@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix skos: . vivo:departmentOrSchool "Arts, Faculty of"@en, "Anthropology, Department of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Prattis, James Ian"@en ; dcterms:issued "2011-06-07T23:16:48Z"@en, "1970"@en ; vivo:relatedDegree "Doctor of Philosophy - PhD"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description """The problem addressed here is the examination of the dilemmas of decision making in different substantive contexts. The contexts include peasant farmers deciding whether or not to accept an agricultural innovation, sophomore students gambling, and fishermen on British Columbia's West coast making up their minds where they intend to fish. The major conclusion is that the structure of decision making is a constant cross-cultural variable. This implies that socio-cultural factors can most profitably be viewed as a framework within which a similar structure of decision making occurs. From this consideration it follows that decision making is a basic building block for the study of social behavior. Identification of the basic structure of decision making is in terms of a theory of risk taking which relates the type of decision strategy used in any situation to considerations of the resources, information and utilities that particular individuals possess with regard to the event the decision is about. From an initial substantive concern with Third World farmers deciding to adopt or reject agricultural innovations I generalise to a number of statements about individuals and risk. Risk taking refers to behavior in situations where there is a desirable goal and a lack of certainty that it can be achieved with attendant possibilities of loss. Three main sources are used to test the assertions about risk taking — first a laboratory experiment, then fieldwork and finally secondary sources. The argument made to justify this procedure is that these situations constitute particular empirical settings in which the propositions about risk taking could legitimately be tested. The argument rests on the assumption that the same scope conditions are met in each substantive setting. The scope conditions considered here place an individual decision maker within parameters of resources and subjective utility with regard to some outcome, information and incentive conditions for any risk. The propositions predict the type of decision strategies that would be employed for given values of the above parameters. This level of abstraction, which is not tied to situational boundaries is, I submit, a necessary prerequisite for effective cross-cultural analysis. Thus my thesis is not about peasant farmers, or fishermen or gambling, the work attempted here is concerned with individuals and risk. The tools used draw upon a tradition of model building extant in economics with reference to decision theory. Thus the work attempted here is part of the growing formal tradition in economic anthropology. The model built was not a perfect fit with data but adequate enough to give one confidence in the set of methodological assumptions which were fundamental to its construction. Also the testing procedure employed has implications for the manner in which anthropologists may conduct enquiry, as it establishes that laboratory contexts are as legitimate a source of verification as field contexts. It is from these two considerations that I offer a test case for methodology in economic anthropology."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/35286?expand=metadata"@en ; skos:note "DILEMMAS.GF DECISION MAKING: A METHODOLOGICAL TEST CASE IN ECONOMIC ANTHROPOLOGY by JAMES IAN PRATTIS B.Litt, Oxon, 1967 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in the Department of ANTHROPOLOGY AND SOCIOLOGY We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA July, 1970 In p r e s e n t i n g t h i s t h e s i s in p a r t i a l f u l f i l m e n t o f the r e q u i r e m e n t s f o r an advanced degree at the U n i v e r s i t y o f B r i t i s h C o l u m b i a , I ag ree tha t the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and s t u d y . I f u r t h e r agree t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y pu rposes may be g r a n t e d by the Head o f my Department o r by h i s r e p r e s e n t a t i v e s . It i s u n d e r s t o o d tha t c o p y i n g o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be a l l o w e d w i t hou t my w r i t t e n p e r m i s s i o n . Department o f The U n i v e r s i t y o f B r i t i s h Co lumbia Vancouver 8, Canada ABSTRACT The problem addressed here is the examination of the dilemmas of decision making in different substantive contexts. The contexts i n -clude peasant farmers deciding whether or not to accept an agricultural innovation, sophomore students gambling, and fishermen on British Columbia's West coast making up their minds where they intend to fi s h . The major conclusion is that the structure of decision making is a constant cross-cultural variable. This implies that socio-cultural factors can most profitably be viewed as a framework within which a similar structure of decision making occurs. From this consideration i t follows that decision making is a basic building block for the study of social behavior. Identification of the basic structure of decision making is in terms of a theory of risk taking which relates the type of decision strategy used in any situation to considerations of the resources, in-formation and u t i l i t i e s that particular individuals possess with regard to the event the decision is about. From an i n i t i a l substantive concern with Third World farmers deciding to adopt or reject agricultural innovations I generalise to a number of statements about individuals and risk. Risk taking refers to behavior in situations where there is a desirable goal and a lack of certainty that i t can be achieved with attendant possibilities of loss. - i i -Three main sources are used to test the assertions about risk taking — f i r s t a laboratory experiment, then fieldwork and f i n a l l y secondary sources. The argument made to justify this procedure is that these situations constitute particular empirical settings in which the propositions about risk taking could legitimately be tested. The argument rests on the assumption that the same scope conditions are met in each substantive setting. The scope conditions considered here place an individual decision maker within parameters of resources and subjective u t i l i t y with regard to some outcome, information and incentive conditions for any risk. The propositions predict the type of decision strategies that would be employed for given values of the above parameters. This level of abstraction, which is not tied to situational boundaries i s , I submit, a necessary prerequisite for effective cross-cultural analysis. Thus my thesis is not about peasant farmers, or fishermen or gambling, the work attempted here is concerned with individuals and risk. The tools^used draw upon a tradition of model building ex-tant i n economics with reference to decision theory. Thus the work attempted here is part of the growing formal tradition in economic anthropology. The model built was not acperfect f i t with data but adequate enough to give one confidence i n the set of methodological assumptions which were fundamental to i t s construction. Also the testing procedure - i i i -employed has implications for the manner in which anthropologists may conduct enquiry, as i t establishes that laboratory contexts are as legitimate a source of verification as f i e l d contexts. It is from these two considerations that I offer a test case for methodology in economic anthropology. - i v -TABLE OF CONTENTS Page Abstract i Chapter 1: Introduction 1 Economic Anthropology: Controversies in Methodology — The Anthropology of Choice — Progression of Thesis. Chapter 2: The Problem and Its Conceptualization 19 Introduction — Farmers and Innovations — LWHU State of Nature — Some Propositions — Supplementary Propositions — Social Mobility — Social Marginality — Co-operative Structure — Whither Next? — Decision Theory — Theoretical Propositions — Testing — Summary. Chapter 3: Test Case: The Experiment 56 Introduction — Experimental Design — Decision Task and Incentive Conditions — Wealth Gradient — Playing for Real Stakes — Information — Subjective U t i l i t y — Data Collection — Testing — Farmers, Risks and Experimental Data — Implications for Field Test. Chapter 4: Test Case: Fishermen as Risk Takers 96 Introduction — The Fishing Industry in British Columbia — Gillnetting — Constraints — Fieldwork — Data Collection — Risk — Wealth — U t i l i t y — Information — Strate-gies as Risks — Rivers Inlet — Johnston Straits — Closure of Northern Fisheries — Investments on Technology as Risk — Case No. 1: Ernie McAllister — Case No. 2: Neil Anderson — Test of Propositions — Strategy Risks — Investment Risks — Congruence Between Investment and Strategy Risks — Relevance of Cultural Factors — Summary. - v -Page Chapter 5: Farmers as Risk Takers 157 Introduction — Case No. 1: Malayan Rubber — LW Farmers and Risk — Wealth and Risk — Information and Risk — U t i l i t y Considerations — Moturiki Community Development Scheme — Food Fishing in Alert Bay — Spanish American Farmers and Hybrid Corn — Some Implications. Chapter 6: Conclusion 187 Introduction — Methodological Implications — On Models and Research Cycles — Editing — Intervening Variables — (1) Education — (2) Social Marginality — Scope Conditions: The U t i l i t y Parameter — New Problems — Theoretical Implications — Practical Impli-cations — Conclusion. Bibliography 216 Appendices: Appendix I: Duplex Gambles Master Sheet 223 Appendix II: Instructions to Subjects Appendix IV: Chi Square Median Test on Propositions. 224 Appendix III: Transformation of Experimental Data, Tables IV-VII. 256 260 Appendix V: Scale for Wealth Stratification for Alert Bay Gillnetters. 276 - v i -LIST OF ILLUSTRATIONS Page TABLES Table I: Summary of Scope Conditions 65 Table II: Data from Experiment 67 Table III: Bets Categorised by Riskiness 68 Table IV: Bets Categorised Per Subject 256 Table V: Frequency of Bets Played Per Cell 257 Table VI: Raw Data Matrices 258 Table VII: Corrected Data Matrices 259 Table VIII: Chi Square Median Test: Proposition 1' 74 Table IX: Chi Square Median Test: Proposition 2' 79 Table X: Chi Square Median Test: Proposition 3' 83 Table XI: \"Sure Thing\" Bets 86 Table XII: Low Payoff Bets 87 Table XIII: Relative Weighting of Risk Components 89 Table XIV: Summary of Experimental Test 90 Table XV: Risks on Technology and Strategies 106 Table XVI: Information Circuits 113 Table XVII: Strategy Risks i n Rivers Inlet 117 Table XVIII: Strategy Risks in Johnston Straits 118 Table XIX: Northern Closure and Strategy Risks ' 120 Table XX: Investment Risks and Range of Exploitation 124 - v i i -Page Table XXI: Risk Taking by Ernie McAllister 134 Table XXII: Risk Taking by Neil Anderson 139 Table XXIII: Data on Alert Bay Gillnet Fishermen 140 Table XXIV: Strategy Risks 141 Table XXV: Proportion of High and Low Risks Taken by Alert Bay Gillnet Fishermen 142 Table XXVI: Investment Risks by Japanese and Yugoslav Gillnet Fishermen 150 Table XXVII: Adoption of Hybrid Corn by Spanish American Far-mers 181 FIGURES Figure 1: States of Nature 22 Figure 2: Propositions 1 - 3 28 Figure 3: Propositions 4 - 6 31 Figure 4: General Model 45 Figure 5: Payoff Matrix for Innovations 47 Figure 6: Duplex Gamble 53 Figure 7: Gambling Wheel 59 Figure 8: Data Matrix 66 Figure 9: Data Matrix 70 Figure 10: Information Circuits Model 145 Figure 11: Revised Information Circuits Model 202 - v i i i -Page GRAPHS Graph 1: LW Preference for Low Risk Strategies: Proposition 1 (SR Bets) Graph 2: LW Preference for Low Risk Strategies: Proposition 2 (FR and VR Bets) Graph 3: Risk Taking with Respect to Wealth (SR Bets) Graph 4: Risk Taking with Respect to Wealth (FR and VR Bets) Graph 5: Risk Taking witn Respect to Information (SR Bets) Graph 6: Risk Taking with Respect to Information (FR and VR Bets) 71 72 77 78 81 82 Map 1: MAPS Coastal Fisheries of British Columbia 99 - 1 -CHAPTER 1 INTRODUCTION The main stimuli for this dissertation were considerations of underdevelopment in the Third World and a concern with methodology in economic anthropology. In a sense I am combining idealism with a philosophical basis for enquiry, as both led me to focus on dilemmas of decision making. My own personal experiences in community development pro-jects in Greece and Sarawak l e f t me with a broad general interest in development problems. However, the translation of this interest into workable models was hampered by the vast number of variables seemingly pertinent to the development process, and the fact that my i n i t i a l attempts to explain the development process merely offered another plausible account rather than a definitive study. One theme that remained consistent from my own f i e l d ob-servations and earlier work (Prattis: 1967) was that I suspected an implicit rationality lay behind the decisions made by the peasant farmers I was working with. The parameters or framework within which decisions were made varied, but I f e l t that the process of weighing up the advantages and costs attached to any alternative, subject to external constraints, was predictable. It was to this type of problem that I turned my attention. - 2 -Rather than deal with the whole range of peasant decision making I decided to confine my attention to decisions made about new agricultural practices. Thus my i n i t i a l substantive concern is with the c r i t e r i a by which Third World farmers decide to accept or reject innovations. This concern involves a movement away from broad general perspectives about the development process and attempts rather to solve a particular problem in addition to simply identifying a problem area. To remain with a broad, .general perspective places the anth-ropologist at a disadvantage with regard to the application of his creed. In the practical world of development planning and the anth-ropologist's contribution to i t , Moerman has stated \"Anthropological notions of cultural systems and of institutional interconnections are so imprecise that he has no special competence, aside from immerse-ment in native l i f e , to make the predictions that his job description and self-definition often demand\" (Moerman 1968: 84). To alter this portrayal of the applied anthropologist re-quires an increasing concern with particular \"tool k i t s \" that can be successfully applied to specific problems. And i t is this concern with tool kits that brings me to the second major stimulus in my research — methodology in economic anthropology. - 3 -Economic Anthropology: Controversies in Methodology Anthropologists have long argued among themselves and with economists regarding the range of applicability of economic theory, so much so that there has been a polarization into two distinct camps — the formalists and the substantivists. Polanyi's division of the term \"economic\" into two distinct meanings is basic to the whole controversy that has divided scholars in economic anthropology for over thirty years. Polanyi discriminates the substantive from the formal mean-ing of economic. By the former he refers to man's interchange with his natural and social environment \"in so far as this results in sup-plying him with the means of material want satisfaction\" (Polanyi 1957: 243). The latter derives from the logical character of the means-ends relationship, \" i t refers to a situation of choice, namely, that between the different uses of means induced by an insufficiency of those means\" (1957: 243). This categorization covers what is known as formal economics, and restricts one's range of study to choice in-duced by scarcity situations. The implication is that the communi-ties anthropologists study do not always have the same scarcity pro-blems possessed by industrial economies. Dalton goes on from Polanyi's distinction and argues that the substantive aspect of economics is the sphere of enquiry for eco-nomic anthropologists (Dalton 1961, 1969). - 4 -He opposes the use of formal economic theory on the grounds that i t grew out of and was made to account for situations of indus-t r i a l economies and is thereby inapplicable to any other situation. \"The differences between primitive economic organisations (i.e. where market transactions and produce are absent or present only in petty amounts) and our own are so great that a special set of concepts, leading ideas and terms are neces-sary to analyze these subsistence economies... the concepts of conventional economics relating to economic organisations are not f r u i t f u l l y applicable outside of market systems\" (1969: 65). This i s the key to Dalton's whole argument; yet i t remains unclear as to what he means by market transactions. It is also not clear whether the distinction he is making is between the form of mar-ket transactions (i.e. physical market place) or the structure of exchange. If he refers to the former, he has missed the point that i t is not necessary to have a physical structure to indulge in market transactions. If he is referring to the latter, then he is clearly wrong as there i s no evidence to indicate that exchange i n market and non-market economies differs i n terms of structure. In both setups the actors involved in exchange are balancing payoffs against costs in terms of the particular u t i l i t y dimensions paramount at the time of the exchange. Dalton seems to be confusing substantive situations with analytical principles. Robbins devastates the materialist definition of economics which was later employed by Polanyi and Dalton and argues that whether or not a good or service is material has nothing to do with whether i t is economic (1935). - 5 -He defines economics as the science which studies human be-havior as a relationship between ends and scarce means which have a l -ternative uses. This definition has frequently confused anthropolo-gists of the substantivist persuasion as they feel i t is applicable only to western industrial society. However Robbins is quite clear about the range of applicability of his definition: \"We do not say that the production of potatoes is economic activity and the production of philosophy is not. We say rather, that insofar as either kind of activity involves the relinquishment of other desired alternatives, i t has i t s economic aspect.\" (Robbins 1935: 15) This consideration and weighing up of alternatives further implies that economic anthropology is about choice. In fact Firth makes the problem of choice a major point of emphasis in the study of social organisation. He states: \"The working arrangements by which a society is kept in being, the ways in which relations between groups are made operative and become effective rest upon individual choice and decisions.\" (Firth 1964a: 46) Implicit in the writings of the early economic anthropolo-gists such as Firth (1939), G-oodfellow (1939) and Herskovits (1940, 1952) is the notion of rational choice b u i l t around a principle known as the \"calculus of maximisation\" (LeClair and Schneider 1968: 6). This assumes that people make decisions and choices in a rational manner, between known alternatives. It also assumes that the choice is made according to determinable principles. From this one can infer that any human activity directed to choosing between alternatives has the component of economizing. Economizing is the - 6 -allocation of scarce resources among alternative ends, which means that an individual can economise with respect to any scarce resource. There are a number of conclusions which can be drawn from the discussion so far. F i r s t of a l l the logic of the premises of for-mal economics is that economic theory i s not so much concerned with analytical principles for western, industrialized society but with a way of looking at behavior. Secondly the assumptions used by an economist, about re-sources, wants and choices, in his formal analysis are sufficiently abstract so as to be applicable to any human society. For instance the theory of maximization, which i s at the heart of neoclassical economic theory says nothing about what is maximized. It i s generally assumed that profit is being maximized, but this re-presents one application of the theory and not maximization theory i t s e l f . One could say that i t is u t i l i t y which i s being maximized --the value which an individual places on certain ends. The end can vary -- maximize profit, r i t u a l solidarity or status, and decisions taken with respect to that end are constrained by the u t i l i t i e s attached to competing ends, but maximization theory is held to be universally applicable (Robbins 1935; Burling R. 1962). The third important inference to make is that Polanyi's implication that some primitive societies do not have the same scarcity problems as possessed by industrial economies is clearly wrong. I - 7 -stated earlier that one can economize with respect to any scarce re-source, and the process of economizing is the same whether i t is ap-plied to material commodities or intangible goods such as prestige. This implies that the structural implications of scarcity are a major focus for enquiry in economic anthropology. Systematic rebuttals of the substantivist'.s position have been provided by Burling (1962), LeClair (1962) and Cook (1966). , They show the logical and empirical deficiencies in the substantivists 1 argument and Cook points out very succinctly a gross misunderstanding of the nature of formal economics by both Polanyi and Dalton. Dalton's argument that formal economics is based on 19th century -industrialization in Britain merely alludes to a particular system of p o l i t i c a l economy and the 19th century European market sys-tem, and totally ignores the subsequent refinements, modifications and new additions to economic science. Dalton seems not to acknowledge the fact that though one may use concepts from other disciplines this does not necessarily entail the use of the same operational definitions for such concepts. The substantive argument tends to confuse analytic principles with the specific contexts to which they are applied. Dalton's con-cern with distinguishing separate economic contexts with separate ana-l y t i c principles misses the point of sci e n t i f i c enquiry. I would submit that our concern should primarily be with principles, and that we use different substantive contexts to test the range of applicability of the principles. - 8 -Within the wider problem area of development I have elected to concentrate on the choice processes that underlie the acceptance and rejection of innovations by peasant producers. This is an i n i t i a l substantive context from which I intend to generate a set of more gen-eral statements. The need for wider generality can be met by borrowing con-cepts from economics, as considerations of choice behavior are funda-mental to modern economic theory. I reject the substantivist orientation as i t s concern with unique contexts prevents generalisability. The concern with choice behavior is not new in anthropology (see below), though i t lacks ex-planatory power on a more general cross cultural level. The synthesis of the two stimuli to my research is in terms of development interests bringing me to the subject area of choice behavior, while methodology suggests that I borrow analytic tools from economics in order to study i t . The Anthropology of Choice If economic anthropology is concerned with problems of scar-city and choice then i t would seem necessary to construct models of choice behavior which can predict the type of decisions made as con-straints of scarcity vary. - 9 -Raymond Firth emphasised as long ago as 1939 that the study of decision making should be a major focus for enquiry, but i t is only in recent years that systematic attempts have been made to explicate the process of decision making and i t s structural implications. The concept of choice is fundamental to Barth's notion of a transactional model (Barth 1966). He maintains that in any social re-lationship we are involved in a flow and counterflow of prestations, of appropriate and valued goods and services. He uses the notion of flow and counterflow of prestations to determine the extent to which particular statuses can interact. The term 'transaction' denotes those sequences of interaction which are systematically governed by reciprocal behavior. Barth is concerned with the incentives and constraints on the choice behavior of an individual actor in terms of strategies opted for and the calculations of gains and losses. This calculation can be along both social and material dimensions. By considering di f -ferent dimensions of value