‘ G in...- E 9. V 35 ABSTRACT BEHAVIORAL CORRELATES 0F PERCEIVED STRESS IN THE URBAN ENVIRONMENT: SPATIAL RESTRICTION IN METROPOLITAN DETROIT by Robert M. Pierce The aims of this study are threefold: l) to determine if there is a relationship between a characteristic of the environment, Spatial restriction, and behavior indicative of social well-being, 2) to indi- cate where in the urban environment that relationship is manifested and 3) to explain why the relationship is found in certain parts of the city and not in others. Behavioral scientists, primarily psychologists and sociologists, have in recent years increasingly focused their research interests upon various forms of density as stimuli which influence behavior detrimentally. Results of animal research are consistent with this conclusion, but aggregate as well as laboratory findings on human behavior are far more inconclusive. At present it cannot be definitively stated that density does or does not affect human behavior. It is the initial task of this research, then,to continue seeking answers to questions previously asked regarding the hypothesized relationship between place and behavior. This research makes three major contributions to density-related behavioral research. First, it expands the concept density by including several forms of concentration under the guise of spatial Robert M. Pierce restriction. This broadens the focus of the problem, which has previously been analyzed in terms of population or room density only. Previous analyses have been carried out at the macro level of an entire city, or the micro level of the laboratory. Both forms of research help to determine if density is related to behavior, but neither provides an indication of where in the city we will find high levels of spatial restriction and related behavior. The second major contribution of this research is an intra-city, spatial approach to the problem. This spatial approach is particularly important because it indicates where in the city high restriction abounds and to what degree certain forms of behavior co-occur with this environmental characteristic. The third major contribution of this study is the development of an environment-perception-behavior framework which attempts to explain the spatial patterns of restriction-related behavior. By analyzing the environmental perceptions of individuals occupying high restriction locations, it is possible to determine how the individual evaluates such surroundings. It is hypothesized that restrictive environments are perceived as stressful by their occupants, and that those perceiving their environment in this threateninq manner will engage in behavior that is potentially harmful to themselves and others as a response to this cognition of Stress. By gathering evidence in the field of perception of the environment, it is believed that this research makes a significant contribution to density-related behavioral research. This study not only indicates if and where density is associated with behavior, but offers perception of environmental stress as an Robert M. Pierce explanation of why the elements present in a location may or may not affect behavior. Two levels of analysis, aggregate and individual, were used to test the hypothesized associations between spatial restriction, stress and behavior. Detroit was selected as the study site for this analysis. Census tract data measuring room, household, and structural densities were correlated with selected behavioral indicators of social well-being.v Among these social indicators were family abandonment by parents, marital unrest, children running away from home, and failing to complete high school. Various forms of density, particularly population and structural concentrations, proved to be significantly associated with the behavior -- but primarily in the inner-city and among the black segment of Detroit's population. Results of the aggregate analysis suggested that the physical environment in the form of spatial restriction does co-occur with behavior. It is a behavioral problem to be reckoned with, but is secondary to the difficulty of being black and poor. Analysis of individual perceptions of the environment does much to explain the aggregate findings. Spatial restriction, more than any other element in the person's residential location, is the major contri- butor to perceptions of environmental stress. Where restriction and frequencies of behavioral social indicators are high, perceptions of stress are equally great. Populations residing in restrictive environments perceive their surroundings in significantly stressful terms. These results suggest that perception may indeed have a role in responses to the environment. The results of the individual analysis also confirm that in the final assessment, the question of stress and social well-being cannot Robert M. Pierce be considered solely in terms of the physical environment. It is equally if not more stressful to be black and poor. Any future attempts to implement programs designed to reduce stress and such behavior must address this fact. BEHAVIORAL CORRELATES OF PERCEIVED STRESS IN THE URBAN ENVIRONMENT: SPATIAL RESTRICTION IN METROPOLITAN DETROIT By Robert MIsPierce A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Geography l97h ACKNOWLEDGMENTS Five of the most important years of my life have been spent at Michigan State University. They have been important not only because they have helped me advance towards a career goal, but also because they have provided me an immeasurable degree of personal growth. This brief statement cannot adequately indicate the thanks I owe my department, advisor and many others, but it is hoped that they will more fully realize the keen sense of gratitude I feel. The Geography Department at Michigan State has been more than generous with its financial support during my stay. Without its help I could not have completed my formal training. To Dr. Lawrence Sommers goes a special thanks for his confidence in my teaching abilities. The experience I have gained through his support has been of considerable importance to my professional future. Doctors Gary Manson, James Zuiches and Joe Darden have each been instrumental in successfully completing my degree. Dr. Manson gave much of his time and Dr. Zuiches and Dr. Darden raised those difficult questions which resulted in a more exacting dissertation experience. In the final analysis, credit for whatever real success I have achieved as a graduate student must be given to my advisor Dr. Stanley Brunn. Dr. Brunn has provided me invaluable guidance in my program development and the com- pletion of the dissertation. I shall not forget the many challenging sessions with him in which I was forced to think, reevaluate and improve. All students need some degree of encouragement and Dr. Brunn provided that stimulus. As a final acknowledgment, I must certainly thank my wife Barbara. She had to endure long hours of hard work on my behalf as well as a considerable degree of neglect during a very difficult year. I wish to all students as compassionate and understanding a companion as I have had. METROPOLITAN :5-"' caJm u DETROIT Suurce: R. Sinclair, The Face of Detroit 9 In ..’ 5- uno- ' u..— l - I ray. "IUCIHL CITIIS d ‘l CHICAfl TABLE OF CONTENTS LIST OF TABLES . LIST OF FIGURES CHAPTER I. INTRODUCTION SCOpe of Problem The Role of Perception Definition of Terms . The Spatial and Interdisciplinary Perspective. The Role of Geography. Significance of the Individual. Study Objectives 2. REVIEW OF THE LITERATURE Animal Research. . . Correlational Studies. . Perception and Territoriality . Laboratory Research . . Conclusions from the Literature Review . Implications of Previous Research for this Study. Theoretical Ties Between Spatial Restriction and Behavior . . Statement of Hypotheses . 3. ECOLOGICAL ANALYSIS OF SPATIAL RESTRICTION AND BEHAVIOR. . . . . . . . Selection of Study Site . Data and Methods of Analysis A Spatial Analysis of Place Characteristics Variation in Spatial Restriction . Spatial Variations in Behavior. Conclusions from Analysis The Association of Race, Spatial Restriction and Social Indicators. A. PERCEPTIONS OF ENVIRONMENTAL STRESS. Stress as a Meaning Page viii 53 SA SA 67 73 80 95 98 IIO III Chapter Page The Measurement of Stress. . . . . . II2 Application of Semantic Differential to Study . . . . I18 The Questionnaire . . . . . . . . . . I22 Administration of Questionnaire and Sampling Framework . . . . . . . . . . I28 Perceptions of the Environment Among Selected Detroit Residents . . . . . . . . . I36 Comparison of Perceived Stress . . . . . . . . Ih3 The Nature and Degree of Association Between Environment and Perception. . . . . . . I52 The Association of Environmental Stress and Behavior. . . . . . . . . . . . I6l- 5. CONCLUSIONS, IMPLICATIONS AND RECOMMENDATIONS. . . . . I67 Comparison of Research Findings. . . . . . . . . I70 Suggestions for Further Study . . . . . . . . . I73 Implications of Findings . . . . . . . . . . . I75 Recommendations . . . . . . . . . . . . . . I77 APPENDICES Appendix A. QUESTIONNAIRE . . . . . . . . . . . . . . I8I B. CODING OF SURVEY DATA. . . . . . . . . . . . I83 C. AGGREGATE VARIABLE ABBREVIATIONS . . . . . . . . . IBA D. INDIVIDUAL VARIABLE ABBREVIATIONS. . . . . . . l85 E. SIMPLE CORRELATIONS OF AGGREGATE ENVIRONMENTAL AND BEHAVIORAL CHARACTERISTICS - BLACK TRACTS . . . . I86 F. SIMPLE CORRELATIONS OF AGGREGATE ENVIRONMENTAL AND BEHAVIORAL CHARACTERISTICS - ALL TRACTS. . . . . I88 G. MEANS AND STANDARD DEVIATIONS OF ALL AGGREGATE CHARACTERISTICS. . . . . I90 H. SIMPLE CORRELATIONS OF ALL INDIVIDUAL CHARACTERISTICS . . I9I I. MEANS AND STANDARD DEVIATIONS OF ALL INDIVIDUAL CHARACTERISTICS . . . . . . . . . . I92 BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . I93 Table I0. II. LIST OF TABLES Correlations of Density, Income and Percent Nonwhite with Juvenile Delinquency and Mental Illness . Stepwise Multiple Regressions of Income, Percent Nonwhite, and Density on Delinquency and Mental Illness . Variables Used in Aggregate Analysis Spatial Restriction and Aggregate Behavioral Factors: Rotated Factor Loadings Factor Scores for Selected Inner-City and Suburban Census Tracts: Low Density Behavior and Room Density Aggregate Suburban Behavioral Stability and In-Migration Factor Scores . . Racial Contrasts in Housing and In-Migration: Detroit MetrOpoIitan Area Inner-City Contrasts in Residential Change in Black Environments Residential Stability in Black Environments Black Migration in High Pathology Locations Aggregate Social, Economic and Restriction Factors: Rotated Factor Loadings Page 29 29 55 69 81 87 89 92 93 95 97 Table I2. I3. I5. l6. I7. 18. I9. 20. 2'. 22. 23. 2h. 25. 26. 27. 28. 29. Restriction-Pathology Associations By Race . Stepwise Regression of Aggregate Behavior in Black Environments Variance in Aggregate Behavior Explained by Restriction Elements in Black Environments Simple Correlations of Black Mobility with Spatial Restriction. Stepwise Regression of Mobility in Black Environments Semantic Differential of the Measurement of Perceptions of Stress. Characteristics of Sample Census Tracts. Comparison of Sample and Tract Population Characteristics Environmental Perception Factors of Semantic Differential Dimensions. Comparison of Environmental Perception Factor Scores with Aggregate Density and Behavior Analysis of Variance: Positive Perception of Environmental Quality - Five Groups Analysis of Variance: Positive Perception of Environmental Quality - Ten Tracts Comparison of Means Test for Positive Perception of Environmental Quality - Five Groups Comparison of Means Test for Positive Perception of Environmental Quality - Ten Tracts. Minimum Stressful Density Levels Simple Correlations of Environmental Perception Factors with Social and Restriction Characteristics . Stepwise Regression of Social and Restriction Characteristics with Positive Perception of Environment Stepwise Regression of Social and Restriction Characteristics with Perceptions of Environmental Quality Change. vi Page IOO IOA I05 I06 108 I20 I3I I35 I37 IhO Ihh HIS 1&7 I48 I49 I53 I56 I58 Page Table 30. Spatial Restriction Factors: Rotated Factor Loadings I60 3I. Simple Correlations of Environmental Perception Factors with Individual Behaviors I62 32. Stepwise Regression of Individual Behaviors I64 with Social and Restriction Characteristics vii Figure I0. II. I2. I3. IA. LIST OF FIGURES Statement of Problem . Census Tracts - Detroit Urbanized Area Generalized Land Use I965. Factor Three - Room Density Room Density - Median Persons per Room Factor F0ur - Apartment Dwellings. P0pulation Density Factor One - Low Density Behavior. Factor Five - Divorce Environment. Factor Two - New Residents Black POpulation Schematic of Environmental Meaning Factor One - Low Density Behavior. Categories of Census Tracts Selected for Surveying viii Page 58 63 75 76 78 79 82 85 86 88 II6 I30 I32 u a and f‘. I!- CHAPTER I INTRODUCTION An ”environmental” problem exists if we can detect a disruption between the designed environment and the system of behaviors which is to be accommodated by that environment. The most prominent of man's designed environments is the city and it is the city which may also be the cause of numerous physical and social maladies. While Western cities have over- come the majority of problems of santiation and hygiene which beset the first phase of urban growth, health problems of a different nature have arisen. These problems consist of an increasing number of disturbances in interpersonal relations and disruptions of social behavior which seem to stem from the process of living in the city itself.I The link between the environment of the city and the incidence of diseases such as cholera and typhoid fever is rather clear: in the early twentieth century the quality of drinking water in many American cities was sufficiently poor to cause high rates of such illnesses. More recently it has been noted that urban air pollution is probably responsible for Wirth, Louis, ”Urbanism as a Way of Life,” American Journal of Sociology, vol. #0 (l938), pp. I-Zh; Simmel, Georg, ”The Metropolis and Mental Life,” in P.K. Hatt and A.J. Reiss, eds., Cities and Society, (Glencoe: Free Press, I969), p. l00. increasing rates of respiratory illness.2 But determining whether the urban environment is responsible for the occurrence of such behavioral maladies as mental illness, high crime and family instability is much more difficult; hence the task of the behavioral scientist is far more demanding than that facing the physical etiologist. In attempting to account for the presence of a series of behavioral disorders, the scientist is presented with a myriad of possible explanations. For example, high rates of crime, suicide, mental illness and family discord may be explained as much by social and racial factors as the spatial character of the urban setting. Scope of Problem It has been asserted that many of the ills of the city are due to overpopulation. Apparently too many people occupying too little space leads to infringements upon basic needs for privacy. These avenues of research have led to the suggestion that population concentration is in- jurious to an individual's physical and mental stability. The population concentrations found in American cities have led some to doubt the possi- bility of a desirable quality of life for the inhabitants of densely settled urban settings. The amount of space available to the urban resident is only one of many characteristics of his setting which could possibly affect his physi- cal and social well-being. Air pollution, noise, crime and poor delivery of social services could all detrimentally affect the individual. The scope of this dissertation is an environmental one, for its ultimate aim is to determine whether there is a link between place and Z McDermott, Walsh, ”Air Pollution and Public Health,” Cities: Their Origin, Growth and Human Impact, readings from Scientific American, (San Francisco: W.H. Freeman and Co., l973), pp. l32-lh0. behavior. Are the disruptions in behavior among urban residents due to the nature of their location or can they be explained by other characteristics? In particular, can it be shown that the amount of space made available to the urban resident, compared with other elements in the city environ- ment, contributes to the occurrence of behavior potentially detrimental to individuals and populations. The Role of Perception - It is conceivable that the city directly influ- ences behavior in the sense that the city as a stimulus elicits particular responses. However, it is more likely that the manner in which the individual mentally structures and evaluates his surroundings mediates the stimulus and response, i.e., environmental behavior. Geographers call this mediation process perception. _To what extent perception in turn influences behavior has not yet been determined, but we do know that it is often misleading to view the problem in terms of direct environmental influence. A more realistic assessment of the possible relationship between the urban environment and human behavior should include an analysis of the intervening role of perception. Few research efforts have attempted to analyze the process through which the population concentration dimension of the urban environment has become associated with behavioral disruptions. Those who have made such endeavors have invariably turned to the concept stress to link environment with behavior. But stress is an ambiguous concept; most etiological studies invoking stress as a causal factor in the incidence of such pathol- ogies as mental illness have defined it differently. Definitions may be given in terms of stimuli, such as electric shock or the loss of a loved one, that lead to changes in the individual; or the definition may be placed in terms of the outcome from such stimuli (hypersturcture, increased blood pressure or personality disintegration). In some cases stress has referred to the emotional state or experience, a changing personal or social situation. Accompanying such definitional difficulties are a num- ber of questionable assumptions about stress: I) all unpleasant environ- mental conditions are stressful; 2) what is stressful for one person will inevitably be stressful for others; 3) events stressful for an individual will lead to behavioral disorder. While there appears to be no universally accepted criterion or de- finition of stress, there does appear to be some consensus in the social psychological literature that stress is closely related to perception. Lazarus has proposed that the distinguishing feature of psychological stress is the cognition of ”threat-cues" by the individual.3 Researchers have often concluded that certain elements in the environment are ”stress- producing” or that given stimuli are ”stressful” without realizing that there first must be evidence that someone finds them so. If we can determine that individuals or populations perceive cer- tain stimuli as stressful, i.e., that an environment with high population concentration means stress, then changing or removing such a stimulus con- ceivably will alter one's perception of his environment. Stress, then, can be defined within the context of one's perception of his urban envir- onment and provides a useful construct for analyzing urban environments from the perspectives of their residents. It is defined in this study only in environmental and cognitive terms. It is fully realized that 3 Quoted in Glass, David C. and J.E. Singer, Urban Stress: Exper- N ' S ' r , (New York: Academic Press, I972). .Ub \ . pr— .