeks or more often. After a name was identi— fied, the respondent was asked to estimate the frequency with which communication took place. Categories used were: several times a day, once a day, two to three times a week, and once or twice every two weeks. When respondents could not recall a full name, they were asked for as much detail as possible about the person. When there was any doubt, a primary piece of information asked Of respondents was the apartment T30 number of the communicatee or as precise a description of the apartment location the respondent was able to give. Although most individuals could remember full names and also volunteered apartment numbers, a full range of permutations of identifying data were gathered regarding the completeness Of names and the geographic location of the contactee. When contacts had incomplete identifying data, the author "triangulated" from available data to identify the specific individual under question. While this was a difficult process, all but three contacts were able to be identified from the secondary identifying information provided. These unidentified contacts were eliminated from the individuals' con- tact lists and analysis was based only on completely identified links. A small number of respondents (3) overtly refused to disclose their communication contacts. These communication network data. were coded according to re- spondent numbers and frequency values and processed by the NEGOPY com- munication network analysis program. The NEGOPY program is essentially a heuristic pattern recognition procedure that identifies the network structure of the system under investigation at two primary levels of analysis-~individual and group level (Richards, Farace, and Danowski, l973; Richards, T974). At the individual level summary data is com— puted and presented that indexes the individual's structural character- istics: variables such as number of links, link strengths, reciproca- tion, integration and so on. At the more macro-level the program identifies sub-groups in the system and the linkages among them, along with summary statistics regarding the structural properties of the network at this level of analysis, such things as group intraconnection, T3T interconnection, role positions in the network, and so on. In this research the concern is with the individual level of analysis and therefore, data from the network analysis computed at this level Of analysis will be reported. A problem in most network analyses in which the individual is asked to self-report on communication behaviors is reciprocation. A link is unreciprocated when a person names another who does not name the individual in return. Thus, it is unclear if the reciprocation is a result of oversight, the result of major differences in the percep- tions of the individuals, or perhaps the result of conscious fabrica- tion. The latter is most likely to occur in an organization where in- dividuals are rewarded and punished for whom they communicate with, and most of these organizations are "work“ organizations. This is not likely to be as serious a problem in the community network with which we are concerned here. However, the first two explanations for unre- ciprocation are more applicable. While there is no empirical data to support this proposition, it is likely that the misperception of com- munication relationships is a much less serious cause of unreciproca- tion than problems of memory. For these reasons all links were used in the analysis in which either both individuals named each other, or one person named another who did not reciprocate the nomination. The network analysis program allows this Option in the analysis of data. Through these procedures, even the individuals who did not name any communication links are given network variables according to the pat- terns Of contacts that others name with them. T32 In addition to intra-community links, the extra«community com— munication links of each respondent were also measured. Respondents were asked: “How many people do you talk with ppr_living in Grange Acres ... at least once or twice every two weeks on the average?“ While for intra-community networks frequency of contacts were measured, for extra—community links only the number Of persons with whom the respond- ent communicated at least once every two weeks were measured. Since the NEGOPY network analysis program could not be used with these data, there was little value in making more precise frequency discriminations for this research. Interviewers were instructed to probe for communica- tion with family, friends, and others, through mail, phone, and face— to-face modes. The resulting measure of extra—community contact is the pepper_of persons the focal person communicates at least once every two weeks. The variables. Two interpersonal communication network variables are used in the analysis: one indexes amount of communication with others, while the other indexes the zone integration of the respondent, thus forming the basis for inferring the content entropy of information flowing through the interpersonal communication network. The amount Of interpersonal communication measure is formed from both the intra-community and the extra-community linkages. To construct the index, the total number of internal and external links are summed. Frequency is not entered into the calculation for two reasons. First, all links occur at a frequency of greater than once every two weeks, contributing to low variance in frequency across the reported links. Second, no more precise frequency data were asked of respondents for the T33 extra-community links beyond an enumeration of those persons whom they communicate with at least once every two weeks. Thus, the interpersonal communication measure reflects the total number of persons the target respondent communicates with on the average of once every two weeks or more Often. Table 6 shows descriptive statistics for the two subcomponents Of the amount of interpersonal communication variable: intra- and extra- community links. It can be seen that there is greater interpersonal communication with the surrounding environment than within the community. Table 6. Descriptive Statistics for Intra and Extra-community Links Variable Median Mean S.D. Range Intra links 6.l0 6.93 0.34 0 - 20 Extra links 8.00 l2.42 l3.90 O - 98 Content entropy of interpersonal communication is indexed by measuring the zone integration of the first-order interpersonal network zone. A more direct measurement of this construct was attempted in the second pilot test, but time factors necessitated that the items be dropped in later survey phases. As was pointed out in Chapter IV, zone integration provides a basis for inferring the entropy of message types flowing in the network as well as the individual's ability to predict the nature of interpersonal communication over time. As the integration of the zone increases, the content entropy decreases. Zone integration T34 was calculated by the NEGOPY network analysis computer software, using the following formula: 2 (number of links) 2': n(n-l) where Z1 is the zone integration value, and n is the number of nodes in the zone. The values for this variable range from 0.00 to l.00, with the low values indicating a low degree of zone integration, and the high values indicating high integration. A person with a highly integrated network zone has communication contacts who also talk to one another, while the person with low integration has contacts who do not communicate with each other. Media Exposure Items It was felt that the measurement of media exposure ought to be as precise as possible, since theoretic and Operational precision are intimately interlinked. Therefore, items indexing media exposure were designed to enable the subsequent determination Of the approximate number Of minutes per month that the respondent exposed to each of the media. Correspondingly, respondents were asked for judgments of the Tcrequencies of exposure as well as the duration of exposures. These two factors are used to construct the indices of time spent using each mEdiLnn per month. Below are the items: l. How often do you get a chance to read a daily newspaper ... like the Lansipg_State Journal, or the Detroit Free Press? 2. When you do get a chance to read the daily newspaper ... on the average ... how many minutes do you spend reading? 9. T0. T35 . On the average day ... how many hours do you watch TV between six in the morning and noon? . On the average day ... how many hours do you watch TV between noon and seven in the evening? . 0n the average day ... how many hours do you watch TV between seven in the evening and two in the morning? . In an average day ... how many hours do you have the radio on? . How often do you get a chance to read a book? . When you do read a book ... on the average ... how many minutes at a time do you usually read? How Often do you read a magazine? When you do read a magazine ... how many minutes at a time do you usually spend reading? Each respondent receives a number of minutes exposure to each medium per month. Then, each medium score is summed to yield an overall total media exposure score. Table 7 provides descriptive statistics on media exposure values. Values are reported in terms of average number of hours per day. Table 7. Media Exposure Descriptive Statistics (hours per day) Variable Mean Standard Deviation Range (l) newspapers 0.75 0.75 0 - 4.00 (2) television 6.l0 3.66 0 — l7.00 (3) radio 2.T8 3.l3 0 - T6.63 (4) books 0.48 0.80 0 - 5.00 (5) magazines 0.T9 0.28 0 - l.50 T36 The average respondent spends three quarters of an hour per day reading a newspaper, roughly six hours per day watching television, about two hours per day listening to radio, approximately three quarters of an hour per day reading books, and about twelve minutes per day read— ing magazines. Media entropy measures. Two mediated entropy variables are used in this research: media channel entropy, and television content entropy. Both are operationalized in identical ways, once the basic raw data is transformed. The following information theoretic formula is used to operationalize both variables in the research through use of a FORTRAN computer program designed for general entropy calculations: H = '2 P1 '092 p1 rel Tog2 N where Hrel is the relative entropy value; pi is the relative probability of exposure to an alternative media component; log2 is the basic func- tion measuring number of bits Of information; and N is the number of alternatives in the larger set. The result of this Operation is a value ranging from 0.00 to T.00, indexing the amount of entropy in the particular exposure pattern. The low values indicate low relative entropy (or high relative negative entrOPY), the high values indicate high relative entropy (or low rela- tive negentropy). These values are the amount Of entropy in the set of alternatives relative to the maximal amount possible, if all alternatives were equally used. T37 For the calculation of the media channel entropy index, the number of minutes of exposure per month per each medium are calculated for each respondent. These values are used to compute the respective relative probabilities of exposure to each medium through dividing each exposure value for a medium by the total exposure value. In calculating the television content entropy measure, the amount of exposure to each of the following program types is indexed: game shows, soap operas, situation comedies, crime and detective shows, movies, news, and sports programming. Relative exposure to each was measured with Likert type scales with the following response alternatives: never, not often at all, not very often, somewhat often, and very often. The respondent's scores for each item are totalled and then the value for each respective item is divided by this total, yielding a probability value of exposure to each content type. These probabilities then form the basis for the calculation Of the television content entropy measure using the above formula. Functions of interpersonal communication. The following questions form the basis for the construction of indices to measure functions of interpersonal communication. The items were originally written by the author* and these particular wordings evolved through the pilot testing. All are seven—point Likert scales: T. Very often I talk withpeople to relax. *Many of the ideas underlying the development Of the scales to measure functions of interpersonal communication and television use are the result of the author's participation in a seminar on the uses and gratifications approach in mass communication research, offered by Professor Bradley S. Greenberg, Department of Communication, Michigan State University. T38 2. Talking with people very often helps me to calm my nerves. 