Barth points out how a l l parties can gain from a transaction. To illustrate his arguments Barth analysed the social organ-ization of a Norwegian seine boat to show how the traditional and ex-pected relationship between statuses -- captain, netboss and crew --was changed as a function of the transactional nature of social rela-tions on board ship (Barth 1966). - 10 -On the seiner he noted that the traditional subordination of crew to captain was relaxed. The crew spent a great deal of time on the bridge and also monitored the equipment that was used to locate fish and navigate. This went against maritime convention which speci-fied the bridge as the exclusive domain of the captain, unless a crew-man was on watch. In his analysis Barth focussed on the goals of the captain and how these were effected by the use of his authority. His primary goal was to maximize fish intake and to this end he had invested in a new type of seine net. This new technology, which enabled the captain to better achieve his goal, required a committed and well trained crew as the co-ordination of activity was greater than that required with the old types of driftnet. Mistakes or slowness on the part of the crew at crucial periods of the workcycle could lead to loss of catch or damage to gear. As a result of these new expectations of crewmen, the captain and crew entered into a kind of bargaining relationship. As part of the transaction, the crew were allowed priveleges which impinged upon the captain's traditional authority in return for the extra performance required with the new technology. The captain could maintain his traditional authority, but only at the cost of his primary goal -- maximization of catch. The change in technology and the different expectations of labour from the crewmen which resulted, permitted a set of transactions which called - 11 -for a partial relinquishment of authority by the captain and increased performance on the part of the crew. Barth maintains that the necessity of choosing between alter-natives (in this case authority and size of catch) can operate to change values as new decisions become systematised. He also emphasises that this new consensus of values has to adjust to further experiences and events. Blau takes as his starting point the notion of exchange as a means of characterising social interactions. He defines social exchange as \"voluntary actions of individuals that are motivated by the returns they are expected to bring, and typically do in fact bring, from others\" (Blau 1964: 91). He specifies that exchange takes place when specific resources are differentially distributed, and interaction permits the flow of these resources between actors. He illustrates this with an analysis of social processes in a work group (1964). The exchange commodities considered in his analysis are ad-vice and social status which he relates in the following manner: asking for advice raises the status of the advisor, giving advice lo-wers the status of the advisee. Blau extends this assumption in terms of a consideration that i f advice on a work problem is highly valued then other favours are insufficient to discharge one's obligations for i t . I t then becomes necessary to reciprocate for advice with respect and compliance, in other words accept a lower status vis-a-vis the ad-visor. The resources diff e r e n t i a l l y distributed in this example are status, as a function of compliance, and advice. - 12 -Recalling Barth's remarks about a l l parties profiting from a transaction, the exchange in the work group results in a mutual pro-f i t for the participants. If one labels Person A as the advisee and Person B as the advisor in the workshop situation then Person A profits from the advice that enables him to perform his job better, and Person B profits from the superior status he is accorded in exchange for his expert advice. Other authors have looked at choice behavior as fundamental to a general notion of exchange and have viewed the process of social interaction in cost-benefit terms. Homans in his essay \"Social Behavior as Exchange\" makes his major point with respect to considering aspects of social interaction from the point of view of calculation of \"costs\", \"rewards\" and \"pro-f i t s \" , and views particular decisions as the outcome of balances or imbalances in the calculations (Homans 1958). The generalisability of the notion of exchange is documented by Gouldner (1960). Drawing upon Malinowski and Mauss, Gouldner hypo-thesizes that a norm of reciprocity is universal -- a dimension to be found in a l l value systems. \"We owe others certain things because of what they have pre-viously done for us, because of the history of previous in-teraction we have had with them. I t is this kind of obli-gation which is entailed by the generalised norm of recipro-city \".(I960: 171). - 13 -Gouldner emphasizes the important distinction between comple-mentarity and reciprocity. He notes that the former connotes that one's rights are another's obligations, and vice versa. On the other hand, reciprocity denotes that each party has rights and duties. Gouldner argues that i f there were only rights on the one side and duties on the other, then there would be no exchange as there can only be stable patterns of reciprocity qua exchange insofar as each party has both rights and duties. The significance of reciprocity as exchange for role systems i s that i t tends to structure each role so as to include both rights and duties. This all-pervasive aspect of exchange is attested to by Levi-Strauss: \"Goods are not only economic commodities, but vehicles and instruments for rea l i t i e s of another order, such as power, influence, sympathy, status and emotion and the s k i l l f u l game of exchange consists in a complex totality of conscious or unconscious manoevres in order to gain security and to guard oneself against risks brought about by alliances and by rivalries.\"(1967: 54) A l l the authors I have mentioned implicitly view social be-havior as an \"economy\" with calculations of rewards, costs and optimum points as functional to choice and decision making. Blau (1964) and Barth (1966) go further in that their view of exchange suggests that one can view social structure not as a given, but as the outcome of choice behavior. From this viewpoint i t would seem imperative to construct models of how people make decisions under varying circumstances, as this endeavour would then explicate what may be a basic building block for the study of social behavior. - 14 -Though the discussion has moved from choice to exchange i t rests on the same assumption of calculus of maximization and ration-a l i t y in choice. It also assumes that the tools of economics -- decision theory and maximizing models -- can also be the tools of the anthropologist. These are the \"in principle\" p o s s i b i l i t i e s , and annumber of anthropologists have endorsed the idea of using formal decision theory to solve anthropological problems (Buchler and Nutini 1969: Davenport 1960). However a balance between anthropological data and formal models has to be kept clear in that more s t r i c t l y anthropological considerations serve to identify the constraints and parameters within which decision models can be applied. This means that ethnographic description of particular value systems and cultures permits one to designate the culturally perceived alternatives that are appropriate to particular decision situations, and to demarcate the principles which are determinate (or seem to be) for choosing between the alternatives. And i t is in this way that the anthropologist, with his tra-ditional tool kit of analysing value systems and culturally appropriate ways of doing things, can make an enormous contribution to the cross-cultural study of decision making. - 15 -The basic assumption that links formal models with anthropo-logy is that men act in order to get things they value, through the rational use of the rules and resources of their physical and cultural environments. The rules, rewards and man's perception of both can change and vary through time and space, but I am submitting that the process of rational calculation remains constant. However,methodological pronouncements and statements of what should be done can get one only so far. While I regard methodology as a necessary philosophical hurdle, i t is of l i t t l e use when not applied. It is obvious where my methodological preferences l i e and i t should be evident that I am taking part in a growing formal tradition in economic anthropology. I t is now necessary to get on with the job. The methodological controversy in economic anthropology has many \"in principle\" advocates on both sides, but very few practitioners. Perhaps the debate can be put into a perspective by viewing economic theory as being made up of axioms which have standard interpretations in a market context, but consists of empirically uninterpreted variables in the non-market context. With this view, I agree with Orans (1968) that the real ques-tion is whether or not empirical interpretations can be given to the axioms of economic theory so that they cover a domain wider than that of the market. - 1 6 -Th-is means that instead of talking about domains of relevance for one approach as against another, one must get on with the job, con-struct models and test them to see i f they work. Then one can see which set of methodological assumptions one can have confidence in. Progression of Thesis In Chapter 2, in keeping with my i n i t i a l concern with under-development in the Third World, I pose a problem of farmers making de-cisions about innovations and offer a solution in terms of a model of risk taking that predicts the sort of decisions particular types of farmers w i l l make under specified conditions. From this s t r i c t l y substantive orientation I then move to a higher level of generality in that I generalise from my i n i t i a l propo-sitions to a series of statements about individuals taking risks, i.e. making decisions under conditions of uncertainty and prospective loss. With this more theoretical level of operation, the testing procedure I use to assess the validity of my statements includes both laboratory and f i e l d settings, with supplementary evidence from secon-dary sources used by way of corroboration. In the laboratory I ran an experiment whereby I manipulated the conditions under which subjects took risks on gambles (Chapter 3); and in the f i e l d I observed the constraints upon decision strategies taken by gillnet fishermen on the west coast of British Columbia (Chap-ter 4) . - 17 -I used evidence on the way peasant farmers took decisions with respect to innovations as a supplement to these tests (Chapter 5). In each case \"testing\" meant the observation of indicators of the concepts in the propositions and the assessment of the degree to which their observed occurrence corresponds to the theoretical predictions. However, for the tests to be at a l l valid, I have to show that the explicit scope conditions specified by my propositions are met in each test situation. That this is possible is a function of the level of abstrac-tion I have chosen to use. My set of theoretical propositions do not refer to farmers and agricultural practices or students and gambling decisions, but individuals taking risks under constraints of uncertain-ty, state of their resources and the incentive conditions (both positive and negative) that apply to the risk. This means that data on farmers and innovations, students and gambling, and fishermen deciding where to fish, can be used as test cases and are comparable, once i t has been shown that the scope condi-tions specific to the abstract formulation have been met. Thus in a rather clumsy manner I w i l l be attempting to adopt a procedure suggested by Homans when he states \"The current job of theory in small group research is to make the connection between experi-mental and r e a l - l i f e studies, to consolidate the propositions that empi-- 18 -r i c a l l y hold good in the two fields, and to show how these propositions might be derived from a s t i l l more general set.\" (1958: 606) The conclusion, as i s the function of a l l conclusions, is intended to pull the disparate threads of discourse together and to spell out the implications of the thesis for economic anthropology. - 19 -CHAPTER 2 THE PROBLEM AND ITS CONCEPTUALIZATION Introduction From the literature on Third World farmers one can establish a number of propositions which link the decisions made by a subsistence farmer about agricultural innovations to his state of resources, incentive conditions attached to an innovation, and information and u t i l i t i e s attached to the outcome \"increase in productivity.\" From tbese statements specific to peasant farmers, I generalize to three statements about individuals and risk taking. In the interests of parsimony i t would perhaps be better to start with the general statements and proceed immediately to testing procedures. However, given the lack of precedents for the type of methodology implemented here and the explicational nature of my research, I think i t is necessary to show the step-by-step progression from one level of generality to another. It is also important to i l l u s t r a t e the source from which my statements originate. Farmers and Innovations The i n i t i a l substantive problem I am dealing with is the explication of the factors that influence farmers 1 decisions on agricul-tural innovations. The conceptual approach I intend to use is that of decision theory, so both my substantive and conceptual concerns are within limited boundaries. - 20 -But even with such limited boundaries, i t is possible to have a multiplicity of variables that seem to be relevant. There are many things going on in the decision making process with many possible combinations of factors. At the outset some are regarded as absurd, others as possibilities which indicates that I have a number of pre-conceived notions of the sorts of things that are relevant. The variables I am assuming to be relevant to the problem can be categorised under three headings: (1) the state of nature that the individual decision maker is in, (2) the incentive conditions attached to the innovation, and (3) access to information. My choice of variables could be considered arbitrary, but in actual fact is not, as they are a l l implicit to decision theory con-siderations. My major justification is that I w i l l be testing for the relevance of each variable and am prepared to reject any proposition that cannot be verified in empirical data. In my i n i t i a l formulations I do not use explicit socio-cultural variables (though these are impli-c i t in the state of nature consideration) chiefly on methodological grounds. This is because I am attempting to ascertain just how much of decision-making can be accounted for in particular situations by using a limited number of variables. Before I can proffer specific propositions I have to define what I mean by a state of nature. F i r s t of a l l i t is a classification, a set of categories into which I can place decision-making individuals. - 21 -But this taxonomic exercise is not merely to garner different coloured butterflies to my bosom -- an exercise that Leach parodies in 'Rethink-ing Anthropology' (1961) -- but is a preliminary step in ordering the complex phenomena that affect decision-making into states of nature, the characteristics of which I can then make statements about. The variables I use to define states of nature are (1) wealth and (2) subjective u t i l i t y attached to an outcome. The u t i l i t y part of the state of nature is a numerical measure of the strength of a far-mer's preference for a particular outcome. The outcome in this parti-cular substantive context remains constant -- \"increase in productivity.\" The implications of the u t i l i t y consideration are that the individual farmer w i l l react (in terms of decision strategies used) according to the way he sees a particular situation. I t is reasonable to assume that as subjective u t i l i t i e s for the outcome 'increase in productivity' vary, so w i l l decision strategies on events directly associated with that outcome, (events such as the consideration of an innovation). I am interested in two implications of the wealth consider-ation: f i r s t l y , individual farmers have varying buying power and are diff e r e n t i a l l y able to absorb losses according to that wealth; second-ly, different wealth levels w i l l have differential access to information about innovations. Thus, the wealth part of the state of nature defines the resources that any individual can mobilise to effect a particular outcome. - 22 -If we posit three levels to each one of these variables, (wealth and subjective u t i l i t y for increasing productivity) -- low, medium and high -- there are nine possible states of nature, as below (Figure 1). Figure 1: States of Nature Subjective U t i l i t y Wealth L M H L 1 4 7 M 2 5 8 H 3 6 9 1 The reason for defining a state of nature in these terms is partly to get round the problem of ethnocentrism, and partly to reduce the phenomena I am interested in to manageable proportions. I t is an analytic device useful for cross-cultural comparison, because any par-ticular state of nature either does or does not occur. Where i t does my statements should hold, where i t does not my statements are empiri-cally irrelevant. I t would be natural to expect man as a culture-bearer to make decisions in light of his cultural values. I am assuming that the subjective u t i l i t i e s that any individual in any culture w i l l have -- with respect to increasing agricultural productivity -- w i l l be a direct function of a number of factors, of which cultural values is only one, though a major one. - 23 -I think i t well nigh impossible to estimate the nature of this function, but one does not have to 'observe' values in order to take them into account. In other words an individual's estimation of subjective u t i l i t i e s is his perception (as an individual culture bearer) of a particular outcome. That this perception w i l l vary with di f f e r -ential wealth stratification accounts for the wealth variable in the state of nature. The use of a state of nature prevents one from making gross statements about the significance of socio-cultural and economic fac-tors with reference to the decision-making process, as i t forces one to be specific about the relationship between parameters and situa-tions . In other words, this form of conceptualisation sets boundaries to the relevance of particular variables, because each state of nature has certain characteristics and conditions which implies that general-isations applicable to one state of nature are not necessarily appli-cable to any other state. Incentive conditions refer to the payoffs and costs attached to an innovation. This is calculated partly in terms of expected yields and capital outlays, but also considers the likelihood of the innova-tion either living up to i t s expectations or f a i l i n g . The latter con-sideration involves the calculation of a further cost -- a livelihood cost. - 24 -Access to information refers to the differential distribu-tion of information to particular categorised individuals. (This is discussed in detail later in the chapter.) LWHU State of Nature The following remarks are pertinent to state of nature No. 3. They refer to a set of conditions whereby a farmer has a low wealth ranking, yet attaches a high u t i l i t y to increasing his productivity and is generally analogous to subsistence agriculture. A major assumption I would make following Schultz (1964) is that peasant production operates at or near an optimum in terms of existing production factors and state of knowledge, in this state of nature. That a farmer has a high preference for increasing producti-vity implies that he w i l l have taken up a l l or most of the 'slack' in terms of existing production p o s s i b i l i t i e s . The low level of productivity that is characteristic of much peasant agriculture is often attributed to peasant inefficiency, and their refusal to adopt new technologies and practices to cultural re-sistance to change; whereas I would submit that production is highly efficient in terms of the given techniques and state of knowledge and the refusal to change a rational economic calculation. Generally, farmers have a given set of production factors which have been unchanged over generations, with the result that a form - 25 -of equilibrium has been reached with reference to maximum production from given areas of land with given factors of production. The implications of this assumption can be categorised in terms of two related properties: (1) one cannot assume that better allocation of the existing production factors would increase agricultural production; (2) farmers bound by traditional agriculture have usually exhausted a l l profitable opportunities to invest in the agricultural fac-tors at their disposal. The farmer in this state of nature i s in a situation where the risks from his occupation are high, but they are familiar and well known through past experience and can thus be expected. There is un-certainty with respect to these familiar risks in that the farmer can never be sure that he w i l l be able to produce enough for his needs. The penalties for not doing so include hunger and debt. Innovations involve new and unfamiliar risks, and i t is highly unlikely that a farmer w i l l decide to allocate from his existing stock of resources the necessary wherewithall to obtain an innovation unless; the conditions attached are highly favourable to him. Some Propositions The propositions that follow apply only to state of nature No. 3, where the farmer cannot afford to lose. - 26 -Proposition 1: Innovations with a high payoff factor which c a l l for a minimal diversion of stocks from existing resources w i l l tend to be adopted. The whole idea of high investment returns at low costs is an obvious one, but i t is rare that innovations presented to LWHU farmers have these conditions attached. In those instances where these condi-tions are present, innovations w i l l be accepted. Proposition 2: Innovations that offer a low return tend not to be adopted. Acquisition of an innovation w i l l require that a farmer takes from his existing stock of resources (capital, land, labour, time) that which i s necessary for him to obtain the use of a new production fac-tor. As his existing resources are barely sufficient for his present needs, i t is obvious that special incentives are required in order to persuade him to risk part of his existing stock of resources (i.e. de-plete his survival kit) for some new production factor. I t is necessary that inputs and factors of production have to change in order to get increased productivity, but they must be of such a nature that somehow overcomes the high-risk element involved in a farmer in this state of nature investing in new production factors. Farmers are not going to invest in new unproven factors un-less they see that the return relative to the cost is so much higher than their existing methods of production. They are unwilling to take risks for marginal benefits. - 27 -Proposition 3: Individual farmers seek to minimise the cost factor of an innovation, hence they are more concerned with cost reduction than high returns. The process of a decision has to balance between the possible gains and the possible losses. If a decision is made to adopt an in-novation and the innovation f a i l s , the attendant costs may be high. Not only has the farmer lost the resources invested in the innovation, but also he had to deplete his 'survival k i t ' with the result that the odds are now against his surviving. The costs have to be assessed relative to the goals and values of the decision-making individual, and in this particular state of na-ture I submit that the deterrent value of the costs for making a wrong decision exceeds the attraction value of the possible gains. We are dealing with a situation where the individual cannot afford to lose. This means that though the 'High investment returns' thesis is an obvious one, proposition 3 sets a limiting condition to i t . I t may prove to be a more f r u i t f u l line of enquiry (in terms of policy im-plementations) to find the conditions whereby the loss factor can be reduced. These propositions can be represented schematically as be-low (Figure 2): - 28 Figure 2: Propositions 1 - 3 Payoff Adopt Innovation Information Flow The concern with information flow and access to information circuits indicates that the decision-making process is one of a per-son struggling to integrate several sources of information into a single choice. One could