\. I :- .Fv '5'.- .u' ~A¢ stress has physical connotations, but it is believed that environmental stress is a perceptual phenomenon that takes place through the mental evaluations of a location by its occupants. In summary, the problem of this dissertation is one of place and behavior, and a concern for the mediating role of perception. Using terminology from geography and psychology, the problem can be stated graphically (Figure l). The environment, in the form of spatial restriction, FIGURE I Statement of Problem lStimulusI )IMediationf--—-—%fl§§iiiggfl / \ V Environmentk )[Perceptifle Behavior] \/ \/ Spatial \- 3 Social [Restriction ’j7 Bilafla Indicators is a stimulus which is mediated by the perception of the individual. If that perception of the environment is stressful, he is likely to respond in ways that are potentially harmful to himself and others. Definition of Terms In this analysis spatial restriction is regarded as the independent parameter and behavioral social indicators as the dependent parameter. Spatial restriction is defined as the density of people and structures which combine to affect the amount and quality of contact between indivi- duals and populations. It is to be distinguished from p0pulation density or concentration, the number of inhabitants per unit of area, in that spatial restriction also takes into account the manner in which territory is occupied. It not only refers to the number of individuals occupying a square mile of territory, but also the proportion of that mile used for residential purposes and the structural denSity of the area. Operationalizing the amount of space available to each individual in terms of simple popula- tion density ignores the effect of room size, structural density and building height. A group of I00,000 people living in a series of single-family homes is a much different environmental situation than if that same number was confined to a few high rise apartments. Twenty individuals living in a house surrounded by acres of open land are spatially restricted, but the density of inhabitants for the total area would actually be quite low. In essence, spatial restriction is a term used to refer to various forms of density, e.g., room, household, structural as well as population concen- tration. By using several forms of density, spatial restriction becomes . a more encompassing parameter of population concentration. The behavioral descriptor harmful or ”pathological" appears repeat- edly in the psychological literature. In some cases it refers to behavior suggesting a personal breakdown, a distuption of personality development, or behavior that is in some way ”deviant” or ”abnormal”. In other cases it has referred to behaviors suggesting social disorganization, Normality, in its usual sense, is a statistical concept. When applied to behavior it has traditionally referred to patterns that fall within an acceptable range in a given context. Thus behavior can only be regarded as normal or abnormal when its cultural context is taken into 5 account. The disadvantage of this use of the term is readily apparent. Michelson, William, Man and His Urban Environment: A Sociological Approach, (Reading, Mass.: Addison-Wesley, I970), p. 56. 5 Howard, A. and R.A. Scott, ”A Proposed Framework for the Analysis of Stress in the Human Organism,” Behavioral Science, vol. l0 (I965), p. l5l. I.) -§.. PaU «1. O I. I \ H b It eliminates the possibility of a cross-cultural or universal measure- ment of what is "normal” behavior. Even within the United States there is a sufficient range of cultures and norms to eliminate the use of the term. What is abnormal in the suburbs may be well within the acceptable standards of behavior in the central city. Behavioral norms, therefore, are not appropriate criteria in the selection of activities which can be considered as disruptions between the designed environment and its inhabitants. Physical breakdowns are more easily delimited, but the focus of this study cannot be exclusively confined to the narrow problem of body tissues. Pathologies in the city extend far beyond increased incidences of lead poisoning, emphysema and death and injury due to automobile traffic. Individuals respond to their surroundings in many ways and physical malfunction is only one. In order to make comparisons by population and locale, behaviors must be selected so that they indicate in some way the ”social well-being” of a population.6 Social indicators, which are social, economic and behav- ioral statistics describing the conditions of major aspects of a society, define such activity. These behavioral social indicators are direct mea- sures of welfare, and subject to the interpretation that, if they change in the ”right” direction, while other things remain equal, things have gotten better or people are ”better off.”7 Behaviors commonly used as social indicators include: crimes of violence, narcotics violations, divorce, children not residing with their parents, suicides, dropping out of high 6 Smith, Daivd M., The Geography of Social Well-Being in the United States, (New York: McGraw-Hill, I973). 7 Toward a Social Report, U.S. Department of Health, Education and Welfare, Washington, D.C., l969, p. 97. school and mental hospital admissions.8 Social indicators form the para- meter of behavior that this study focuses upon, and forms of spatial res- trlction will then be analyzed to determine to what degree they can account for and explain the variation of this behavior within a metropolitan area. The Spatial and Interdisciplinapy Perspective One task of socio-medical research is to identify human tolerances for different physical and psychological strains imposed by the environ- ment over various periods of time. Some environmental effects can be inferred from mortality statistics, but an overall hygienic assessment should also include health problems which cause temporary or chronic illness, prolong a period of convalescence or are felt as a malaise, and are generally imcompatible with otherwise rising material standards of living. Statistics reflecting the lingering and less dramatic strains of the environment are extremely limited, but if we are to have a healthier city, more detailed knowledge will be needed as a basis for dealing with the intricate effects of the city on the social health of its inhabitants.9 How should a study of this nature be approached? We know that popu- Iation concentration varies spatially within a metropolitan location,'0 however, we do not know jf_there are behavioral influences of spatial restriction; or where they take place. Psychologists have attempted to measure behavioral influences of spatial restriction in laboratory tests, 8 Smith, op.cit., pp. 70, 9h, IO“. 9 Charlestan, Gosta, ”The Individual, The City and Stress.“ in L. Levi, ed., Society, Stress and Disease, (London: Oxford University Press, l97l), p. lh3. Newling, Bruce E., "The Spatial Variation of Urban Population Densities,” Geographical Review, vol. 59 (I969), pp. 2&2-252/ but concern must be maintained for the manner in which restriction and behavior vary from place to place in the city. Only through this spatial perspective can we fully determine where behavior due to restriction occurs. A spatial approach to the problem of behavior will tell us where it occurs but not tell us why it occurs. Answering the latter question requires the analysis of the individual's symbolic and affective view of his sur- roundings. Evaluation of the environment requires man to interpret the physical and social components of his stimulus field, an area of inquiry shared by the disciplines of geography and psychology." Geography has long dealt with the stimulus properties of the environment -- weather, tOpography, city form -- but has only recently considered the symbolic responses men make to those stimuli. Psychology has placed more emphasis on personal attributes arising out of functional and symbolic transactions between men and other stimuli, but those stimuli have been primarily confined to the laboratory or an artificial setting. However, the concerns of both disciplines are similar -- the environment as a stimulus, and man's response to it. (Both geographers and psychologists analyze subjective evaluations of the environment. In geography, man's evaluation of his surroundings is increasingly being linked to his behavior in various settings. The main focus is the individual's mental representation or image of the environment, which is developed on the basis of experience and which serves as his data source for decision making. An individual's attitudes, values, ways of structuring his experiences, and judgments of physical, " Beck, Robert, ”Spatial Meaning and the PrOperties of the Envir- ¢3nment," in Raymond Proshansky, et al., eds., Environmental Psychology: flan and His Physical Setting. (New York: Holt, l970), pp. I3K:l35. l0 aesthetic, and symbolic objects all affect modes of adaptation to envir- onmental elements. The manner in which knowledge of the environment is sorted and structured in the mind is termed perception. From an environmental perceptual viewpoint, space or territory has meaning according to the purpose for which it is sought and the circum- stances under which it is experienced. In the context of spatial restric- tion, it is clear that judgments of utility or acceptability are not based entirely on objective measures of open space or monetary value, but also include the potential of the space to provide for privacy, isolation, or freedom from whatever one finds distasteful in his environment.'2 That potential will vary as much according to the individual as it will to the physical nature of the space. Linking place with behavior is a most demanding objective. It requires the philosophies and methodologies of disciplines other than geography, particularly psychology. Methodologies differ but these differences offer benefits to the researcher who utilizes the resources of both disciplines. Geography's research is organized by area, is carried out in the field, and is marked often by generality of vision at the expense of precision of insight. Psychology in the main has ignored the broader physical environment, even though it has studied intensively the perception and symbolization of discrete stimuli. Its carefully con- trolled laboratory research has often been marked by precision of insight obtained at the expense of the ability to extrapolate its findings to a different scale. Yet findings and principles drawn from psychology may 2 Sonnenfeld, Joseph, ”Variable Values in Space Landscape: An Inquiry into the Nature of Environmental Necessity,” Journal of Social Issues, vol. 22 (October I966), p. 26. ll be highly suggestive for studies relating man's behavior to the environ- l3 ment in which he lives. The strong ties between the two disciplines, help to solve the problems of if, whgpg, and why, The study of environmental perception continues to be plagued with basic problems of definition and conceptualization. In one of the most recent statements on the status of research in environmental perception and behavior Lowenthal notes the current lack of commonly accepted definitions, objectives and mechanisms. Above all, studies in perception and behavior require a more systematically organized theoretical base.'h One of the most important results stemming from research could be the development and application of a model that adds to the understanding of human behavior. Too little work has been conducted by behavioral geographers in the past which goes beyond the model building stage. As a result we have a series of frameworks which purport to explain human behavior, but little empirical evidence to support these theories. It is believed that through an empirical, nonlaboratory testing of the model such as that developed in this study, an environment-perception-behavior frame- work can be demonstrated that has general application in the study of perception and behavior. This would be a most important contribution to meeting one of the major needs of the field: a systematically organ- ized based of theory. In summary this research will provide an analysis of the roles played by perception and environment in the incidence of behavioral social '3 Kates, Robert W., ”Stimulus and Symbol: The View from the Bridge," Journal of Social Issues, vol. 22 (October I966), p. 26. 