3. Often I get bored and look for someone to talk to. 4. I often like to talk to people that are highly exciting. 5. Talking with people very often helps me to take my mind off things. 6. Often I get to thinking too much and look for someone to talk to. 7. If it weren't for talking to people, I wouldn't know nearly as much as I do today. 8. I like talking to people because I very Often hear new things I haven't heard before. Table 8 shows the means and standard deviations for these items indicating they are generally very normally distributed. Table 8. Descriptive Statistics for Functions Of Interpersonal Communi- cation Items Variable Mean Standard Deviation (T) relax 3.l3 T.62 2) calm 3.33 l.7l 3 bored 3.85 l.97 (4) excite 4.T3 l.8l (5) mind-off 2.97 T.48 (6) thinking 3.9T l.79 (7) knowledge 3.00 T.65 (8) new things 2.73 l.50 Table 9 shows the inter—item correlations among these variables, and Table l0 shows the results of a factor analysis. As inspection will reveal, the factor structure is not consistent with theoretic explica— tion. The relax, calm, mind-off, and new things items all Toad highly T39 Table 9. Correlations Among Functions of Interpersonal Communication Items (l) relax T.00 (2) calm 0.58 T.00 (3) bored 0 28 0.32 T.OO (4) excite 0.T4 0.T9 0.24 T.00 (5) mind-off 0.44 0.53 0.27 0.32 T.00 (6) thinking 0.30 0.32 0.73 0.22 0.40 T.00 (7) knowledge 0.32 0.35 0.38 0.30 0.27 0.47 T.00 (8) new things 0.35 0.47 0.3T 0.35 0.4T 0.4T 0.67 T.00 T 2 3 4 5 6 7 8 and cleanly on factor one. The bored and thinking items load high and clean on factor two, while the excite and knowledge items do not load well on either factor. From the theory, the relax, calm, mind—Off, and thinking items should cluster together on a single factor. Items actually loading on factor one are consistent with this expectation, except that thinking does not load when it should, and new things does when it should not. A second cluster of items should develop including the bored, excite, knowledge, and new things items. Actually, of these, only the bored item loads as expected, while knowledge and excite items load on neither factor, and the thinking item unexpectedly loads on factor two along with bored. (See Table TO, on the following page.) This suggests that the functions constructs are either not properly operationalized, or that the concepts themselves are inappro— priate. It will be assumed until further research that the former is more plausible. While the particular results of the factor analysis T40 Table l0. Factor Analysis of Functions of Interpersonal Communication Items Variable Factor l Factor 2 Relax 0.63 0.T6 Calm 0.76 0.T5 Bored 0.T9 0.75 Excite 0.30 0.24 Mind-Off 0.6T 0.23 Thinking 0.24 0.84 Knowledge 0.45 0.45 New things 0.60 0.36 will be important later when a composite set of indices Of functions cutting across both interpersonal and television components is con— structed, for the test of the hypotheses concerned with relative attribu- tions of functions to discrete mediated and interpersonal components the factor structure is less important. These hypotheses call for an over- all measure Of the extent to which functions are attributed to inter— personal communication in general. Therefore, the functions items are summed to form an index of the extent of attribution Of functions to interpersonal communication. This will develop the variable at the appropriate level of abstraction. After discussing the operationaliza- tion of the television functions index we will return to a consideration of this factor structure in light of the findings for television. Functions of television. In measuring the functions of television parallel questions to those asked of interpersonal communication were developed and asked of respondents. These are also seven point Likert scales: T4T . Very often I watch television to relax. . Television very often helps me to calm my nerves. . Often I get bored and turn on the TV. . I often like to watch TV shows that are highly exciting. . TV very Often helps me to take my mind Off things. . Often I get to thinking too much and turn on the TV. . If it weren't for TV ... I wouldn't know nearly as much as I do today. . I like to watch TV because very Often I see new things I haven't seen before. Table ll shows the means and standard deviations of these items. As with the previous measures they are generally very normally dis- tributed. Table TT. Descriptive Statistics for Functions of Television Items Variable Mean Standard Deviation (l) relax 2.83 T.66 (2) calm 3.66 l.87 (3) bored 3.98 l.96 (4) excite 4.4T 2.00 (5) mind-off 3.l4 T.73 (6) thinking 4.25 l.82 (7) knowledge 2.98 l.8l (8) new things 2.63 l.46 Inter—item correlations appear in Table l2, and results of a factor analysis in Table T3. The factor structure does not correspond well to theoretic expectations. The relax, calm, bored, mind—off, and T42 Table l2. Inter-item Correlations Among Functions of Television Items (l) relax T.OO (2) calm 0.63 l.00 (3) bored 0.42 0.527 T.00 (4) excite 0.23 0.28 0.24 l.00 (5) mind-Off 0.54 0.66 0.49 0.35 T.00 (6) thinking 0.34 0.46 0.60 0.30 0.55 T.00 (7) knowledge 0.22 0.28 0.23 0.T3 0.30 0.29 T.00 (8) new things 0.38 0.29 0.28 0.T9 0.32 0.29 0.63 T.00 T 2 3 4 5 6 7 8 thinking items load above 0.50 on factor one, and have their highest loadings on this factor. The knowledge and new things items are the only items loading high on factor two. The excite item does not load well on either factor. Considering the overall pattern, the learning function items both appear to constitute factor two. The remaining functions items, with the exception Of the excite item, constitute factor one. It would be expected from the theory that the learning items would cluster with the stimulation items——bored and excite--to form the general uncertainty expansion function, and the relax, calm, mind—off and thinking items would cluster together to form the general uncertainty reduction func- tion. Empirical incongruities arise in that the excite item does not load with the learning items and the bored item loads with the uncer- tainty reduction items. Thus, while it appears that there is a proper trend in the relationships among the variables, it is not a clear vali- dation Of the proposed functions definitions. l43 Table l3. Factor Analysis of Functions of Television Items Variable Factor l Factor 2 Relax 0.63 0.22 Calm 0.78 0.l7 Bored 0.66 0.l6 Excite 0.37 0.ll Mind-off 0.78 0.l9 Thinking 0.64 0.20 Knowledge 0.l8 0.74 New things 0.24 0.78 This factor structure is not important in the context of the tests of hypotheses concerning control over environment and the degree of attribution of communication functions to television and interpersonal components and the relative attribution to interpersonal versus tele- vision components, since these hypotheses call for an overall measure of the degre§_of attribution to the respective components. Thus, all tele- vision functions are combined in a linear fashion to yield an overall index of attributions of functions to television. The factor structure is more important, however, in the hypotheses regarding overload and underload and the attribution of communication to uncertainty expansion and uncertainty reduction functions. For this index, the appropriate items from the television functions and the interpersonal functions must be combined into overall uncertainty reduction and uncertainty expansion across both components. While the television factor structure is not strongly aligned with theoretic expectations in a particular way, the interpersonal functions factor structure is unaligned in still other l44 ways. Thus, the items do not appear to measure the constructs in a Consistent and reliable manner. Two basic strategies for analysis are suggested by this observa- tion. One strategy is to abandon testing of hypotheses involving the uncertainty expansion and reduction variables, making the assumption that the level of reliability is sufficiently low to argue against the use of these variables. The second strategy is to go ahead with the test of the particular hypotheses, combining the items in theoretically specified ways, while keeping in mind the unreliability of the scales. Because this is exploratory research in this area, it was decided to pursue the second strategy. The hypothesis testing may shed light on theoretic relationships, even though the scales are not highly reliable. It is assumed that a weak test of the hypotheses is more useful than no test at all. The scales formed to measure uncertainty expansion and uncertainty reduction across both interpersonal and television components are as follows. Relax, calm, mind-off, and thinking items, for both television and interpersonal communication, are summed together to form the uncer— tainty reduction index. The bored, excite, knowledge, and new things items for both television and interpersonal communication are summed to form the uncertainty expansion index. Socio-economic Status Socio-economic status is operationalized from three subcomponents: education, occupation, and quality of the respondent's furnishings. Education was measured by first asking the respondent what the name of l45 the last school attended was, to cue the person for the following ques- tion: "How many years of school did you complete?" The respondent was then asked for the occupations that he/she held, and also the occupations held by the spouse (or former spouse). Each of the listed occupations were coded according to the Hollingshead Occupational Index. In cases where there are more than one occupation held over the lifetime, the average of the component occupational scores is taken. For respondents having spouse occupation(s) listed, the highest value between the two is taken to represent occupational status. Quality of household furnishings was measured by having the interviewer code his/her evaluation of qual- ity on a l through 5 point scale. For each of these subcomponents, the scores are standardized to yield ”Z-scores.” These respective Z-scores are then summed together to yield the socio-economic status index. Ass Age is operationalized by the following question: "May I ask your age?” Actual number of years was coded without grouping into categories. Summary In Part A of this chapter, sample selection, the research setting, contact with the research site, and pilot testing were discussed. Also discussed were interviewer recruitment and training, response rates in the primary sample survey, basic demographic characteristics of the