go into the whole area of information-pro-cessing (with respect to the selection and weighting of different types of information), but as this is highly complex I prefer.';to regard i t as a 'black box' operation and trust that the state of nature and variables employed adequately reflect the internal process. If we visualise information as flowing in distinct circuits, i t is readily apparent that some people w i l l gain access to more c i r -cuits than others and thus possess more information. - 29 -But to make information a useful variable for analysis there must be some cr i t e r i a by which individuals can be differentiated in terms of the amount' of information they have access to. The c r i t e r i a I think most useful are wealth and education. I discussed wealth ear-l i e r i n terms of a varying a b i l i t y to absorb costs. With reference to information circuits I w i l l be hypothesising that people at different wealth levels have differential access to information. This assertion breaks down into two considerations: (1) The absolute cost of obtaining information about innovations places the poor farmer at a relative disadvantage with respect to the rich farmer. (2) The number of information circuits available and the amount of information-dispensed w i l l increase with wealth. I am assuming that wealthy farmers w i l l tend to associate more with other wealthy farmers and mutually reinforce one another with respect to information that is exchanged between them, but not exchanged with poor farmers. Also the information media associated with selling farm tech-nology and innovations w i l l concentrate i t s agents and advertising ef-fects on those individuals who can most readily afford them. The relationship between information and innovation is a d i -rectly proportional one. As rich farmers have greater access to information circuits they tend to innovate more because an obstacle to risk taking has been removed, as increased information reduces the un-certainty aspects attached to any innovation. - 30 -Education could be tied in in a variety of ways to the de-cision-making problem I am interested in; however, my concern with the education variable is the way i t operates in terms of allowing i n d i v i -duals different access to information circuits. I am hypothesising that as education increases so does the a b i l i t y to gain access to more information. Between wealth and education there are certainly some feed-back effects, as i t is reasonable to expect that higher forms of educa-tion are most likely to result from a higher wealth base. I am also assuming that a higher level of education w i l l result in a greater range of exposure to information about innovations and a greater awareness of changing market conditions and productive techniques. It could be argued that an increase in education would drive individuals away from the farms. This may be so, but the point of the model is that of those remaining (or i f those leaving were to remain), the individuals with higher education w i l l have access to more infor-mation circuits than those with less education. But there are some problems here. In addition to consider-ations of formal education, one has to consider the qualitative weight-ing that can be given to relevant experience of the situation in hand. The propositions dealing with information are: Proposition 4: Ab i l i t y to gain access to information increases with wealth. - 31 -Proposition 5: Access to information increases with education. Proposition 6: Innovative behavior increases with information. These propositions are concerned with comparisons between states of nature. Assuming that subjective u t i l i t y remains constant we are dealing with the LWHU, MWHU, and HWHU states of nature. These three propositions can be represented schematically as below (Figure 3): Figure 3: Propositions 4 - 6 No. of Information circuits H - a b — V * High Education^ Information -A _ - - c * ^HIGH / X M B H H - - - - - - - d x N E f Medium N H Wealth—^M >Education-)M - - - - - - - e ^ Information 0 A S \\ ,s w : i / 8 \\ v 1 i Education-^M - - - - - - - h ^ Low LOW-\"\"E \\^ ^ ^ ^ j f l n f ormation<^ The diagram has 9 different circuits ( a - i ) , which indicates a possibility of 9 different gradations of information. In terms of the propositions that relate wealth and education to information I can say that a > b > c ; d ^ e ^ f ; and g )> h > i . - 32 -There are a number of logically possible ways to obtain an ordering for the whole sequence, depending on the relative weightings given to wealth and education. By assuming that education is slightly more important than wealth in terms of access to information, one of the orderings I could get would be to have d / c and g > f and for the whole sequence a > b > d _ > c 7 e 7 £ 7 f > h> i . This ordering enables me to break the nine categories down into gross categories of high, medium and low that I am using for the other variables. The ordering I have is an assumed one and the sig-nificance of testing procedures w i l l be to see whether the assumptions for ordering are valid or not. Supplementary Propositions Although the propositions I have stated do not contain ex-p l i c i t reference to socio-cultural variables, this does not mean that I totally disregard these latter factors. The testing of the model w i l l function to establish the rele-vance or irrelevance of factors such as social mobility, social margi-nality and co-operative structure in terms of individual risk-taking behavior. Where my propositions are inadequate to explain the behavior they are \"supposed\" to I w i l l insert additional propositions to see i f a better f i t between model and data results. - 33 -Social Mobility As an index of what I shall c a l l 'social mobility' I intend to use a measure of the degree to which \"role slots\" are present in the society which are recruited by achievement c r i t e r i a , rather than by ascriptive c r i t e r i a . This rests on a conceptualization of roles as; slots or positions in a social structure for which individuals may compete. My concern with this index is with the rules of competition which define the flow or circulation of persons between positions. The index measures the degree to which 'achievement role slots' are present in a particular society. Thus a society with a high index of social mobility indicates that a greater number of people can compete for roles than in a society with a low social mobility index. The implications of social mobility for innovation are as follows. If .we visualize the competition for a particular role as requiring the attainment of characteristics (a n); i t follows that persons aspiring to that role have to acquire those characteristics. Societies low in social mobility are characterized by exclusion rules which prevent everyone in the population from acquiring the necessary prerequisites for particular roles. Societies high in social mobility do not have these barriers. The characteristics (a h) may consist of prestige symbols, expected behavior patterns and, for want of a better word, 'entry fees'; and can be acquired in a variety of ways. If prestige symbols have to be attained, this means that an individual has to involve himself in - 34 -assymetrical exchange relations with others, with reference to ideas and goods, in order that others accord him prestige. Classically this can be done by exchanging goods for enhanced prestige, and the process usually requires increased productivity on the material level. Other characteristics may require capital outlays for 'entry fees', and ex-pected behavior patterns may involve a considerable amount of expense. Particular role characteristics may be directly related to an income level or standard of prosperity. In a society where roles are open to competition (i.e. where there is high social mobility) people without this optimal level of wealth have: (a) the incentive and (b) the opportunity through in-creased productivity to attain these role characteristics. If one is a farmer one way of increasing productivity is to adopt agricultural innovations. I would predict that in a society where roles are more open to achievement, the incentives to increase productivity are present in a far greater degree than in a society whi allocates roles on a more ascriptive basis. In the latter societies, the incentives to aspire to certain roles, and hence by logical exten-sion to increase productivity, are constrained by the fact that only certain categories of individuals are eligible for particular roles. Proposition 7: The proportion of innovative individuals in a society increases with the society's index of social mobility. - 35 -Social Marginality Social marginality implies that individuals are at the mar-gin of a given culture or are in a social position i n which they strad-dle more than one culture. When this is accompanied by some form of sectarian identity, the marginal individual — not bound by wider group norms yet being able to draw upon the resources of his own immediate community — has a very wide range of activity. The lack of constraining ties and the possibility of mobili-sing more economic resources should place the marginal individual in a situation where the opportunities and resources for innovation are much better than for the non-marginal individual. The role of marginal individuals in diverse economic pursuits in many underdeveloped countries gives support to the notion that there is a relationship between innovation and social marginality (Hoselitz and Moore 1966) — the Chinese in South East Asia and Indians in East Africa are perhaps the most obvious examples. But marginality in i t s e l f is not a sufficient criterion, as i t could be argued that marginal individuals may be more prone than others to succumb to anomie. It is when a cohesive and marginal group operates within a wider cultural context can we then expect high i n -novative behavior. - 36 -Socially marginal trader groups are innovative because they are f;reer to try novel economic behavior and because they restructure the culture of deference and the rewards system rather than accomodate themselves to i t . I stated earlier that decisions were the outcome of balan-cing up the perceived costs and payoffs of particular outcomes (p. 27). Social marginality implies that there are certain costs to being mar-ginal i n terms of relationships with the wider society, which may mani-fest i t s e l f through social and legal discrimination. Thus marginality vis-a-vis the wider society has inherent costs with the result that day to day existence fraught with costs is a reality. Because of this factor I would submit that the cost ele-ment attached to any alternative is not as big a deterrent to innovating as i t would be for a non-marginal individual. Thus reinforced by his group and with a differential appraisal of costs, the socially margi-nal individual is more likely to innovate than the non-marginal i n d i v i -dual. These considerations suggest the following proposition: Proposition 8: The proportion of innovating persons in so-c i a l l y marginal business groups is higher than in non-marginal business groups. Co-operative Structure A supplement to proposition 3 (Individual farmers seek to minimise the cost factor of an innovation)is the notion that forms of - 37 -productive organisation which spread the risk and cost factors over a collectivity greater than a single farmer w i l l result i n more innova-tions being adopted. This is by merit of prevailing conditions of col-lective insurance. Recent work carried out in experimental psychology indicates a definite tendency for individuals to take more risky decisions when they are in a group context as against their acting alone. This phenomenon has been labelled as the \"risky s h i f t \" concept (Kogan and Wallich 1967) and suggests the following proposition. Proposition 9: Individuals making decisions in the context of a co-operating group of individuals w i l l adopt more innovations than an individual decision-maker acting alone. Whither Next? It would appear that that would be as far as I could take my statements about farmers and the sorts of risks they w i l l tolerate with reference to new production factors. However, though I could regard this as a substantive limita-tion on my endeavours, conceptually i t is possible to go much further. Desire for more generality leads one from the immediate sub-stantive concern for peasant producers deciding on innovations to a higher level of generality in terms of a theory of risk-taking, whereby data about particular sets of individuals constitute separate empirical tests of the theory. - 38 -Cancian in a somewhat confusing yet stimulating article makes similar types of suggestions (Cancian 1967). He develops a theory of s t r a t i f i c a t i o n and risk taking and tests i t oh agricultural innovation, though he states that i t could be applied to any situation involving s t r a t i f i c a t i o n and risk taking. His i n i t i a l theory which predicts that a ranking based on the possession of a valued resource has a negative relationship to risk taking, is modified to develop a predicted curvilinear relationship between resource and risk. In the application of the theory to agriculture he takes wealth as the valued resource and the early adoption of a new agricultural prac-tice as a risk. His i n i t i a l assumption is that \"the better your finan-c i a l position, the more you have to lose and the less you have to gain from taking chances; therefore, insofar as adopting an innovation is risky, the richer you are, the less likely you are to adopt.\" (Cancian 1967: 913). He defines risk as \"a characteristic of situations of ex-change in which the rate of return on investment of resources is un-certain; the greater the uncertainty, the greater the risk.\" (1967: 913) We do not know what specific \"characteristic\" Cancian refers to, but his use of the term is synonymous with the presence or absence of uncertainty. He does not consider as important the incentive conditions attached to the innovation (yield vis-a-vis cost) which means he does not differen-tiate between expensive innovations with long term yields and cheap innovations with immediate results. - 39 -It is evident that considerations of relative cost and yield would influence a farmer's decision whether to adopt an innovation, which implies that the indicators of innovation that Cancian uses in his comparative study are not in fact very good ones. From Marsh and Coleman's study (1955) he takes the rate at which Kentucky farmers adopted bluestone lime, and uses Gross's (1942) study of Iowa farmers adopting hybrid corn seed for comparative purposes. The lime was an expensive f e r t i l i z e r with results expected after a three-year period of use, while hybrid corn seed was an innovation with immediate results promoted by the Department of Agriculture as part of the post depression agricultural effort. The adoption rates of these are not s t r i c t l y comparable unless some method of standardiza-tion or control for variation of relative costs and yields is used. Cancian does not use any controls either for these two instances or for the five other case studies that he uses. So Cancian does not have an accurate assessment of risk tak-ing behaviour i n the different farming communities he was concerned with, and in fact his theory of risk taking i s a misnomer, as his theory is about information diffusion and is not directly about risk taking. His i n i t i a l theory predicts an inverse relationship between wealth and early adoption of innovations and he labels this the \" i n -hibiting effect\" of wealth. Where the conditions of his theory are not applicable (per-fect d i v i s i b i l i t y of risks; knowledge equally spread over a l l wealth ranks) then we have what he calls the \" f a c i l i t a t i n g \" effect of wealth; - 40 -a positive relationship between wealth and early adoption. He then develops a \"curvilinear effect\" in order to explain conservatism on the part of the middle wealth rank in terms of early adoption. He offers a number of rationales for this curvilinear effect but by way of under-cutting his own efforts states, \"None of the arguments offered con-stitutes a satisfactory•theory explaining the curvilinear effect\" (1967: 916) and \" i t s (curvilinear effect) theoretical status remains far from satisfactory because two distinct and inconsistent arguments for the effect were made during the development of the modified theory.\" (1967: 925) The point to emphasize in Cancian's concern for different types of \"effects\" is that some of the conditions of his original theory do not apply when he refers to f a c i l i t a t i n g and curvilinear effects, which means that he is dealing with discrete sets of scope conditions. (Scope refers to the set of phenomena a particular statement has ref-erence to. To test the validity of a statement in a scope other than that i n i t i a l l y ; specified for the statement is not to offer a test at all . ) This further implies that his statements with respect to one set of scope conditions are not applicable to any of the other discrete sets. Cancian seems to ignore this aspect of comparative methodology and as a result one may doubt the ..value of his compari-sons . - 41 -In operational terms Cancian visualizes the adoption process as having at least two stages: \"Stage 1 in which inclination to risk is important, and stage 2 in which inclination to risk is substantially less important\" (1967: 917). Suddenly, we are introduced to \"inclination to risk\" as def-ini t i v e of adoption stages. Cancian does not offer a measure for this or even a proposition relating i t to either wealth or adoption. He states that risk is high in stage 1 and low i n stage 2, on the basis of there being more information available in stage 2. But this cate-gorization of risk into high and low, as a function of information, has very l i t t l e to do with a farmer's actual decision to adopt. An expensive innovation is s t i l l a financial risk i n stage 2, whereas subsidized agricultural practices with high, immediate yields would be low risks in stage 1. Cancian does not consider investment factors as important in his calculation of adoption rates and this is a major shortcoming in his conceptualisation of risk taking. His comparative analysis supports the curvilinear effect of middle class conservatism, and he argues that there are instances where the inhibiting and f a c i l i t a t i n g effect are operative. However, due to confusion of scope conditions and lack of standardization of measures, too much importance should not be attached to the results. However, I found his suggestions for further research ex-tremely interesting, mainly because they endorsed the conceptual approach attempted here. - 42 -It should be pointed out that his is presented as an introduc-tory exercise, a working paper. But, be this as i t may, I gained a great deal of encouragement from his suggestions that as part of further re-search \"direct tests may be possible i n laboratory situations\" (1967: 926) and that \"theories of risk taking may be more important than theories of information diffusion for the study of the process of adoption of new agricultural practices.\" (1967: 927) I found his conceptualization of risk unsatisfactory and his lack of concern for incentive conditions an oversight. His concern with the implications of information diffusion had already been incor-porated into my model, while his consideration that future research lay with risk taking theory was fundamental to my concern with states of nature and u t i l i t i e s . To i l l u s t r a t e the relative merits of my model i t is necessary to further explicate i t s basic building blocks, and the f i r s t is the conceptualization of risk. If we were to think about risk i t would likely conjure up images of uncertainty of achieving desirable goals and of the penalties that may ensue from the failure to attain the desired goals. In fact I use these two aspects — lack of certainty and prospect of loss — as definitive of risk-taking. If we look at the decisions common to everyday l i f e , the per-vasiveness of the risk-taking concept becomes apparent. The choice of a career or of a marriage partner is fraught with risk: the decision to buy a house, to invest in bonds, to drop out of society — as long - 43 -as they contain a t t r i b u t e s of uncertainty and p o t e n t i a l loss — can be considered as r i s k s . To talk about r i s k - t a k i n g , then, i s to r e f e r to behavior i n s i t u a t i o n s where there i s a desirable goal and a lack of certainty that i t can be attained, with attendant p o s s i b i l i t i e s of l o s s . Risk taking i s thus a s p e c i a l category of decision-making, and to explicate the factors involved i n the proeess I intend to use some of the tools developed by decision theorists (Edwards 1954; Edwards, Lindman and P h i l l i p s 1965; Von Neumann and Morgenstern 1953). Decision Theory The class of models which most decision t h e o r i s t s have been concerned with have been t o t a l l y devoted to choice outcomes with r e f e r -ence to what i s deemed to be r a t i o n a l c r i t e r i a . An i n d i v i d u a l i n a decision s i t u a t i o n evaluates the probabi-l i t i e s and payoffs attached to a p a r t i c u l a r outcome and then r a t i o n a l l y s e l e c t s the a l t e r n a t i v e that maximises something of value to him. The value attached to any decision i s a r r i v e d at by taking the values of the possible outcomes, mu l t i p l y i n g them by t h e i r respec-t i v e p r o b a b i l i t i e s of occurrence and then summing up these products. For example i n the case of an i n d i v i d u a l having to^decide whether to invest c a p i t a l i n a r i s k y enterprise, consider the values attached to two outcomes — the value a t t r i b u t e d to the enterprise i f the r i s k pays o f f , and the value a t t r i b u t e d to the enterprise i f i t does not. (This l a t t e r value w i l l be negative.) - 44 -If we denote the f i r s t value by U g and the latter value by and let P represent the likelihood of the risk paying off and (1-P) be the probability of the risk not paying off, then the value of the enterprise (V) w i l l be: V = U .P + U f.