'h Lowenthal, David, ”Research in Environmental Perception and Behavior, Perspectives on Current Problems,” Environment and Behavior vol. A (September l972), pp. 333-342. l2 indicators as they vary from place to place in the urban setting. If it can be shown that spatial restriction is an important behavioral stimulus, this information can be used in designing new towns and cities, and in formulating urban renewal projects. Such a discovery would be highly significant to those behavioral sicentists and urban planners interested in improving the quality of life in our cities. The Role of Geography - With its emphasis upon behavior, perception and interdisciplinary analysis, the question may be raised as to the geo- graphical nature of the study. Numerous statements have attempted to delimit and define the discipline by its subject matter, but the spatial perspective of the geographer has remained a common theme. This interest in distribution and in spatial variation appears to be the common denom- inator that binds the discipline together. The basic subject of this dissertation is human behavior. Two avenues of study will be taken in its analysis: The analysis of place as it influences behavior and the distributional analysis of behavior. The latter refers to the study of behavior as it varies from place to place, and falls within the traditional confines of geography. The study of behavioral influences of location has also long been of interest to geographers and developments in this branch of the field have been fundamental in its evolution. At the turn of the century American geo- graphers were making deterministic statements regarding the environment's impact on human activity. Sauer was prominent among those leading the discipline away from such an interpretation, suggesting that geography cast off its adherence to determinism as other social sciences had already begun, and rely instead on historical and comparative methods without I3 recourse to causation.IS Today, we see a return to the study of environ- mental influences as a new research tradition in the discipline.'6 The study of behavior in space aims to determine how physical and social elements of a particular location affect human activity, but necessarily considers the all-important function of the individual in its analysis. In this manner deterministic statements can be avoided. The distinction is very similar to the evolution of behavioralism in psychology. This is not to infer, however, that mankind is above the impress of his surroundings To what degree the urban American is affected by his environment is the problem of this study -- a subject that is traditionally geographical. Significance of the Individual - During recent years many countries have been obliged to take stock of increasing rates of alcoholism, crimes of violence and attempted suicide. Sociological and social-psychological research has shown that clusters of these disturbances are found most commonly in overcrowded underprivileged sectors of the large cities.'7 Explanations for this phenomenon have taken various forms, including the suggestion that physical conditions in the city, particularly in the slum, act adversely upon the resident. Glass and Singer, in their analysis of stress in the city, have concluded that life in the city is an endless round of obstacles, conflicts, confrontations and inconveniences. The '5 Chappell, John E. Jr., ”0n the Causation in Geographical Theory,” Proceedings, Association of American Geographers, vol. I (I969), p. 37. '6 Kohn, Clyde, ”The l960'si A Decade of Progress in Geographical Research and Instruction,“ Annals, Association of American Geographers, vol. 60 (June I970), pp. 2ll-2l9. '7 Carstairs, George M., ”Overcrowding and Human Aggression,” Hugh D. Graham and T.R. Gurr, eds., The Histogy of Violence in America, (New York: Bantam Books, I969), p. 598. IA urban dweller is confronted daily with noise, litter, air pollution and overcrowding. Some of these conditions are pervasive. Others occur only at home, or at work, or in transit. ”Their incidence is profoundly disturbing...such conditions produce behavioral and physiological con- sequences inimical to the health and well-being of man.”' But an observed departure from the normal, while it may have a statistical connection with the city itself, does not necessarily have a causal connection. It is therefore difficult to determine whether a specific disturbance derives from the physical conditions of the city or is more attributable to the general pattern of urban society -- or has no connection with the city whatsoever. Observation of urban life cannot but lead to the conclusion that despite the unpleasantness of conditions, it does go on. Either the urban environment is not as potent as popu- larly believed or else man has the capacity to adapt or ward off the detrimental effects of his surroundings. While it is correct that urban residence does not automatically inflict behavioral disorder, it nevertheless seems true that there is some connection between the two. A crucial factor in the case of spatial restriction may be the absolute number of individuals who are forced or required to interact with one another. When the amount of space per person interferes with any activity or causes too much or forced social interaction, one will be conscious of being spatially restricted. To the extent that this does not occur, spatial restriction will be less l9 apparent. l8 Glass, David E. and J.E. Singer, Urban Stress: Experiments on Noise and Social Stressors, (New York: Academic Press, l972), p. S. '9 Freedman, Jonathon L., ”Conceptualization of Crowding,” Population Distribution and Policy, (Washington, D.C.: U.S. Government Printing Office, I973). PP. 507-508. l5 Whether speaking in terms of the general urban environment or the element of spatial restriction, it seems clear that the role of the individual may be significant in determining whether disruptions will occur in the environment-behavior relationship. In each of the posited explanations, variations in individual adaptation levels are the keys to the appearance of social indicators. This reemphasizes the need to consider the role of individual perception, and the importance of avoiding the use of a stimulus-response model. Study Objectives - Spatial restriction has been defined in strictly objective terms; yet the above discussion intimates that it is the indi- vidual's subjective evaluation of his surroundings which is the basis for understanding variations in behavioral responses. Subjective indicators are needed to supplement objective measures of spatial restriction for the very reason that an individual's response to a set of circumstances is dependent, not only on those circumstances as viewed obejctively, but on a whole set of values, attitudes and expectations that he brings into the situation.20 Thus being spatially restricted is not merely having so many inhabitants and residential structures per square mile. It is also very much a subjective experience. Just as objective indicators can be misleading and will remain so until measures of the human meaning attached to them are obtained, so are subjective indicators by themselves insufficient as guides to policy.2' Therefore, this analysis of spatial restriction 0 Marans, Robert and Willard Rogers, Toward an Understanding of Community Satisfaction, Mimeo, Institute for Social Research, (Ann Arbor: University of Michigan, December I972), p. 7. 2' Ibid., p. IO. l6 will be conducted on two levels: objective and subjective. The objective analysis will examine behavioral correlations with measures of physical space, e.g., persons per room and persons per square mile, and the sub- jective analysis will inSpect correlations between behaviors and the evaluation the individual attaches to his surroundings. By conducting the study from a dual perspective, we can determine I) if physical measures of spatial restriction are perceived as confining; 2) what aspects of restriction, objective and subjective, are closely associated with selected forms of behavior; and 3) if behavioral correlations with objec- tive measures of spatial restriction can be explained by evaluations of the environment by urban residents, or by other factors, such as race or poverty. CHAPTER 2 REVIEW OF THE LITERATURE In order to develop research hypotheses and determine appropriate methods for testing them, it is necessary to review previous research which has concentrated on behavioral responses to spatial restriction. A review of previous research serves as a reference point which can be used as a guideline for meeting the author's own research objectives. In addition to developing hypotheses and a methodology, a survey of research findings can help provide meaning to findings obtained in this study. This chapter investigates how others have approached the problem of restriction and behavior and the present status of the hypothesized link between the two themes. From this review a series of research hypotheses will be developed for testing. Three basic lines of research have been related to the issue of spatial restriction: aggregate level correlational studies, research on the human use of space, and laboratory studies concerning the effects of spatial restriction on human behavior. Each has produced varied results while treating important methodological and conceptual issues. What are the physical and social consequences of a continued, massive concentration of human populations? As yet, the science of human behavior is not sufficiently developed to answer this question l7 l8 with precision, or even with confidence.' However, it is possible to learn from studies of animals, both in their natural environment and under experimental conditions. Certain regularly occurring consequences of severe spatial restriction are reported in animal populations. This research warrants some attention. Animal Research - In I962 John Calhoun published his poineering study which detailed the ways in which extreme population density, defined as the number of Norway rats occupying a specified amount of territory, affected their behavior.2 In his experiment he provided the rodents with sufficient food and water, but held constant the amount of space they occupied. A population increase meant the amount of space for each member decreased. Under these conditions Calhoun observed the effects of what he described as ”pathological togetherness.” As population density increased through natural reproduction, conflict cycles among males began to disrupt the social structure of the population and its spatial distribution. Dominant males horded females in inaccessible pens located adjacent to a large central pen. Subordinate males and other females crowded into the central area and it was this portion of the population that demonstrated pathological social patterns. In this ”behavioral sink'I females abandoned their litters, failed to make nests and failed to bring their pregnancies to full term. Cannibalism occurred and some males in the crowded central area demonstrated sexual deviance, Carstairs, George M., "Overcrowding and Human Aggression," in Hugh D. Graham and T.R. Gurr, eds., The Histofiy of Violence in America, (New York: Bantam Books, I969), p. 59%. 2 Calhoun, John, "P0pulation Density and Social Pathology,“ Scientific American, vol. 206 (l962), pp. l39-lh8. I9 with some becoming homosexual or pansexual. Others demonstrated a behavior which Calhoun called ”probing.” Probing rats were sexually hyperactive males and often cannibalistic, forcing intercourse on females in their nests, something normal male Norway rats never do. Research dealing with other animals has indicated that rats are not alone in being adversely affected by a high density environment, but Calhoun's work has received the greatest attention and is representative of this form of spatial re- striction research. The applicability of data from animal research to the analysis of human spatial restriction is open to question. There is strong evidence against a direct association between rodent and humna behavior. Rodham cites many studies which support his view that rat behavior is predomi- nantly stereotyped and reflexive, the result of stimulus-response, and not cognitive. He proposes that cognition allows man-made degrees of behav- ioral freedom and is a necessary ingredient for the development of culture.3 Biologist Dubos sees little application of Calhoun's research to human situations.“ Getis and Boots have concluded that little has been done In the study of rat behavior that has direct bearing on analysis of human spatial behavior. In part this is due to the fact that little attention has been paid to the spatial behavior of rats. With the excep- tion of Calhoun, few psychologists have considered such factors as density and spacing. Thus the geographers do not reject the possibility of human 3 Rodham, C., ”Cultures, Rats and Men,” American Journal of Psycholpgy, vol. 58 (l9h5), pp. 262-266. a Dubos, Rene, ”The Social Environment,” In Environmental Psychology; Man and His Physical Setting, Raymond Proshansky, et al., eds., (New York: Holt, I970), pp. 202-208. 20 implications from the spatial aspects of rat behavior. They do note, however, that enormous field work difficulties face researchers inter- ested in making such ties.S Stea, in a recent review of the geographical nature of this animal research, is much more positive. He has concluded that it has prominent implications for spatial learning by human beings. ”While it is true that much more of rat behavior may be termed 'instinctive' considerable spatial learning does take place, and the parallels between the two organisms in this respect may be much closer than in others."6 It would be misleading to suggest that evidence is primarily against behavioral associations between rodents and human beings. Others have drawn much broader implications than has Stea. Tolman was among the first to argue that an examination of rat behavior would exhibit basic laws and principles which underlie and therefore “explain" man's cultureized intelligence, motivation and instability.7 Carey, in reaching the most sweeping conclusion, has surmised that Calhoun's research ”conjures up the ghastly image of the turmoil-ridden core city surrounded by the relatively more placid suburbs."8 Although it is not the purpose of this chapter to debate the above issue, it is beneficial to examine non-human responses to spatial S Getis, Arthur, and Barry N. Boots, ”Spatial Behavior: Rats and Man," Professional Geographer, vol. 23 (January l97l), pp. lI-Ih. 6 Stea, David, ”Rats, Men and Spatial Behavior, All Revisited or What Animal Geographers Have to Say to Human Geographers,” Professional Geographer, vol. 25 (I973). pp. lO6-Il2. 7 Tolman, E.C., “A Stimulus-Expectancy Need- CathexisPsychology,” Science, vol. lOl (l9h6), pp. l60-I66. Carey, George, ”Density, Crowding, Stress and the Ghetto,” American Behavioral Scientist, vol. I5 (March/April I972), p. 502. 2l restriction because they suggest other aspects of human behavior that might be investigated. It would appear that direct human implications of such research are limited by problems of validity which arise whenever one generalizes from communities of rats to societies of men. Correlational Studies Until recently there were few studies relating the behavior of human beings to crowded environments, and these produced rather incon- sistent results. Scientists interested in the epidemiology of mental health were only indirectly concerned with the affects of crowding on human behavior; they preferred to relate the distribution of a pathology to population characteristics. One value of such research is that it provides clues to the etiology of the disorder. When a high rate of occurrence is found to be associated with certain population character- istics, the implication is that these characteristics contribute to the production of the disorder; one such characteristic is high population density. The classic epidemiological study of mental illness was con- 9 ducted by Faris and Dunham. Their analysis of the spatial distribution of schizophrenia and manic-depressive psychosis in Chicago indicated that the highest rates of mental disorder were found in those parts of the city that they characterized as socially ”disorganized." A transect from the inner city to the urban fringe revealed decreasing rates of hospitalized schizophrenia. Highest rates of illness were found among those social and ethnic groups residing in the most densely populated 9 Paris, Robert E. and Warren Dunham, Mental Disorders in Urban Areas, (Chicago: University of Chicago Press, I939). 22 areas of the city. The initial findings of Faris and Dunham have been replicated elsewhere. In each case the hypothesized inverse relation- ship between social class and the incidence of mental disorder and the positive association between pathology and high density were substan- tiated.lo Other epidemiological studies of mental illness have dealt directly with its spatial distribution and relationship to high popula- tion concentration. Downs and Simon reported that various diseases and maladies, including psychoneuroses, cluster in those areas of Baltimore where population density is high." Similar findings of spatial clustering are reported in St. Paul'2 and Boston.'3' Mental illness has not been the only subject of epidemiological research related to spatial restriction. Schmitt computed five differ- ent measures of population density for the census tracts in Honolulu and '0 See : Hollingshead, A.B. and F.C. Redlich, Social Class and Mental Illness, (New York: John Wiley, I958); Kaplan, Bert, Robert B. Reed, and W. Richardson, ”A Comparison of the Incidence of Hospitalized and Non-Hospitalized Cases of Psychosis in Two Communities,” American Sociological Review, vol. 27 (August I962), pp. h72-479; Jaco, E.G., The Social Epidemiology of Mental Disorders, (New York: Russell Sage, I960); Robins, L.N., et al., ”The Interaction of Social Class and Deviant Behavior,” American Sociological Review, vol. 27 (August I962), pp. h80-h92; Dunham, Warren, Community and SchIZOphrenia: An Epidemiol- ogical Analysis, (Detroit: Wayne State University Press, l965); and Parker, Seymour, and Robert Kleiner, Mental Illness in the Urban Negro Community, (Glencoe, Illinois: Free Press, I966). '1 Downs, J. and K. Simon, “Characteristics of Psychoneurotic Patients and their Families as Revealed in a General Mortality Study,” Milbank Memorial Fund Quarterly, vol. 22 (l95h), pp. h2-h6. '2 Lemkau, Paul, "Position Paper on Mental Health and Housing,” Mimeo, prepared for Environmental Control Administration of the Depart- ment of Health, Education and Welfare, and the American Public Health Association, I970. '3 Loring, William C., “Residential Environment: Nexus of Personal Interaction and Healthful Development,” Journal of Health and Human Behavior, vol. 5 (Winter I96“), pp. l66-I69. 