primary sample and generalizability. Part B provided operational defini- tions for the variables to be used in the test of hypotheses. In the following chapter, the hypotheses are subject to empirical testing. CHAPTER X RESULTS Overview In this chapter the overall objective is to present the results of the empirical testing of the operational hypotheses specified in the previous chapter as subject to testing. First, analysis procedures will be briefly overviewed, second, descriptive statistics and inter- item correlations will be reported, and third, the actual presentation of research results will occur. While in the previous chapter, descrip- tive statistics and inter—item correlations were reported for the particular individual items used in constructing the indices, in this chapter the descriptive statistics and inter—item correlations for the final indices will be presented. Hypothesis testing follows these data. Analysis Procedures The hypotheses are tested using Pearson simple correlations, and partial correlations. However, before deciding to use linear analysis models, scattergrams were plotted for all zero-order relations among variables. No evidence of curvilinearity was found for any pair of variables. Thus, it was felt appropriate to use linear assumptions about the relationships in further data analysis. This was in some ways a meta—validation of the overall theoretic approach, for the hypotheses l46 l47 originally specified linear relationships, implicit in the working of the propositions. All analyses reported were performed on an SPSS canned statistical analysis package, roughly half on a CDC 6500 and the remainder on an IBM 370. Descriptive Statistics and Inter-item Correlations Before moving to tests of the hypotheses, the descriptive statis- tics for the final indices will be presented, as well as their inter- item correlations. In the last chapter this kind of information was presented only for items that would be combined into indices, while in this chapter the values for the resultant indices are presented. Table l4 shows the descriptive statistics for indices used in testing hypotheses. The maximal range for each variable is included to provide a context for the reported means and standard deviations. It will be noted from inspection that the variables are generally normally distributed. Two significant exceptions are the media entropy measures: the channel entropy index and the television content entropy index. The channel entrOpy index is very close to the bottom of the possible range, indicating very low entropy for exposure to alternative media channels. This suggests that there is little equality in exposure across media for the average respondent. The large proportion of time spent viewing television accounts for this phenomenon. The television content entropy measure is skewed in the opposite direction of the channel entropy measure, indicating very high entropy in exposure to types of television programs. Individuals expose 148 Table l4. Descriptive Statistics for Indices Standard Min.-Max. Variable Mean Deviation Possible Overload 6.45 3.46 O - 2l Underload 5.47 3.l2 O - 2l Control 3.48 l.3l l - 7 Interpersonal communication l9.35 l5.26 O - Integration 0.30 0.25 0.0 - l.0 Media exposure 9.7l hr/da 4.70 hr/da O - 24 TV exposure 6.l0 hr/da 3.66 hr/da 0 - 24 Channel entropy 0.l9 0.l4 0.0 - l 0 Content entropy 0.94 0.04 0.0 - l.0 Interpersonal functions* 27.l5 9.l0 0 - 56 TV functions* 28.22 l0.26 0 - 56 Uncertainty expansion* 27.96 8.67 0 — 56 Uncertainty reduction* 27.40 9.2l O - 56 Socio-economic status 2.35 2.87 -9.0 - +9.0 Age 72.l5 8.3l —————-— *The larger the value, the less the extent of attribution. (As the values of all other variables increase, the greater the amount of these variables.) l49 relatively equally to the range of alternative program types measured in the present research. Thus, it may be the case that either the old person exhibits very low selectivity in exposure to television, or that the old person consciously decides to select programs relatively equally across the categories. It cannot be determined through data gathered in this research whether the individual merely leaves the television on a particular random channel all day, thus indirectly exposing to a wide range of programs relatively equally, or whether the individual actively selects programs that vary in type of content. Table 15 reports the inter-item correlations among indices. Many of these relationships will be the focus of hypothesis testing to follow, and will be examined in detail through zero-order and partial correlations. All interrelationships among variables are reported here to summarize the patterns in the data, and provide the reader with additional reference for non-hypothesized links among variables. Tests of Hypotheses H]: The greater the uncertainty discrepancy at a point in time, the greater the amount of subsequent communication (mediated and interpersonal). To test this hypothesis, four comparisons are necessary: under- load with amount of mediated communication, underload with amount of interpersonal communication, overload with amount of mediated communi- cation, and overload with amount of interpersonal communication. Although not specifically hypothesized, television exposure will also be examined parallel to the total media exposure variable. 150 mF wF mF NF oo.F mo. mF.u FF.: oo.F NF.u mo.- oo.F mm.. FF oF m w FF.- oF.u mm.s FF.: oo.u Fo.u #0.: 00. mm. on. em. FF.u oo.F mm. Fm. Fm. wo.u oo.F mv. mm. mo.- oo.F om. mF.u oo.F #0.- oo.F m m Fm.s mF.u FF.- «0.- mo.: No.: we. ow. Kw. FF. oo. oo.F mm. mF. Fm. 0F. mm. FF.I mm. oo.F m q FF.: oo. wF. mo. mF. no. .mo. mo. do. 00. oo.F mF.u ¢o.- FF.- mo. m0.. m0.: mo. oF.u oo.F m mo.u mF. FN.u wF.u ow.- NF.- mo.u No. wF.u mo.u FF. oo.F N No.1 No.- mF. NF. oF. FF. vF. Fo.u FF. mF. Fo.u Fo.- Fo. oo.F F 00. F0. «0.- mo. «0.- mo.- NF. 00. #0.: no.. mF.- FF.- oo. oo.F mam mzpmum oFeocoumquUOm :onusme xchmmeoc: :ochmaxm Achmmeocz mcon023F >F mconoczm FmCOmLmqgmch xaogpcm pcmwcoo >F qugpcm chcmgu mgzmoaxm >F mesmoaxm mFume :onmmech conmochesoo chomcmacmch Fogpcoo umoFLm>o umoFLmuc: mmoFucF mcoe< mconmFmLLou .mF anmF l5l Underload correlates 0.04 (p<:0.3l) with amount of total media exposure and 0.00 (p<:0.47) with television exposure. Controlling for SES, there is no change in the correlations. Underload correlates —.l5 (p<:0.03) with amount of interpersonal communication, which is in the opposite direction predicted. Controlling for SES the relationship remains unchanged. Thus, for underload the hypothesis is not supported; the associations with media exposure are negligible and the correlation with amount of interpersonal communication is not in the direction pre- dicted. Examining the overload relationships, overload correlates 0.l5 (p<:0.03) with amount of total media exposure, while television exposure correlates 0.l7 (p<:0.0l). Partialing for SES introduces no change in the magnitude of the associations. The correlation between overload and amount of interpersonal communication is 0.0l (p< 0.45), and again the correlation is maintained when SES is partialed out. Thus, for overload there is a significant correlation with media exposure in the predicted direction, but a zero correlation with amount of interpersonal communi- cation. Considering the four comparison points, the hypothesis is sup- ported only for the relationship between overload and amount of media exposure. H2: The greater the uncertainty overload at a point in time, the less entropy of subsequent communication relationships (mediated and interpersonal). To test this hypothesis the relationship between overload and three measures of entropy will be examined: media channel entropy, television content entropy, and zone integration, a measure of interpersonal con- tent entropy. The correlation between overload and media channel l52 entropy is -.07 (p<:0.20) between overload and television content entropy is 0.l4 (p<:0.04) and between overload and interpersonal content entropy is -.0l (p<:0.45). Partialing for SES introduces no change in the values of these associations. The overload/entropy relationship is in the pre— dicted direction, but weak and insignificant. The overload/television content relationship is not in the direction predicted. The overload/ interpersonal content entropy relationship is essentially zero. Thus, overall the hypothesis is not supported. H3: The greater the uncertainty underload at a point in time, the greater the entropy of subsequent communication rela— tionships (mediated and interpersonal). This hypothesis will be tested similarly to the previous one. The correlation between underload and media channel entropy is 0.l2 (p<:0.06) between underload and television content entrOpy is -.05 (p<:0.27) and between underload and interpersonal content entropy is -.07 (p<:0.l8). Controls for SES introduce no changes in the correla- tions. Thus, the underload/media channel entropy relationships is in the predicted direction, but rather weak. The underload/television con- tent entropy relationship is very weakly in the opposite direction pre- dicted, as is the interpersonal content entropy/underload relationship. Hence, the hypothesis is not supported. H4: The greater the control over environment, the less the uncertainty discrepancy. This hypothesis calls for the examination of the relationship be- tween control, and underload and overload. The correlation between control and underload is -.ll (p<:0.08) while the correlation between control and overload is 0.0l (p<:0.43). Partialing for SES introduces l53 no change. Thus, there is a trend in the predicted direction for the control/underload component, but no relationship for the control/ overload component. Therefore, the hypothesis is not supported. H5: The greater the control over environment, the greater the interpersonal communication. Control correlates 0.l7 (p<:0.0l) with amount of interpersonal communication. Controls for SES increase the correlation minutely to 0.l8. Controls for uncertainty discrepancy introduce no change. Hypothesis 5 is supported. H6: The greater the control over environment, the less the media exposure. Control correlates -.08 (p<:0.lS) with amount of total media exposure, while control correlates -.l8 (p<:0.0l) with the amount of television exposure. The associations are unchanged when SES and un- certainty discrepancy are partialed out. Thus, the relationship between control and amount of media exposure is in the predicted direction, but not statistically significant, while the control/television exposure relationship is significant and in the expected direction. However, since the hypothesis as stated was concerned with overall media exposure, not specifially television, the hypothesis does not receive support. H7: The greater the control over environment, the less the attribution of communication functions to interpersonal communication and television. Testing this hypothesis requires examining the control variable in relation to the two indices which represent the sum total of attribu— tions to interpersonal communication and television respectively, across all functions. The correlation between control and attributions to interpersonal communication is -.l7 (p< 0.0l) and between control and l54 attributions to television is -.20 (p<<0.0l). Controlling for the amount of interpersonal communication and SES the correlation between control and attributions to interpersonal communication is -.l6 (p< 0.02). Likewise, controlling for amount of television exposure and SES, the correlation between control and attributions of functions to television is —.l3 (p<:0.05). Controlling for uncertainty