(l-P). V can be differentially calculated by a series of models ac-cording to whether the values of the outcome (i.e. the u t i l i t i e s ) and the probabilities attached to the outcome are subjective or objective. (An objective view states that there is only one unique real number that can represent a particular outcome or probability irrespec-tive of the variety of decision-making individuals. The subjective view represents a degree of belief on the part of the individual decision-maker and is not uniquely determined.) A l l the models used by decision theorists have i n common the notion that the u t i l i t i e s and probabilities w i l l be considered by the decision-maker in adherence to specific logical rules that define rationality. This is why descriptive theories of decision-making have a l -ways been closely linked with normative models. When logical consider-ations indicated that people should behave according to certain prescribed principles, those principles were incorporated into descriptive models against which performance could be compared. While I think i t is useful to employ normative models as starting points for descriptive theory, preoccupation with normative aspects, seems in the case of decision theory, to have led to the neg-- 45 -lect of possible psychological aspects of the decision making process and has often'completely ignored the situational*context. • The result has-'been that when these models have been applied to gambling situations, only a moderate level of success in predicting choices between bets has been achieved — between 55% and 70% accuracy. (Kogan and Wallich 1967: 118). This is slightly better than predictions that would emanate from a randonu generator. The model builders i n attempting to erect a comprehensive mathematical theory of decision-making have been unable to account for the limited set of decisions which constitute the domain of human gamb-ling behavior. This indicates that further explication of the nature of decision-making is required before mathematical formalisation be-comes meaningful in an applied sense. The reason for this may be found i n the consideration that extraneous influences that bear upon the decision-maker, while not changing the calculated value of any alternative w i l l directly i n f l u -ence any decision with respect to that particular alternative. It is important to think of decision-making as multidimen-sional, and that the four general components considered in the models I have discussed are only part of the process. I would go as far as to say that any general decision-making model, of the.sort described, w i l l meet with no more than limited suc-cess i f i t ignores the variety of motivational and situational factors that enter into the decision making process. - 46 -My concern with states of nature is an attempt to remedy this defect. Similarly my interest i n variations of information is a selected situational factor. These broad considerations can be repre-sented diagramatically as. below, where D is the decision-making indivi-dual (Figure 4): Figure 4: General Model INCENTIVE CONDITIONS /S. A. U U„ P RISK The linkages between the different parts of the model w i l l be provided by propositions that relate a particular decision to var-iations in one or more of the other categories. To show the relevance of the decision theory format just described, consider the farmer deciding whether to take a risk on an - 47 -innovation. Recall that the value*of a decision is defined in terms of the value of the outcome \"risk pays off\" times the probability that i t w i l l pay off, plus the value of the outcome \"risk does not pay off\" times the likelihood that i t w i l l not. i. e i V = U .P + n . (l-P) s f If we consider an innovation as having incentives analogous to this equation, we get U = expected yield or payoff i f innovation works P = chance that innovation w i l l come up to expectations = calculated losses i f the innovation f a i l s l-P = chance that the innovation w i l l f a i l . Thus i f we consider an innovation as either succeeding or fa i l i n g , and a farmer as deciding to either accept or reject i t , we can .construct a 2x2 payoff matrix as below to account for every pos-sible situation (Figure 5): Figure 5: Payoff Matrix for Innovations INNOVATION Succeeds Fails Adopts Y-I -I-L Rejects 0 0 - 48 -In the 1,1 c e l l , i f the farmer adopts an innovation and i t succeeds he w i l l have a payoff of (Y-I), where Y i s the yield from the innovation and I is the i n i t i a l capital investment in the innovation. The 2,1 c e l l where the farmer rejects an innovation that would have worked, we have a statement of opportunity cost. The pay-off in this c e l l would be zero. In the 1,2 c e l l where a farmer adopts an innovation which f a i l s , the payoff is negative in that the farmer has lost his invest-ment capital (I) and i f unable to absorb this loss may suffer cost to his livelihood (-L). The 2,2 c e l l has a payoff value of zero. In decision-theory terms the value attached to the outcome \"innovation succeeds\" is the sum of payoffs in the f i r s t column, which is (Y-I). i.e. U = Y-I. s The value attached to the outcome \"Innovation f a i l s \" is the sum of payoffs i n the second column, i.e. = -I-L. We can consider U as gains that would accrue to the farmer s and as losses he may have to absorb. On a theoretical level what I intend to show is that there are certain combinations of expected gains and calculated losses which determine which decision strategies individuals w i l l employ in di f f e r -ent situations. - 49 -T h e o r e t i c a l P r o p o s i t i o n s D e c i s i o n s t r a t e g i e s can be c a t e g o r i s e d i n t o low and high r i s k s according to whether the d e c i s i o n maker chooses that outcome which maximises gain, minimises l o s s or a f f o r d s a long shot. High r i s k s t r a t e g i e s i n c l u d e maximisation of gain — choosing the a l t e r n a t i v e that would g a i n the most i f the outcome i s p o s i t i v e ( i . e . i f the r i s k pays o f f ) and long shots — choosing the a l t e r n a t i v e w i t h the lower p r o b a b i l i t y of success. Low r i s k s t r a t e g i e s seek to minimise l o s s and i n v o l v e choice of the a l t e r n a t i v e that would l o s e the l e a s t i f the outcome i s negative, i . e . the r i s k does not pay o f f . With these con s i d e r a t i o n s I have g e n e r a l i s e d from my i n t e r -mediate p r o p o s i t i o n s about farmers and innovations to three p r o p o s i t i o n s r e l a t i n g i n d i v i d u a l s to r i s k t a k i n g s t r a t e g i e s as a f u n c t i o n of v a r i a -t i o n i n wealth, u t i l i t i e s and i n f o r m a t i o n . (The supplementary p r p p o s i t i o n s are not p a r t of t h i s process, they remain as t h e i r name suggests — suplementary.) The f i r s t three p r o p o s i t i o n s a l l r e f e r to a preference shown by LWHU farmers f o r low r i s k s t r a t e g i e s ; \"minimal d i v e r s i o n of stocks from e x i s t i n g resources\" ( p r o p o s i t i o n 1); \"no r i s k of resources f o r marginal r e t u r n s \" ( p r o p o s i t i o n 2); and \"farmers seek to minimise the cost f a c t o r of an i n n o v a t i o n \" ( p r o p o s i t i o n 3). These three statements can be g e n e r a l i s e d to one p r o p o s i -t i o n : - 50 -1. Antecedent Conditions 1. Every decision maker can be classified on a two way matrix using wealth and subjective u t i l i t y for some outcome as para-meters . 2. Consider the LWHU state of nature. Proposition 1': Individuals w i l l show a preference for low risk strategies with respect to decisions involving the outcome to which HU is attached. The statements about wealth, information and innovative be-havior can be generalised to the following two propositions: 2. Antecedent Conditions 1. Same as 1 above. 2. Consider states of nature where U remains constantly high. Proposition 2': Individuals w i l l show an increased tolerance for high risk strategies, with respect to the outcome to which HU is attached, as wealth increases. (Wealth in this formulation i s defined with reference to the specific outcome i n question and could be considered as the re-sources possessed by an individual to attain a particular end.) 3. Antecedent Conditions 1. Same as 1 above. 2. Same as 2 above. - 51 -Proposition 3': Individuals w i l l show an increased tolerance for high risk strategies, with respect to the outcome to which HU is attached, as i n -formation increases. (An increase in information reduces uncertainty and removes a deterrent to high risk strategies.) Testing I have constructed a very simple model of risk taking and information processing. The f i r s t proposition refers to one particular state of nature, the second and third refer to changes across states of nature. Given a state of u t i l i t y and a wealth ranking, and control-ling for information, the types of risks an individual w i l l take are predictable. But to be at a l l useful the model has to be tested in empirical data. It i s possible to use three main sources to test my assertions about risk-taking — f i r s t in the laboratory, then in the f i e l d and f i -nally from secondary sources. In following this through I anticipate objections to the possibility of generalising from one situation to any other. But this objection can be rendered irrelevant by consider-ing precisely the methodological steps that should be taken. I should emphasise that my research is concerned with theo-re t i c a l considerations. - 52 -Theories consist of systems of assumptions and conditions which specify the scope or range of appropriate application of parti-cular statements. The assumptions assert relations among abstract concepts, and scope conditions specify the circumstances under which the relationship specified in the assumptions is expected to hold true. Thus the laboratory, f i e l d and library are particular re-search settings in which predictions from the theory may legitimately be subjected to empirical test. In each of these three cases theory testing means the obser-vation of indicators of the concepts in the propositions, and the mea-surement of the degree to which their relationships correspond to the theoretical predictions. In each setting the consequences of testing are the same — so long as the scope conditions have been met, either failure or suc-cess of the theory's predictions has direct implications for the assump-tions of the theory. The scope conditions that I am concerned with place an individual decision maker within parameters of wealth and subjective u t i l i t y for some outcome, information and incentive conditions for any risk. The model is attempting to predict the type of decision strate-gies employed for given values of the above parameters. It is important to note that i f there i s no theory, and i f there are therefore no explicit scope conditions, then no generalisa-tions of the results of a laboratory study, or of any other study, is permissable. - 53 -The reason for this is that without theory, there are no grounds for asserting that the results of any given empirical study were produced by a small set of specifiable factors. The results may as well have been the consequences of the interaction of an i n f i n i t e number of non-recurring circumstances and therefore would themselves be non-recurring. If any parallels are to be drawn between a number of empiri-cal studies then there must be a theory which shows that the studies have in common the factors specified by the scope conditions. So long as the explicit scope conditions are met, whether inside the laboratory or in the natural environment, the assumptions of the theory w i l l enable prediction of empirically testable conse-quences. My testing procedure begins in the laboratory, moving thence into the f i e l d and fi n a l l y uses secondary source material for corro-borative evidence of the f i r s t two testing procedures. There is a good strategic reason for starting one's set of tests from a labora-tory base. In any natural setting there are a large number of variables that one cannot control for, whereas in the laboratory one can select a small number of variables and systematically control their variation to see what happens. Thus i t is more likely that any theory w i l l be able to make more precise predictions for the laboratory setting. And this is where the value of the laboratory li e s — i t can provide the 'purest' possible test of the predictions of the theory, dealing with - 54 -a small set of variables and conditions at a time. Confirmatory re-sults from such a rigorous test can then increase one's confidence i n applying the theory to a natural setting. If the laboratory test is confirmatory of the theory's pre-dictions, and then the theory is applied to a natural setting (where the same scope conditions are met) any difference that occurs in the latter test may then be attributed to variables other than the ones controlled for in the experiment. I was (and s t i l l am) interested i n more s t r i c t l y socio-cultural variables with respect to decision-making, and formulated a series of propositions relating structural aspects of group membership such as social mobility, social marginality, and co-operative structure to risk-taking behavior. My reasons for excluding these additional variables from the immediate model included the fact that they made model building very d i f f i c u l t , and also that I was not too convinced of their rele-vance. However, the primary reason was because the testing procedure I intend to use may, in fact, show their relevance. By f i r s t testing my statements in the laboratory and then in the f i e l d , a lack of congruence between the two tests could indicate that there are other constraints operating which were not controlled in the laboratory. It is then imperative to bring i n additional propositions relating these additional constraints to risk taking be-havior in order to see i f there is a better f i t between data and model. - 55 -However, i f the results from the different tests agree, i t means that the constraints outlined in the supplementary propositions are not overly relevant to risk-taking in the situations considered. Summary In a somewhat pedantic manner, I have arrived at a set of general propositions via the intermediary considerations of farmers making decisions about agricultural innovations. This was necessary in order to show that my second order propositions did not appear out of thin air but were based i n i t i a l l y on specific substantive consider-ations . The advantage of operating at a higher level of generality is that the scope of my statements is not restricted to the domain of farmers and agricultural innovation, but is bounded by considerations of u t i l i t i e s , information, and resource gradients with reference to any risk. Verification of these general statements requires that they be operationally defined in terms of a number of substantive contexts and tested. Provided that the scope conditions specified by the second order of propositions can be shown to be the same in each substantive context, then the laboratory and the f i e l d are legitimate sources to test the same propositions. - 56 -CHAPTER 3 TEST CASE: THE EXPERIMENT Introduction Recall that in Chapter 2, I stated an intention to test my propositions, with respect to individuals and risk-taking, in labora-tory and fi e l d settings. The argument used was that once the scope conditions were adequately met then this testing procedure could be regarded as valid. An additional argument was made for a precedence of laboratory work in terms of affording the experimenter greater control over the variables. The point to be made is that each substantive situation (stu-dents playing bets; fishermen selecting strategies; farmers deciding whether to accept an innovation) provides a test for the theoretical propositions. These latter statements relate u t i l i t y preferences (for some outcome), to the calculated incentive conditions attached to an event (associated with that outcome) in terms of the decision strate-gies particular categorised individuals are predicted to take with res-pect to the event. In other words the scope conditions place an individual de-cision maker within parameters of wealth and subjective u t i l i t y for some outcome; present him with a decision task associated with the outcome; and subject him to variations in incentive conditions and in-formation attached to the event. - 57 -Thus, i f I can reproduce the conditions of my concern in a laboratory setting, and of the n variables, control (n-1) while mani-pulating one variable in a systematic way, then I w i l l be able to ob-serve what effects different combinations of the variables have on tak-ing risks. In this way, the laboratory becomes one test situation for general statements I am making about risk-taking behavior. Experimental Design In the following experimental design I am controlling for wealth, u t i l i t i e s , information and incentive conditions. The experi-mental design has to reproduce: (1) a decision task whereby the incentive conditions vary with res-pect to a given outcome; (2) a wealth gradient; (3) an information gradient; (4) another consideration is that a l l the propositions refer to states of nature where the u t i l i t y for the outcome remains constantly high (see Theoretical Propositions, Chapter 2). This means that I have to screen a l l subjects who participate in the experiment as being analogous to the HU category. - 58 -Decision Task and Incentive Conditions Recall that decision models comprised loss and gains fac-tors in terms of the equation V = U gxP s + U^xP^; and that I conceptua-lized V as the value attached to the decision, U xP as the u t i l i t y s s function of the risk paying off and U_xP as the u t i l i t y function of 1 JC s the risk not paying off. These considerations can be reproduced in a duplex gambling design developed by Slovic (1968). The gamble has four components, a chance of winning (P ), w an amount to win ($ ), a chance of losing (P,) and an amount to lose w I ($!>• This can be visualised in terms of two discs (Figure 6): Figure 6: Duplex Gamble The l e f t hand disc represents the chances of making gains (60% chance of winning $4), the right hand disc represents the odds on accruing losses (i.e. 40% chance of losing $6). These 4 components can be varied systematically, in terms of the gross categories of high, medium and low that I have been dealing with throughout, and this gives me 27 discrete gambles (Appendix 1). The decision task was for subjects (selected on the basis of HU, see below) to decide which of the 27 gambles they would like to - 59 -play for real stakes. They were asked to rate the bids on a -5 to +5 degree of attractiveness scale. They were told that the zero on the scale represented an indifference point where the gamble appeared nei-ther attractive nor unattractive. Then they were asked i f they would bet on the gamble. Those gambles that they decided they would like to play, they in fact played for real stakes (rather than for hypothetical payoffs and losses) after completing the decision task for a l l 27 gambles. The actual playing was conducted on a roulette wheel marked into percentage segments as below (Figure 7): Figure 7: Gambling Wheel A pointer which could be spun was attached to the centre of the wheel and to determine how a subject would fare on any particular bet two sp'ins at the wheel were necessary. For instance consider the bet: 60% chance of winning $4, 40% chance of losing $6. On the f i r s t spin i f the pointer comes to rest anywhere in the 10-60 area of the wheel then the subject wins $6, i f i t f a l l s outside this area (i.e. in the 70-100 area) then he does not win $6. (The divisions on the disc represent probability in spatial - 60 -terms.) On the second spin, i f the pointer finishes up in the 10-40 area then $6 is lost, i f i t f a l l s outside this area (i.e. in the 50-100 area) then no loss is incurred. So in actual playing i t is possible to win and lose, win and not lose, not win and lose, and not win and not lose depending on the vagaries of the spinning wheel. I should emphasize that the actual process of playing the wheel for real stakes was not important as far as the data collected from the experiment was concerned. The playing of bets was to simulate r e a l i s t i c conditions. The information I wanted lay in the questionnaire -- the re-cord of actual decision making -- as i t was here that the subject made his choices of what bets he would like to play on the wheel for real stakes. The whole questionnaire was completed before any bets were played. Whether or not subjects won.^ or lost was not important data for the purposes of this experiment. While completing the decision task (before playing), the sub-jects were instructed to regard each gamble as discrete in terms of the constraints that were imposed upon them (Appendix I I : Instructions to Subjects). These constraints consisted of: (a) variation in starting stake -- wealth factor; (b) variation in information about the gamble; (c) a $1 ante for every bet they decided to play. - 61 -Wealth Gradient The subjects were given a salary for participating in the ex-periment, but they had to use their salary to bet with. The variation in salary was $1 (low), $3 (medium) and $5 (high). The $1 ante for any bet played by the subjects was irrespec-tive of the subject's salary. The implications of this for the actual playing of bets is that the low wealth ($1) individual has to surrender his whole salary for any bet played, and thus is in a position where he cannot afford to lose; the medium and high wealth individuals can absorb the $1 ante and s t i l l have a chance of remaining in the game. Recall in my i n i t i a l discussion of peasant farmers, a major characteristic of a farmer being in the LWHU state was that he was in a position where he could not afford to lose. In order to calculate wealth or resource gradients for other substantive settings I use this characteristic as a baseline in order to estimate an appropriate gradient vis-a-vis context. In the context of hard up students gambling away their exper-imental salaries I estimated that a $1 salary f u l f i l l s the requirements of the baseline category. - 62 -Playing For Real Stakes It has been a feature of much of the experimental work on decision making that the experimenters have assumed that an individual 1 behavior w i l l be l i t t l e different under hypothetical conditions than in real l i f e situations where he has to accept the consequences of his actions. Slovic showed that when subjects played for hypothetical stakes they maximised gain and ignored potential, hypothetical losses. However when subjects knew they would actually play some of the gambles they were more cautious, preferring relatively higher P , lower P^ and lower $•]_. (Slovic 1969) His study also points to an area that has largely been neg-lected by theorists, namely the effect that the i n i t i a l financial pos-ition may have upon decision making. In this experiment I attempt to accomodate these two factors. F i r s t of a l l the bets were played for real money, and the subjects gam-bled with their own money -- (their salaries). Also, subjects were grouped into Low, Medium and High salary groups so that, a l l other things being equal, I could observe the dif -ferential effects on decision making caused by a change in financial standing. - 63 -Information The 27 gambles were the same for every subject, the only difference being that the information dispensed varied. The L.Inf subjects were given a vague verbal description of each bet in terms of the number of people who play i t and whether or not they were good gamblers (see Appendix II). The gradations in description were a function of the ex-pected value (E.V.) of the bet as below. (See Appendix I) E.V. Information Given y 2 Many good gamblers have played this bet. .5-2 Of the people who have tried this, some have been good gamblers. -.5 - .5 Some people have tried this bet. -2.5 - -.5 Of the people who have tried this, some have been poor gamblers. <-2.5 Only poor gamblers try this bet. The subjects were told that good gamblers were generally successful at the wheel while poor gamblers were generally unsuccessful. M.Inf subjects were given a verbal description but with more specifics, e.g.: Bet-A: Good Chance of High Winnings, Poor Chance of Moderate Losses 0 ? W ) . C$w) ( P ^ ($i) They were informed of the range of possi b i l i t i e s of the ver-bal description. - 64 -A good chance of winning/losing -~ 70% - 1007<> A f a i r chance of winning/losing = 40% - 707> A poor chance of winning/losing = 07o - 407> . and High winnings/losses; = $5 - $10 Moderate winnings/losses.- = $2 - $5 Low winnings/losses- = $0 - $2 The H.Inf subjects received the same information as the M.Inf subjects, but they could also have any 2 of the 4 bet components spe-cified as to the precise monies and probabilities. So the task for the subject was to make a decision whether to play a gamble or not in terms of the information he has, his i n i t i a l salary and the knowledge that the bet w i l l cost him $1 to play. It was emphasised to the subjects that they regard each bet as discrete in terms of the above constraints, while they completed the questionnaire. Subjective U t i l i t y The scope conditions specified that decision-makers have a high subjective u t i l i t y for the outcome associated with the event they are making decisions about. In the experiment the decision situation involves gambles for money, therefore the outcome to which HU must be attached is more money.* * Abbreviations are used throughout for experimental categories. - 65 -To screen my subjects (male students at U.B.C.) into an HU category, I obtained a l i s t of male students who were registered at the University Placement Office on the assumption that students so registered would have a high u t i l i t y for increasing money stocks as they were looking for part time jobs. (I took only those students who had not found jobs). The second part of the screening was to ask subjects to in-dicate their preference rating for increasing money stocks from the experiment. Only those who had a high preference rating (i.e. subjec-tive u t i l i t y ) took part in the experiment. Table I: Summary of Scope Conditions Wealth LW $1 MW $3 HW $5 Information LOW ; vague description : Med ium description.. High description and data Incentive Conditions P„$ - gains factor \" w 27 different P-^ $^ - loss factors combinations Investment Factor -$1 -$1 -$1 Subjective U t i l i t y A l l Subjects screened as HU for increase in money stocks - 66 -Data Collection The data from the experiment was put into a simple 2 way matrix (using wealth and information as parameters) for easy reference (Figure 8) . Figure 8: Data Matrix WEALTH L M H L 1 2 3 Information M 4 5 6 H 7 8 9 (It should be pointed out that subjective u t i l i t i e s are held constant at HU.) The bet components in cells 4, 5, 6 were converted into num-bers by taking an average of the range to which the verbal description applied. For instance: \"High Chance of Moderate Payoffs\" referred to a 707o - 1007o probability range, moderate payoffs indicate a range of $2-$5. This data was-recorded as 857» chance of $3.50. In cells 7, 8, 9 the same calculations were used, except for the two components for which information was sought. As the master sheet consisted entirely of averages of the range of pos s i b i l i t i e s , the bet components for these cells were the same as for cells 4, 5, 6. The type of data collected is recorded in Table II. - 67 -Table II\".'