23 correlated these with the rates of juvenile delinquency, adult crimes, suicide, mental hospital admissions and other measures. In two studies, separated by approximately ten years, he found strong positive correla- tions between population per net acre and almost all dependent measures: .30 infant mortality, .70 tuberculosis, .83 veneral disease, .7h mental hospital admissions, .SO illegimate births, .63 juvenile delinquency, .55 imprisonment. The single exception was a .Oh for suicides.'h These studies of mental illness and other pathologies suggest the possibility of an association between measures of population concentration and behavioral disorders. But is there meaning in this relationship? Or is it merely the result of the clustering of the ill in areas of the city left to them by others, locations which just happen to have large numbers of those people most likely to experience pathologies? Work on mental illness conducted by epidemiologists further suggests the strong possibility that social class, rather than location or spatial restriction is the primary explanatory variable. Members of the lower end of the social spectrum suffer in greater numbers from varying degrees of mental illness. They also tend to be concentrated in those portions of American cities where restriction is greatest. The social geography of urban America tends to indicate that race and poverty covary with spatial restriction, and, in turn, pathology. If the poverty were eliminated, would the diffiCulties of living in a spatially restrictive environment also disappear? The correlational studies of Schmitt, Chombart de Lauwe, Winsborough, and more recently Mitchell, and Galle, Gove and McPherson approach these '“ Schmitt, R.S., ”Density, Delinquency and Crime in Honolulu,” Sociologyjand Social Research, vol. kl (March/April I957). pp. 27h-276. De 2% issues through aggregate analysis. The most frequent approach is to correlate various census measures of population density, such as number of persons per acre or per dwelling unit with indices of social and medical pathology, rates of crime, tuberculosis and suicide. To account for the potential influence of non-density variables, the effects of characteristics such as income level, education and race have usually been controlled through the technique of partial correlation. The results of such research have been mixed. Schorr concluded in his analysis of social pathology that poor housing and crowded conditions leads to pessimism, passivity and a general state of dissatisfaction as well as sexual deviancy among residents of urban environments.'5 Winsborough, however, has demonstrated that the positive correlation between density and pathology disappears when certain measures of social status are used as control variables.'6 Culture appears to be a mediating influence in Spatially restric- tive environments. Measures of spatial restriction in parts of Hong Kong reach levels far beyond any found in cities of the United States, yet two researchers have concluded that such conditions in Hong Kong do not necessarily lead to behavioral social indicators. Mitchell saw little association between variations in restriction and certain forms of behavior. The only element of social organization that appeared to be affected by restriction pressure was parental control of children. Where room densities were higher, parents found it more l5 '6 Winsborough, H., "The Social Consequences of High Population Density," Law and Contemporagy Problems, vol. 30 (I965), pp. l20-I26. Schorr, A., Slums and Social Insecurity, (Washington, D.C., I963). 25 difficult to maintain household decorum.I7 Schmitt, in an earlier study, concluded that population density per acre is not associated with behavioral anomalies. Other factors, such as cultural traditions and the nature of residential land use, seem to mediate the relationship between population density and human behavior.'8 Conversely, in another cultural setting Chombart de Lauwe found measures of room density to be significantly associated with family disorganization among Parisians.‘9 Two of the most recent aggregate level studies of human behavior in spatially restrictive environments indicate that no definitive answer to the question of environmental influence has been obtained. Like previous studies, they reach opposite conclusions. However, their recency as well as their thorough evaluation of important conceptual - and methodological issues central to this study, suggests a detailed analysis. Galle, Cove and McPherson carried out ecological analyses of human behavior under spatially restrictive conditions.20 Basing their study upon the findings of Calhoun, the authors were interested in solving two problems: obtaining a means of comparing human with rat behavior and determining what effect population density has upon '7 Mitchell, Robert, ”Some Social Implications of High Density Housing,” American Sociological Review, vol. 36 (February l97l), pp. l8-29. '8 Schmitt, R.C., ”Implications of Density in Hong Kong,“ Journal of the American Institute of Planners, vol. 29 (I963), pp. 2IO-2l9. '9 Chombart de Lauwe, P., Famille et Habitation, (Paris: Editions du Centre National de la Recherche Scientifique, I959). 20 Galle, Omer, Walter Gove and J. Miller McPherson, “Population Density and Pathology: What are the Relationships for Man," Science, vol. I75 (I972). PP- 23-30. 26 human behavior. Using the seventy-five community areas of Chicago as their units of analysis, Galle et al. selected five human alternatives for the rat behaviors that were judged to be ”pathological”: l) fertility, 2) mortality, 3) number of welfare recipients under eighteen years of age in each community, A) rate of juvenile delinquency, and 5) admission rates to mental hospitals. Selection of mortality and fertility rates affords a direct comparison between human and rat populations, but the other variables are, according to the authors, human surrogates for animal behaviors. Calhoun observed that females congregating in the densely populated central pens failed to make nests for their young and often abandoned their litters. Families receiving public assistance are typically disrupted, having only one parent in residence, and the family is unable to provide for the children in a socially acceptable manner. Because of the similarity in behaviors, welfare rates were used as a human surrogate for the observed rat pathology. As an analog for asoclal, aggressive behavior among rat populations, juvenile delinquency rates were used in each Chicago community. As a substitute for withdrawal and psychotic behavior among Calhoun's Norway rats, age-adjusted rates of admissions to mental hospitals were used. Addressing themselves to the question of intercorrelation between density and socio-economic status and race, Galle, Cove and McPherson statistically controlled for both parameters in their analysis. Their regression analysis of density, defined as population per acre, and pathological behavior led them to a reevaluation of the concept density, as applied to human activity. Initial results of their analysis indicated that socio-economic Status and race were more associated with patholo- gical behavior than density. However, given the ability of an individual 27 to find privacy in his own room, despite living in a high population density location, the authors concluded that mere population density may not be an adequate measure of ”density” when analyzing human behavior in spatially restrictive locations. Density could be viewed as a measure of a person's ability to attain privacy as well as the actual amount of space theoretically available to him. Galle, Cove and McPherson refer to the inability to find privacy because of continuous contact with others as ”interpersonal press," a form of density. Density, then, was broadened in its interpretation, and was composed of four variables: I) the number of persons per room, 2) the number of rooms per housing unit, 3) the number of housing units per structure, and A) the number of resi- dential structures per acre. As a consequence of their reinterpretation of the concept density, the authors conducted a second analysis of the relationship between the environmental characteristic, socio-economic status and pathological behavior. They found that the measure of density was related to each of the pathological behaviors and in each case a significant relationship remained between the density components and the pathologies when social class and race were used as controls. The number of persons per room proved to be the most highly related to pathological behavior. In conclusion the components of density vis-a-vis the number of people per unit area of analysis, form an important intervening variable that signi- ficantly affects the relationship between social class, race and pathol- ogy. Freedman, employing basically the same methodology in his analysis of population density, juvenile delinquency and mental illness 28 in New York City, reaches the opposite conclusion.2' ”Density” is operationalized through two measures: the number of people per residential acre and the number of people per room in each residence. For each of his New York City Health Areas, Freedman obtained measures of median income, median years of education, percentage substandard housing, unemployment, change of residence and percentage non-white; these were used as control variables. Three major types of analyses were conducted. Simple correlations between density and the control variables were initially computed followed by the relationships between density and rates of youth crime and admissions to mental hospitals. Partial correlations between density and both crime and mental illness were then calculated removing the effect of median income and percentage non-white. Finally, a stepwise regression analysis that included all the independent variables was performed This gave an indication of the relative importance of each variable in explaining variation in juvenile crime and mental illness. It is clear from Table I that simple correlations between the dependent and independent variables are significant. These findings are consistent with general intuition and observation of the most densely settled areas of the city. There tends to be more juvenile delinquency and mental disturbance in the more densely settled neighborhoods.22 But Freedman also noted that higher correlations between the dependent variables and income and race. This initial analysis led him to suspect 2' Freedman, Jonathon L., ”Population Density, Juvenile Delinquency and Mental Illness in New York City,” Population Dsitribution and Policy; (Washington, D.C.: U.S. Government Printing Office, I973), pp. SIG-523. 22 Ibid., p. Sl9. 29 that social class and race might be more important or more basic factors than density. High density neighborhoods in New York also tend to be low income and to have a relatively high percentage of non- white residents. TABLE I Correlations of Density, Income and Percent Nonwhite with Juvenile Delinquency and Mental Illness Juvenile Mental Delinquengy Illness Density per Acre .36 .56 Density per Room .93 .53 Median Income .67 .60 Percent Nonwhite .75 .69 Source: Freedman, Jonathon L., “POpuIation Density, Juvenile Delinquency and Mental Illness in New York City," Population Distribution and Policy, (Washington, D.C.: U.S. Government Printing Office, I970), p. 5I9. This suspicion was confirmed through a stepwise multiple regression analysis, the results of which are shown in Table 2. Once income and TABLE 2 Stepwise Multiple Regressions of Income, Percent Nonwhite, and Density on Delinquency and Mental Illness Juvenile Mental Delinquency Illness Increase In Percent Increase in Percent 'Variance Explained Variance Explained Income h5.00 36.00 Percent Non-white IS.00 Ih.OO Density Per Acrea .39 5.00 Density Per Rooma .28 2.00 a Separate analyses for the two measures of density Source: Freedman, Jonathon L., ”P0pulation Density, Juvenile Delin- quency and Mental Illness in New York City,“ Population Dis- tribution and Policy, (Washington, D.C.: U.S. Government Printing Office, I970), p. 520. 30 race have been entered into the regression formula, neither density per acre nor density per room explained much additional variance. Density accounted for virtually none of the variance in juvenile delinquency and only a relatively small proportion of the variation in mental hospital admissions. When the other independent variables were entered into the analysis, the results remained substantially the same. From this analysis, Freedman concluded: .those who blame all of the ills of the cities on population density are greatly overstating the case. At the extreme we might argue that high density has virtually no effect, but a more conservative statement would be that density is vastly less important than income level and ethnic backgrounds . . .. Therefore, the glib statement that population density is 'killing the cities' and causing a breakdown in our social, mental, and moral fiber seems highly implausible and certainly unproven. Who is correct? As with most correlational statements, the relationship or lack thereof, is difficult to interpret. The question of how spatial restriction affects human behavior is extremely important, but remains largely unanswered by these researchers. Correlational studies of the type discussed do not ordinarily allow one to draw definitive conclusions about causation. Galle, Cove and McPherson turn to a discussion of individual behavior in explaining their aggregate findings. Aggregate correlational studies serve the purpose of suggesting relationships and help to answer the question of environmental influence; but to be fully effective investigations need to include an examination of the individual's relationship with his surroundings. In this manner aggregate findings are better understood and become more meaningful. 23 Ibid., p. 522. 3I Perception and Territoriality The work of Hall and Sommer on proxemics and human territoriality reflect the individual scale of analysis. While they do not focus directly upon the behavioral effects of restriction, their inquiry into the perception and use of space is closely related to a consideration of problems arising from spatial restriction. Research into personal space has shown that human beings carry with them portable Space “bubbles“ which affect interpersonal communications and regulate spacing and density within residential and work settings. Hall was among the first to study this dimension of personal space. He used the term ”proxemics” to refer to the study of man's perception and use of space as a component of culture.27 The main implication from his work on proxemics is that people differ in their habits, attitudes and values concerning the use of space and interpersonal distance, and that differences along these dimensions are largely culture-bound. The size of an indivi- dual's space ”bubble”, the minimum amount of Space which he must maintain between himself and other objects and beings in order not to feel threatened, varies according to the situation and personal characteristics. It is ultimately determined by the person's cultural background. Thus, the spatial distance considered comfortable for conversation is much less in Latin America than it is in the United States.25 Robert Sommer has also defined personal space in territorial terms. Personal space is "an area with invisible boundaries surrounding a person's 2“ Hall, Edward T., The Hidden Dimension (Garden City, N.Y.: Doubleday and Co., l966). 25 Hall, Edward T., The Silent Laoguagg, (Garden City, N.Y.: Doubleday and Co., I959). 32 body Into which intruders may not come."26 It involves an emotionally charged zone around individuals which may regulate spacing and it also concerns the process of delimiting and personalizing territory. Percep- tion of spatial relations among objects is significantly influenced by the type of activity which occurs in a given area. One of the newest research directions in geography rephrases the Man-Environment tradition of the discipline in behavioral terms. This is particularly true of urban geography, where increased attention is being focused upon the spatial and physical characteristics of the city as 27 they are related to the behavior of the metropolitan inhabitant. Recent work by urban and behavioral geographers has not dealt with the subject of spatial restriction, but the interest in man's perception of his surroundings and how his behavior is associated with the image of his environment is couched within the stimulus-perception-reSponse framework of this study. Wolpert regards perceived differences in the urban environment as the vehicle which motivates decisions to migrate on the part of a cer- tain proportion of the urban population.28 Migration is viewed as an adaptive response to perceived differences in environmental quality. Other things being equal, one will move to another location if it appears more promising as a place where he can realize a more satisfying life-sytle. 26 Sommer, Robert, Personal Space - The Behavioral Basis of Design, (Englewood Cliffs, N.J.: Prentice Hall, I969). 27 Taafe, Edward J., Geography, (Englewood Cliffs, N.J.: Prentice Hall, 1970). p. 87. 28 WOIpert, Julian, ”Migration as an Adjustment to Environmental Stress," Journal of Social Issues, vol. 22 (October I966), pp. 92-l02. 