discrepancy, the original correlations are unchanged. Thus, hypothesis 7 is fully sup— ported by the data. H8: The greater the control over environment, the greater the relative attribution of communication functions to inter- personal communication than to television. The test of this hypothesis involves comparing the correlations reported above for attributions of communication functions to television and interpersonal communication to determine which is more negative. The hypothesis predicts that there will be greater attribution to inter— personal communication than to television, as control increases. Thus, the negative correlation between control and interpersonal attributions should be less strong than the negative correlation between control and television attributions. Inspection of these correlations reveals that this is the case. Control correlates -.20 with television, while correlating -.l7 with interpersonal communication. Ideally, this dif- ference would be subject to a test for statistical significance. However, the only standard test for comparing two correlations--the Fisher's R to Z transformation--is valid only when the correlations are between sets of variables measured within two independent samples. Since this is not the case with these data, the significance of the 155 difference can only be intuitively examined. Judging from a difference of 0.03 between the correlations, although it is in the predicted direc— tion, the magnitude is very weak. Thus, hypothesis 8 is not supported. H9: The greater the overload at a point in time, the greater the subsequent attribution of communication functions to uncertainty reduction. It will be noted from the explication of Operational definitions in the previous chapter that the uncertainty reduction index was con— structed by combining the uncertainty reduction items for television and for interpersonal communication into a single value. Likewise, for the uncertainty expansion variable, whose test with underload will follow, a parallel procedure was followed. The correlation between overload and attributions of communication to uncertainty reduction is 0.l6 (p<:0.02). Controls for control over environment and SES yield a corre- lation coefficient of 0.l7. Considering the magnitude of the correla- tion between overload and attribution of functions to uncertainty reduc- tion, the hypothesis is supported. H10: The greater the underload at a point in time, the greater the subsequent attribution of communication functions to uncertainty expansion. The magnitude of the correlation between underload and attribu- tion of communication functions to uncertainty expansion is 0.05 (p<:0.28). Partialing for control over environment and SES yields a value of 0.03. Thus, the relationship is in the predicted direction, but is very weak, suggesting no support for the hypothesis. H1]: Control over environment is a stronger predictor of changes in communication behaviors than chronological aging. l56 To test this hypothesis, we must reexamine the hypotheses involv- ing the control over environment variable, substituting chronological age in the respective relationships, and then examining the relative predictive power of control and age. The hypotheses concerned with control over environment are 4, 5, 6, 7, and 8. Table 16 shows these comparisons. Table l6. Comparison of Control Over Environment with Age in Predic— tive Power Hypothesis Control 'r' Age 'r' 4a -.ll* 0.00 4b 0.0l —.O7 5 0.l7* 0.00 6 -.08 —.l8 7a -.l7* —.l0 7b —.20* -.l7 8 0.03 (diff.) 0.07 (diff.) *Control is a stronger predictor. There are seven possible points of comparison within the five hypotheses. Examining the relative predictive power of control and age dichotomously, in four of seven comparisons, control is a better pre- dictor than age. Thus, control has the advantage in this regard. We can also calculate a continuous measure of relative predictive power by l57 subtracting the age correlations from the control correlations, within each comparison, and then adding these values across all seven com- parisons. Control is stronger by 38 points, and age is stronger by 22 points. Thus, overall control is a better predictor by l6 points. Considering these results the author hesitates to claim strong support for the hypothesis. While control is the better predictor of changes in communication behaviors, the margin of the advantage is less than overwhelming. Thus, it is concluded that hypothesis ll is supported by the data, although not strongly. Summary of Results The overall results of the tests of hypotheses show that of the eleven hypotheses, four were fully supported, one was partially sup- ported, and six were not supported. If we judge the theory overall, if by chance (MK? out of twenty hypotheses would be expected to be supported purely as random fluctuation, roughly half being supported suggests that the theoretic approach has merit. Figure 8 shows a sum- mary listing of the propositions by whether or not they received support. As the table indicates, nearly all the hypotheses involving re- lationships among control over environment and other communication variables receive some margin of support. Four of six are fully sup- ported by the data, and one is in the right direction, with a fully supported subcomponent, while one shows no relationship with a corre- lation of roughly zero. 158 .mchmmp mmmmcpoqx; mo cheszm .cochmaxm zchmpcmocz 0p mconu::F :onmchzs .506 we conzaFcppo pcmzommnzm mgp pmpmmcm mcp .mEFp cF choa m pm umoFLmuc: mcp cmpmmgm mgh .coFmF>mFmp 0p cusp :onmoFczeaou chomcmapmch 0p mconoczF :on -moFczesou Fo :onanpppm m>FmemL mcp mewmcm mgp .pcoscocF>cm pm>o Foppcoo mgp pmpmmcm mcF .zocmamcumFu xpcpmpcmuc: mgp mmmF mgp .pcoECOLF>cm pm>o Focpcoo mgp pmpmwcm mgF .FchomLma -pmch use cmmecmev maFgmconmFmp :onmoF::EEoo pcmzcmmnzm Fo xqocpcm mcp memocm msp .mer :F choq a pm uwochucz xpcpmppmuc: asp pmpmmcm och .Fchompmg -mecF vcm ummeumEv mapzmcopmewc :onmchze -Eop pcmzcmmnzm Fo xaocpcm mcp mmmF asp .mEFp :F choa a pm umonm>o zpcpmpcmucz mgp memmcm sz .FchomLmamecF new emprumEv conmuF::EEoo pcmso -mmazm Fo pczoEm mgp memmcm mzp .mEFp :F choa a pm FocmqmpumFu xchmppmucz asp memmcm mgF umppoaazm poz .m mesmpd .mcme FmoFmoFococgu cusp mcoF>msmn :onmoFc:EEou :F mmmcmco Fo cequnmLa cmmcogpm m mp pcmscocF>cm Lm>o Foppcoo .conoznmL Apcpmpcmocz op mconoc:F conmoFczeeou Fo :on -anpppm pcmzcmmnzm mcp cmpmmpm mcp .wEFp :F chog a pm umoFLm>o mzp cmpwmpm 85F .coFmF>mFmp new :onmchaesoo chomgmapmpcp 0p mconoczF :onmchzseoo Fo coppanpppm mgp mmmF esp .pcmscocp>cm pm>o Focpcou mzp memmgm mgF .FcoFmF>mFmp LOF FFcov mczmoaxm «Fume msp mmmF mcp .pcmECoLF>cw cm>o Foppcoo mgp memmcm mgF .con -moF::EEoo FmCOmcmamecF mzp mewmcm wcp .pcmscopF>cm Lm>o Focpcop mgp pmpmmpm mgF umppommam l59 The hypotheses involving uncertainty discrepancy receive support in only one of five cases. Not only do the results show no relation- ship, but many indicate directions of associations that are not as predicted. In the following chapter these findings will be interpreted in light of the theory presented in Part II. CHAPTER XI INTERPRETATION OF RESULTS AND IMPLICATIONS Overview In this final chapter the results of hypothesis testing of com— ponents of the information theory of communication functions will be discussed. First, an interpretive review of the pattern of results found with the two major groups of hypotheses will be developed—~those involving control over environment and those involving uncertainty discrepancy. Second, factors influencing this pattern will be dis- cussed, particularly with respect to the unreliability of measurement, and the limitations of testing a process theory with a cross—sectional design. Third, we will review what the present research suggests is now known in the area of informational aging. In the second part of this chapter, we will consider directions for future research in light of the factors discussed in the first part of the chapter. Three areas of methodological improvement will be sug- gested: l) improved operational definitions for the uncertainty dis- crepancy and control constructs, 2) the sampling of a wider age distribution--from early childhood to advanced old age, and 3) account- ing for time in the research design. l60 l6l A. The Results in Light of Methodological Considerations The Findings As the first eight chapters indicated, control over environment is a key variable in the information theory of communication functions. It is critical in these areas: separation of ego space from environ- mental space, amount of uncertainty discrepancy experienced, amount of interpersonal and mediated communication, attribution of communication functions, and informational aging. The pattern of results reported in the previous chapter, with five of six control-related hypotheses having some degree of support, gives the suggestion of validity to this major core of the theory. 'Suggestion of validity' is used here in light of the rather weak, although statistically significant magnitudes of the associations. As was pointed out in Chapter IX-B, uncertainty discrep- ancy and attributions of communication functions measures give indica- tions of low reliability. This may account for the rather weak relation- ships that are supported. The relationships involving the uncertainty discrepancy variables were largely insignificant, with a mixed pattern of insignificance-- some components in the right direction, others essentially zero, and still others in the opposite direction predicted. As was stated earlier, in pilot tests a more precise, yet more difficult scaling technique for this construct was used without success. Respondents were asked for two estimates on each dimension of uncertainty, actual and preferred. The same underlying continuum was used for each component, allowing for the precise measurement of uncertainty discrepancy through subtracting 162 these two values. This technique created great confusion for respond- ents, which led us to the simpler, yet potentially more unreliable technique used in the primary phase of this research. It will be recalled that for the measurement of the target sample we transformed each component into an assertion, pitching the valence of uncertainty discrepancy in one direction, and then assumed a "disagree" response indicated the desire for the opposite uncertainty state. That is, if the respondent responded negatively to the state- ment: "I would like my daily activities to be less usual ... less routine, we assumed that the respondent wanted more_routine, with the magnitude of the disagreement indicating the degree to which m9rg_ routine was wanted. This may, however, be an invalid assumption, for it may not be the case that a "disagree” response indicates a desire for the opposite state of affairs, but that it indicates disagreement with a change, i.e., satisfaction with the status quo. Possibly the indi- vidual responding I'very strongly disagree" is only indicating a strong preference for the status quo, and not a desire for the opposite state of affairs. Another source of unreliability may also be that the seven compon- ents selected for the measurement of uncertainty discrepancy are inap- propriate to index the construct: l) they may not capture a large enough portion of key life dimensions on which there is likely to be uncertainty, 2) they may clearly be the wrong items entirely, 3) they may not be operating at the appropriate level of abstraction, or 4) they may be too specific, hence the more pervasive and strong uncertainties in the affective region are not tapped, although this is 163 likely to be a level at which the proposed theoretic relationships.may