. Data From Experiment Ce l l Bets Rating of Components of Bets Information Sought No Played Bets (Averages of Range) by Subjects 1 X X 2 X X 3 X X 4 X X X 5 X X X 6 X X X 7 X X X X 8 X X X X 9 X X X X In order to categorise the bets into levels of riskiness I obtained a random sample of 20 from my ninety experimental subjects and asked each one to evaluate the gambles on the master sheet in terms of whether they thought the bet was slightly risky, f a i r l y risky, or very risky. From these 20 sets of evaluations, I assigned bets to one of these three categories in terms of 80% agreement among evaluating sub-jects, i.e. i f 807> of the subjects agreed to assign bet i to category j then bet i was assigned to category j . Had there been instances of less than 807° agreement on any bet then I would have assigned i t to one of the three categories in terms of the E.V. of the bet. But this did not arise. Thus my scaling of bets into categories of riskiness was determined empirically (Table I I I ) . An even more important implication is that the objective values given to the variations in incentive conditions corresponded closely with the subjective evaluations by the subjects. Had there been large diver-- 68 -gences, in subjective evaluation of the bet dimensions, between subjects then this would have shown i t s e l f i n the categorisation of bets into degrees of riskiness. This is significant in terms of comparative purposes, as the data on the fishermen (who constituted the f i e l d test) was on their subjective evaluation of alternatives. Table III; Bets Categorised by Riskiness Slightly Risky Fairly Risky Very Risky [SR] [ER] [VR]. Bet No: 7 1 2 10 3 5 12 4 6 14 11 8 16 15 9 18 17 13 23 20 19 26 21 22 24 25 27 Total 8 9 10 The data was f i r s t of a l l recorded per c e l l in terms of (a) bets categorised by subject (Table IV) and (b) frequencies of bets played (Table V). These raw data were then transferred into a 2 way matrix for each level of riskiness (Table VI). The data in these matrices were corrected for the unequal number of bets i n each risk category (viz: there were 8 SR bets, 9 FR bets, 10 VR bets). V I N T ^ . . A (N ) . . V I ' I J r b ' i j til ttl is the c e l l entry in i row and j column; B i s the number of bets taken in that c e l l ; (N T).. x (N . ).. is the number of possible bets I i ] rb i ] that could have been taken in that c e l l where N^ . i s the number of individuals in the c e l l and N , is the number of risky bets in that c e l l . rb This means that i n order to get the new c e l l entry, the number of bets taken i n each c e l l is divided by the number of possible bets that could have been taken. Multiplication by 100 converts the index into a percentage (Table VII). These stages of data transformation are represented in Tables IV-VII (Appendix III). Testing Recall that my theoretical propositions were as follows: 1'. Individuals w i l l show a preference for low risk strategies with respect to decisions involving the outcome to which HU is attached. Conditions: LW and HU. 2'. Individuals w i l l show an increased tolerance for high risk stra-tegies, with respect to the outcome to which HU is attached, as wealth increases. Conditions: HU. 3'. Individuals w i l l show an increased tolerance for high risk stra-tegies, with respect to the outcome to which HU is attached, as information increases. Conditions: HU. - 70 -The data matrix reproduced below is to f a c i l i t a t e reference to c e l l numbers (Figure 9): Figure 9: Data Matrix Information Proposition 1' WEALTH L M H L 1 2 3 M 4 5 6 H 7 8 9 Given low wealth, individuals w i l l show a preference for low risk strategies with respect to decisions involving the outcome to which HU is attached. In operational terms proposition 1' means that individuals in cells 1, 4, 7 w i l l take fewer risky bets than individuals i n cells 2, 5 and 8 respectively. The accompanying graphs plot SR bets, then a combination of FR and VR bets, over a wealth increment from low to medium; holding i n -formation constant. The graphs indicate that the ordinal relationships required to support proposition 1' do i n fact hold — but they also show much more. Graph 1 shows that there is a significant difference between L.Inf and M.Inf categories, but not such a large difference between M.Inf and H.Inf categories. From Graph 2 i t is reasonable to infer - 71 -GRAPH 1 L.W Pre ference for Low Risk Strategies: Proposit ion 1 Low Info. Med. Info. High Info. Wealth in S . 1 3 1 3 1 3 % of S.R. Bets 31.3 51.3 81.3 91.3 85 93.8 100' 80-60 % of SR. BETS TAKEN 4 0 20 H.INF. M. INF 1 L INF 3 WEALTH IN S GRAPH 2 L W P re fe rence for Low Risk S t r a teg ie s : Propos i t ion 1 Low Info. Med. nfo. High Info. Wealth in S 1 3 1 3 1 3 % of F.R.. & VR. Bets 0 3.3 178 25.6 25.6 38.4 100 80 -60-^ % o f and V R. ' BETS TAKEN 40' ..-H. INF 20-^ M. INF L.INF i i \\ 1 ~ 3 WEALTH IN % - 73 -that the three information levels show differences, but that within the L.Inf category the ordinal relationship may not indicate a signi-ficant difference. On the basis of these results one could accept the data as providing support for proposition 1. However, there i s a need for further testing in that one could argue that the differences between cells may not be significant. Recall that I visualised low risk stra-tegies as attempting to minimize the loss in a decision situation. I can get an additional test to proposition 1 by showing whe ther or not individuals in different cells made their decisions as a function of minimizing $-^ . The method is to calculate the maximum pos sible loss for each subject in the cells I am interested in (i.e. the bets played by each subject are summed as i f each one had a negative outcome). The next step is to calculate the range of maximum possible losses for each c e l l , and to do a chi square median test on the cells I want to compare, to ascertain whether the differences between the cells are s t a t i s t i c a l l y significant. The null hypothesis to be tested in each case is that the median amount of loss w i l l be the same for each c e l l . (Appendix IV) - 74 -Table VIII: Chi Square Median Test: Proposition 1' Chi Squ. Significance level Null Hypothesis Cell 1 cf Cell 2 7.46 p = .01 Reject Cell 4 cf Cell 5 5.00 p = .05 Reject Cell 7 cf Cell 8 0.20 Not Significant Not Reject The chi square of 7.46 for cells 1 and 2 implies that there are.very significant differences between the medians of both cell s , and that there is only a 1% possibility that these results are due to chance. The chi square for cells 4 and 5 similarly enables one to reject the null hypothesis at the 5°L level of significance. But the chi square of 0.20 for cells 7 and 8 means that the null hypothesis cannot be rejected, and implies that at high levels of information there is no significant s t a t i s t i c a l difference between the way LW and MW individuals seek to minimise their losses. The graphs provide data to support proposition 1' as they indicate that the ordinal relationships required of the data do in fact occur. However the chi square median test shows a facet concealed by the graphs, viz: that LW individuals (as distinct from MW individuals) do indicate a preference for low risk strategies except when informa-tion i s high. From both tests we can infer that proposition l 1 holds at low and medium levels of information but not under conditions of high information. (In terms of this last finding i t is significant that in - 75 -the f i e l d and from secondary sources I did not discover one empirical example of a LW individual having access to high information. This seems to indicate that while LW/HINF is a logical possibility and can be set up in an experiment i t is not to be found in the empirical world. This lack of empirical correspondence suggests that there are a number of intervening variables which makes LW incompatible with high informa-tion. However, what this result states is that i f LW individuals do have HINF (with respect to a particular outcome) then they w i l l take similar types of risks as do MW/HINF individuals. The important task is to delineate the intervening variables that prevent the LW/HINF category from occurring.) Proposition 2' \"Individuals w i l l show an increased tolerance for high risk strategies, with respect to the outcome to which HU is attached, as wealth increases.\" In the accompanying graphs I have plotted the proportion of bets taken as wealth increases (holding information constant) for each level of information. The f i r s t graph deals with SR bets, the second with a combination of FR and VR bets. In operational terms proposition 2' means that for each level of information, increased wealth should result in an increase in the number of bets taken for each risk category. - 76 -The f i r s t graph shows some interesting results. MW/L.Inf subjects take 20% more SR bets than LW/L.Inf subjects, but there is only a 6% difference between MW/L.Inf and HW/L.Inf. To a lesser extent this relationship holds for M.Inf and H.Inf subjects; i.e. that there is a bigger increase in risks taken as wealth moves from low to medium, then when wealth moves from medium to high. From the second graph at L.Inf, HW and MW subjects took the same proportion of risky bets, (which was higher than LW individuals), though the M.Inf and H.Inf levels showed a more nearly linear increase with respect to wealth. These results seem to indicate that a sort of \"step function\" is operative in that there is a much greater increment in risky bets taken between LW and MW individuals than there is between MW and HW individuals. The data i s sufficient to provide overall support for the ordinal relationship between wealth and risk taking, but supplemen-tary tests may indicate more about the precise nature of this relation-ship . By doing the obverse of the procedure used to test proposition 1', a chi square median test can be established. If we consider high risk strategies as those that seek to maximize gain ( i f the outcome is positive); then by summing the maximum possible payoff for each sub-ject one has an index of his tolerance for high risk strategies. The null hypothesis being tested i s that the median amount of gains w i l l be the same for each c e l l . GRAPH 3 R I SK TAKING w.r.t. W E A L T H Low Info. Med. Info. H igh Info. Wealth in 8 1 3 5 1 3 5 1 3 5 % of S.R. 93.8 Bets Taken 313 51.3 57.5 81.3 91.3 91.3 85 96.3 1 1 1 ! 1 1 3 5 WEALTH IN $ - 78 -GRAPH 4 RISK TAKING w.r.t. W E A L T H Low Info. Med. Info. High Info. Wealth in S 1 3 5 1 3 5 1 3 5 % of F.R. & VR. Bets 0 3.3 3.3 17.8 25.6 30 25.6 38.4 46 100i 90-80-70-60-% of FR.&VR. 50i BETS TAKEN A 0 \" 30' 20-10-\"MM*-. H.INF M.INF U N F WEALTH IN S 1 - 79 -Table IX: Chi Square Median Test; Proposition 2' Chi Squ. Significance level Null Hypothesis 1 cf 2 cf 3 15.2 p = .01 Reject 1 cf 2 3.4 close enough to P = .05 Rej ect 2 cf 3 1.8 not significant Not Reject 4 cf 5 cf 6 6.8 P = .05 Reject 4 cf 5 ' 5.0 P = .05 Rej ect 5 cf 6 0.1 not significant Not Rej ect 7 cf 8 cf 9 3.2 close to p = ..05 Rej ect 7 cf 8 0.1 not significant Not Reject 8 cf 9 1.8 not significant Not Reject The f i r s t six results confirm the idea of a step function. Overall, as wealth increases, individuals in the M.Inf and H.Inf categories do show a tolerance for higher risk strategies. But when this is broken down one finds for L.Inf a s i g n i f i -2 cant difference between cells 1 and 2 (X = 3.4), and a non-significant result between 2 and 3. Similarly for M.Inf, the difference between 4 and 5 is significant, while the difference between 5 and 6 is not. o The test for H.Inf level, as wealth increases, gives an X of 3.2 which is close enough to the .05 level of significance (3.8) to make the inference that, overall, subjects with H.Inf do tend to indicate differential tolerance levels for risk as wealth changes. - 80 -When this i s broken down, however, one finds that there are not s t a t i s t i c a l l y significant differences between cells 7 and 8, and 8 and 9, viz: H.Inf conditions produce non-significant differences i n tolerance for risk, as wealth increases. For each information gradient the graphs and chi-square test support the notion of tolerance for risk increasing with wealth. More specifically there is a cutting off point at MW where though the graphs show a difference, the chi square test indicates that i t is not s t a t i s t i c a l l y significant. One reason for this may be that the experi-mental distinction between MW ($3) and HW ($5) is inadequate. However the important inference to make is that there is a big jump in tolerance for high-risk strategies as wealth moves from low to medium,except under conditions of high information. Proposition 3' Individuals w i l l show an increased tolerance for higher risk strategies, with respect to the outcome to which HU is attached, as information increases. In operational terms this means that as information increases (holding wealth constant) more risky bets w i l l be taken for each wealth level. Graph number 6 indicates that these ordinal relationships hold for FR and VR bets. With respect to SR bets graph number 5 seems to indicate two things: (a) a threshold for risk at the MW level, (b) very l i t t l e difference between the MW and HW general categories. - 81 -GRAPH 5 RISK TAKING w.r.t. I NFORMAT ION Low Wealth Med Wealth High Wealth Info. L M H L M H L M H % of S.R. Bets Taken 31.3 81.3 85 51.3 91.3 93.8 57.5 91.3 96.3 - 82 -GRAPH 6 R ISK TAKING w.r.t. INFORMAT ION Low Wealth Med. Wealth High Wealth Info. L M H L M H L M H % of FR. & VR. Bets 0 178 25.6 3.3 25.6 38.4 3.3 30 46 100-1 90-80-70-INFORMATION (N. B.: Ord ina l Not Metr ic Interva ls ) - 83 -The important implication from the f i r s t graph is that L.Inf subjects took 50% fewer SR bets than M.Inf subjects for each wealth level. The supplementary test to assess the significance of these di f -ferences is exactly the same as for proposition 2', the only difference being that different cells are compared. The null hypothesis being tested is that the median amount of gains w i l l be the same for each c e l l . Table X: Chi Square Median Test: Proposition -3' Chi Squ. Significance level Null Hypothesis 1 cf 4 cf 7 13.5 .01 Reject 1 cf 4 16.1 .01 Reject 4 cf 7 1.8 ___ Not Reject 2 cf 5 cf 8 16.8 .01 Rej ect 2 cf 5 16.1 .01 Reject 5 cf 8 0.2 ___ Not Reject 3 cf 6 cf 9 16.8 .01 Reject 3 cf 6 9.7 .01 Reject 6 cf 9 1.8 _ Not Reject These results lend support to the threshold notion, as at each W level there i s not a s t a t i s t i c a l l y significant difference be-tween M.Inf and H.Inf. - 84 -There are two possible implications of this — one that the experimental distinction between M.Inf and H.Inf was not in fact a good dividing line, or that increments in information after the med-ium level are not significant in terms of increased tolerance for high-risk strategies. Of these alternative inferences, i t is reasonable to accept the second one, as the test of the f i r s t proposition indicated that H.Inf has an effect on LW individuals such that they do not take significantly different risks when compared to MW/H.Inf individuals (this was in terms of a preference for low risk strategies). Whereas M.Inf operates on LW individuals to produce a significantly different pattern of risk preferences when compared to LW/M.Inf individuals. Due to these different effects i t is reasonable to assume that High Infor-mation is different from and greater than Medium Information. The implication of this result is that information increments after the range of odds, costs and payoffs have been specified (M.Inf) do not significantly influence an individual's tolerance for high risk strategies. Risk was defined as having attributes of uncertainty and loss. Information acts to lower the uncertainty, but the experiment shows that there is an optimal point at which tolerance for high risk is not affected by additional increments in information. This cutting off point seems to be a specification of the range of odds and payoffs specific to a particular decision. In order to extract the maximum - 85 -information from the controlled situation extant in the experiment, I intend to test some of the f i r s t order propositions to see i f the ex-perimental data can provide additional information as to the domain of relevance for my assertions. Farmers, Risks and Experimental Data From Chapter 2 recall that I made a number of propositions relating to farmers and innovations, e.g.: 1. Innovations with a high payoff factor which c a l l for a minimal diversion of stocks from existing resources w i l l tend to be adopted. 2. Innovations that offer a low return tend not to be adopted. 3. Individual farmers seek to minimise the cost factor of an innova-tion. Conditions: LW/HU. By translating these propositions into more abstract statements relating to risk I should be able to use the experimental data to test the f i r s t series of statements. I conceptualised the decision to accept or reject an inno-vation in risk-taking terms, so the propositions above with regard to farmers and innovations can be translated into statements relating individuals and risk in the following way: A. Risks with a high payoff factor and a low cost factor w i l l tend to be taken. B. Risks with a low payoff w i l l tend not to be taken.* * Proposition 3 has been directly tested. - 86 -Conditions: LW/HU. A can be operationalised by considering gambles whose u t i l i t y functions for payoffs ($ .P ) have an E.V. } 2 and whose u t i l i t y func-w w tion for losses (^'^j) have an E.V. < 1. What I am testing is whether or not LW individuals take \"sure thing\" bets. The calculation of relative E.V.s limits consideration to bets 7, 12, 14, 16, 18 and 23. In the table below I have compared LW subjects at M. and H.Inf levels. (The data collected for low informa-tion levels does not permit a calculation of E.V.) Table XI: \"Sure Thing\" Bets Bet No. 7 12 14 16 18 23 Subjects Per Cell Cell 4 LW/M.Inf 9 10 10 8 10 7 10 Cell 7 LW/H.Inf 7 8 10 7 10 8 10 Of the 12 cases in the table, there are 5 cases of a l l sub-jects playing the bets, 1 case with 90%, 3 cases with 80% and 3 cases with 70%. These data are sufficient to provide support for proposition A. Proposition B: Risks with a low payoff w i l l tend not to be taken. Operationally this means that subjects i n LW categories w i l l not take bets 6, 8, 11, 13, 15, 17, 19, 21 and 27. The table below rep-resents data collected from LW/M.Inf and LW/H.Inf cells. (Information - 87 -supplied to L.Inf subjects did not specify whether payoffs were low, medium or high). Table XII: Low Payoff Bets No. of Subjects Playing Bet Bet No. LW M.Inf (4) (7) LW H.Inf 6 0 0 8 0 0 11 5 5 13 0 0 15 1 3 17 0 1 19 0 0 21 1 1 27 No. of Subjects Per Cell _0 10 _0 10 With the exception of bet 11 for c e l l 4 and 7, and bet 15 for c e l l 7, the data in the table supplies general support for the con-tention of proposition B. Bet 11 referred to a good chance of low payoffs, and a low chance of low losses. Bet 15 referred to a good chance of low payoffs, low chance of moderate losses. - 88 -In both cells 4 and 7 50% of the subjects took bet 11 which indicates a tolerance for low payoffs when (a)the loss factor i s negligible a r e ^-ow) a n ^ ^) when the chance of the low payoff is good (P i s high). w In c e l l 7 30% of the subjects took bet 15, indicating that some LW subjects, under conditions of H.Inf, would take a low payoff bet when (a) the chance of the low payoff is good (P high) and (b) w the loss factor has a low and a moderate P^. These cases indicate that there are instances where low pay-offs are attractive to LW subjects — and the attractiveness seems to be a function of the probabilities of winning and losing. These results led me to test for the overall relevance of the chance factor for LW subjects. I did this by estimating the relative weightings given by LW subjects to each of the 4 risk dimensions ($ , P , $.. , P,). This can w w l I be done by correlating each subject's evaluation .of bets with the le-vels of each risk dimension across the 27 gambles. I am assuming that the correlations between risk dimensions and evaluation of bets is directly proportional to the weights in a linear regression equation characteristic of each subject. The model underlying this analysis uses A(G) = u + w,$ + w P + wn$, + w.P, (Slovic 1968: 12) l w 2w 3 1 4 1 where A(G) is the evaluation of the gamble, u is the mean attractiveness of a set of gambles, and w_^ are weights reflecting the relative impor-tance of each dimension. - 89 -This model assumes additivity of components and also that the import of probabilities and payoffs is a linear function of their objective values (Slovic 1968: 13). The value of this format is that the model permits indepen-dent assessment of the relative influence of a l l 4 incentive dimensions, Table XIII: Relative Weighting of Risk Components Subject No. Correlations Between Risk Dimensions Cell 4 LW M.Inf Cell 7 LW H.Inf and Ratings of Attractiveness $ P $. P , yw w v l 1 1 0.55 0.66 -0.40 -0.66 2 0.38 0.68 -0.44 -0.68 3 0.28 0.74 -0.32 -0.74 4 0.32 0.77 -0.40 -0.78 5 0.40 0.65 -0.52 -0.65 6 0.60 0.49 -0.28 -0.49 7 0.16 0.47 -0.39 -0.44 8 0.42 0.66 -0.47 -0.66 9 0.47 0.68 -0.41 -0.68 10 0.42 0.76 -0.36 -0.76 1 0.32 0.66 -0.38 -0.64 2 0.34 0.76 -0.39 -0.75 3 0.58 0.50 -0.17 -0.50 4 0.54 0.66 -0.22 -0.65 5 0.45 0.72 -0.39 -0.72 6 0.28 0.79 -0.17 -0.77 7 0.33 0.74 -0.42 -0.74 8 0.38 0.75 -0.12 -0.73 9 0.28 0.87 -0.03 -0.87 10 0.35 0.77 -0.38 -0.77 In both cells 9 out of 10 subjects weighted the chance fac-tors (P ; P T ) higher than the monetary factors, from which one can i n -w 1 fer their relative importance to a LW operator. This finding helps to account for the exceptions to proposition B, in that the payoff factor - 90 -(low $ ) was not a deterrent when the P was high and P.. low. These w w i. re s u l t s are summarised i n Table XIV. Table XIV: Summary of Experimental Test Accept/Reject Conditions of A p p l i c a b i l i t y Exceptions Prop. 1' Accept L.Inf; M.Inf; H.Inf Prop. 2' Accept LW - MW; L.Inf; M.Inf; MW - HW; H.Inf Prop. 3' Accept L.Inf - M.Inf; LW; MW; HW M.Inf - H.Inf Prop. A Accept LW None Prop. B Accept P (low - med); ? ± (Med -hxgh); $ 1 (med - high) P w high; P-L low; $j low From Table XIV one can see that the major exceptions to my assertions about r i s k taking usually include a s i t u a t i o n where condi-tions of High Information are present. I argued above ( page 84 ) that i t was reasonable to accept that High Information was greater than Medium Information. This leaves one with the consideration of a threshold notion, whereby once an optimal point i s reached with respect to information received, then further increments i n information a f t e r t h i s point do not make any s i g n i f i c a n t differences to d e c i s i o n making. From the experiment i t i s possible to i d e n t i f y this threshold as knowledge of the range of odds and payoffs attached to any a l t e r n a t i v e . The test of proposition 2', which r e l a t e d wealth to r i s k , indicated that HW i n d i v i d u a l s did not take s i g n i f i c a n t l y d i f f e r e n t - 91 -types of risks to MW individuals. One could accept this result as i t stands, though one could also argue that the experimental distinction between MW and HW was not in fact a good one. However, the important thing to note is that the results are quite conclusive in terms of supporting the notion that LWHU individuals prefer low risks. This was further endorsed by the test of proposition A which produced no exceptions and proposition B — the exceptions to which in actual fact support the above contention. With reference to proposition B I had argued that LWHU i n -dividuals would not take risks for marginal benefits. The exceptions to this proposition indicated that some LWHU people w i l l take risks for marginal payoffs when the remaining incentive conditions are strongly in their favour, i.e. when conditions of high P , low P w 1 and low $^ are present. So the testing of propositions 1', A and B gives strong support to the notion that low risk strategies are preferred by LWHU individuals. Other findings from the experiment suggest that the chance factors involved in a decision are the most important ones for LW individuals. Implications