33 Perceived differences in place utility evolve from cognition and evaluation of stimuli. The individual's ”action space” represents the spatial parameter of the mover-stayer decision process. Through daily travel experience, the person gains direct knowledge of various locations within his urban area. This knowledge is transformed through evaluation, primarily influenced by what desires the person has regarding a satisfactory living environment. If one beleives that his residential surroundings are deteriorating in quality relative to other locations in his city, he Is likely to move. Perceptions of other areas as more enticing -- and within the individual's ability to reside there -- may also prompt a decision to change residence. Thus spatial aspects of intra-urban migration, according to Wolpert, are measures of the utility of one place relative to perceived values of other locations. Other geographers have been interested in the role of perception as it relates to behavior. Mental map studies have been used in the design process as a basis for the reconsideration of urban aesthetics, and as a means of analyzing the underlying dynamics of human behavior. In the latter context, Porteous has noted that the individual's per- ception and cognitive structuring of the city has been used as a partial explanation for shopping behavior and choice of travel paths.29 Sonnenfeld has described differences in the ability to adapt to the Arctic environment among native and non-native population by measuring 30 expressed levels of contentment with that environment. 29 Porteous, J. Douglas, "Design with People: The Quality of the Urban Environment,” Environment and Behavior, vol. 3 (June l97l), pp. l6l-l6h. 3° Sonnenfeld, Joseph, ”Environmental Perception and Adaptation Level In the Arctic,” Environmental Perception and Behavior, David Lowenthal ed., Department of Geography Research Paper No. IO9, (Chicago: University of Chicago Press, I966), pp. h2-59. 3A Within geography, the research area with the greatest concentra- tion of perception studies deals with the perception of natural hazards.3' This research has been concerned with the persistence of settlement in hazardous areas and the ways in which man has attempted to adapt to the hazards. The types of hazards investigated have included floods, coastal storms, drought, snow, hail, landslides, and tsunamis. More recently, the effects on the physical setting due to man's actions, such as air or water pollution or weather modification, have been added to this list. White, Burton and Kates have been the major investigators and they have outlined five major avenues of geographical study: Minimal understanding of conditions upon which social policy might be based would involve research helping to, I) assess the extent of human occupance by hazard zones, 2) identify the full range of possible human adjustments to the hazard, 3) study how man perceives and estimates the occurrence of hazard, A) describe the process of adoption of damage-reducing adjustments in their social context, and, 5) estimate the optimal set of adjustgpnts in terms of anticipated social consequences. The individual's view of the hazard has been sought In answering these questions. Variations in perception of hazards are considered to be a major factor In the decision-making process. If a resource manager perceives a risk or hazard-threat, he is likely to be more receptive to an Innovation which might help minimize the risk. Variations in per- ception among hazard zone occupants have been explained in terms of 3' Saarinen, Thomas F., Perception of Environment, Commission on College Geography Resource Paper No. 5, (Washington, D.C.: Association of American Geographers, I969), p. I9. 32 Burton, Ian, Robert Kates, and Gilbert F. White, "The Human Ecology of Extreme Geophysical Events,” Natural Hazard Research Working Paper No. l, (Toronto: Department of Geography, University of Toronto, 1968). p. 36. 35 three main factors: the relation of the hazard to the dominant resource use, the frequency of occurrence of the hazard, and variations in degree 33 of personal experience. The emphasis, direct or implied, by geographers in studying human perception of hazards has been upon behavioral adjust- ments which occur as a result of the perceived hazard or the lack of it. Golant and Burton have analyzed avoidance behavior as a response to environmental risk and to perceived risk and have used a semantic differ- ential to explore preconceived ideas people have about potential natural hazards; the implication is that the perceived meaning a person attaches to the stimulus will be associated with the way he responds to it.3u Beck, Tuan and Kates each note the considerable Importance that perception, or the subjective interpretation of one's surroundings, has for behavioral responses to environmental stimuli. Tuan has stated that it Is not enough to examine the environment as a series of discrete stimuli If we wish to understand its impact on man in the course of his day-to-day living. Changing the physical setting alone will not revolu- 35 tionize people's lives. However, if we can determine that Individuals or populations perceive certain stimuli as threatening, then changing or removing such a stimulus conceivably will alter one's interpretation of his environment and his response to it. Kates suggests that ”the central task of the behavioral and social sciences in the study of environment is to relate the stimulus 33 3“ Golant, Stephen and Ian Burton, "Avoidance-Response to the Risk Environment,” Natural Hazards Research Paper No. 6 (Toronto: Department of Geography, University of Toronto, I968). 35 Tuan, Yi-Fu, ”Environmental Psychology: A Review," Geographical Review, vol. 62 (April l972), pp. 2h5-256. Saarinen, op.cit., p. 20. 36 properties of the environment to their symbolic human manifestations and in turn to define the stimulus properties of the symbolic environ- 36 ments that man creates.” For the social scientist to understand the variance in behavioral responses to environmental stimuli, he must ascertain what meaning those stimuli have for the individual. Beck points out that interpretations of stimuli vary within and between cultures. Meaning is derived from a satisfaction of needs and needs have spatial qualities. Meaning ”clothes the perceptual world" so that the individual comes to structure his environment according to his I.37 understanding of experienced Stimul It is through the perceived meaning of a spatially restrictive environment that such surroundings could affect one's behavior. Laboratory Research Experimental investigations directly concerned with the effects of spatial restriction on human behavior reflect the most recent approach to the study of spatial restriction phenomena. Such studies have generally been of two types: those which define restriction In terms of group size and those which manipulate it in terms of room size. The research of Ittelson, Proshansky and Rivlln, Hutt and Valzey, and Griffit and Veitch represents the first type of Investigation, while those of Freedman, and Freedman, Klevansky and Ehrlich represent the 36 Kates, Robert W. ”Stimulus and Symbol: The View from the Bridge,” Journal of Social Issues,_vol. 22 (October I966), p. 23. 37 Beck, Robert, "Spatial Meaning and the Properties of the Environment," Environmental Percepgion and Behavior, David Lowenthal, ed., Department of Geography Research Paper No. IO9 (Chicago: University of Chicago Press, l966). 37 second.38 Results from the first set of studies indicate that members of larger groups are more aggressive and asocial than those of smaller ones, regardless of whether the setting is a psychiatric ward, a play- ground, or a psychological experiment. The second set of experiments, however, demonstrates that when group size is kept constant but room size is varied, the task performance of subjects in the small room is not adversely affected. Freedman did observe in the small room condi- tion, however, that interpersonal relations within female groups were more Intimate and friendly than in male groups.39 In the most recent experimental study of spatial restriction, Stokols et al. observed that variances in room size do affect task performance and the quality of social interaction."0 An additional finding was a marked difference in the impact of room size upon the sexes; this verified the conclusions of Freedman. Men reported higher feelings of aggressiveness and restriction when density was increased 38 Ittelson, W., H. Proshansky, and L. Rivlin, "The Environmental Psychology of the Psychiatric Ward," in Environmental Psychology, H. Proshansky, W. Ittelson and L. Rivlin, eds., (New York: Holt, I970); Hutt, C., and M. Vaizey, "Differential Effects of Group Density on Social Behavior," Nature, vol. 209 (March 26, I966), pp. I37I-l372; Griffit, W. and R. Veitch, "Hot and Crowded: Influences of Population Density and Temperature on Interpersonal Affective Behavior,“ Journal 5g;fersonality_and Social Psychology, vol. l7 (I97l), pp. 92-9 ; Freedman, Jonathon, "The Effects of Crowding on Human Behavior,” unpublished manuscript, Department of Psychology, Columbia University; Freedman, Jonathon, S. Klevansky, and P. Ehrlich, "The Effect of Drowdin on_Human Task Performance,” Journal of Applied Social Psychology, vol. I (1971). pp- 7-25. 39 Stokols, Daniel, ”A Social-Psychological Model of Human Crowding Phenomena," Journal of American Institute of Planners, vol. 38 (I972), p. 7h. “0 Stokols, Daniel, et al., "Physical, Social, and Personal Determinants of the Perception of Crowding,” Environment and Behavior, vol. 5 (March l973), pp. 87-ll5. 38 by reducing room size. Women reported an opposite reaction. The authors concluded from their experiments that perceptions of ”stress” vary between individuals, and that personal and cultural variables can mediate these perceptions as well as their behavioral ramifications. Perceptions of limited physical space are not always translated Into identical feelings of restriction. Rather, the impact of reduced space for the individual appear6»to be determined through the interaction of the physical and social setting and the characteristics of the individual. It remains to be determined what these personal social and cultural factors are. In order to understand more fully the relationship between spatial restriction, perception and behavior, the authors suggest two directions for future research: ”...explore more fully the dimensions of stress which overlap with the experience of crowding and...specify more clearly the mechanisms by which perceived spatial restriction and its related symptoms of stress arise. Second, it would be useful to delineate specific behavioral indices of crowding stress, and to state more precisely the conditions under which behavioral responses to the AI experIence will occur." Conclusions from the Literature Review From the stages of research outlined above, a preliminary picture of spatial restriction phenomena begins to emerge. The animal studies portray restriction as a situation which develops over time. The physical condition of spatial limitation, which places constraints upon certain social acitvities (e.g., allocation of food and sexual behavior), represents the necessary condition for these phenomena. As I. ' Ibid., p. il3. 39 population density increases, spatial constraints become more acute until they lead to social disorganization and physiological pathology. Situations of restriction among non-humans, then, are characterized both by the element of spatial restriction and by the manifestation of its deleterious effects on organisms over time. The research on human populations, however, indicates that spatial restriction is not inevitably associated with behavior. The correla- tional studies suggest that in Asian societies cultural traditions offset the supposedly detrimental effects of high population density. Other aggregate analyses of United States populations reach opposite conclusions examining restrictive conditions In the same cultural envir- onments. Studies examining the perception of spatial adequacy and the delimitation and use of personal space provide further evidence that cultural norms mediate the perception and adjustments of interpersonal space. Such research also suggests that the type of activity in a given area largely determines whether the amount of available space is perceived as adequate. The laboratory investigations of human restric- tion demonstrate that when group size is held constant and the physical consequences of spatial restriction are controlled (e.g., high tempera- ture, stuffiness, limited movement), high density exerts virtually no ill effects on human task performance. At present, it cannot be stated that there is a definitive conclu- sion regarding work analyzing the behavioral influences of spatial restriction. With the exception of non-human research, inconsistent results have been obtained. Similar methodologies have produced differ- ent results in both correlational and laboratory studies. The salient question In this study is does spatial restriction adversely affect #0 human behavior? Results of studies of human groups and individuals are inconclusive. Implications of Previous Research for this Study Although previous studies provide some insight into the nature of human restriction phenomena, interpretations from various lines of inquiry are made difficult by methodological and conceptual inadequacies. The findings of aggregate studies are plagued by the causal ambiguities of correlational research. Emphasis upon aggregate analysis severely limits the scope and applicability of such ”behavioral" research. Findings can be discussed only in terms of a population, or a segment thereof. Aggregate data tangentially deal with the question of restric- tion-produced behavior, but we are not able to discern the indivdual who is, or is not detrimentally affected by a spatially restrictive environ- ment, only the characteristics of his population. Clearly an analysis of individual behavior is needed if we are to better understand how one reacts to certain environmental conditions. Aggregate studies have also failed to develop conceptual and Operational definitions of spatial restriction and behavior that facilitate comparisons of findings. Every analysis of crowding or density and behavior has used different meanings for the the terms. This has resulted in a lack of consistency in their use, further hindering a testing of research reliability and validity. As Stokols notes, we have yet to determine specifically what are the behavioral reactions to a restrictive environment -- in part because in each case different behaviors are examined under different conditions."2 “2 Stokols, op. cit., (I972), p. 7h. Al The inherent nature of laboratory studies leads to considerable deficiency in this method of research. Conclusions based upon studies of students in controlled laboratory settings limit the inferences one can draw from such work about the average individual or population. By controlling the physical consequences of spatial restriction, e.g., higher temperature, stuffiness and limited movement, the psychologists engaged in this form of spatial restriction research have virtually negated the possibility of extrapolating their findings to more realistically restrictive environments. What is needed is not only a movement away from aggregate analysis, but also one that avoids the confines of the labora- tory and offers the potential of describing ”real-world” situations where crowding does or does not detrimentally affect human behavior. With few exceptions most analysts of spatial restriction have failed to direct their attention to the temporal or spatial aspects of the problem. The analysis of aggregate or individual data is usually based upon behavior observed at one point in time. We know little about the temporal or developmental aspects of spatial restriction and the ways people respond to it over time. The temporal aspect of this research may be particularly important in determining whether such conditions as length of exposure may influence certain behavioral responses to the experience of restriction. Social scientists have arrived at fairly broad agreement that a single pattern of urban population density is repeated in a large number of Western cities: population density declines with distance from the 43 city center. This finding provides us with a referent which can “3 Clark, Colin, ”Urban Population Densities,” Journal of the Royal Statistical Sociepy, Series A, vol. llh (l95l), pp._A90-k96; idem, Population Growth and Land Use, (London: MacMillan and Co., I967); 42 serve as an indicator of where we are likely to find restriction-related behavior. If a major need of restriction analysis is to state more pre- cisely the conditions under which behavioral responses to the experience will occur, then one important aspect of this need is to determine where we will find these conditions. Galle et al. have done much to unravel the use of the term density as it applies to human behavior and argue that it invdlves far more than the number of people per square mile. The work of geographers studying the perception of the environment and the laboratory study of Stokols have also reaffirmed the notion that spatial restriction is much more than an objective characteristic. By Operationalizing restriction as a cognitive rather than a physical characteristic, the authors have demonstrated the important role perception has in the occurrence of behavioral responses to such conditions. The individual will not respond directly to