indeed hold. Another factor influencing the internal validity of the design is the time frame for the research. While the theory takes a highly process orientation, the research is conducted at a single point in time. This may obscure the detection of expected relationships in the actual data, because individuals are at different stages of adaptation to uncertainty discrepancy. Some persons may be just beginning to experience uncertainty imbalance, and hence the causes of this im« balance may still appear positively correlated to the effect. For example, in the early stages, interpersonal entropy would be positively associated with the uncertainty overload. At the other end of the time frame, an individual with the same level of uncertainty discrepancy, in the same direction, may show the inverse balancing configuration between uncertainty discrepancy and such things as the amount and entropy of interpersonal and mediated communication components of infor— mation space. For example, in this case, at later stages interpersonal entropy would be negatively correlated with uncertainty overload. In short, at a single point in time some individuals may be at the early uncertainty detection.stage, before action is initiated to change ego space, while others may have already largely accomplished change in ego space. Hence, these two examples indicate that controlling for direction and magnitude of uncertainty discrepancy, while simply not_ controlling for time, may yield "artificial" zero correlations. Some of these individuals are at the stage of awareness of uncertainty discrepancy, but no action to reduce discrepancy has been taken, while 164 others are in a state of full action to reduce uncertainty. Hence, some correlations would be negative while some would be positive--yielding overall correlations of zero. However, the theory itself is still adequate in explaining both these states of affairs from a process perspective. In brief, the cross-sectional design is limited in ade- quately testing the theory, although resources dictated a single point in time design in the current research. To summarize, then, two main factors may account for the pattern of non—supportive and weakly supportive results--unreliable measurement, and lack of control due to the inability to account for change over time. Of course, another factor may also explain the results, that the logic of the theory is invalid and that basic changes in conceptualiza- tion rather than methodological procedures are required. While this may indeed be the case in the present research, evidence suggests otherwise. The encouraging pattern of the control over environment hypotheses, and the measurement error in the uncertainty discrepancy scale and other indices, indicated through field observations and statistical procedures, suggests that it is reasonable to proceed with the working assumption that the theory is sound in logic, while the current methods of analysis are inadequate. As future research directions are discussed in Part B of this chapter, the approach will be taken that until major portions of the theory are unsupported while having valid and reliable measurement, the logic will be retained. To close this section, a brief summary is given to what we now have some initial evidence to suggest are valid relationships. Let us give a sketch of a person's communication behaviors, although keeping 165 in mind the unreliability factors. First we will discuss a person with high perceived control over environment. The high control individual, with a large ego space, will talk to many people and use little tele- vision. Overall, the high control individual will attribute less uncertainty balancing functions to his/her communication behaviors, perhaps slightly more with interpersonal than television. Considering all the individual's communication behavior, we can predict more about these communication behaviors by knowing the magnitude of the person's perceived control over environment, than by knowing the person's age. 0n the other hand, the low control individual will consume greater amounts of television messages, and talk less to people in a face-to- face mode, while attributing more uncertainty balancing functions to both mediated and interpersonal communication. Evidence also exists to suggest that regardless of level of control, persons experiencing uncer- tainty overload consume greater amounts of mass communication. Those who are experiencing underload have interpersonal networks of low integration, suggesting a high interpersonal content entropy level to balance the underload. Also, persons experiencing uncertainty overload attribute communication behaviors to uncertainty reduction functions. While these findings characterized above are stated quite strongly, we must caution that these are not heavily supported knowledge claims. The sizes of the correlations, reliability of measures, and internal validity of the research design make this caution imperative. However, the considerations discussed in this section clearly suggest implica- tions and directions for future research in testing the theory. These are discussed in the following section. 166 B. Future Research Suggestions Three general strategies for more adequately testing the theory will be outlined: l) improvement in the operationalization of vari- ables, 2) improvements in the representativeness of the samples used for research, and 3) development of greater correspondence between the process nature of the research design and the process nature of the theory. Improved Operations Greater improvements are indicated for the operationalization of the uncertainty discrepancy scales. One strategy that may yield better payoffs is to alter the component dimensions of life uncertainty that are used in research. Two separate directions are suggested in creat— ing more valid and reliable scales of uncertainty discrepancy. One is the expansion of the number and raflg§_of key life components, with more detailed pilot testing necessary to derive these key components. Second, a weighting factor may be introduced into the operationaliza- tion, such that respondents rate how important each life component is to them. This would control for individual differences in the relative import of various components, reducing the error variance in quanti- fying uncertainty discrepancy. This measurement might be performed with a ratio scaling technique such as: "If the average component of your life has ten units of importance, how many units of importance does (each life component) have?" Each of these values would then be multi- plied by each respective uncertainty component prior to final index construction. Thus, the uncertainty discrepancy scale would in a sense 167 be tailored specially for each respondent. At another level, perhaps more abstract measures of uncertainty discrepancy are called for. Only two global questions might be asked that would substantially reduce the complexity of the scale: "Would you prefer more or less uncertainty in your life?" and "If your ability to predict the weather is ten units of uncertainty, how many more/less units of uncertainty would you prefer?“ Or perhaps, "If your ability to predict the weather is ten units of uncertainty, how many units of uncertainty are there currently in your life? How many units would you like there to be in your life?" Thus, these two alternative ways to measure uncertainty discrepancy seek to capture an overall average uncertainty across all life components, perhaps eliminating the neces- sity to measure and aggregate many subscales. Furthermore, instruments might be designed to distinguish the level of specificity of uncertain- ty discrepancy. This would also be likely to enhance capabilities for more precise analytical operations. The latter more abstract kinds of operations may be more appro- priate if the theory and notions of informational aging are tested from a developmental perspective, with wide ranging age distribution samples. Then the uncertainty scales cannot contain components that are specific to a particular age segment. The questions that are asked of old people must be able to be asked of children. Focussing on old people's life components in relatively concrete terms is likely to result in life components that are valid or of importance only to old people. Thus, more abstract measures for indexing uncertainty discrepancy are 168 mandated by this process view and the concern for the external validity of the research. The operationalization of the construct control over environment may also be improved. A useful approach may be to attempt a more direct assessment of the size of ego space than was done here. Rather than asking the respondent to react to Likert-type assertions, metric-multi- dimensional scaling and unidimensional ratio scaling could be coupled into a unique measurement technique. First, key environmental compon- ents would be selected, and then these would be presented to respondents, who would make ratio estimates of dissimilarity distances among all pairs of components. Subsequently, these data would be subject to a metric-multidimensional scaling. In addition, the individual would also be asked to unidimensionally ratio scale the perceptions of control over each of the spatial elements that were used in the judgments to be multidimensionally scaled. Once the M05 is completed and the results of the control scaling overlaid on the space, the area within which the individual has greater than 50% perceived control could be measured for its yglumg_by reference to the coordinate system of the multi-dimen- sional space. The resultant measure of volume of ego space could then be entered into the analysis as a more valid and reliable measure of the construct control over environment. Age Distributions As early chapters of Part I suggest, the information theory of communication functions is designed to be generalizable to the entire age distribution of the population. Furthermore, the notions of 169 informational aging specify a developmental approach to studying communi- cation behavior, calling for the examination of persons at all stages of development, both in relative and absolute terms. Future research ought to draw more representative samples of the population from the early years of childhood to the advanced years of aging. This will in- crease the likelihood that a sample is obtained having people of widely variant informational aging. Or, if this sampling is not feasible, at least the conduct of research with samples drawn from three separate parent age populations, perhaps children, middle adults, and old people, although less representative, would be useful. The latter approach may be an important undertaking for pilot research. Time The theory calls for a process view of the operation of the prin- ciple of optimal information, attribution of communication functions, and informational aging. Earlier, we specify time-ordered sequences of changes in the interrelationship of these variables, suggesting a time- lagged feedback and adjustment system in relationships of uncertainty discrepancy and the structure of information space. A longitudinal design would more fully enable an adequate test of theory, in control- ling for the time-phasipg_of information processing across a sample of individuals. A cross-sectional study merely shows time-phase variance as error, increasing with the time delay of the feedback loop. On the other hand, a longitudinal design would allow time-phasing to clarify the seemingly muddled relationships involving the uncertainty discrep- ancy constructs. 