for Field Test In the f i e l d context of B.C. gillnet fishermen I w i l l be subjecting the same propositions to test. The sample of fishermen used w i l l be characterised by HU for the outcome \"increase fish i n -- 92 -take\", and I w i l l be analyzing their risk taking behavior subject to constraints of information, wealth and the incentive conditions at-tached to alternatives. As a result of the experimental test I w i l l be expecting significant differences in risk taking behavior between low wealth and medium wealth fishermen, in terms of the former showing a dis-tinct preference for low risk strategies. I w i l l also expect low wealth gillnetters to take advantage of high return/low cost invest-ments but not to go for marginal payoffs except when the remaining incentive conditions are strongly in their favour. In other words they w i l l only go for marginal payoffs when there is a good chance that they w i l l be successful, and a low chance that they w i l l lose anything by i t . It w i l l also be interesting to see whether the chance fac-tor (P , P^ that was so important to LW subjects i n the experiment w J_ is as important to LW fishermen in the f i e l d . These considerations discussed above a l l refer to a combin-ation of factors which amount to a preference for low risks. The main expectation of the f i e l d test (based on the experimental results) is that LWHU fishermen w i l l be low risk takers. The wealth gradient for fishermen is calculated from the same baseline or property used in the experimental calculation, i.e. a LW individual is defined as one who cannot afford to lose. It could be argued that the experimental distinction between MW and HW was not in fact a good one. This makes i t imperative to have a better opera-tional definition for the f i e l d test. To this end, I consulted the Department of Welfare's minimum wealth standards of e l i g i b i l i t y for financial assistance i n addition to assessing from fishermen their \"folk\" view of the dividing lines between wealth levels. From these two sources I constructed a sliding scale which matched off a man's income with the number of dependants he had to support. The experiment showed significant increases in tolerance for high risks as wealth moved from low to medium. It did not show significant increases as wealth moved from medium to high, (possibly due to the poor distinction between MW and HW). So i n the f i e l d I w i l l expect MW fishermen to take significantly more high risks than LW fishermen, and w i l l be trying to assess the difference that HW makes to decision strategies chosen. One of the most interesting results from the experiment was the establishment of a threshold for the influence of information upon decision making. I w i l l be assessing whether or not this operates in the f i e l d . If i t does then my main concern w i l l be with an infor-mation gradient up to and including the specification of the range of odds, payoffs and costs (medium information in the experiment). Field data should show that increments in information after this level do not affect decision strategies. The experiment showed that increments in information from low to medium were accompanied by an increased - 94 tolerance for high r i s k s . In the f i e l d I w i l l expect fishermen to i n c r e a s i n g l y take high r i s k s as information increases to the medium l e v e l . Another r e s u l t from the experiment was that high informa-t i o n operates on a LW i n d i v i d u a l i n such a way as to make his p r e f e r -ence for low r i s k s t rategies no d i f f e r e n t from that of a MW i n d i v i d u a l . This f i n d i n g c o n f l i c t s with the information c i r c u i t s model (with re-gard to technological investments) i n Chapter 2. This model ruled out the p o s s i b i l i t y of a LW/H.Inf configuration e x i s t i n g . The task w i l l be to see whether there are empirical examples of this configuration i n the f i e l d . I f there are, then this refutes the information c i r c u i t s model with regard to investment r i s k s . I f there are not, one implica-t i o n i s that the experimental design dealt with a l o g i c a l p o s s i b i l i t y which has no empirical referent. Results contrary to those expected as a consequence of the experimental findings require that the supplementary propositions (re-l a t i n g s o c i a l marginality, s o c i a l mobility and co-operative structure to r i s k taking) be brought i n to see i f they can account f o r them. These l a t t e r factors were not c o n t r o l l e d f o r i n the experiment and constituted a set of constraints which I thought might bear upon r i s k taking behavior. Their use i s predicated on the f a i l u r e of the secondary set of propositions (1', 2', 3') to account for the phenomena of r i s k taking. Generally, as a function of the experimental findings, I expect LW fishermen to be low risk takers, and that increments in wealth to the medium level w i l l be accompanied by a progressively increased tolerance for high risk strategies. I also expect fishermen to take more high risks as information increases up to the medium level. Beyond this, increments in information should not make any difference to decision making. - 96 -CHAPTER 4 TEST CASE: FISHERMEN AS RISK TAKERS Introduction The move from the laboratory to the f i e l d setting involved fieldwork i n a fishing village on Northern Vancouver Island. This provided another substantive setting i n which to test my theoretical statements. Given that I can demonstrate that I am dealing with the scope conditions which applied to the laboratory setting, then I can assume that both situations are valid tests for the same set of propositions. The scope conditions specify that fishermen can be categor-ised according to parameters of wealth and subjective u t i l i t y for an outcome (the outcome in this context is \"increase fish intake\"). I w i l l be analysing decisions associated with that outcome, in terms of their variation subject to constraints of information and the in-centive conditions that apply to the alternatives the fisherman has to decide between (cf my statements in the Introduction of Chapter 3). The Fishing Industry in British Columbia The fishing industry i n B.C. is the province's fourth l a r -gest industry. In 1968 the wholesale value of fish and li v e r products produced in B.C. amounted to $119.3m. Of this figure salmon products accounted for $100 m. (Department of Fisheries of Canada 1969). - 97 -7,548 boats and 12,133 fishermen exploit the west coast fis h -eries (1968). These are supported by 1200 workers on packers and other auxiliary vessels and approximately 5,000 persons i n fish-processing, packing and handling plants. However the cyclical nature of the fish runs (particularly salmon) means that the majority of these 18 1/2 thousand individuals are fully employed in the fishing industry for only six months of the year. Despite this fact, the importance of the industry is that in many areas i t is the only opportunity for gainful employment. Information from the Department of Fisheries indicates that the number of fishermen each year does not vary directly with the catch. Approximately the same numbers of men fish in poor years and bad years. As the fishermen know beforehand whether i t is going to be a poor year for fishing (from Department of Fisheries catch estimates), one reason for this \"inelasticity\" may be the fact that the fishermen can-not find suitable alternative employment. Both primary and secondary sectors of the fishing industry are seasonal, though the primary sector appears to be basically more variable than the processing, packing and handling sector. The primary object of this multi-million dollar industry is the Pacific salmon. Five principal species of salmon are caught: sockeye (oncorhynchus nerka), coho (oncorhynchus kisutch), spring (on-corphynchus tschawytscha), pink (oncorhynchus garbuscha) and chum (on-corhynchus keta). - 98 -These species d i f f e r considerably i n s i z e , migratory patterns and i n commercial value. A l l f i v e species come i n from the P a c i f i c to spawn i n fresh water, and are most abundant between Southern B r i t i s h Columbia and Western Alaska. Within this length of coa s t l i n e the geo-graphic range of each species overlaps considerably, so that some spe-cies at the same time occupy not only the same geographic t e r r i t o r y but frequently the same stream. The spawning and migratory habits have been studied over the years, and a f a i r l y accurate p i c t u r e of the salmon cycle has been b u i l t up. So much so that the Department of Fi s h e r i e s issues a serie s of projections f o r the timing, s i z e and place of the d i f f e r e n t salmon runs for each year, with a s u r p r i s i n g degree of accuracy. The salmon are caught on t h e i r way to the spawning grounds by three main methods — seining, g i l l n e t t i n g and t r o l l i n g . The pro-cess I studied i n the f i e l d was that of g i l l n e t t i n g . The accompanying map indicates the main coastal f i s h i n g areas (Department of Fi s h e r i e s S t a t i s t i c a l Map of B.C. F i s h e r i e s ) . G i l l n e t t i n g G i l l n e t t i n g boats range i n s i z e from 16 feet to 40 feet and are generally operated by one man. G i l l nets take a l l v a r i e t i e s of salmon but more p a r t i c u l a r l y sockeyes, cohos and pinks. The p r i n c i p a l waters fished with the g i l l nets are the estuaries of the l a r g e r r i v e r s such as the Nass, Skeena and Fraser and those i n l e t s which have salmon streams flowing into them, e.g. Rivers I n l e t , Smith I n l e t , Knight I n l e t , Burke Channel, etc. • _ 99 -M A P 1: C O A S T A L F I S H E R I E S O F B R I T I S H C O L U M B I A - 100 -The gillnet snares the salmon by the g i l l s while they are en route to spawning grounds. The net is wound around a drum in the stern of the boat and the drum is operated by a power take-off from the main engine and is controlled by a clutch operated by the fisher-man as he stands at the stern. The net is hung from a cork line at the surface and is weighted down by a lead line which runs along the bottom of the net. It hangs like a curtain in the water. A buoy is attached to one end of the net and after the net is set out over the stern, the boat and net d r i f t with the current or tide. The net is taken in by being rewound on the drum. As the net comes in over and between sets of rollers, the fisherman pulls out the salmon. Constraints The Inshore Salmon Fisheries on the West coast is rigidly policed by the Federal Department of Fisheries. Regulations govern the size and type of nets and the manner in which they can be used. Certain areas, usually near the mouth of rivers, are closed to commer-c i a l fishing at a l l times. Federal law also calls for a minimum weekly closed period of 72 hours, and conservation frequently demands that the fishing week be reduced to two days or less. - 101 -These regulations are a safeguard to the salmon as a resource, as the contemporary fishing fleet, working with nylon nets, electronic navigation instruments and capable of ranging a l l over the coast to find fish can wipe out a salmon run in a matter of days or hours. The closures thus place time as a prime factor of consideration for the fisherman. If he has only one or two day's grace a week to exploit the salmon resource then the way in which he uses his time during the openings is very important. Fieldwork The fishing village of Alert Bay on Cormorant Island provided the base for the f i e l d test. Located off the north-east tip of Vancouver Island in Johnston Straits, Alert Bay straddles the main waterway be-tween Vancouver and Prince Rupert. Two administrative structures — a village council and a band council — served the population of 1500, two-thirds of whom were Indians. The village was incorporated as part of the Mt. Waddington District in 1966 and had domain over 800 people, the majority of whom were white. The Nimpkish band council administers two Indian Reserves with a joint population of 700. Alert Bay has two further significant features in that i t serves as the cultural centre for the Kwakiutl nation, and as a ser-vice centre for Northern Vancouver Island. - 102 -With respect to i t s s i g n i f i c a n c e as a c u l t u r a l focus f o r the Kwakiutl, A l e r t Bay has witnessed a long procession of anthropologists from Boas through Codere to the host of museum o f f i c i a l s and l a t t e r -day exponents who throngrthe summer potlatches. As a ser v i c e centre, the v i l l a g e boasts a 70-bed h o s p i t a l , boat-building wharves, 15 r e t a i l stores, a l i q u o r store and eight R.C.M.P. o f f i c e r s . In terms of both service and occupation A l e r t Bay's depen-dance upon the f i s h i n g industry i s extant. The inshore salmon f i s h i n g opens i n June and continues to November. During the intervening months, the population of A l e r t Bay i s swollen by as many as 1500 fishermen, and the r e t a i l and service industries are l a r g e l y dependant on this transient custom. In the winter most A l e r t Bay fishermen take jobs as loggers. There i s no shortage of winter employment p o s s i b i l i t i e s , as the logging camps are always short-handed. Three f i s h i n g companies operate out of A l e r t Bay — B.C. Packers, Canadian Fish Co., and Nelson Bros. Ltd. Two of these — B.C. Packers and Nelson Bros. — are owned by the same parent company, but they operate independantly of one another. The companies' operating concern was with f l e e t s of seine boats; t h e i r main relevance to a g i l l n e t t e r was i n terms of the p r i c e paid for f i s h and the services given to r e t a i n a fisherman's a f f i l i a t i o n . Minimum prices f or landed catch are negotiated between the U.F.A.W.U. and company owners, though landing prices vary from place to place. \\ - 103 -Each fisherman i s a f f i l i a t e d with one of the companies, which means that i n return for c e r t a i n services — d e l i v e r i n g nets, f l o a t i n g loans, etc. — the fisherman i s obligated to s e l l h i s catch to a par-t i c u l a r company. Data-Collection P r i o r to the salmon opening i n June I spent several months making contacts with both Indian and white fishermen from A l e r t Bay with the r e s u l t that I was known personally by about s i x fishermen when I began fieldwork. The Department of F i s h e r i e s , the United Fishermen and A l l i e d Workers Union and the Native Brotherhood were consulted both before and during my period of fieldwork. In A l e r t Bay, I l i v e d at the wharf on a boat (with the un-l i k e l y name of \"Lintu\") which meant that even when I was not out at sea with a p a r t i c u l a r fisherman I was i n continuous i n t e r a c t i o n with the f i s h i n g community. In c o l l e c t i n g information on g i l l n e t t i n g a c t i v i t i e s I ranged over the whole coast from Rivers I n l e t , where I spent two and a h a l f weeks at Wadham's Landing, to the Fraser River where I observed the f i n a l week of f a l l f i s h i n g i n November. As the A l e r t Bay fishermen followed the f i s h runs, I had to be with them. The process of p a r t i -cipant observation on each boat gave me a minimum of four days exclu-sive i n t e r a c t i o n with each p a r t i c u l a r fisherman, and t h i s was augmented by regular contact throughout the r e s t of the f i s h i n g season. - 104 -In any average week I would spend three days ashore and four days at sea. The time spent ashore, particularly in the fishcamp at Wadham's Landing, was used to f i l l up any information gaps I had on any fisherman and to cross check the r e l i a b i l i t y of the data I already had. In Alert Bay my sample of independent gillnetters was composed of 12 Indian fishermen and 14 Anglo-Canadian fishermen. The composition of the sample was such that I should be able to observe whether or not cultural differences accounted for any difference i n decision-making. Of these 26 boats I went out on 10 and directly observed the decision-making of the individual captains. Of the remaining boats I used data gathered from observing their boats while on the fishing grounds and supplemented this by talking to the captains on shore, to ascertain whether the inferences I had made about their decision-making behavior were in fact accurate. This method worked extremely well in some cases but not in others. For instance, a group of boats — Annie L, Beatrice S, Eddie May, Flora M, Miss Lorna, and Sea View — always fished together and in the same way. Thus, one had to go out on one boat to observe a l l six with respect to fishing strategies employed and perform follow-up i n -terviews on the remaining five to obtain the necessary background mat-e r i a l . I worked on the assumption that another gillnetter faced with the same alternatives as the captain I was with could be classified as the same type of risk-taker,if his strategies were the same as on \"my\" boat. - 105 -But there were two boats for which this assumption broke down. I could not be sure that the a l t e r n a t i v e s were the same as f o r my \"skipper\", and as the shortness of the f i s h i n g season prevented me from going out with these p a r t i c u l a r fishermen, I dropped them from my sample on the grounds that I had inadequate data for them. Risk I am dealing with two categories of r i s k , though I treat them the same a n a l y t i c a l l y . F i r s t of a l l I am concerned with invest-ment r i s k s taken on new technology (I do not include general maintenance costs), and secondly with strategy r i s k s with respect to e x p l o i t i n g f i s h as a resource. The reason f o r this d i s t i n c t i o n i s that I want to discover i f there i s consistency i n r i s k - t a k i n g , across the board, as i t were; or whether c a p i t a l r i s k s taken on new technology r e s u l t i n lower ex-p l o i t a t i o n a l r i s k s being taken. For instance, an investment i n radar would reduce the uncer-tainty aspect to f i s h i n g i n fog; i n depth sounders would give more information about proximity to reefs and shore, whether or not this reduction of uncertainty leads to the taking of high r i s k s i s a fac-tor I wish to e s t a b l i s h . Risks on both technology and strategy choice involve a c a l -c u l a t i o n on the part of the fisherman of payoffs, costs and the chance factor as indicated below (Table XV): - 106 -Table XV: Risks on Technology and Strategy Technology Strategy Payoff Calculated increase i n f i s h intake A b i l i t y to compete better for f i s h resource Calculated intakes from d i f f e r e n t loca-tions Costs C a p i t a l Cost Calculated loss i f payoff does not accrue Possible loss of gear P h y s i c a l danger Time Cost of t r a v e l l i n g Chance Required confidence l e v e l f o r fisherman to invest c a p i t a l C a l c u l a t i o n of the chance of a l t e r n a t i v e locations g i v i n g ex-pected y i e l d s . C a l c u l a t i o n of chance of i n c u r r i n g losses due to hazards, etc. High r i s k i s defined as the decision to maximise gain or to take a long shot — low r i s k s trategies are those that attempt to mini-mize costs. The d i s t i n c t i o n between the two categories of r i s k i s simply a h e u r i s t i c device to (a) sort out the data and (b) permit me to test the information c i r c u i t s model (Chapter 2) as that model i s spe-c i f i c to r i s k s on technology. Wealth For a wjalth c l a s s i f i c a t i o n I used a combination of Depart-ment of Welfare standards and fishermen's estimates of r e l a t i v e wealth to define the low wealth category. - 107 -Department of Welfare standards involved a sliding scale whereby a minimum income vis-a-vis the number of dependants was cal-culated to establish whether a person qualified for welfare payments. Using the same principle I constructed another sliding scale but gave slightly higher figures than the minimum necessary to stay off welfare as a result of the \"folk\" view of what constituted low wealth (Appendix V). Recall that in my discussion of farmers taking risks, and of the experimental subjects taking risks, the low wealth individual was in a position where he could not afford to lose. In order to meet the necessary scope conditions low wealth fishermen must be similarly categorised — i.e. they are i n a position where they cannot afford to lose. Medium wealth ranged from the dividing line discussed above to +$3,000, high wealth was categorised as +$6,000. U t i l i t y I inferred a fisherman's u t i l i t y for increasing fish intake from two sources of data: (a) from verbal statements about fishing intent and (b) a subjective evaluation (on the part of the fishermen) of increasing fish intake. The latter involved a scale score 1-10, and followed ques-tions from me as to the importance of increasing fish intake. I asked the fishermen to rate themselves on the scale and categorised HU as having a score between 7-10, MU 4-7 and LU 1-4. This method e l i c i t e d a number of verbal statements that often l e f t no doubt as to the parti-cular fisherman's preference level for the outcome \"increase in fish intake\" - 108 -Information (a) With respect to fishing strategies decided upon, there were a number of information sources that could be used. (1) Department of Fisheries projections. (2) Other fishermen. (3) Radio. (4) Performance of other boats. (5) Depth sounder. (6) Fish jumping. The f i r s t few categories give information of a general nature, the latter two of a more specific kind. The Department of Fisheries projections give estimates of timing, place and size of different salmon runs. Other fishermen act as a general pool of information exchange — an exchange often character-ised by distortion. Much of the time during closure is spent informa-tion-broking. In the fish camps and taverns a gaming-like situation exists in that fishermen are seeking as much information as they can while attempting not to divulge any themselves. Different channels of infor-mation are regarded as reliable, but most queries from one fisherman to another are met by evasion and distortion. It was as though there were different networks of information within ithe community of fishermen, and entry into one or more networks seemed to be a function of a variety of factors. Kinship played a role - 109 -amongst both Indian and white fishermen in that information was dis-pensed freely and accurately between kin. But there seemed to be a limit to kinship as a useful criterion — one's peers in an informa-tion network had to.be more or less equal in terms of performance with respect to fish,intake. Implicit contractual ties also defined particular networks. This sometimes operated as a \"buddy\" system whereby two or more fisher-men actively helped and directed one another on the fishing grounds, but more generally i t consisted of a reciprocal pooling of reliable information. Here again the limits of the network seemed to be de-fined by performance. Highliners did not dispense accurate informa-tion to non-highliners and vice versa. (The term \"highliner\" refers to a top fisherman in terms of recorded catch.) One network that I saw operating at Rivers Inlet consisted of the skippers of the Virginia M, Miss Christie, Aunna, Ruby S, Slave ( a l l Alert Bay boats) plus one gillnetter from Sointula. The network was defined by a l l six skippers at one time or another during the closure period eating a meal on each one of the other boats. They did not take food on anyone else's boat, and nobody outside the six ate on their boats. This was a \"highliner\" network, in that a l l six were considered to be top fishermen. So whether the network was based on kin or contract, the limits to i t seemed to be defined by performance. - 110 -The radio was also an important source of information, though different fishermen had different evaluations of i t . Most fishermen operated with the radio on, but the information dispensed over the radio was rarely regarded as reliable. If a fisherman was doing well he would hardly broadcast i t , as he would have 100 boats breathing down his neck in less than an hour. If a \"buddy\" system was operative statements over the radio would be coded -- so that only the \"buddy\" would know what was meant. Some makes of depth sounder (Ekolite, Marconi) were equipped with a fish-finder function which would give readings on a trace-out from which a fisherman could infer the presence or absence of fish. Perhaps the most highly valued piece of information was the sighting of fish jumping. As salmon generally run in schools, one or two jumpers usually means there