an environment that physically provides him little space unless he finds that situation restrictive. Variations in response are due as much to the perceiver as they are to his surroundings. An Stewart, John Q. and William Warntz, “The Physics of Population Distri- bution," Journal of Regional Science, vol. I (Summer I958, pp. 99-l23; Winsborough, Halliman H., ”A Comparative Study of Urban Population Densities," (Ph.D. dissertation, University of Chicago, l96l); Muth, Richard, ”The Spatial Structure of the Housing Market,“ Ropers and Proceedings of the Regional Science Association, vol. 7 (l96l), pp. 207-220; idemI "The Variation of Population Density and its Components in South Chicago,” Papers and Prooeedings of the Regional Science Association, vol. l5 (I965), pp. l73-I83; Berry, Brian J.L., James W. Simmons and Robert J. Tennant, ”Urban Population Densities: Structure and Change,” Geographical Review, vol. 53 (July l963), pp. 389-205; Berry, Brian J.L., "The Internal Structure of the City,” Law and Contemporary Problems, vol. 30 (Winter I965), pp. lll-Il9; Newling, Bruce E., ”Urban Growth and Spatial Structure: Mathematical Models and Empirical Evidence," Geographical Review, vol. 56 (April I960), pp. 2l3-225; Berry, Brian J.L. and Frank E. Horton, Geographic Perspectives on Urban Systems, (Englewood Cliffs, N.J.: Prentice Hall, Inc., I970), p. 276. A3 objective and subjective approach to the problem appears needed and potentially beneficial. It is also clear that in both aggregate and individual levels of analysis, steps must be taken to statistically control for the influences of social class, race and ethnicity. From studies of personal Space and aggregate analyses of Hong Kong and Paris, it is apparent that cross- cultural studies of Spatial restriction and behavior suffer in the comparison, especially from a perceptual point of view. There is a definite need to maintain a cultural constant in a study of behavioral responses to spatial confinement. Theoretical Ties Between Spatial Restriction and Behavior There are further ramifications of Hall's research for this Study. If an individual's personal space is breached, how will he respond to the invader? If he feels threatened, how will he defend himself? Hall has not been directly concerned with extending his studies of individual proxemics to the group level, but he has speculated on the possible implications of individual proxemic relations for urban planning and architecture and for the problems of spatial restriction in cities. It is clear from his research that responses would vary cross-culturally, but it seems equally clear that any investigation of behavioral responses to spatial restriction should include consideration of individual varia- tions in personal spatial adequacy. On the question of personal space invasion, Sommer suggests that knowledge of personal space and another mode of social organization, (kxninance behavior, combine to limit aggression. Dominance behavior Feféws to a hierarchy of dominant-subordinant relationships, as exem- PIIfied by social ranking. Territoriality and dominance behavior are 1.1. viewed as interdependent and complementary. Both processes limit aggres- sion because an individual either refrains from going where he is likely to be involved in a dispute or, based on his knowledge of who is above or below him, engages in dominant-subordinant behavior rather than in actual combat.hh However, Sommer's research still does not completely answer the question, what will happen if and when personal territory is knowingly or unknowingly violated? Carstairs, in his description of violence in America, suggests that human aggression is predominant in those cities where spatial 45 restriction is greatest. In these parts of urban America it is difficult to avoid infringing on another's personal space on an individual or group basis. Sommer also notes in his own research that in modern Western society there appears to be an aura of instability where particularly important societal boundaries remain fuzzy and undefined. The study of proxemics thus suggests a potential deleterious response to high spatial restriction due to the invasion of personal space. Conflict and behavioral instability in restrictive environments may also arise because of the invasion of another form of territoriality in the residential environment. Goffman, in his analysis of face-to-face interaction, has referred to various regions of interaction which affect the degree of formality on the part of the ”performers.""6 Front stpge is the location where a particular performance is or may be in progress. When one's activity occurs in the presence of others, some aspects of the “A Sommers, op.cit., p. l2. #5 Carstairs, op.cit., p. 59%. A6 Goffman, Erving, The Presentation of Self in Everydgprife, (Garden City, N.Y.: Doubleday and Co., I959). 'IS activity are expressively accentuated, and other aspects, which might discredit the image of the performer, are suppressed. Back stage is the location where those suppressed facts make an appearance. Back stage, relative to front, is the location where the performer can relax, where he can drop the image he portrays, forgo speaking his lines and step out of “character.” Basic to the existence of this behavioral region is the element of privacy. Commonly it is cut off from the remainder of the residential environment by a partition and guarded passageway. The back region is the one location where the performer can reliably expect that “7 Without this element of no member of his ”audience” will intrude. privacy and, therefore, the presence of a back stage, the individual is unable to adequately prepare himself for interaction with his peers as well as strangers. Removal of back stage not only deprives the performer of an opportunity to act in a relatively informal, familiar way, but reduces the ability of the individual to develop the guarded image neces- sary for him to perform adequately on front stage. A poorer quality of interaction develops because the person is unable to find a place of privacy in which to prepare himself for interaction with others. Reduction of the quantity of space, therefore, affects the quality of space. Most American homes, regardless of class, make some attempt at partitioning the living room front stage from that of the rear portions of the home. When this physical separation is absent, or fails to prevent others from entering the back stage, disillusionment and torment can result. When individuals witness actions that are not meant for them, their impressions of the performer, and consequently their "7 lb id., p. II3. A6 relationship with that person, may change radically. From the performer's perspective, inability to maintain control over his back stage leaves him in a position of not knowing what character he will have to project from one moment to the next, making it difficult for him to be successful in any role he plays. By properly scheduling performances, it is not only possible to separate one's audiences (thus avoiding different and inconsistent fronts before the same people), but also allow a few moments in between performances so as to extricate oneself psychologically and physcially A8 from one personal front, while taking on another. Goffman writes in an allegorical fashion, but his concept of back stage has real life application to spatially restricted environments. Gerald Suttles demonstrates the utility of Goffman's concept in his description of domestic congestion in the Addams Area of Chicago and the consequences of the inability to carry on certain acitivites in privacy: Even where there is the strongest desire for privacy girls must often dress, make themselves up, run around in their slips, and sometimes use the bathroom in someone else's presence. Boys sleep in the same room as their older and more worldly brothers, wear each other's clothing, and perform bodily functions in the same enclosures. Husbands and wives cannot keep their intimaceis or arguments a private matter that orners can at least pretend didn't happen. At time everyone may use the same towel, eat off the same plate, or indulge in a common obscenity. Fathers are seen in their underwear, mothers while in labor, sisters during their period....Sometimes there are disclosures that could lead to various consequences: abortions, incest, illegitimacy, adultery, and narcotics scars. More common, A7 however, are those skeletons for which every family is supposed to have a closet: defecation, intercourse, parental arguments and dressing.89 From a theoretical as well as a real-life persepctive, it appears that restriction of living space can and does have a marked affect on inter-personal relationships particularly in family settings. Sharing the same bathroom and bedroom congests relationships so that parental affection and authority lose meaning to children who have observed the back stage behavior of their parents. Relations become quite blunt since they cannot be ordered according to who has priority in a given space. Inevitably, conflict over spatial usage results in a definition 50 of personal power rather than situational rights. Statement of Hypotheses As has been noted, the various avenues of spatial restriction research have produced rather inconsistent results. In part, this inconsistency can be explained by poor conceptualization of the environ- mental element. Given the theoretical work of Goffman, Hall and Sommer, and the observations of Suttles, the key issue appears to be the degree to which individuals and populations are forced into contact with one another by the density and spatial arrangement of their environment. High population density alone may not affect the degree of contact between individuals, especially if they are able to retreat to the back stage portions of their dwellings. But if retreat into the home does not provide the privacy essential to stable social interaction, serious “9 Suttles, Gerald, The Social Order of the Slum, (Chicago: University of Chicago Press, l97l), p. 9l. 50 Ibid., p. 9i. A8 consequences may result. This is not to suggest that density may not in some way be related to behavior; only that density in and of itself may not be a sufficient etiological factor. High population density is likely to have a considerable influence on the other elements of spatial restriction; e.g., structural density, room density and density of residential space. Thus while others have suggested that pathological behavior is a function of density, this writer has broadened the hypo- thesis by suggesting that behavioral pathology is a function of spatial restriction: Hypothesis One Populations residing in spatially restricitve locations will exhibit greater frequencies of social indicators behavior than those which do not. A series of additional questions, or sub-hypotheses emanates from the initial premise. Having suggested that population density is only one of a number of elements of spatial restriction, the initial hypo- thesis can also be used to test the degree to which these place charac- teristics are interrelated. Do those areas which are high in population density also demonstrate greater degrees of spatial restriction? To what extent do the other spatial restriction elements, such as room density and structural density, form similar spatial patterns? A geo- graphical analysis of the characteristics of place should provide answers to these questions. Social indicators cannot be merely the result of variations in spatial restriction. If it were, individuals provided with the same amount of space would all exhibit similar degrees of behavior. Responses to external stimuli can be affected by the individual's interpretation of them. Unless the person views a situation or object as threatening, A9 he is unlikely to be detrimentally affected psychologically and, in turn, behaviorally. Thus explanations for responses to spatial restriction can also be obtained through analysis of cognitive interpretations of external stimuli. Until the recent research of Stokols, this psychol- ogical element of restrictive stimuli was virtually ignored. Behavior was discussed as primarily a function of room or population density. This first hypothesis has broadened that interpretation; the second introduces the role of perception as an intervening variable in the spatial restriction-behavior relationship. Hypothesis Two Spatially restrictive environments will be per- ceived as stressful. When the amount of space available to an individual or population is reduced to the point where infringement upon personal space or back stage is probable, the residents of such an environment are likely to feel threatened. Lazarus has described an individual In a state of stress as one who perceives threat cues in his environment. Stress usually envolves the self-concept of some cognitive level: either that some part of the person's body is suffering or being attacked, or his self-identity is being threatened, attacked or involved in some hazardous and potentially harmful situation or circumstance. Hence, stress is usually ”felt" by the individual, an emotional state is aroused and his defenses and coping mechanisms are mobilized, even though he may not be aware of the source of such threat other than a vague feeling that "something is wrong" and that his guard must be up if he is to survive what he perceives is an Impending attack. Stress, then, is both external and internal in relation to the individual. External stimuli, when perceived as threatening, become 50 stressful. Through their simultaneous and continuous demands these stimuli can eventually overwhelm the individual and lead to some forms of behavior that are indicators of low social well-being. Stress forms the link that ties environment to behavior. A Spatially restrictive environment is potentially stressful, but unless that stress is realized, behavior due to restriction is unlikely to occur. This thesis is expressed in hypothesis three. Hypothesis Three Individuals who perceive their environment as stressful will exhibit greater frequencies of behavioral social indicators than those who do not. Each person structures and evaluates his surroundings within his own mind. If he perceives his surroundings as stressful, and is unable to avoid them, he will have to cope with the situation. People expend energy and commit resources to coping with perceptions of personal threat, and usually they successfully adapt. In most cases the success- ful adaptation permits stable behavior to be maintained during exposure to stressfully perceived stimuli. However, there is evidence that a cost is paid in the adaptation. Irrespective of whether adaptation was successful, the greater the commitment to coping with feelings of stress, the greater the adverse aftereffects. Time becomes an important factor, for if the perception of stress is continuous, the individual will have to commit dwindling mental and physical resources just to maintain stability. All persons possess varying abilities to adapt, but eventually those abilities are overcome. If a threat to an individual's self- image cannot be overcome, self-defensive measures are necessary to make the threat tolerable. The degree to which a person must Invest in 5i such defense measures is the degree to which his.energy.and resources are being used beyond what would be demanded had mastery been achieved.5' Success or failure in coping with stress can lead to behavioral diffi- culties. The greater the degree of perceived threat, the fewer resources the individual has available to him to effectively interact with others and face additional demanding situations. Behavior, if it is associated with spatial restriction, is also a function of stress. But restrictive stimuli are not the only elements in the individual's physical and social environment which can be per- ceived as stressful. Hypotheses two and three focus on a single element in one's surroundings and suggest a series of additional questions and sub-hypotheses which relate to stress and behavior. To what extent is the stress felt in a community the result of poverty, racial discrimination and crime? It is highly possible, even likely, that the stresses of racial discrimination and poverty are more immediate than lack of residential space. Given the adjustment abilities of human beings, can we expect that recent arrivals in restrictive locations will be more susceptible to perceptions of stress than long-time residents who conceivably have adjusted to their surroundings? Or will the opposite be the case -- that longer periods 6f residence in a stressful environment are associated with dwindling coping resources and, consequently, behavior? Also seemingly associated with perceptions of stress and inci- dences of behavior would be a radical change in one's residential 5' Howard, A. and R.A. Scott, ”A Proposed Franework for the Analysis of Stress in the Human Organism,” Behavioral Science, vol. l0 (I965). p. ISO. 52 environment. Are residents more able to cope with spatial restriction if they have been raised in such an environment? Are individuals more affected by a restrictive environment If they have recently left surround- ings which provided them more open space and greater privacy? Time, residential background, race and poverty are but a few of an infinite number of elements which can conceivably affect the manner in which a person views his surroundings and responds to them. The need for a spatial analysis of the problem of spatial restriction and behavior necessitates a testing framework that lies beyond the controlled environ- ment of the laboratory. We cannot expect to isolate a single causal variable in the incidence of selected behaviors, but we can expect an increased understanding of the role Spatial restriction and perception have in mediating behavior. CHAPTER 3 ECOLOGICAL ANALYSIS OF SPATIAL RESTRICTION AND BEHAVIOR The purpose of this research is to relate the nature of places to human behavior by means of an intervening variable, environmental perception. Chapter Three describes the characteristics of ”place,” in this case census tracts, and indicates the degree to which a set of attributes of that place, forms of density, co-occurs with a particular set of behaviors designated as indicators of social well-being. Ecological, or aggregate, data are particularly useful in describing areal and group properties but behavioral scientists have often made the mistake of inferring individual behavior from these prOperties. Robinson has described this error as the ”ecological fallacy.”' Because of this limitation of aggregate data, individual characteristics assume particular importance in a behavioral study. In this chapter aggregate data will be used to measure the spatial association between levels of restriction and forms of behavior at the census tract level. Individual data evaluating the meaning residents of selected tracts attach to their environment will be analyzed in chapter four. These data on perceived meanings add to the understanding of association on an aggregate level while avoiding the ecological fallacy. ' Robinson, W.S., ”Ecological Correlations and the Behavior of Individuals,” American Sociological Review, vol. l5 (I950), pp. 35I-357. 