170 Elaboration of the Research Program In addition to the methodological guidelines we have specified for further research, more substantive expansion of the research program can also be undertaken, contingent of course upon an indication of further validity of the theoretic model. The following thrusts are considered important: l) the expansion of measurement of functions of mediated communi- cation to include the assessment of parallel functions across all communication modes, both interpersonal and the range of mediated modes--not just television; 2) improvement in the conceptualization and measurement for the entropy variables within all components of information space, particularly at the level of form entropy, for the measures of channel and content used in the present research are high— ly exploratory and form entropy is not measured; 3) greater attention to content_preferences and their differential relationships to control, uncertainty discrepancy attributions, and so on, and 4) basic work in the laboratory, testing for propositions regard— ing informational aging in a highly controlled setting. Summary In concluding Part III with this chapter, we have discussed the overall findings of the research in light of theoretic and methodological considerations. It has been suggested that while the logic of the theory may be invalid, evidence suggests it is more appropriate to view the logic as valid and to assume the lack of strong support for all aspects of the theory is a function of methodological factors. Two key factors are low reliability of measures of theoretic constructs, and the problem of reduced control inherent in a cross-sectional study. Building on these notions, Part B of this chapter suggested directions for improving 171 the tests of the information theory of communication functions by: l) improving operational definitions, 2) expanding the age distribution of samples, and 3) using longitudinal research designs. We close by suggesting that if the theory is given greater support in future research, areas of greater substantive focus will be clear. L I ST 0F REFERENCES LIST OF REFERENCES Adorno, T. W., Frenkel-Brunswik, E., Levinson, D. J., and Sanford, R. N. The Authoritarian Personality. New York: Harper, 1950. Aronoff, C. "Old age in prime time,” Journal of Communication, 1974, 24:86-87. Ashby, W. R. "Principles of the self-organizing system," in W. Buckley (ed.), Modern Systems Research for the Behavioral Scientist. Chicago: ATdine, 1968. Atkin, C. K. “Anticipated communication and mass media information- seeking," Public Opinion Quarterly, 1972, 36:188-199. Back, K. W. "Influence through social communication," Journal of Abnormal Social Psychology, 1951, 4659-23. Barnes, J. A. "Social networks," Addison-Wesley Module No. 26, 1972, p. 1-29. Berelson, B. "What 'miésing the newspaper' means" In P. F. Lazarsfeld and F. N. Stanton (eds.) Communication Research, 1948-9. New York: Duell, Sloan, and’Pearce, 1949. Berlo, D. K. The Process of Communication. New York: Holt, Rinehart, and Winston, 1960. Berlo, D. K. "The process of social research." Unpublished manu- script. East Lansing: Michigan State University, Department of Communication, 1969. Berlo, D. K. “Essays on human communication." Unpublished manuscript. East Lansing: Michigan State University, Department of Communica- tion, 1970. Beyer, G. H. and Woods, M. E. "Living and activity-patterns of the aged," Research Report No. 6. Ithaca, New York: Center for Housing and Environmental Studies, Cornell University, 1963. Blumler, J. G., Brown, J. R., and McQuail, D. "The social origins of the gratifications associated with television viewing. Unpub- lished manuscript. University of Leeds, 1970. 172 173 Bower, R. T. Television and the Public. New York: Holt, Rinehart, and Winston,_1973. Bracey, H. E. In Retirement. Boston: Houghton Mifflin, 1966. Bruner, J. S. ”On perceptual readiness," psychological Review, 1957, 64:123-152. Bultena, G. "Age grading in the social interaction of an elderly male population," Journal of Gerontology, 1968, 23:539-543. Cassata, M. "A study of the mass communications behavior and the social disengagement behavior of 177 members of the Age Center of New England." Unpublished doctoral dissertation. Indiana University, 1967. Cartril, H., and Allport, G. W. The Psychology of Radio. New York: Harper Brothers, 1935. Chaffee, S. H. and McLeod, J. M. "Communication as coorientation: two studies." Paper presented to the annual meetings of the Association for Education in Journalism. Boulder, Colo., 1967. Cohen, A. A. "Coping with uncertainty, information usage and ticket splitting." Unpublished doctoral dissertation, Michigan State University, 1974. Collins, B. E., and Raven, B. H. "Group structure: Attraction, coali- tions, communication, and power," in G. Lindzey and E. Aronson (eds), Handbook of Social Psychology. Reading, Mass.: Addison- Wesley, l969. Coster, G. Yoga and Western Psychology. London: Oxford University Press, 1934. Cowgill, D. 0. and Baulch, N. "The use of leisure time by older people," The Gerontologist, 1962, 2347-50. Cumming, E. and Henry W. H. Growing Old: the Process of Disengagement. New York: Basic Books, 1961. Cushman, D. P. and Whiting, G. C. "An approach to communication theory: toward consensus on rules," Journal of Communication, 1972, 22; 217-238. Cutler, N. E. and Bengtson, V. L. ''Age and political alienation: maturation, generation, and period effects," Annals of the Ameri- can Academy of Political and Social Science, 1974, 4153160-175. 174 Cutler, N. E., Bengtson, V. L., and Dowd, J. J. "Numbers and needs: the changing population structure of Los Angeles," Paper prepared for the Mayor's Conference on Children, Youth, and Senior Citi- zens, Los Angeles, April, 1974. Danowski, J. A. "An uncertainty model: friendship communication net- works and media related behaviors." Paper presented to the annual meetings of the International Communication Association, New Orleans, l974a. Danowski, J. A. “Alternative information theoretic.measures of television messages: An empirical test." Paper presented to the annual meeting of the Association for Education in Journalism. San Diego, l974b. Davis, R. H. “A descriptive study of television in the lives of an elderly population." Unpublished doctoral dissertation, University of Southern California, 1972. Dorfman, D. D. and McKenna, H. "Pattern preference as a function of pat- tern uncertainty," Canadian Journal of Psychology, 1966, 29: 143-153. Eckblad, G. "The attractiveness of uncertainty,” Scandinavian Journal of Psychology, 1963, 431-13. Einstein, A. Relativity: The Special and the General Theory. Trans. R. W. Larson. New York: Holt, 1920. Ericson, P. "Relational communication: complementarity and symmetry and their relation to dominance-submission." Unpublished disser- tation. Department of Communication, Michigan State University, 1973. Festinger, L. A Theory of Cognitive Dissonance. Evanston, 111.: Row and Peterson, 1957. Festinger, L., Schacter, S., and Back, K. Social Pressures in Informal Groups: A Study of Human Factors in Housing, New York: Harper, 1950. Glick, I. O. and Levy, S. J. Living with Television. Chicago: Aldine Publishing Co., 1962. Graney, M. J., and Graney, E. E. "Communications activity substitu— tions in aging," Journal of Communication, 1974, 24:88-96. Greenberg, B. S. "Children's reactions to TV blacks." CUP Report #14. East Lansing, Michigan: Department of Communication, Michigan State University, 1971. 175 Greenberg, B. S. "Gratifications of television viewing and their corre— lates for British children." in J. Blumler and E. Katz (eds.). Annual Review of Communication Research, III. Beverly Hills, Ca1if.: Sage PUblications,T1974. Hance, K. G., Ralph, D. C., and Wiksell, M. J. Principles of Speaking. Belomont, Calif: Wadsworth Publishing Co., 1969. Hanneman, G. J. "Message uncertainty as a predictor of arousal and aggression," Unpublished doctoral dissertation. Department of Communication, Michigan State University, 1970. Heider, F. The Psychology of Interpersonal Relations. New York: Wiley, 1958. Herzog, H. "What do we really know about daytime serial listeners?" in P. F. Lazarsfeld, and F. N. Stanton (eds.) Radio Research: 1942—1943. New York: Duell, Sloan, and Pearce,71944. Hoar, J. R. "Reading, listening, and viewing behaviors of the aged: an inventory of the mass communications habits and preferences of 200 aged persons in Oxford, Miss." Unpublished doctoral disserta- tion, University of Iowa, 1960. Horton, D. and Strauss, A. "Interaction in audience-participation shows," American Journal of Sociology, 1957, 625579-587. Horton, D. and Wohl, R. R. "Mass communication and para-social inter- action," _§ychiatry, 1956, 123215-224. Jones, E. E., and Gerard, H. B. Foundations of Social Psyphology. New York: Wiley, 1967. Jones, A., and McGill, D. W. "The homeostatic character of information drive in humans." Journal of Experimental Research in Personal- ity, 1967. 2g25-31. Katz, 0. "The functional approach to the study of attitudes,” Public Opinion Quarterly, 1960, 245163-176. Katz, E., Blumler, J. G., and Gurevitch, M. "Uses and gratifications research." Public Opinion Quarterly, 1974, 21; 509-524. Katz, E., and Foulkes, E. "On the use of the mass media for 'escape': clarification of a concept," Public Opinion Quarterly, 1962, gg; 377-388. Katz, E., Guretich, M., and Haas, H. "0n the use of mass media for important things." Paper presented to the 20th International Congress of Psychology, Tokyo, 1972. 176 Katzman, N. "Communication flow and social entropy," Ekistics, 1970, 1223212-219. Kelley, H. "Attribution in social interaction," in E. E. Jones §i_§i, (eds.) Attribution: Perceiving the Causes of Behavior. Morriston, New York: General Learning Press, 1972. Kleemeier, R. W. "Leisure and disengagement in retirement,“ Gerontologist, 1964, 45180-184. Laumann, E. 0. Bonds of Pluralism: The Form and Substance of Urban Social Networks. New York: Wiley-Interscience, 1973. Lazarsfeld, P. F. and Kendall, P. Radio Listening in America: The People Look at Radio Again. New York: Prentice-Hall, 1948. Lefcourt, H. M. "Internal versus external control of reinforcement: a review," Psychological Bulletin, 1966, 655206-220. Lyle, J., and Hoffman, H. R. "Children's use of television and other media," in E. Rubenstein §t_gi, (eds.) Television and Social Behavior, Vol. 4, U. S. Government Printing Office, 1972. Lowenthal, M. F. "Social isolation and mental illness in old age," American Sociological Review, 1964, 29:54-70. Malinowski, B. The Dynamics of Culture Change. New Haven, Conn.: Yale University Press, 1945. Mark, R. A. "Coding communication at the relationship level," Journal of Communication, 1971, 2ig221-232. Martel, M. U. "Age-sex roles in magazine fiction," in B. L. Neugarten (ed.), Middle Age and Aging. Chicago, 111.: University of Chicago Press, 1968. Martin, W., Bengston, B. and Acock, A. “Alienation and age: a context specific approach," Social Forces, 1973, 52567-84. McLeod, J., Atkin, C., and Chaffee, S. "Adolescents, parents, and television use," in G. COmstock and E. Rubinstein (eds.) Television and Social Behavior: Television and Adolescent Aggres- siveness. Washington, D.TC.: U. S. Government Printing Office, 1972. McNeil, L. W. and Rule, B. G., "Effects of conceptual structure on information preference under sensory-deprivation conditions," Journal of Personality and Social Psychology, 1970, 11335-39. 