are quite a lot more there. Fishermen watch one another very closely -- to see how they are doing on a comparative basis -- and one can infer from observing other boat's catches more about the general nature of fi s h running in the area than simply from one's own haul. Sources (1) to (4), even i f each one is tapped, only gives the fisherman information of a general nature. The sounder and jumpers give more specific information about the presence of fish. So for stra-tegies the former defines low information, the latter high information. (N.B. High information for strategies i s thus analogous to the medium information level in the experiment. Recall that medium in-- I l l -formation in the experiment specified the range of odds, payoffs and costs -- high information in the f i e l d similarly gave a range of possi-b i l i t i e s . There was no effective way that information could be increased for the fisherman. The most that he could get was a function of fi s h jumping and sounder readings. Thus for the f i e l d test there was no cor-responding high information level with respect to strategies decided upon.) Another important source of information that the fisherman could draw upon was his own experience and s k i l l . This s t i l l , however, gave information of a general nature with respect to tides, currents and salmon runs. (B) The information circuits model differs from my preceding re-marks in that with this model I was stating that access to information circuits with respect to new technologies varied with wealth and edu-cation (Chapter 2). The implication was that individuals with greater access to information circuits would be more innovative, i.e. take high risk strategies. I use the same definition for wealth as previously. Educa-tion, with respect to this model, is defined in terms of formal edu-cation. Schultz and others have discussed the general benefits of education with regard to productivity (Schultz 1963; F i r t h 1964b) -- I am attempting to show i t s precise relevance for investment risks taken by inshore fishermen. Low education is defined as below Grade 6, medium education consists of standard secondary education, high education is categorised as a function of post secondary training -- university, etc. - 112 -Information Circuits The following are the main information sources that can be tapped: (1) Fishing journals. (2) Department of Fisheries. (3) Manufacturing companies. (4) Fishing companies. (5) Union. (6) Other fishermen. I constructed a scale for access to information circuits by assigning 1 unit for each information source tapped by particular fishermen. There were three fishing journals current on the West Coast (one unit for each); the Department of Fisheries ran courses every year which gave out information about new exploitative and technical devices and also issued projections for the fish runs (two units for this). One unit was assigned for any i n i t i a t i v e taken by a fisherman in approaching manufacturing companies with regard to technology; also the fishing companies operated as information brokers. Consideration of both the union and other fishermen as additional sources of infor-mation gives me a scale of 0 —^10. - 113 -Table XVI: Information Circuits Fishing Journals subscribed to 3 Department of Fisheries Courses Salmon run projectsion 2 Manufacturing companies (a) electronic (b) net 2 Fishing Companies 1 Union 1 Other Fishermen 1 I define access to information circuits as high when a score is greater than 7, medium with a score between 4 and 7 and low as hav-ing a score less than 4. Every fisherman used (6) — other fishermen — as one source of information, but this was generally regarded as unreliable. Dis-tortion of information and evasion to questioning was the general rule. This was not surprising i n terms of the very fierce competition for the salmon as a resource. Each source could generally give information that others could not. For instance the department of fisheries offered catch projections, the journals offered general information on technology, the manufacturing companies offered demonstrations of technology, and the union gave information as to variation in landing price for fish. Even where duplication of information exists, the point of the scale is to measure the number of information sources a particular - 114 -fisherman uses on the assumption that more than 7 sources gives him higher information than 5 sources, and 5 gives him greater information than 3. The point to be made is that i f a person seeks out a number of information circuits with regard to technology even though there may be some duplication, he w i l l be more likely to take high risks on i n -vestment. I w i l l also be trying to assess whether the information category a particular fisherman is in is the one he is \"supposed\" to be i n as a function of his wealth and education. Case Histories of Risk-Taking by Alert Bay Gillnetters I intend to present in detail an analysis of two types of fishermen with respect to the risks they took on technology and stra-tegies. The exercise is to make clear the c r i t e r i a that I use to ar-rive at categories of risk-taking. For the remaining 24 fishermen I present the end result of the analysis in tabular form — to write 26 case histories in f u l l would make this section longer than the rest of the thesis. Strategies as Risks In order to infer whether a particular strategy i s high or low risk one has f i r s t of a l l to be aware of the characteristics of salmon runs and secondly of the alternative means of getting at them. - 115 -There were three major situations where I was on hand to ob-serve which strategies different fishermen elected to choose — (a) Rivers Inlet fishing; (b) Johnston Straits fishing; and (c) the closing of the Northern Fisheries. This was supplemented by other instances, but these three situations are the major ones that were used to categorise an individu-al's risk-taking behavior. Rivers Inlet Rivers Inlet stretches 40 miles inland from the Queen Char-lotte Straits. The main channel has a number of breaks provided in the form of reefs, shoals and islands, and the mouth of the inlet has quite a number of offshore islands and reefs. Runs of sockeye coming i n from the Pacific, turn the tip of Vancouver Island, and a large proportion head directly across Queen Char-lotte Straits into Rivers Inlet, using the offshore islands as a guide. Inside the inlet the schools tend to follow the shore, reefs and islands as pointers to fresh water. Over shoals and reefs the salmon rise clo-ser to the surface and are less likely to swim under the net the closer a set is made to the reef. The schools tend to run more predictably in bad weather con-ditions and with the tide. Rain has the result of making the salmon bunch together and head faster for fresh water — i t also brings them closer to the surface. - 116 -From a knowledge of salmon runs, i t is perhaps an understate-ment that a l l the ideal spots are hazardous. To fish the sides of the channel and the reefs meant the possibility of fouling on the rocks. Fishing outside meant an 8-10 foot groundswell and the choice spots involved locations close to the reefs and islands. In addition to these locations, another ideal spot to fish was the boundary; in this case the hazard to be feared was other fish-ermen. Boundary fishing was almost a law unto i t s e l f . At the head of Rivers Inlet a demarcated boundary indicated the extent to which commercial fishing was permitted. This was where the fish were heading to the main salmon stream. Because of this fact 100-150 boats would be packed tightly together along the boundary, which meant there were very real dangers of fouling nets and ruining gear. Also a rip current would frequently drag a boat and net across the boundary whereupon a fisheries patrol would be ready with a fine and confiscation of catch. With the number of boats, and narrowness of the boundary, the special feature about boundary fishing was the possibility of a very large catch, and also the very real possibility of losing equip-ment (fouling on other boats) and catch (crossing the boundary with a net out constituted an offence against federal fishing regulations and resulted in a fine as well as confiscation of any fish that had been caught). Rivers Inlet had choice spots and conditions where fish could be caught, but they were a l l hazardous. The interesting thing, analy-- 117 -t i c a l l y , is what a particular fisherman does when he gets positive indicators (viz. fish jumping) off a hazard. Does he make a set (high risk or go somewhere safer (low risk)? Fishing inside in the main channels away from reefs and shoals involves low risk inasmuch as the danger of losing equipment (either from fouling or hazards) is much less though the catch w i l l be smaller. During the fish runs there is always the prospect of getting a f a i r catch, but the large catches involve playing the hazards. Table XVII; Strategy Risks i n Rivers Inlet Fishing Locations High risk Low risk Boundary x Reefs and Shoals x Outside - off reefs x Centre Channel x Protected shores x Johnston Straits The village of Alert Bay was a central focus point for any boat wishing to fish Johnston Straits. The same remarks about salmon habits hold, though there are two additional hazards in the Straits. One is the preponderance of large kelp beds, the other is the use of the Straits as a major waterway. The former is significant in that salmon seem to \"play\" around kelp beds at the estuary of certain streams and they offer a - 118 -hazard inasmuch as that, though fish are to be found close to kelp, the process of extraction usually brings up a whole lot of kelp which may take 2-4 hours to pick out. With a fishing week very often of 1-2 days, this was time that could cost a fisherman one or two sets. The considerable amount of t r a f f i c through the Straits, es-pecially in the daytime, constituted a hazard to nets. The narrowness of the channel and the size and disregard for nets of the large tugs and liners made the prospect of losing gear a very real one. The Straits do have a number of relatively safe spots clear from the t r a f f i c lanes — Growler's Cove, Robson's Point — which can be considered low risk areas in that losses due to fouling can be re-garded as negligible, though only moderate catches can be taken. Shore fishing in Johnston Straits is generally regarded as a prerogative of seine boats. Though some g i l l netters may dispute this, there is very l i t t l e that they can do when a 200 ton seiner nudges them away from a beach location. Table XVIII: Strategy Risks in Johnston Straits Fishing Locations HR LR Shoals and reefs x Kelp beds x Traffic Lanes x Safe Areas x Shore fishing x - 119 -Closure of Northern Fisheries In August a situation arose that affected a l l Alert Bay g i l l -netters. Due to an inadequate amount of spawning in the salmon streams the entire Northern Fisheries was closed as a conservation measure. In the f i r s t week of the closure only four areas were l e f t open: (1) The Rivers Inlet area was open for 3 days (Areas 9, 10 and 11). (2) Johnston Straits was open for 2 days (Areas 12 and 13). (3) Central Inshore Fisheries (Areas 14, 15, 16 arid 17) was open for two days. (4) The Fraser River area was open for 2 days. Of these openings, Rivers Inlet was the poorest prospect i n terms of expected volume of salmon. The summer sockeye run was over, and no more would be expected t i l l the f a l l . The Straits showed promise of a moderate run, as steady re-turns had been reported for the previous week. The central part of the inshore fisheries were practically devoid of salmon. A run of sockeye was expected at the Fraser, though the pre-vious week's catches had been very low on the average, which was not a good indicator, of an impending run. To fish Rivers Inlet or central areas with Johnston Straits and the Fraser open would be a useless exercise. The decision facing the Alert Bay gillnetters was either to fish Johnston Straits and re-main close to Alert Bay or go south to the Fraser and fish 2 days. - 120 -The calculated advantage in going south would be that there was a low to moderate chance of a good catch. But against this had to be offset the cost of travelling there (2 days journey from Alert Bay), the hazards from t r a f f i c (which were worse than for Johnston Straits), and the warning from the Department of Fisheries that they were not sure whether the run would materialise. Fishing Johnston Straits offered an advantage in that though no high catches would be expected, there was a fair chance of moderate hauls (the week prior had shown a steady run). The cost factor was reduced by lack of travelling costs. From this one could infer that a high risk would be to go south; a low risk to stay put in Johnston Straits. Table XIX: Northern Closure and Strategy Risks Johnston Straits Fraser Payoffs: Moderate Payoffs: High Chance: Moderate Chance: Low Costs: Hazards and Traffic Costs: Hazards and Traffic and Travel LOW RISK HIGH RISK Investments on Technology as Risk The options on technology open to the gillnetter are very wide and cover a variety of functions. These options can be broken down into general categories of technology in terms of either (a) lo-cational technology or (b) exploitative technology. - 12* > (a) Depth sounders with fish-finder functions are a recent entry to the market and as well as being a safety device, readings on the trace out can give information about the presence of fish. However these are expensive additions to a fisherman's repertoire as the cheapest ones cost $1300 (Marconi). Ship-to-ship and ship-to-shore radio telephones can give the fisherman an extensive range of information from which he can draw in order to make decisions. Someone fishing the Fraser can find out what is going on at the Nass River, over 400 miles away. Again, this tech-nology is not cheap, as each type of radio telephone costs upwards of $150, and some are much more expensive. Compass and charts are standard navigational devices; though quite a number of boats are beginning to carry radar. As well as a navigational aid, radar functions to increase fishing time during fog conditions. Most boats without radar stayed in port during fog, while those with i t had the fishing grounds virtually to themselves. (b) The main exploitative device used on the gillnetters is the power drum, which can be either hydraulicly or gear-operated. The hydraulic drum is much smoother, and a fisherman can make and pick up sets much faster. The conversion from a gear-driven drum to a hydrau-l i c system costs in the region of $300. Most nets are made out of nylon with either a monofilament web (single strand) or a polyfilament net (several strands). The l a t -ter are f a i r l y standard throughout the industry whereas the monofilament - 122 -nets are quite new introductions. They cost half as much again as the standard nylon nets, but their advantage is i n terms of longer l i f e and a better \" l i e \" in the water. To exploit the inshore fisheries effectively a fisherman requires at least six different nets — varia-tions i n colour and web-size are the main considerations. Most marine engines used in the gillnetting fleet were capable of a steady eight knots, but recent years have seen the advent-: of boats with high speed engines and planing hulls. A marine firm (Albion Co. Ltd.) at Haney, B.C. started putting these on the market in 1968. The advantage of a high-speed engine was that i t increased the area of exploitation possible during an opening as travelling time was cut in half. The slower boats had more or less to stick to one major fis h -ing area as they would lose too much time travelling during an opening. This was not a handicap to the high-speed boats, though their operating costs were very much higher. Operating Costs The average operating costs for a gillnetter worked out at around $2,000 per annum. This included general maintenance and repairs as well as replacements for fishing gear. (This figure does not include costs on new technology). The high speed gillnetters, because of their very much greater fuel consumption, had operating costs of around $2,700 per annum. - 123 -One result of the technological explosion i n recent years has been to decrease the hazardous nature of the business (improvement in safety features as well as better exploitational devices), but there is s t i l l a large measure of uncertainty attached to the process — as the prospect of losing l i f e and capital in pursuit of the salmon is an everyday fact of l i f e to the fisherman. Perhaps the major impact of technology has been to change the patterns of mobility and exploitation. In the past gillnetters tended to fish in areas adjacent to their home base and effectively operated as trappers — they attempted to catch the fish as they passed through. But the larger and technologically more proficient boats capable of withstanding rough seas permit the fishermen to range the whole of the inshore fisheries hunting the salmon, i.e. going i n pur-suit of the salmon, instead of lying i n wait for i t . This increased mobility has been coupled with improved com-munication systems and navigation aids, making i t possible for the fishermen to keep i n close contact with fishing conditions i n differ-ent areas and to take advantage of unexpected salmon runs. The dis-tinction between hunting and trapping goes a long way to explain d i f -ferences in technological investment. If a fisherman is concerned mainly with trapping and stays close to his home base then there is often no need for the latest tech-nology. The fishermen would have such an intimate knowledge of the area and conditions, to an extent that would render things like radar and depth sounders worthless to his purposes. - 124 -A hunter, on the other hand, depends far more on technology as his range of exploitation requires the use of devices other than intimate knowledge of a particular area. There is a very close correlation between range of exploita-tion and investment in technology, and i t appears as though two dis-tinct investment cycles are operative. Hunting reinforces capital i n -vestment, while trapping discourages i t . The following table indicates this relationship. Table XX: Investment Risks and Range of Exploitation No. of Fishermen Taking Fisherman's Range of Investment Risks Exploitation L M H H 8 0 2 6 L 18 3 15 0 Despite the vast array of technology available, i t offers to the gillnet fisherman only a marginal advantage in exploiting the fish resource. There are no high payoffs to be expected from technological investment, and the cost of new equipment is generally high. Thus the investment of capital in new fishing technology is always a high risk inasmuch as the calculated payoff from i t is marginal, in terms of increased fish intake, and the capital cost is usually steep. However in an industry as highly competitive as the inshore salmon fisheries, the marginal advantage is a sufficient incentive for some fishermen to take high capital risks. - 125 -Case No. 1: Ernie McAllister I travelled up to Rivers Inlet from Alert Bay with Ernie McAllister on his boat the Virginia M. During the journey I had ample opportunity to find out the types of investment risks he was prepared to take. He had built the boat himself a year ago; the design and technology incorporated indicated a very astute appraisal of conditions in the fishing industry. At 40 feet long and with a 12-foot beam the Virginia M was much bigger than the ordinary gillnetter. This was i n anticipation of the boat being used as a multipurpose vessel. The large beam meant that the owner could fish outside i n unprotected waters in large swells. Also the boat was big enough to go tro l l i n g off the West coast of Vancouver Island (once trol l i n g poles were fi t t e d ) ; and could also be rigged to long line for halibut. The boat had been bu i l t with tougher than normal timbers in order that i t could also operate as a salvage and beachcombing vessel. Ernie had anticipated a possible change i n fishing operation. At present the companies had fleets of collectors and packers that col-lected fish from the boats on the fishing grounds. Ernie saw the day coming when the companies would cut costs by demanding that fishermen ice and f i l l e t their own fish and deliver to a port, thereby excluding the cost of maintaining a packing fleet. To this end, a large hold was incorporated in the design of the boat - 126 -to f a c i l i t a t e what was an eventuality (to Ernie) of having to ice and f i l l e t his own catch. Technologically, Ernie had invested in a very large power drum — much bigger than was necessary for a g i l l net. This was so he could, i f he wanted, put a small seine net on his boat. Thus the Virginia M was designed to exploit the West coast fish resource in terms of gillnetting, t r o l l i n g , seining and long-lining, as well as having supplementary use as a salvage vessel. He had also invested in a Marconi depth sounder which had a fish-finder function, as well as two radio telephones (ship-ship; ship-shore), and an automatic p i l o t . He was trying for the f i r s t time a monofilament web on his outside side. He calculated that this total investment of $26,000 would give him a marginal advantage over most other fishermen in exploiting the West coast fish resource; so in terms of my definition of risk-taking he was taking high investment risks. Range The size and strength of Ernie's boat, gave him a range of mobility denied to smaller vessels. He fished from the Fraser to the Queen Charlotte Islands following the fish runs. He had seen his own exploitative patterns change over the years since he f i r s t fished — he had never ventured very far from his home base then, and even with - 127 -better technology through the years the tendency was s t i l l to \"trap\" the fish as they passed near to his home base. In the last ten years the technology and information explo-sion had led to him hunting the fish rather than trapping them. In-formation dispersed by the fisheries department on the size and place of runs were f a i r l y reliable; also the technology available permitted a gillnetter to range the entire length of the inshore fisheries. Born into a fishing family, Ernie at 55 was an acknowledged highliner in Alert Bay. He had fished since he was a boy and remembered the days when he operated from a rowing boat with a net which had to be hauled in by hand. Thus his experience and knowledge of the West coast fisheries was extensive, having fished on a l l types of boats. He put himself through university at U.B.C. but on graduation the most that a teacher could earn per annum was $800, so he stuck to fishing. (However, he was a great believer in education, and was very proud of the fact that he had put his daughter through the University of British Columbia). He had very much of a \"protestant ethic\" attitude towards work — \"You have to work hard at i t to be a good fisherman\" -- and he had a high incentive to increase his take as he wanted to pay off the loan on the boat as soon as he could. When I asked him to rate himself on a 1-10 u t i l i t y scale, he laughed and said that I should put him at around 100 (HU). - 128 -His taxable income worked out at around $12,000 per annum. The only dependent he had was his wife who worked part-time at a re-t a i l store in Alert Bay. His wealth category was high. He seemed very well informed on issues affecting the fishing industry from both a union and management aspect. At one time he was secretary of the Alert Bay local of the U.F.A.W.U. Thus i t is not sur-prising that he tapped just about every possible information source. He subscribed to two fishing journals (2), had taken a fisheries course at U.B.C. and used the department's salmon run projections (2). He made a point of seeing his net company and electronics company agent at least once a year on what was new (2), and also drew upon his fishing company, union and other fishermen for information (3). This gave him a score of 9 which indicated that he had high access to information circuits. Strategy Risks He planned to spend the next two weeks at Rivers Inlet as the sockeye run was expected to be at a peak. The journey from Alert Bay took a day and was through Queen Charlotte Straits which were notoriously rough -- other areas such as Johnston Straits or Hardy Bay which did not involve the cost of a long journey or rough waters were rejected as possibilities as the chances were that more fish were expected at Rivers Inlet. - 129 -The week before he had planned to fish Rivers Inlet, but on arriving there and talking around he calculated that the sockeye would peak at Smith's Inlet this week and Rivers Inlet the following week. So he pulled out of Rivers Inlet and went to Smith's, knowing that he would lose at least half a day's fishing which was important as a cost factor. (There was a two-day opening). But his calculation paid off. That week the top boat in Rivers Inlet took 200 sockeye, Ernie at Smith's Inlet took 250, so he was well pleased with his decision. Ernie used a specific set of indicators for f i s h presence. He kept the radio on most of the time but did not evaluate i t too highly as he did not think the fishermen would be telling the truth. He never broadcast over the radio himself. He used i t to pick up oc-casional tidbits pf information between collectors and packers. He was very observant of tide and weather conditions. He f i g -ured that f i s h ran more predictably in rough weather and with a f a i r tide or current and these were the conditions he chose over any other. Other boats would often be furiously setting or picking up, but not Ernie. On the average he would make 5-6 sets a day, but he would be very careful as to when and where he did so. He took into ac-count a number of information factors that added up to a calculation of when the fish would likely be running and in which direction. As the fish ran with the tide, the best time to set was just before the tide changed and then pick up an hour after slack water --- 130 -a 3-hour set. A swell of 8-10 feet and rough weather conditions (rain and 15-20 mile per hour wind) would drive the fish directly into the inlet when fishing outside. Good weather conditions and no tides per-mitted the fish to \"play\" and not go directly into the Inlet. He preferred overcast and rough conditions as the fish could less easily see the net, thus his preference for outside fishing as the inside waters were always more protected. He made his f i r s t set inside the Inlet, off a reef close to Wadham's Landing. He could have fished the main channel, but he calcu-lated that any fi s h coming into the main channel would use the reef as a guide and also come closer to the surface. On this set he constantly had to tow the net away from the reef. He opted for this location on L.Inf -- we had not sighted any jumpers and the readings from the depth sounder were ambiguous. When he pulled in the set he found he had one fish. Other boats in the main channel, nearby had pulled in an average of twelve. Ernie swore at himself and immediately set off to sea. He realised his net was too dark for that part of Rivers Inlet; the fish would see i t in the water and swim round i t . At Rivers Inlet one needed at least three nets to effectively f i s h the waters -- a very light green one for the boundary (known as a head net), as the s i l t brought down by the salmon streams made the water cloudy and milky. A darker green was necessary for the central parts of the inlet, while a very much darker green was needed for fishing outside. (This latter net had a monofilament web.) - 131 -Ernie had planned to fi s h the inside reefs (which were f a i r l y well down the i n l e t ) , and the outside reefs -- so had kept his outside net on the boat on the calculation that he could save time, and also exploit both areas effectively. He had wondered whether his net would be too dark for the inside reefs, but had decided to take a chance on i t . He did have an inside net in the l o f t at Wadham's Landing Fish camp, but to put this on would have prevented him from fishing outside.