53 5A Aggregate or ecological analysis of spatial restriction and behavior replicates research reported in the review of the literature. However, the inconsistency of findings and previous approaches suggests that further work, focusing especially on the spatial dimension, is needed. This will provide the added benefit of an intra-urban description of a spatial restriction and behavior, rather than a discussion of the city as a whole. In this manner it will be possible to determine how characteristics of place vary from location to location within a metropolitan area. Knowing the location of those parts of the city which are highly restrictive will in turn provide the spatial rationale for the selection of survey sites for the analysis of individual perception and behavior. Selection of a Study Site As one of the largest urban concentrations in North America, Detroit displays significant differences in population concentrations within its confines. In addition as the fifth largest Urbanized Area in the United States, Detroit is representative of those types of urban areas that are the focus of growth on the continent. Primarily for these reasons, it was selected as the study site. Significant differences in spatial reStriction are needed if we are to determine any behavioral influences of this characteristic of place. By including the central city and its contiguous suburbs in the analysis, Detroit should prove an excellent study site that represents metropolitan North America. Data and Methods of Analysis In order to determine the nature of place as it varies within the Detroit Urbanized Area, several variables were selected to measure 55 various forms of spatial restriction. Others were selected as additional locational characteristics which might account for the spatial variation of behavior. The place characteristics presented in Table 3 are regarded as physical, social and temporal stimuli in the environment which may account for the presence of certain forms of behavior at various locations within the urbanized area. TABLE 3 Variables Used in Aggregate Analysis Race and Social Class Percentage Negro Median Income Median Number of School Years Completed Percentage of Male P0pulation Unemployed Percent of Families below Poverty Income Level Forms of Spatial Restriction Number of Personppper household Percent of housing units having l.0l or more persons/room - Percent Overcrowded Number of persons per square mile - Population Density* Median number of Persons per room* Number of Rooms per housing unit Percentage of total area in Residential land use* Number of Housipg units per square mile* Number of Housing units per square mile of residential area* Others Percentage of all residential structures constructed in the previous five years - New Structures Distance of Census Tract from Central Business District - CBD Distance* Behaviors Normal Family Index - Percentage of children under l8 who are residing with parents High School DrOpouts - Percentage of l6 to 2l years olds who are not high school graduates and not enrolled in school Matriarchy Index — Percentage of household heads that are female 56 TABLE 3 (cont'd.) X19 Marital Unrest Index - Ratio of divorced and separated persons to those now married* X20 Welfare - Percentage of all families receiving public assistance income or public welfare income X2] Fertility - Number of children born per IOOO women ever married. X22 Non-Migrant Index - Percentage of p0pulation living in the same house for at least the past five years X23 New Residents - Percentage of residents living at present location 2% years or less Source: U.S. P0pulation Census, I970 for Detroit SMSA * Calculated by author The latter set of characteristics, X16...X22, are behavioral responses to environmental stimuli. As in any project of this kind, choosing the appropriate Spatial unit of analysis for dependent and independent data is a compromise between the ideal and reality. Given the behavioral focus of this study, it is particularly important to avoid units of analysis that are so large they incorporate a considerable degree of pOpulation and thus behavioral heterogeneity'within their limits. All other things being equal, the larger the unit of observation, the greater the degree of aggregation and possible variations in the individuals engaging in those behaviors being studied. Locational conditions measured by averages and by the relations between averages also become less meaningful for similar reasons. Census data, in either block or tract form, reduce the problems of within variation because their relatively small areal size increases p0pulation homogeneity. Census tracts from the I970 census have been selected for this analysis because they are less numerous and avoid problems of mapping and computer analysis that the smaller block units present. There are nearly 57 IOOO tracts in the study area, and several times that number of blocks. Figure 2 presents the census tracts of the study area. It should be noted that the Detroit Urbanized Area, as shown in Figure 2, is not the same as that officially delimited by the Bureau of the Census. Tracts 93I and 93A.OA, located beyond the western boundary of the study area in Northville and Plymouth Townships, have not been included for study because they contain several state mental hospitals and correctional institutions. Ideally a selection of variables which may be affected by elements in the environment would include additional social indicators. Inclusion of other behavior, e.g., homocide, suicide and drug addiction, has been limited not only by lack of published resources, but also by the lack of a common spatial denominator. One of the most difficult tasks facing a researcher, who must draw upon more than one source for his data, is the strong likelihood that the characteristics will come in various units of observation. For example, crime data for the study area are reported for precincts and squad car areas, while suicides and attempted suicides are recorded by zip code areas. With 93A census tracts in the study area, the task of converting non-census material to tract form or vice-versa becomes formidable. Even if the conversion is made, the accuracy of the data is highly questionable. A common spatial unit can be calculated for all phenomena, but the units of analysis do not necessarily describe the population engaged in these behaviors. The usual means of conversion is to calculate the areal proportion of each census. 58 CENSUS TRACTS DETROIT URBANIZED AREA Figure 2 €I r-e 59 tract found in a precinct or zip code area and assign to it that per- centage of the behavior in question. However, we have no guarantee that simply because ten percent of a police precinct is found in a census tract, ten percent of its crimes were committed in that tract or if several census tracts are found within a zip code area (as is the case in the study area), that an equal number of suicides was attempted in each. Behavior varies within each form of observation, and unless the location of each act is known, it is impossible to accurately describe it in another unit of analysis. Because of these difficulties, it was felt that for purposes of quantitative analysis, it was preferable to limit types of behavior to those available in census tract form. By so doing, other social indicators such as crime, suicide, drug addiction and mental illness were not studied; however, their inclusion would cast considerable coubt on the accuracy of any findings generated through this form of analysis. Selection of both dependent and independent variables then is limited to data provided in the U.S. Census for the Detroit Urbanized Area. However, a large number of behavioral characteristics can be obtained from this source, and while they may not include many of the aforementioned behaviors,. they do contain several measures of family instability and behaviors which previous researchers have used as indicators of social well-being. 0f the dependent variables used in this level of analysis, only measures of migration might in and of themselves po£_be considered in some degree indicators of social well-being. However, the non-migrant index, which measures the degree of residential stability in a census tract, and the new residents figure, which describes population influx 60 into an area, may denote one of the most significant behavioral responses to spatially restrictive stimuli in the urban environment. If the individual finds his environs Spatially restrictive and in turn stressful, it would initially seem that he has two choices in reSponding to such a situation: remain and adapt to the circumstances, or avoid the stress by moving to a more satisfying and congenial locale. Wolpert has used this premise as the basis of a model which describes the decision-making process of the individual who migrates in reSponse to a stressful environment:2 In essence, he relates migration behavior simultaneously to variations in pe0ple and places. The migration question, however, raises an additional question regarding environment and behavior. It is reasonable to assume that if the individual is unable to c0pe with his surroundings he will eventually leave. But a considerable portion of the urban population is limited in its residen- tial choice by numerous social and institutional barriers. What are the behavioral consequences of an individual's or population's inability to avoid the stress of offensive Stimuli? High or low rates of migration, depending upon their location within the urbanized area, may provide strong indicators of at least one major reSponse to stressfully per- ceived environmental stimuli. In chapters One and Two some effort was made to distinguish between crowding and pOpulation density. To test the validity of this argument and to measure spatial restriction in as comprehensive a manner as possible, several structural and human density variables have been 2 Wolpert, Julian, “Migration as an Adjustment to Environ- mental Stress,” Journal of Social Issues, vol. 22 (October I966), pp. 9l-l02. 6i selected as environmental characteristics. Spatial restriction has a number of components among which is the number of persons per room within a dwelling unit. Where the number of persons per room exceeds one, this type of spatial restriction is referred to by the term ”overcrowding.” Overcrowding, or room density, is the form of spatial res- triction which Galle, Cove and McPherson, and Carey,3 found to be the most highly associated with'pathologicaf'behavior. A second type of density is a direct function of housing type: how many people live under the same roof, whether it be a single family dwelling or an apartment com- plex. This is called structural or building density, and also can influence the degree of contact between individuals. Within the structure spatial restriction can be further refined so as to view the potential of the particular housing arrangement to bring people into desired or undesired contact with one another. Therefore, not only do the number of persons per household represent a form of potential spatial restriction, but the number of rooms available to that household population affects their degree of contact as well. Another element of spatial restriction refers to the number of people living on a certain amount of territory -- population density. This can be further refined by considering the degree to which that land is in residential use. It is possible to have a low population density in any particular census tract, implying a low degree of spatial restric- tion; but if the residents of that tract are limited to only ten percent of that territory, this information paints a much different picture. Carey, George W., “Density, Crowding, Stress and the Ghetto,” American Behavioral Scientist, vol. l5 (March/April I972), pp. A95-5IO. 62 Thus structural and human density in residential space also represent the degree of spatial restriction in the territory available for their occupation. Therefore, it is not incorrect to argue that population density is a form of spatial restriction. It is misleading to argue that it is the only form. This argument applies to any other form of the parameter as well. By including the various components of spatial restriction, it will be possible to determine which, if any, are significantly associated with behavior. The most recent available source that distinguishes residential from non-residential land use in Detroit was completed in I965. This map of generalized land use for the Detroit S.M.S.A. is shown in Figure 3. While highly generalized, it provides the basic distinction that is needed to calculate the proportion of the total tract area which is used for residential purposes.“ However, because considerable residential development has occurred on the northwestern fringe and much of the suburban periphery, a means of converting graphic data to census tract form and updating them to I970 had to be found. A base map and grid were simultaneously overlain on Figure 3 The base map depicted residential land use in each census tract, while A The U.S. Census provides data that can be used to calculate room densities, but not residential and total area of each tract. A planimiter was used to calculate total tract area, with an average of three measure- ments being used as a final figure in order to minimize error. 