177 McTavish, D. G. "Perceptions of old people: a review of research methodologies and findings," The Gerontologist, 1971, 11390-102. Merton, R. K. "Social theory and social structure," in On Theoretical Sociology. New York: The Free Press, 1967. Meyershon, R. "A critical examination of commercial entertainment,” in R. W. Kleemeier (ed.), A in and Leisure. New York: Oxford University Press, 1961, pp. 3-272. Millar, F. "A transactional analysis of marital communication patterns: and exploratory study." Unpublished doctoral dissertation. Department of Communication, Michigan State University, 1973. Miller, A. H., Brown, T. A., and Raine, A. S. "Social conflict and political estrangement, 1958-1972." Paper presented to the annual meetings of the Midwest Political Science Association, Chicago, 111., 1973. Miller, G. R. Speech Communication: A Behaviora1_Approach. New York: Bobbs-Merri11,—1966. Miller, G. R. and Steinberg, M. Between People: A New Analysis of Interpersonal Communication. PaTo Alto, California: Science Research Associates, 1975. Miller, J. G. l'PSychological aspects of communication overloads," in R. W. Waggoner and D. J. Carek (eds.), International Psychiatry Clinics: Communication in Clinical Practice. Boston: Little Brown, 1964. Miller, J. G. "Living systems: the group," Behavioral Science, 1971, 465350-387. Mischel, W., Zeiss, R., and Zeiss, A. ”Internal—external control and persistence: validation and implications of the Stanford pre- school internal-external scale," Journal of Personaliiy_and Social Psychology, 1974, 293256-278. Monge, P. R. "The study of human communication from three systems paradigms." Unpublished doctoral dissertation, Department of Communication, Michigan State University, 1973. Mortensen, C. 0. Communication: The Study of Human Interaction. New York: McGraw-Hill, 1972. Moss, G. E. Illness, Immunity, and Social Interaction: The Dynamics of Biosocial Resonation. New York: Wiley, 1973. 178 Munsinger, H. and Kessen, W. "Uncertainty, structure, and preference,“ Psychological Monographs, 1964, 2831-24. Nagel, E. "A formalization of functionalism,” in Logic Without Meta- physics. New York: The Free Press, 1957. Newcomb, T. M. “An approach to the study of communicative acts,” Psychological Review, 1953, 693393-404. Nordenstreng, K. "Comments on gratifications research in broadcasting,‘I Public Opinion Quarteriy, 1970, 343130-133. Olsen, M. E. "Motion picture attendance and social isolation,” Sociological Quarteriy, 1960, 15107-117. Opinion Research Corporation, "The public appraises movies." A Survey for Motion Picture Association of America. Princeton, N. J., 1957. Palmore, E. B. "The effects of aging on activities and attitudes," Gerontologist, 1968, 83259-263. Parker, E. B. and Paisley, W. G. "Patterns of adult information seek- ing." Stanford, Calif.: Institute for Communications Research, Stanford University, 1966. Pastorello, T. "Adjustment in later life as a function of the quality of social relations." Paper presented at annual meetings of Gerontological Society, San Juan, Puerto Rico, 1972. Petersen, M. "The visibility and image of old people on television,” Journalism Quarterly, 1973, 595569-573. Radcliffe-Brown, A. R. Structure and Function in Primitive Society. Glencoe, Ill.: The FFeeiPress,71952. Richards, W. 0. "Network analysis in large complex systems: techniques and methods--tools." Paper presented to annual meetings of the International Communication Association, New Orleans, 1974. Richards, W. D., Farace, R. V., and Danowski, J. A. "NEGOPY: program description," Unpublished manuscript. East Lansing, Michigan: Department of Communication, Michigan State University, 1973. Richards, W. D. and Lindsey, G. I'Social network analysis: an overview of recent developments." Paper presented to the American Cyber- netics Society, Philadelphia, Penn., 1974. Riley, M. W. "Aging and cohort succession: interpretations and mis- interpretations," Public Opinion Quarterly, 1973, 22535—49. 179 Riley, M. W. and Riley, J. W. "A sociological approach to communica- tions research,” Public Opinion Quarterly, 1951, 153445-460. Robin, E. P. "Death and dying--in children's readers," Paper presented at annual meetings of the Gerontological Society, Portland, Ore., 1974. Robinson, J. P. "Toward defining the functions of television," in E. I. Rubinstein ei_oi, (eds.) Television and Social Behavior, Vol. IV. Washington, D. C.: U. S. Government Printing Office, 1972. Rogers, Edna and Farace, R. V. "Analysis of relational communication in dyads: new measurement procedures," Human Communication Research (in press). Rokeach, M. Open and Closed Mind. New York: Basic Books, 1960. Rosenberg, G. S. "Age, poverty, and isolation from friends in the urban working class," Journal of Gerontology, 1968, 225533-539. Rosengren, K. E. "Uses and gratifications: an overview." Unpublished manuscript. Sweden: University of Lund, 1972. Rosengren, K. E., and Windahl, S. "Mass media consumption as a func- tional alternative," in D. McQuail (ed.), Sociology of Mass Communications. Harmondsworth: Penguin, 1972. Rosow, 1. Social Integration of the Aged. New York: Free Press, 1967. Rotter, J. B. "Generalized expectancies for internal versus external control of reinforcement." Psychological Monogpaphs, 1966, 29; (whole No. 609). Rubenstein, E. A., Comstock, G. A., and Murry, J. P. Television and Social Behavior, Vol. I-V. Washington, D.C.: U. S. Government Printing Office, 1972. Rump, E., "Relative preference as a function of the number of elements in abstract designs,” Australian Journal of Psychology, 1968, 20: 39-48. Schacter, S. "Deviation, rejection, and communication,” Journal of Abnormal and Social Psychology, 1951, 465190-207. Schacter, S. and Singer, J. E. "Cognitive, social, and physiological determinants of emotional state," _sychological Review, 1962, oo;379-399. 180 Schalinske, T. F. "The role of television in the life of the aged person." Unpublished doctoral dissertation, Ohio State University, 1968. Schramm, W. "The nature of news," Journalism Quarterly, 1949, 2o; 259-269. Schramm, W. "Aging and mass communication,‘ in M. W. Riley, J. W. Riley, and M. E. Johnson (eds.), A in and Societ , Vol. 2: Aging and the Professions.‘ New York: Russel Sage Foundation, 1969, pp. 352-375. Schramm, W. and White, 0. M. "Age, education, economic status factors in newspaper reading," Journalism Quarterly, 1949, 263149-159. Schroder, H. M., Driver, M. J., and Streufert, S. Human Information Processing. New York: Holt, Rinehart, and Winston, 1967. Seltzer, M. and Atchley, R. C. "The concept of old: changing atti- tudes and stereotypes," Gerontologist, 1971, 1:226-230. Shannon, C. E. and Weaver, W. The Mathematical Theory of Communica- tion. Urbana 111.: University of Illinois Press, 1949. Smith, R. E. "Changes in locus of control as a function of life crisis resolution," Journal of Abnormal Psychology, 1970, 75:328-332. Steiner, G. A. The People Look at Television. New York: Alfred A. Knopf, 1963. Swartz, H. and Herbik, H. "Eine experimentelle untersuchung uber den einfluss der syntakischen information der anordnung von baukor- pern entlang einer strasse auf stimmungen des betrachters," Zeitschrift fur Experimentelle und Angewandte Psychologie, 1971, 18:499-511. Thornton, D. I'The function of visual information in cross-modality satiation of auditory information drive.“ Unpublished doctoral dissertation. University of Pittsburg, 1966. Thornton, D. and Jones, A. "The nonspecific character of information drive with respect to sensory modalities." Paper presented to annual meetings of the Midwest Psychological Association, 1965. Tobin, S., and Neugarten, B. L. "Life satisfaction and social inter- action in the aging," Journal of Gerontology, 1961, 165344-346. U. S. Bureau of the Census, Census of Population Detailed Character- istics, 1970. Washington, D.C.: U. S. Government Printing Office, 1973. 181 Villard, K., and Whipple, L. It Takes Two to Know One: Beginnings on Relational Communication. NewTYork: John Wiley (in press). Wartella, E. and Ettema, J. S. "A cognitive development study of children's attention to television commercials," Communication Research, 1975, 1:69-88. Watt, J. H. and Krull, R. "An information theory measure for tele- vision programming," Communication Research, 1975, 1344-68. Woelfel, J. "Procedures for the precise measurement of cultural processes," Unpublished manUscript. East Lansing, Michigan: Department of Communication, Michigan State University, 1973. Woelfel, Joseph, Woelfel, John, Gillam J. and McPhail, T. "Political radicalization as a communication process," Communication Research (in press). Zborowski, M. "Aging and recreation," Journal of Gerontology, 1962, 115303-309. Zubek, J. P. Sensory Deprivation: Fifteen Years of Research. New York: AppTeton-Century-CrBfts, l969. APPENDIX INFORMATION LETTER AND QUESTIONNAIRE 182 MICHIGAN STATE UNIVERSITY EAST LANSING - MICHIGAN 48823 COLLEGE OF COMMUNICATION ARTS - DEPARTMENT OF COMMUNICATION - CABLE: COWDEPT July 19, 1974 Dear Resident of Grange Acres, Two weeks ago, a group of us from the Department of Communication at Michigan State University had the pleasure of sharing a Pot Luck dinner with many of you at the Community Center. We heartily enjoyed the good conversations and the good food! As we discussed earlier at the Pot Luck dinner, we are doing a survey of Senior Citizens' activities and opinions. An interviewer will be visiting your home within the next two to three weeks to talk with you about these important matters. We are looking forward to meeting and visiting with you. Each individual's answers will remain completely confidential. However, the final results of the study will be beneficial and important to Senior Citizens around the Country. We sincerely appreciate your cooperation in this effort. Any questions about the survey may be directed to Mrs. Smith or to myself at 485-2565. Thank you very much for your attention. Respectfully, James A. Danowski 183 APT # RES # INTERVIEWER: FINAL DISPOSITION (CODE: c_ pp 5 [Comment]) DATE ATTEMPTS 1 3 4 5 6 COMMENTS: TIME: COMP. TIME MIN HELLO..,I'M FROM MICHIGAN STATE UNIVERSITY. IF R SAYS: '1 don.t know too wE ARE DOING A SURVEY AT THIS TIME OF SENIOR CITIZENS' much. _. SAY: This is not a OPINIONS THROUGHOUT GRANGE ACRES. THE INTERVIEW WILL ONLY TAKE LESS THAN HALF AN HOUR... AND ALL INFORMATION IS ENTIRELY CONFIDENTIAL...OF COURSE. MAY I TALK WITH YOU? Interviewer: If R hesitates, see comments >' Enter R's home and establish comfortable seating. test and there are no right or wrong answers...We just want to know what 192 THINK about important matters. Your answers are very import- ant to us. TO BECIN...LET ME EXPLAIN THE QUESTIONAIRE. THE QUES- TIONAIRE IS DIVIDED INTO DIFFERENT SUBJECTS...AND I WILL ASK YOU QUESTIONS ABOUT EACH OF THEM. THIS IS NOT A TEST...AND THERE ARE N9 RIGHT OR WRONG ANSWERS...WE JUST WANT TO KNOW WHAT YOU THINK ABOUT IMPORTANT MATTERS. TO MAKE IT EASY TO ANSWER THE QUESTIONS...WE WILL USE A CARD WITH A SET OF ANSWERS I WOULD LIKE YOU TO PICK FROM. Interviewer: Hand card to R. HERE IS A CARD To USE FOR THE FIRST SET OF QUESTIONS... TAKE A LOOK AT IT FOR A MINUTE...WE WILL USE THE GREEN SIDE FIRST. NOW...IS EVERYTHING CLEAR? FOR THE FIRST SET OF QUESTIONS...I WILL READ A STATEMENT ...AND I WOULD LIKE YOU TO ANSWER WHETHER YOU AGREE OR DISAGREE WITH THE STATEMENT. THEN I WILL ASK YOU WHET- HER YOUR FEELING IS...VERY STRONG...STRONG...OR, SOMEWHAT STRONG. PICK THE ONE BEST ANSWER FROM THE CARD THAT BEST MATCHES YOUR FEELINGS ABOUT EACH STATEMENT. IF R SAYS: 'Who said you could do this survey?' SAY: We are working closely with MRS. SMITA‘ the manager...and she has given the survey complete approval. You may call her if you like. IF R SAYS: 'I don't have time.‘ SAY: We can finish in 15 or 20 minutes...and your answers are very important to us. IF R PERSISTS, SAY: Can I come back at a more convenient time? It is very important that we talk with you. IF R SAYS: 'Who's in charge of this survey?’ SAY: Mr. Danowski and Prof. Farace from the Dept. of Comm. at MSU. You may call them if you like. (485-2565) THE FIRST SET OF QUESTIONS ARE ABOUT YOUR USE OF TELEVISION... AND NOW WE ARE READY TO BEGIN... 1. There should be more Senior Citizens appearing on regular television programs. 1 Do you agree or disagree with this statement? Do you VERY STRONGLY ...STRONGLY ... or SOMEWHAT? , , lRepeat for each question VSA STA ASW [circle one] NAD Dsw STD 2 3 a 5 6 7 2. 10. THESE NEXT STATEMENTS ARE ABOUT TALKING WITH PEOPLE. 184 -2- ——--. .— IF R SAYS: "It depends..." or .1_T "Don't know..." FOR ALL QUESTIONS. CODE AS "NAD" I would like to see a Soap Opera on TV made I'VSA specially for Senior Citizens. Very Often I watch Television to relax. If R does not have a TV set, GO TO Q 11. 'Television very often helps me to calm my nerves. Often I get bored and turn on the TV. I often like to watch TV shows that are highly exciting. TV very often helps me to take my mind off things. Often I get to thinking too much and turn on the TV. If it weren't for TV...I wouldn't know nearly as much as I do today. I like to watch TV because very often I see new things I haven't seen before. STA Asw NAD Dsw STD VSD '= 1 2 3 4 5 6 7 9 [circle one] VSA STA Asw NAD DSw STD vsp NA 1 2 s 4 5 6 7 ' u 1 VSA STA Asw NAD Dsw STD VSD NA 1 2 3 4 5 6 7 ‘ 9 VSA STA Asw NAD Dsw STD VSD NA 1 2 3 a 5 6 7 9 l VSA STA Asw NAD Dsw STD vso: WA | 1 2 3 a 5 6 7 9 . _fi VSA STA Asw NAD DSW STD VSDl N“ 1 2 3 4 5 6 7 ' 9 VSA STA Asw NAD DSW STD VSD: NA, 1 2 3 4 5 6 7 ' f IvSA STA Asw NAD Dsw STD vsp; .1. I 1 2 3 a 5 6 7 : L g VSA STA ASW NAD DSW STD vspi NA 1 2 3 a 5 6 7 - v ..WE WILL USE THE SAME ANSWER CARD. . , 185 -3- IF R SAYS: ”It depends..." or "Don't know..." CODE AS "NAP” FOR ALL QUESTIONS. !' 11. Very often I talk with people to relax. VSA STA ASW NAD DSW STD VSD LA I 1 2 3 4 5 6 7 9 l 12. Talking with people very often helps me VSA STA ASW NAD DSW STD VSD. NA i to calm my nerves. 1 2 3 4 5 6 7 I E g I . 13. Often-I get bored and look for someone VSA STA ASW NAD DSW STD VSDI u; I to talk to. 1 2 3 4 s 6 7 . 9 | | . 14. I often like to talk with people that are 'VSA STA ASW NAD DSW STD VSDI LA highly exciting. I 1 2 3 4 5 6 7 I 9 15. Talking with people very often helps me to IVSA STA ASW NAD DSW STD VSD NA take my mind off things. 1 1 2 3 4 5 6 7 9 I l 16. Often I get to thinking too much and look NSA STA ASW NAD DSW STD VSD 11 for someone to talk to. ' 1 2 3 4 5 6 7 l i 1 . [s 17. If it weren't for talking to people, I VSA 'STA.-ASW NAD DSW STD VSD ‘ [ wouldn't know nearly as much as I do today. I 1 2 3 4 5 6 7 ; i 1 . = 18. I like talking to people because I very SA STA ASW NAD DSW STD VSD, .‘ often hear new things I haven't heard I 1 2 3 4 5 6 7 1 ; 1 before. NEXT...SOME STATEMENTS ABOUT OTHER ACTIVITIES... 19. I usually plan my daily activities in $13.4. STA ASW NAD DSW STD VSD! 1 advance. . 1 2 3 4 5 6 7 l ” | 20. When I make plans...I am almost certain I FSA STA ASH NAD DSW STD VSD NI! can carry them out. 1 2 3 4 S 6 7 9_| 1536 -4- I I 21. It is not easy to change my life... VSA STA ASU NAD DSW STD VSD. NA too many things stand in the way. | 1 2 3 4 . 5 6 7 9 I 22. Most things that happen to me are ‘VSA STA ASW NAD Dsw STD VSDl NA just a matter of luck. ! 1 2 3 4 5 6 7 . 9 IN THE NEXT GROUP OF QUESTIONS...I WILL ASK YOU ABOUT YOUR USE OF NEWSPAPERS... TELEVISION...RADIO...AND SO ON. YOU WON'T NEED A CARD FOR THESE ANSWERS... 23. How often do you get a chance to read a O NEVER IF NEVER, daily newspaper...like the_Lansing State 1 LESS 1/WEEK GO TO Q 25 Joubnal or the Detroit Free Press? 2 ONCE/WEEK 3 2-3/WEEK 4 4-5/WEEK 1 Probe for Frequency 3 EVERY DAY NA 24. When you do get a chance to read the daily .-—-—---- newspaper...on the average...how many minutes | MINS do you spend reading? . 25. How often do you get a chance to read the 0 NEVER weekly newspaper...the Towne Courier? 1 LESS l/MONTH . 2 1—2/MONTH 1 Probe for frequency 3 ONCE7WEEK | L 4 MORE 1/WEEK , 9 NA I 26. How often do you get a chance to read the 0 NEVER Ingham County News? 1 LESS 1/MONTH ,. 2 1-2/MONTH Probe for frequency ' 3 ONCE/WEEK ‘ 4 MORE 1/WEEK ‘ 9 NA 27. On the average day...how many hours do you watch HOURS TV between 215 in the morning and noon? . -—-—— l (Wynto TO Q 30’} 7 l 28. On the average day...how many hours do you watch , l TV between noon and seven in the evening? ‘—————HOURS 29. On the average day...how many hours do you watch ! TV between seven in the evening and two in the morning 1 HOURS 7 30. 31. 32. 33. 34. 1€37 -5- In an average day...how many hours do you have “the radio on? How often do you get a chance to read a book? Probe for frequency ! When you do read a book...on the average... how many minutes at a time do youllsually read? How often do you read a magazine? | Probe for frequency When you do read a magazine...how many minutes at a time do you usually spend reading? IF R WATCHED ZERO HOURS OF 1!, GO TO Q 42. HOURS NEVER LESS l/MONTH l-2/NONTH 1-2/UEEK 3-4/WEEK EVERY DAY NA \OLnwah-O IIINS } NEVER LESS l/MONTH 1-2/HONTH 1-2/UEEK 3-4/WEEK EVERY DAY NA €)U!$~u)h)h‘C) MINS NEXT...I'LL ASK YOU HOW MUCH YOU WATCH DIFFERENT KINDS OF TELEVISION SHOWS... YOU MAY USE THE BACK OF THE GREEN CARD FOR YOUR ANSWERS... 35. 36. How often do you watch the GAME SHOWS on TV... shows in which peeple play for money and prizes? If R names specific shows...after R is finished...repeat: THEN WOULD YOU SAY YOU UATCHED MOST CAME SHOWS: NOT OFTEN AT ! 0R, VERY OFTEN? IF NEVER, ‘W co T0 Q 33; I ! IF NEVER, GO TO Q 35 1 2 3 4 5 NEV NOA NVO SUO VRO NA 9 How often do you watch the SOAP OPERAS on TV? NEV NOA NVO SUO VRO l 9 1 2 3 4 5 NA 37. 38. 39. 40. 41. 188 -6- How often do you watch SITUATION COMEDIES on TV...shows in which the characters have humorous NEV NOA SWO and funny experiences? How often do you watch the CRIME AND DETECTIVE shows on TV? NEV 1 NOA SW 0 And how often do you watch MOVIES on TV? l NEWS_on TV? [morning news, noon news, evening news, late evening news] How often do you get a chance to watch the ‘ How often do you watch SPORTS PROGRAMS on TV... shows like...Baseball, Football, Wrestling, Tennis, Hockey, Basketball...and other shows like these? NOA SWO NOA SWO NOA SWO 4 VRO 5 NA 9 NEXT...I WILL READ A SERIES OF STATEMENTS ABOUT DIFFERENT ACTIVITIES. HOW MUCH YOU AGREE OR DISAGREE WITH EACH STATEMENT...WE WILL USE THE CARD AGAIN.. . TELL ME USE THE GREEN SIDE FOR YCUR ANSWERS...AND PICK THE ONE BEST ANSWER WHICH MATCHES YOUR FEELINGS... 42. 43. 44. l I would like my daily activities to be i VSA less usual...less routine. 1 1 STA ASW 2 'F J NAD DSW 4 5 STD VSD 6 7 Do you agree or disagree with this statement? Do you SOIIEWHAT . . . STRONGLY . . . 9 VERY STRONGLY;____: Repeat for each S ta tement o I would like to be much more sure that I will visit with my family often. (I 5 6 VSA STA ASW NAD DSW STD VSD 7 1 2 3 l I would like to be much less sure about VSA STA ASW what lies ahead in the near future. 1 2 3 NAD DSW STD VSD 4 5 6 7 1239 -7- \ 45. I would like to be very much more certain that I will have transportation when I need it. 1 2 3 4 5 6 | 7 VSA STA Asw NAD Dsw STD vSD; N' 7 - 9 46. I would like to talk more to people different VSA STA ASW NAD DSW STD VSD‘ NAl from me. 1 2 3 4 5 6 7 ; 9 i 47. I would like to be much more sure 1 could VSA STA ASW NAD Dsw STD VSD] NA! depend on others I know. I 1 2 3 4;: 5 6 7 9 I 48. I would like to be much less sure that my VSA STA Asw NAD ’Dsw STD v health will be the same in six months... 1 2 3 4.- 5 6 i l I SD NA] 7 9 n as it is now. NEXT...I WILL ASK YOU SOME QUESTIONS ABOUT THE PEOPLE YOU TALK WITH IN GRANGE ACRES. THINK NOW ABOUT THE PEOPLE YOUtUSUALLY'TALK WITH' AT LEAST ONCE QR TWICE EVERY TWO WEEKS...ON THE AVERAGE. THINK OF THE PEOPLE YOU TALK WITH MORE THAN JUST TO SAY HI... OR TALK TO FOR ONLY A MINUTE. WITH SOME PEOPLE...YOU MAY TALK MOSTLY OVER THE PHONE...AND WITH OTHERS...MOSTLY FACE-TO-FACE. IF YOU DON'T REMEMBER SOMEONE'S LAST NAME...DON'T WORRY...JUST TELL ME THEIR FIRST NAME...OR WHERE THEY LIVE. NOW...I WILL ASK YOU FOR THEIR FIRST AND LAST NAMES...AND 9: COURSE YOUR ANSWERS WILL REMAIN ENTIRELY CONFIDENTIAL. 51. READY... PROBES: TRY TO THINK OF ALL OF THEM... ANY OTHERS? ARE THERE OTHER PEOPLE YOU TALK WITH AT LEAST ONCE OR TWICE EVERY TWO WEEKS? IF R ASKS: "Why do you need the QQQ£§ZEL.oSAY: WE ARE STUDYING PEOPLE'S COMMUNICATION PATTERNS...AND IN ORDER TO DO THIS...WE TEMPQgARILY NEED PEOPLE'S NAMES. ‘*THIS WILL BE KEPT ENTIRELY CONFIDENTIAL...BUT WE NEED NAMES TO BE ABLE TO MAP OUT COMMUNICATION NETWORKS. _’*NO ONE IN GRANGE ACRES WILL EVER SEE THIS INFORMATION. '*WE WILL QUICKLY CHANGE THEM TO NUMBERS ALD NO ONE ELSE WILL EVER SEE THEM... 190 -3- . .____—_____ —_ E I ASK FOR APARTMENT f ONLY 52. How often do you usually talk with this person? I IF NAME IS INCOMPLETE... _._._—. I BUT RECORD IF FREELY lRepeat for each name4 I HENTIONED BY R. l PROBES: TRY TO THINK OF ALL OF THEM. ANY OTHERS? ARE THERE OTHER PEOPLE YOU TALK WITH AT LEAST ONCE OR TWICE EVERY TWO WEEKS? ITS IMPORTANT THAT WE GET ALL THE NAMES. 1' SEv. IONCE 2-3 - NAME A APT# DAY I’DA‘! WK 2NK 191 -9- NEXT...I WILL ASK YOU ABOUT THE PEOPLE YOU TALK WITH WHO DO NOT LIVE IN GRANGE ACRES. THINK OF THE PEOPLE YOU COMMUNICATE WITH AT LEAST ONCE OR TWICE EVERY TWO WEEKS... ON THE AVERAGE...THROUGH EITHER MAIL...PHONE...OR FACE-TO-FACE. THINK OF FAMILY...FRIENDS...AND ACQUAINTANCES...0R OTHERS. AND AS BEFORE...THESJ SHOULD BE PEOPLE YOU TALK WITH MORE THAN JUST TO SAY HI...OR TALK TO FOR ONLY A MINUTE. 52. How many people do you talk with not living in Orange Acres... at least once or twice every two weeks on the average? PROBE: FAMILY?-..FRIENDS?...0THERS?... I IL...PHONE...FACE—TO-FACE. 'TALLY HERE while respondent mentions names and ‘ 'record total number in box. . # PEOPLE # LIVE MERID , I l 53. How many of these people Live in Meridian Township... ! Haslett...or Okemos? I AND NOW...WE HAVE JUST A FEW FINAL QUESTIONS... 54. What kinds of progrzm: for Senior Citizens would you like to see on television? PLUS TWO PROBES 55. And how long have you liv;d in C:ange Acres? 56. Do you [or your husbaHJ/wife] have a car to drive when you want to go somewhere? ; YES NO NA I I I 57. And what was the name of the 1n:t school you attended? 3 Ask but do not code 192 -10- 58. How many years of school did you complete? 7 8 9 1N2" 1_7 18 19 + 1 2 3 4 5 6] l71_<_)_‘1~1 I13 liil—S 1—_—1 ‘21 What kind of work did you [or your husband/wife (if you were married)] do most of your life? 59. May I ask your age? PROBE for detailed " J " ' ' i . 'IF NOT YET KNOWN-”ASK: HAVE YOU BEEN MARRIED?: I NEVER MARRIED ‘ MARRIED NOW WIDOI-I(ER) DIVORCED . NA \l‘bL-JNH . "L...THAT COMPLETES OUR INTERVIEW. ..THANK YOU VERY MUCH FOR YOUR COOPERATION... THE INFORMATION YOU HAVE PROVIDED WILL BENEFIT MANY SENIOR CITIZENS. I IF R ASkS: 'How will this information be used? SAY: THE RESULTS 1 i WILL BE USED TO IMPROVE THE WAYS IN WHICH SENIOR CITIZENS CAN 1 1 GET THE INFORMATION THEY NEED. I DO N__OT SAY ANYTIII£G_ [-21 SE ..IF R WANTS TO KNOW MORE...SAY: ' I DON'T KNOW THE DETAILS...BUT IF YOU WANT TO KNOW MO" E... : YOU CAN CALL MR. DANOWSKI AT 485- 2565...HE WILL BE HAPPY TO I I ANSWER YCJR QUESTIONS. DO IQTASK THEJOLLOILING. ..BUT CODE FROM YOUR OBSERVATIONS. R's mental acuity? [verbal skill, wandering on answers, 1 memory] Hf‘ O MED HI 5 l R's physical mobility? [unconstrained movement, rec use of limbs, able LO MED HI 1 2 3 4 5 to walk easily] 1 I I' ——-—-—————— , I Quality of R's furnislings? (media equipment furn— ILO MED HI iture, artifacts] l 2 3 4 5 _} I Sex? [FEMALE MALE ‘ 1 2 HICHIGR N STATE UNIV. LIBRARIES I WHIMWHIWWIWWIIIWWNWII 931 01 990590 312