63 GENERALIZED LAND USE I965 RESIDENTIAL - COMMERCIAL - INDUSTRIAL PUBLIC I INSTITUTIONAL RECREATION sOURCE: ‘ DETROIT REGIONAL TRANSPORTATION AND LAND USE STUDY. A SPECIAL nOIECT OF THE PLANNING D AGRICULTURAL l- VACANT DIVISION. SOUTHEAST MICHIGAN COUNCIL OE COVERNMENTS. Mllu o lo 20 40 Figure 3 6A of each tract devoted to residential land use. From these percentages the number of square miles in each census tract given over to residential land use as of I965 was calculated. The U.S. Census provides information showing the percentage of residential structures in each tract which have been constructed since I965. These data, when combined with lot size data, can be used to update residential area by census tract. Lansing and Hendricks report that the average lot size in the city of Detroit is only .l2 of an acre, while in the outlying areas the average lot is nearly three times as large.5 By multiplying the number of structures built since I965 by the average lot size in each tract, the amount of new residential territory that has deveIOped since that year can be determined. When this figure is added to I965 residential area, an up-to-date measure of residential area in each census tract is achieved. Inclusion of this characteristic as a component of spatial restriction will allow the distinction between density per total area and density per residential area, thus providing a truer picture of business and agricultural areas which contain little residential land use. The initial task in the analysis of the hypothesized environment- perception-behavior relationship is to determine the manner in which 5 Lansing, John B. and Gary Hendricks, Living Patterns and Attitudes in the Detroit Regiopl a report of TALUS, the Detroit Regional TranspOrtation and Land Use Study, (January I967), pp. Al-A3. the 87% Ex Nb. de oi C 65 the environment varies from location to location in the Detroit Urbanized Area. Given the complexities of the relationship, a means of analysis is needed which can simultaneously manage a large number of var- iables, compensate for random error and disentangle the complexities of the relationship into major and distinct patterns of regularities. Factor analysis is such a technique. Factor analysis is a mathematical tool which can be inherently spatial. As phenomena co-occur in Space and in time, they form distinct independent patterns of interrelated characteristics. Factor analysis extracts those interrelationships and indicates their spatial patterns when the unit of analysis is locational. Thus a factor analysis of density and behavioral variables for census tracts will provide two essential pieces of information: I) an indication of the extent that various behaViors and components of spatial restriction are interrelated, and 2) the spatial variation of these characteristics of place, the first element in our hypothesized relationship. As noted in the review of literature, many of the components of spatial restriction are strongly associated with income levels and the distribution of non-whites in urban areas of the U.S. High population and structural density areas tend to be low in income and to have a relatively high percentage of non-whites. It is also possible that several other variables are associated with the dependent characteristics, Rummel, R.J., ”Understanding Factor Analysis,” Journal of Conflict Resolution, vol. II (I967), p. AAA. T-c' 66 but are interrelated with each other as well. Under these circumstances, a factor analysis which includes social class and racial characteristics may not provide meaningful patterns of spatial restriction and behavior which are independent of socio-economic status and race in the study area. To compensate for this likelihood, separate factor analyses will be run to test for the impact of race and social class have on the interrelationship of the other characteristics. Factor analysis is basically descriptive and is seldom employed for purposes of statistical inference. While this method of analysis does provide an indication of which variables are interrelated and in what direction, tests of significance and hypothesis testing of this nature are usually reserved for other techniques. Under circumstances other than interrelated data, simple correlations would provide measurable degrees of interrelationship between spatial restriction and behavior.. But clearly it is impossible to know if areas of high spatial restriction might have more family instability because of less available space or because they are low income and have a higher percentage of non-whites. Given this ambiguity, it is necessary to turn to still other methods of statistical analysis where these control variables are held constant and the effect of spatial restriction assessed. A convenient means of achieving this is a multiple correlation and regression analysis which will measure the association between a component of spatial restriction and a single behavioral characteristic, holding all other variables constant. Because the order in which an independent variable is entered into the multivariate correlation and regression equation 'determines the amount of variance it' 67 explains, it is more desirable to perform an analysis in which no variables are forced into an equation, Rather ' all are allowed to enter at the points when they are the most important remaining characteristic, i.e., when they account for the greatest proportion of the variation in the dependent characteristic. This type of stepwise multiple regression adds variables one at a time in the order of the percentage of the total variance they explain. Through this form of analysis it is possible to see the relative importance of each variable according to the additional variance it explains- A Spatial Analysis of Place Characteristics Census data representing the density and behavioral character- istics were subjected to an R mode factor analysis employing an orthogonal rotation of axes. Additional independent variables describing income, education, employment and ethnic characteristics were used as controls to compare their potential for explaining behavior with that of popula- tion density. It has been theorized that areas high in spatial restriction will be positively associated with high frequencies of social indicators. One means of empirically testing the hypothesized relationship between spatial restriction and behavior is through an analysis of factor loadings. Factor loadings can be interpreted like correlation coefficients, i.e., they mea- sure which variables are interrelated in a pattern, or factor, and the degree and direction of that interrelationship. If the hypothesis is confirmed we would expect overall spatial restriction and its components to form a pattern of interrelatedness with the various forms of behavior. 68 Furthermore, we would expect that relationship to be positive. Factor scores resulting from this analysis indicate the perfor- mance of a census tract on a factor of interrelated variables, and can be used for two purposes: I) to describe the spatial distribution of related restriction and behavioral variables; and 2) to select census tracts that will serve as the study area for collection of individual data on perceived meanings of environmental stimuli. A summary of the factor loadings of the rotated factor matrix is provided in Table A. This first analysis seeks patterns among place characteristics and behaviors exclusive of the influence of social class and race. Six patterns of interrelationships were found among the characteristics, explaining over seventy-six percent of their variation in the study area. Factor One, entitled Low Density Behavior is of particular significance not only because it accounts for the greatest proportion of the total variance (twenty-eight percent), but also because it is the only factor which contains a substantial number of restriction components and behavioral pathologies. This is in essence a behavioral stability pattern. Measures of low social well-being all lOad negatively, indicating a population which does pot behave in this fashion. The strong negative association of population density and a mea- sure of structural density, the number of housing units per square mile, with behavior provides initial evidence supporting the hypothesized relationship between spatial restriction and behavior.. These two components load most highly on this factor indicating that they are more interrelated with such behaviors. as welfare recipiency, run-away children, high school drop-outs and forms of family instability. As hypothesized, this initial finding indicates that those areas which 69 LOLODO >3 emum_SO_mu "mumzom .oom. wmmo. ome. Ndum. mnoa. .qu. .>.¢ o>_um_:E:u memo. mono. moo_. mm__. mm~_. _:wu. oucm_cm> mo :o_ucoaoLm umoo. w_:w.u JWNJ. Nmmo. 0—00. .m-.- mOL< mom can a: make. N:N_.I .mm..- ammo.- mmao.- Name. :4 max x ammo. m_o_.- mewm. mac..- memo.- memm.- _z om Lpa a: :o_m. Nmmo.- am... o~om. maoo.- mmmu. cm can meme mmom. mm:o.- :w:.. wmmo.u owe—.I ommx.u memo no; .mmo. o-o.u .om~.I ammu. mum_. _m~N. emu um_o m_mm. momo.n _~m_. mmmo.u _om. mmw_.u mom 3oz comm. :m_o. .muo.u nmmm. «um..- excu>o Com momm. wamo. mN_m.- mmm_.l mauo.- u_c=m: max mama. k_~o. ckum.- ma... m_~m.- cptm_z coz maaa. km~_.- amkm.- mama. ~m:o.- >p___ptpa mmmm. wmmo.u .mmo.n wmmo. Noo_. acme—O3 Noam. Noo~.- Noeo. mos..- om_o. “were: cm: mmmw. ammo.u :moo.u memo.u moNo. o. meCumz _mmn. :moo.u mmmo.u msmo. cums. mmom. .uuacum zoz «Nam. omo_. moo_. mmm_. mmm_. .I a0.5 um _: mmwm. .omo. o:~m.- ammo. muno.u mama. e_;mz meme .maw. ammo. mum..- memo. mmo_.n mwmw. Emeecoz >u__mc:EEOU co_uo_cumom ucOeeOC_>cm mmc___ozo >u_mcoo mucoe_mom co_>mzom Oanm_cm> _m_umam ouco>_o acoeucma< 5001 302 >u_mcoo 304 mmc_um04 Copomd vaumuoa "mcouomu .mco_>mcom mummOme< ocm co_uu_cumom _m_umam : m4m3 0000.30.00 "mumnom .000. 0000. 0000. 0000. 0000. 0000. ..0> .0.0 .500 0000. 0000. 0000. 0000. 0.0.. 0000. ..0> 00.0.000.0 0000. 0000. .000.- 0000.- 0000. 0000.- 0000.- 00.< 000 .00 0: 0.00. 0000.- 000..- 000.. 0000.- 0000.- 0000. 00 .0.0500.m00 0 0000. 0000.- 0000.- .000.- 0000.- 0000.- .000.- 0..z .00 .00 0: 0000. 00.0.- 0000.- 000..- 0000. 0000. 0000. 50 .00 0000.00 0m00. mm::.n ~m_~.u omu..n 00m_.u w.m_.n oumw.u >u_mcoo 000 0000. 0000. 0000. 0000. 0000. 0000. 0.00. 000 00.0 0000. 0000. 0000.- 000..- 0000. 0000.- 0.00.- 00500.000 302 0000. .000. 0000. 0.00.- 000..- 00.0. 0000.- 0c.030.0.0>0 0 00mm. mom..u .00.. 0wa. 0000.- m~0o.u .mmu. 0.520: .00 050 0000. 00.0. .00..- 0000.- 0000. .000.- 0000.- 00.0)00 0000. .000. 0000.- 0000.- .000. 000.. 0000.- ..05050 0.02 0000. 0000.- 0000. 0000. 0..0.- 000.. 000.. 00.0.: 002 0000. 000.. 0000. 0000. 0000.- 0000. 0000.- 00.....00 0000. 000.. 000..- 0000. 0000. 0000. 0000.- 0.0..03 0.00. .000. 0000.- 0000.- .000.- 000..- 0.00.- 000.00 .00..0z 0.00. 0.0.. 0000.- 0000.- 00.0.- 000..- .000.- 0. .0...0z 0000. 0000.- 000..- 0000. 000.. .00..- 0.00. m.> .000 00: 0000. 0000.- 0000.- .000. 0000. 000..- 0000. 05000. 00: 0000. 0000. 00.0. 000.. 0.00. 000.. 0.00. 00 00.00.00 0000. 00.0. 000.. 0000.- 000.. 000.. 0000.- 00.0 000 .0 00.0. 0000.- 0000.- 00... 0000. 0.00.- 0.00.- 0.002 0 0000. .00..- 0000. 00.0. 0000.- 0.00. 000.. 0.10: 0.00 0.00. 000..- 00... 0.0.. 0000.- 00... 0000. 5005.02 >u._m::EEou u:OE:O..>cw co.uu..umo¢ 00.3030 00:00.000 >u.mcoa .o_>m;om o_nm_.m> Ou.o>.o .0.0000 .000 mam—01.000: 302 5000 >u.mcoo 304 .0 u4m3 oeum_:o_mu Humanom _mm.1 mm_._1 .ma.~n mam nmo.1 Nmo.u oo:._1 mo: mmm.u .am.i mum._1 w>_h_h< om:._ am..- _om. «w .13.- 3:. im. .2 mom. moo. mom. m>.h_mom uh_h_moa :u_: omcmcu mo co_uauoLoa >um_m:d mo co_>mzom >u_mcoo 304 pooch maocu comuooocom o>_u_mo¢ 4<:o_>_oz_ mhmzom new >u_mcoo mummome< nu_3 moLoom Louomm co_uaoocoe _mucOEc0c_>cm mo com_cnaeou _N w4mn coup—:u_mo "mumaom m Mm.mm ”.00. cam _o>04 co_uuou0cm ucocomm_t >_ucmu_w_cm_m 0cm memos« :o_:. mnu. mua. ummco>< .aoa. «.mo. mmm. moo. o>_u_m0a oumcocoz wmmm. «mam. «cum. o_m. m:~.: o>mummoz mumcocoz oamm. «mmm.. .omm._ «woo._ «Nae. _:m.- u>_ummoz :3: momcmm wmo. m~:. moo. m:~.u .am.i memo: ucmo_m_cm_m o>_u_m0m ommco>< o>_u~m0m o>mummoz o>_ummuz ma30cc umoucozm 5mm: uumcoco: uumcovoz cm_z menace o>_u u >u__m:d .mucch0c_>cm wo co_uaoucom o>_u_moa Lo; amok memo: mo c0m_cma50u :N uam_ucmu_m_cm_m 0cm mcmoz « om—J. «mam. mNo._ NNN awpa. «mm—.— «Nom. 50:. .NN mm_:. emm~._ «mom. _o_. mom. mmm mmoa. «mmm._ «~n~._ «0.5. «mow. Maw. u mmm Nmoz. «mmm.- «:m~.. «Nun. k—wm. «.0. mmN. u __4 moon. «omm.— «_wm._ «mmm. ammo. mmo. 550. «mm. . No ma_m. «_w~.~ «mme._ «4N..— «muo.— «:_4. «N04. «mum. nmo. 0 mo: owom. «NmM.N «Nom._ ko:N._ «mm—.— «0mm. «m—m. k—aa. m——. mum. i —_ mmmN. «mam.~ «:mN.N «Nmm.— «_mm._ «an. «cum. kmmw. «wwm. «uma. mN~.—I mmo momcmm s~o.. m~o._ nos. mom. ma~.- mm~.- ~mm.- nmw.- m-.- m-._- memo: ucmumm_cm_m mom NNN —NN nmn mmm _—a «m mo: p. wa uumch amouLOcm muomch cub i >u__m:d .muc08c0c_>cw mo co_uauuLum o>_u_moa co» amok memo: mo :0m_cma50u mN u4m2 coup—:u_mu "wuxaom coco. omom. omm_. mnqw. manm. ow~_. mm:~.- some. .mmm. m:o_. o__w. ommo.n mmqo. momm. mmSN. ommm.- ~_mn. memo.- .mm..- m_m~. _mmm. cums. __om. ~m_~. meow. >um_mc:EEOu >u_mcoo om: new; co_uu.cumu¢ 304 _m_ucou_mmm . maz_o mo co_umoa0cm o>_um_aee:u .nmw. oucmmcm> $0 co_ucoa0cm acao. moc< mo¢\mumc: mc_m:o: ommo.- on: new; _m_ucoe_mo¢ w mmmw. scam can chmcom mauo.- >u_mcoo compo—anon ammo. poozocuco>o N mooo.u u_c: mc_m:oz con mEOOm ammo. p.050mao: coo chmcum >u_mcoo scam o_nm_cm> mac—vmoA couumu ooumuox "mLOuumu co_uu_cumo¢ .mwumam on m4mo com :o... em:o.- Rome. mwo.u .mo.u m.o.u «m.m. «owe. u.e:m: max «man. homo. «smm.I mo..- «.mm. m~.. am~.I au~m.u >uco>oe «m.m. «awn. «mma.u m.o.u emu. :.o. cm..- «on:.: eeve: o.nz mm..- ~mo.- me_. _ao. S¢m~.- «NNN. _mo. se.. coLm_z coz eoma. mm..- w:~.I mmo. «New. .o.. «amm. mo..- >u...ucou ooo.. «nmm. «mN~.- omo.- «.Nm. om.. m.o. «amm.- acme.o: ooo.. «omm.u om..- no.. m~.. «omo.u «wno.I unecco cm: sm_m.- mmo.- «ass. __~. «mm~.- «mam.- a. Lm_Lumz «own. Nmo. ea.:.n so..- mno. Snow. mc>zum out ooo.. moo. s:::.. .5..- «mum. ammo. oeooe. to: ooo.. m...: a... .NN. m:.. auscum 302 ooo.. «mo. nmo. «smm.I noco um .: ooo.. :No.I om~.I ocmuz com ooo.. ammm. e.gm: meme ooo.. Soweto: acme umucco oeouc. uuacum ooco acme: o.;m: en» :02 .ucom .03 cm: ce.cumx on: no: so: om .: cue mcoe Eco: mhu<¢h xo<:mm oz< 42m mho. .o. ecu um ucmu.m.cm.m k Leeann >3 oeue.:u.mo "muxoom ooo.. «moo.u ..:N. ooo.- mmu. u.~.u mmo. ~:.. mm..- «sen. .«NNN. ooo.. «ohm. ~mo.: «.mu. mNo. am.. ma~.I mNo. «n.m.I on..- ooo.. ~...I «:.m. «omm.u «~94. noN.I «.om.I 9... am.. ooo.. NNo. «.aa. umo. «o.n. n.o. -..I .o..- ooo.. «~m~.I «soc. «mo.u ~oo.- mNo. m... ooo.. :ou.u «emu. No.. «amm.u no..- ooo.. a:o.u «.m~.I m.~. moo. ooo.. mmo. om~. :o.. ooo.. emmm.u -oo~.I ooo.. _ecmc. ooo.. o u.c:m: xucu gene: con 3: can can a: men; ace um.o 302 com max I>oe 0.0: A.o.ucouv u x.ozwmm< o cue u.c:m: mam >uco>0m neon: o.m: I88 an..- m.~o. nmu. n.~. NMN. oo~.I mm~.I emo..I mo~. mm.. om~.I mom.u «#0.. ammo.u ma..u mm..n an..- ma.. m~.. u..o.u mo..- uomo.u *mmo. .o.. *omo.- *mNo. o.:. «.m. was. “NM. nmm.- ~.m.- nm~. on:. own.. wma.- .N.. *mao. co..- m:..: m.m.n unmo. «ouo. mm.. -..I mm..u omm. muu. «woo. mm.. mma. .o.. .No. owe. cmc.u oca.l amm. Jam. mm..- .am.u \~.o.u *Nmo.u o.m.u .m~.I ooo.- mum. mma. «ma. so:.u Nom.I «me. Now. mam.- *.mo.I emu. 05.. mm.. *mmo.I mm..- mun. emu. se.. oo~.I ms~.i Nu.. m.:. woe. ammo. mom. u.:. ma:.a o.~.u own. m.m. mac. mm~.I m... ~m.. om~.u oo~.- :.m.u .m:. :.o. se.. mam.u mu..- m.o. o.:. m-.- mom. :wm. woe. .mm. .om. ooo.- mm~.- mmm. ~.u. mm~.- omm.- an..- «mu. cum. :.~. :mm. :wm. mom.u o:~.: :.m. com. on..u mum.u ooo.. *.mo. wo~.I .m..I :N..u *~.o. mono. nom.n ~4~.I om..u mm.. mmu. ooo.. mum. nm.. m.m. .mm. mm~.- Nm..- mom. ~m.. mmN. o-.- ooo.. meg. .mm. mum. muo.u owN.I .mm. awn. mm..- now.u ooo.. moo. mum. o.m.i ca..- ~s.. «mm. .m~.I ~.m.u ooo.. _mm. mmo.- 5mm.- sum. New. ~.m.- omm.- ooo.. mom. mam. ~mo.: om:.I emu. :mo. ooo.. mom. mom.I o.m.- com. new. ooo.. o:~.- mm~.- mn.. can. ooo.. mos. om~.- moo.- ooo.. noo..- mmn.u ooo.. :mm. ooo.. cmcm.z >u... 0cm» umoLco o_ mc>num pecun- uuacum coco econ: e.gm: so» :02 nucoe I.oz cm: cm.cumz on: no: 302 um .: can mean Ecoz o com u.c:m: max >uco>oe neon: 0.0x cmcm.: coz >a__,utoc 0co.oz unecco Ln: o. cm.cum: mc>com no: osou... out auscum 302 note um .1 ocmoz cue e.gm: neon EmmEcoz mku<¢h 44< I mu.hm.¢whu<¢zm wh3 naum.:u.mu ”mumaom oueoo.mcoo mo .o>u. .oo. um acmu.m.cm.m koz * ooo.. om..- mum. noNo. ~.m. .mm.- mm.. uomo. .om.- ~.m. mNN. ooo.. «mmo. um.o.- «moo. xoao. «.o..- umoo.- o... mm..- «N..- ooo.. m.~.- MN“. com.- «No.. «.mo. cm~.- oua. .mm. ooo.. .o~.- wnm. *No..- mom. .a..- oo~.- mm..- ooo.. ..o.- *woo. m~.. om..- mug. m.:. ooo.. *~.o.- *mmo.- m.m. mam.- mow.- ooo.. *mmo.- cm~.- mmw. mo.. ooo.. n.~.- :mm. ::M. ooo.. umm.- nm~.- ooo.. new. ooo.. (mm :4 mpg zom ea a mono omo mom o3co>o u.c:m: >uco neon: can oz com com a: mean no; um.o 302 com me: ->0e o.mz A.o.ucouv u x.ozmme< (me Log 2: :4 mom can 2am Lea 3: em a neon memo eon omo um.o no: 302 u3cu>o cue u.com: max >uco>oe gene: o.mz l90 APPENDIX G MEANS AND STANDARD DEVIATIONS OF ALL AGGREGATE CHARACTERISTICS Variable §§§2__ Standard Deviation Normfam 78.958A l6.66OA Pers Hshld 3.272l .5576 Per Negro 22.2622 35.5AAI Hi Sc Drop 17.0878 l2.620l New Struct IO.6276 l7.50AA Med Income l2002.h807 hl57.8hl9 Med Schyrs ll.4353 l.37OA Matriar ID l3.0730 9.3280 Mar Unrest .l599 .3685 Welfare 5.78%7 7.0A82 Fertility 3.2589 .7365 Non Migran 55.5092 lh.5llh Male Unemp 5.775l A.O962 Poverty 8.375I 8.8970 Rms Hsunit 5.2327 .8AO3 Per Ovcrwd 8.0535 A.265l New Res 29.l53h l2.8058 Dist CBD I0.8lh5 6.Al52 Pop Dens 9335.99OA 69A5.6l6l Pers P Rm .55h8 .l027 "U per SQ“ 3295.5396 3538.l50l Per Res LU 7I.A925 A8.305h HU per RSA 9879.0836 3I389.I79N SOURCE: Calculated by author I3! oucoo.meou .o .O>o. .o. may on aceo.m.co.m « Lecuom >9 ovum—no.00 "mumoom ooo.. .oo.- mmo. .oo.- oooo ooutoo ooo.. «o_o.- «mom.- _moo ooutoo ooo.. «mos. so mtoo ooo.. etoo mtoo oooo _moo so atom aoucoe ooucom moon mean A.... ~_o. soo_. ooo. ooo. moo.- moo.- «m_~.- «m_~.- neo.- o_o. «oum.- oooo ooutoo «mo~.- .oo.- Rmom. amoo. «om..- ~_o. m~_. «o-.- Semm.- «moo. ooo. «mo..- .moo oouLuo o... Nmo. o.o. «.oo.- EoNo. Room. mmo.- ooo.- omo. oo..- «.o.- .Noo.- so mLoo «.o.. oNo. «omm.- om..- so_~. «Nam. mmo.- moo. «m_m. sou..- o_o. «.NN. eLoo mtoa ooo.. ooo. soo~.- oN..- moo. ems..- Ro-.- moo. Seam. «ooo.- omo. «se.. oSz moe_e ooo.. mmo. «.o. oeo. «.o. moo.- moo.- om..- nmo.- mso. sow..- mum >OLo ooo.. Smuo. «mom. Room. mo.. «oo~.- «.oo.- «mo_. omo. «o.o.- mstoo 2:: ooo.. scam. «mow. mo.. «ma~.- «mmm.- soon. moo. «oo~.- em E:z ooo.. «who. _No. _N..- moo.- ooo.- __o.- ooo.- mom 2:: ooo.. «mom. oNo. «om~.- *_o~. ooo.- m_o. o__oo soz ooo.. om.. «Nm~.- «moo. __mo. nmo.- um _mo_Lmz ooo.. «om_. “.o. oo..- «m_~. mm_o oouuo ooo.. emom.- ooo. «ooo. o_;mtmczo ooo.. ooo.- ooo.- moo cameos ooo.. o.o.- xum ooo.. some o>z moo metoo so mum o__;o om mm_o o_;m moo xom «one moe.h >0ce 8:2 2:2 5:2 5:2 .mu.cmz oaouo coc3o cameo; mo_»m_¢o»o<¢.oz_ 43¢ no mzo.e