- So he was definitely out to maximise gains; he appreciated the cost aspect of using his outside net inside but s t i l l went ahead. On his way to the mouth of the Inlet, he saw three jumpers, immediately stopped and made a set -- the cost factor of his net type was s t i l l operating. He only caught 18 and as he reckoned he should have had at least sixty he immediately roared out to sea and fished the reefs off Table and Egg Islands off the mouth of Rivers Inlet for the remainder of the opening. Time By deciding to fish outside, i t meant that Ernie had to jour-ney back inside the Inlet to deliver fish, as i t was too rough for the collectors to come outside. This time loss -- costing him at least one set per day --could have been used to get his net changed so he could fish the inside reefs, but he calculated that weather conditions and runs outside would - 132 -give him more fish.. Also the inside fishermen were not doing particu-la r l y well. Weather conditions were quite rough and there was a large groundswell (8-10 feet), perfect conditions for outside fishing (ac-cording to Ernie). He consistently fished the reefs on both high and low information. No other boats were around him -- but this did not seem to bother him. He was interested in the way other boats were doing only in terms of rationalising whether his own decision to fish a particular location was good or not. Off the reefs outside, Ernie averaged 40 fish per set which was above average. A constant watch had to be kept on the net's d r i f t , tide and wind conditions to keep from foundering on the reefs. On one piece of high information Ernie took the plunge again. Over the radio he picked up a c a l l from a troller that was operating off the North East coast of Vancouver Island, who was excitedly claiming that he had run into a \"fucking great school of sockeye\". Ernie calcu-lated that the run coming round the tip of Vancouver Island would head directly to Rivers Inlet and be further out from the islands he was fishing off. Conditions were very rough, near gale force winds and rain with a 12-foot groundswell, but Ernie pulled further out to sea, made a set and took 70 and 50 sockeye on two sets, which was exceptional for daylight sets (generally expected to average 25-30). So despite his disastrous start Ernie's f i r s t week at Rivers Inlet gave him a catch well above average. Each decision he had made - 133 -both on low and high information was high r i s k i n that he was attempt-ing to maximize f i s h intake even when the costs could have been con-siderable. The second week at Rivers I n l e t he fished the boundary, put-tin g on his head net. During the closed period between week I and week I I , he learned that.the high boats were those that fished the boundary, (though the lowest boats also fished the boundary). He was f u l l y aware of the cost factor of getting fouled by other boats but he stated that \"where there's f i s h , y o u ' l l f i n d Ernie M c A l l i s t e r \" and as an afterthought \"Anyway my boat's bigger than any-one e l s e ' s . \" I did not see him f i s h i n g Johnston S t r a i t s , but i n the f i r s t week of the Northern closures he headed south to the Fraser. This de-c i s i o n was i n terms of high information. He knew the f i s h e r i e s depart-ment were doubtful about the size of the run and that p r i o r to the Nor-thern closures very small catches had been taken. He was also aware of the cost of t r a v e l l i n g and of Fraser junk and t r a f f i c hazards. But he calculated that a l l things being equal he could only get a moderate catch i n Johnston S t r a i t s , while there was a s l i g h t chance of making a k i l l i n g at the Fraser. He agreed that i t would be safer and more \" r e l i a b l e \" to stay i n Johnston S t r a i t s but observed that \" i f you didn't got for the big catches, you would never get them.\" - 134 -Table XXI: Risk Taking by Ernie McAllister Summary: Ernie McAllister HEd/ H. Inf Circ./ H. Inv. R./ HW Strategy Risks Location Factors of Decision Smith's Inlet Peak expected, lose 1/2 day's fishing by leaving RI. Rivers Inlet Week I Rivers Inlet Week II 1. Peak Expected, a l l fishing locations chosen hazardous. 2. Using outside net inside, hoped to optimise two different locations; cost of net being inappropriate. 3. Fishing inside off reef, cost of fouling on rocks. 4. Fishing by jumpers inside; cost factor of inappropriate net. 5. Fishing off reefs outside -cost in time of delivering fish. 6. Fishing further out to sea -large school, dangerous conditions. 1. Fished boundary - last weeks high boats were here, cost of fouling. INF H H L L H H H Strategy Max Gain Max Gain Max Gain Max Gain Max Gain Max Gain Max Gain Max Gain Risk H H H H H H H H N. Closures 1. Went to Fraser. Week I H Max Gain H - 135 -Case No. 2: Neil Anderson I went out with Neil on his boat, the Storm Winds, for one day at Rivers Inlet and for two days in Johnston Straits. He valued his boat at $4,500 and had bought i t eight years ago. I t was 34 feet long with an 8 foot beam, with a gear rather than a hydraulic drum. In the eight years that he had had the Storm Winds he had not invested in any new technology -- had merely maintained equipment and replaced nets. He rejected the idea of investing in new technology on two grounds -- the prohibitive price and the fact that i t would only mar-ginally increase his fish intake. Thus, a hydraulic drum was not con-sidered as a replacement for his gear drum -- though with the former he could make and pick up sets much faster. I t was cheaper:, to maintain his gear drum than to invest $400 in a hydraulic system. Similarly, depth sounders, radar, etc., had not been adopted chiefly because of their prohibitive price. This concern with keeping costs down was exhibited in his choice of nets -- he had a combination inside/outside net for Rivers Inlet (a dull green net) thus cutting down the cost of owning two nets for two fishing areas, though i t was not as good (in terms of optimising catch) as two separate nets. He exploited a range between Namu and the Fraser River. The major source of information about technology used by Neil was other fishermen, though he occasionally dropped in to see his net - 136 -company to shoot the breeze. He did not use the fisheries office or companies as information sources, neither did he subscribe to any jour-nals though he occasionaly read the union newspaper i f one was l e f t lying around in the taverns. Thus his score of three puts him in the low category with respect to access to information circuits. He had been brought up on an Indian Reservation near C h i l l i -wack -- his father was white, his mother Salish. He had turned his hand to a variety of professions -- logging, clerk, nightwatchman --and had been in the fishing industry for fifteen years. A l l his exper-ience had been gained on gillnetters. He was hoping for a good season, and rated his desire to in-crease his fish intake as high. Against this had to be weighed his fear of fouling his nets or boat as he could not afford heavy mainte-nance costs. He had grade 10 education which puts him in the MEd category, though his wealth ranking was low. His taxable income worked out at around $6,000, and he was supporting a wife from whom he was separated. He drew unemployment benefit during the winter, and one of his preoccupations was \"fishing for stamps\". In order to qualify for this benefit, he had to show that he had been employed f u l l time as a commercial fisherman during the fishing season. This preoccupation be-came evident in the second week of the Northern closures. During this week Rivers Inlet was l e f t open, everything else was closed (including Johnston Straits) except the Fraser. - 137 -The run at Rivers Inlet was over, so one could not expect many fish there. The Fraser River run had not materialised and boats had done very badly the week before. So Neil could either stay tied up in Alert Bay, go to Rivers Inlet where there were few fish, or go to Fraser where there were few fish. I t was cheaper to go to Rivers Inlet than to the Fraser in terms of travelling costs and there were no t r a f f i c hazards. To stay tied up in Alert Bay might have been a good strategy as there were no fis h anywhere -- but he would have forfeited a stamp and run the risk of being disqualified for winter unemployment benefit. So in this situation he went to Rivers Inlet, caught 5 fis h and collected his stamp. His strategy was aimed entirely at minimizing costs. I t was cheaper^ and safer to go to Rivers Inlet as against the Fraser; i t was also a cost-minimizing strategy to go to Rivers Inlet, though he knew he would not catch much, because staying tied up may endanger his unemploy-ment benefits. At Rivers Inlet he never fished the boundary as he was afraid of getting his net torn and snagged; neither did he fis h the reefs. He calculated that he would always be able to catch something by playing i t safe in the main channels and by fishing outside under calm conditions. He relied heavily on what other boats did around him -- when they set, he set, when they picked up he picked up. He liked other boats around him, because then he f e l t that not everybody could be wrong about fishing a particular area. - 138 -He had the radio on a l l the time and was constantly asking other fishermen how they were doing. He did not have a depth sounder. He always chose a low risk strategy when information was low. Once when he saw four jumpers off a reef he almost made a set, but a strong cur-rent into the reef put him off so he went 300 yards down the inlet be-fore setting, and was very nervous about drifting into the reef. He only l e f t the set in for a half-hour before picking i t up. Apart from this one instance, he never took a high risk,irres-pective of information. On another occasion, I observed two jumpers near a shoal -- he ignored them and stayed in the main channel. In Johnston Straits Neil did not fish the beach for fear of seiners, nor near kelp beds. He fished Growler's Cove and Robson's Point, both of which were safe areas. He also travelled back to Alert Bay during the day. There were two reasons for this: day sets in John-ston Straits were always unproductive, averaging 6-10 per set; and also : there was a tremendous amount of day t r a f f i c . I t did not seem worth his while to dodge the t r a f f i c hazards for a mere six fi s h per set, though some day sets would take as many as twenty fish. The f i r s t week of the Northern Closures, Neil elected to fish the Straits rather than go to the Fraser on the grounds that his costs (in travel and time) and in terms of hazards, would be less in the Straits than at the Fraser. - 139 -. Table XXII: Risk Taking by Neil Anderson Strategy Risks Location Factors of Decision Rl 1. Fished main channel, knew more fish at boundary/ reefs, afraid of costs to fouled gear, etc. 2. Fished safe spots -- main channel. 3. Fished near reef when jumpers sighted. 4. Ignored jumpers when near reef. J.S. 5. Never fished shore. 6. Never fished kelp. 7. Fished Growler's Cove. 8. Fished Robson's Point. 9. Returned to Alert Bay during day. North Clo- 10. Fished Areas 12 and 13 sures These two cases represent polar extremes in terms of risk-tak-ing and could be regarded as \"ideal types\". However, not a l l the other fishermen conformed in risk-taking to either one of these stereotypes. Some took high investment risks and low strategy risks, others took high investment risks and a mixture of low and high strategy risks. The results of the analysis of the other 24 boats are presented in tab-ular form (Table XXIII) -- the case histories of Ernie McAllister and Neil Anderson were merely to indicate the c r i t e r i a by which I categor-ised individuals and their risk-taking behavior. Inf Strategy Risk H Min Loss L L Min Loss L H Max gain H H Min Loss L H & L Min Loss L H & L Min Loss L L Min Loss L L Min Loss L H Min Loss L H Min Loss L Boat Skipper 4J Xi crj CO c o •r-1 4J cd O W c o •I-l 4-1 cn C +J j-i a) • H B D 4-1 co o co M M CU CO •rH > T-l fl o fl erf H H S-l o Strategy Risks LINF HTNF H L H L CJ to chance of losing $4. To determine how you would do on this gamble, you w i l l spin a pointer on the disc to determine your winnings. Only i f the pointer stops in the shaded area w i l l you win $6. (The shaded area covers 607> of the total disc and accurately portrays the chance element.) To ascertain your losses, you w i l l again spin the pointer on the disc. You w i l l lose $4 only i f the pointer stops in the shaded area, which represents 807, of the disc. - 228 -Your earnings from each bet w i l l consist of any positive amount of money gained minus the $1 that you have to pay the experimen-ter for playing the bet. For instance, i f you played the gamble above, and spun the pointer and i t stopped in the 607> chance of winning area, you w i l l have won $6. If you then spin the pointer and i t stops outside the 807o chance of losing area, you w i l l not lose $4; so your net gain is $6 -$0, i.e. $6. Subtracted from this sum is the $1 stake necessary to play the bet. So you would receive $6 - $1, i.e. $5. Your total earnings w i l l consist of any positive amount you may have accumulated. To stay in the \"game\" however you must always have at least a $1 stake before you can play the next bet. This means that i f you won $7 on the f i r s t gamble played, but lost your stake on the next one you played -- then you would be out of the game. This is because you would not have the $1 stake that i s necessary to enable you to play any subsequent bets. ARE THERE ANY QUESTIONS Ref: L.Inf Bet 1: A NUMBER OF PEOPLE HAVE TRIED THIS BET; A FEW OF WHOM HAVE BEEN POOR GAMBLERS. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? - :2t9 -Bet 2: THOSE PEOPLE WHO HAVE TRIED THIS BET RARELY DO WELL. 1. Do you think this is an attractive gamble? , 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 3: OF THOSE WHO HAVE TRIED THIS BET, SOME HAVE BEEN POOR GAMBLERS, SOME HAVE BEEN GOOD ONES. 1. Do you think this is an attractive gamble? ' 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 4: SOME PEOPLE HAVE TRIED THIS BET 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 5: OF THE FEW PEOPLE WHO HAVE TRIED THIS BET, SOME HAVE BEEN POOR GAMBLERS. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 6: THIS BET IS TRIED BY PEOPLE WHO ARE GENERALLY CONSIDERED TO BE POOR GAMBLERS. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? - 230 _ Bet 7: QUITE A FEW PEOPLE HAVE TRIED THIS BET, SOME HAVE BEEN GOOD GAMBLERS. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 8: ONLY A FEW PEOPLE HAVE TRIED THIS BET, SOME HAVE BEEN POOR GAMBLERS. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 9: VERY FEW PEOPLE HAVE TRIED THIS BET, ALL HAVE BEEN POOR GAMBLERS. 1. Do you think this i s an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 10: BOTH GOOD AND POOR GAMBLERS HAVE TRIED THIS BET. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 11: SOME PEOPLE HAVE TRIED THIS BET. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? 231 -Bet 12: SOME GOOD GAMBLERS HAVE TRIED THIS BET. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 13: VERY FEW PEOPLE HAVE TRIED THIS BET. 1. Do you think this i s an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 14: OF THE PEOPLE WHO HAVE TRIED THIS BET, MOST ARE CONSIDERED TO BE GOOD GAMBLERS. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 15: A FAIR NUMBER OF PEOPLE HAVE TRIED THIS BET, A FEW OF WHOM HAVE BEEN POOR GAMBLERS. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 16: SOME GOOD GAMBLERS HAVE TRIED THIS BET. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? - 232 -Bet 17: A NUMBER OF PEOPLE HAVE TRIED THIS BET. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 18: MANY PEOPLE HAVE TRIED THIS BET, SOME OF WHOM HAVE BEEN GOOD GAMBLERS. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 19: THERE HAVE BEEN SOME PEOPLE WHO HAVE TRIED THIS BET, THOUGH THEY ARE CONSIDERED TO BE POOR GAMBLERS. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 20: OF THE PEOPLE WHO HAVE TRIED THIS BET THERE HAVE BEEN MORE POOR GAMBLERS THAN GOOD GAMBLERS. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 21: FEW PEOPLE HAVE TRIED THIS BET, THOSE WHO HAVE ARE CONSIDERED TO BE POOR GAMBLERS. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 22: POOR GAMBLERS SEEM TO BE THE PEOPLE WHO HAVE TRIED THIS BET THE MOST. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness seal -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 23: A NUMBER OF GOOD GAMBLERS HAVE TRIED THIS BET. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness s;ale -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 24: OF THE PEOPLE WHO HAVE TRIED THIS BET, SOME HAVE BEEN POOR GAMBLERS. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness seal -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 25: GOOD GAMBLERS HAVE RARELY TRIED THIS BET. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness seal -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 26: QUITE A NUMBER OF PEOPLE HAVE TRIED THIS BET. 1. Do you think this i s an attractive gamble? 2. How would you rate this on a degree of attractiveness seal -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? - 234-; -Bet 27: GOOD GAMBLERS HAVE STAYED AWAY FROM THIS BET. 1. Do you think this i s an attractive gamble? 2. How would you rate this on a degree of attractiveness scale -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Ref/M.Inf INSTRUCTIONS TO SUBJECTS I would like to thank you for helping me in this experiment. The experiment you are about to participate in calls for you to make evaluation about different gambles that w i l l be presented to you. The gamble has four components, an amount to win ($ w); the chance of winning (C w); an amount to lose ; and a chance of loosing ( C L ) . I t can be represented by visualising two discs: - 235 -The l e f t hand disc represents the chances of winning a certain amount (60% chance of winning $4); the right hand disc represents the chances of losing a certain amount (40% chance of losing $5). Each gamble has these four components. The experimental task is for you to decide which of the gam-bles in the questionnaire you would play for real stakes. In order to do this you w i l l be given a salary at the beginning of each session. Your salary may be small, $1, moderate $3, or high $5. By a random assignment your salary w i l l be $1, $3, $5, in this session. You w i l l need your salary to bet with, with the result that you may end up making a lot of money, none at a l l , or somewhere in between. There is a financial obligation placed upon you in that for any bet you decide to play, you w i l l have to pay the experimenter $1 out of any possible winnings. This means that i f you decided to play a particular bet and i n fact won $10 you would receive $10 - $1, i.e. $9. This $1 obliga-tion remains constant for every bet you decide to play. To make i t clear what I want you to do, examine the bet be-low. You w i l l see a verbal description of a gamble, and a series of questions: BET A: GOOD CHANCE of HIGH WINNINGS, FAIR CHANCE of MODERATE COSTS. (C L) ($ L) (C w) ($ w) - 236. -1. Do you think this is an attractive gamble? 2. How would you rate i t on a degree of attractiveness scale? 0 -5 +5 very indifference very unattractive attractive 3. Would you bet on this gamble? The f i r s t question concerns your evaluation of the gamble. The second question refers to your rating of the gamble. The indiffer-ence point at zero represents a situation where the gamble appears nei-ther attractive nor unattractive. The third question calls for your decision. The verbal descriptions of each component cover a range of values as follows: A GOOD CHANCE of winning or losing has a range of 70% to 100% A FAIR CHANCE of winning or losing has a range of 40% to 70% A POOR CHANCE of winning or losing has a range of 0% to. 40% HIGH winnings or losses range from $5 to $10 MODERATE winnings or losses range from $1 to $5 LOW winnings or losses range from $0 to $1 (These ranges are written on the blackboard for your reference.) This means that Bet A: GOOD CHANCE OF HIGH WINNINGS; FAIR CHANCE OF MODERATE COSTS could be either (1) 100% chance of winning $10; 40% chance of losing $1.50 OR (2) 71% chance of winning $5; 69% chance of losing $4.50. You w i l l note that one verbal description covers a wide range of p o s s i b i l i t i e s . (2) is very different to (1), yet they are both covered by the same verbal description. - 237. \" As you go through the questionnaire, regard each gamble as discrete in that you w i l l be considering each one in terms of having your salary in your hand and being obligated to pay the experimenter $1 for every bet played. To determine your earnings for taking part in the experiment, those gambles that you decide to bet on in the questionnaire, you w i l l in actual fact play for real stakes. Say for instance, you decided (in the questionnaire) to bet on the gamble: FAIR CHANCE OF HIGH WINNINGS; HIGH CHANCE OF MODERATE LOSSES and that this verbal description meant: 60% chance of winning $10; 807> chance of losing $2. To determine how you would do on this gamble, you w i l l spin a pointer on the disc to determine your winnings. Only i f the pointer stops in the shaded area w i l l you win $10. (The shaded area covers 607> of the total disc and accurately portrays the chance element.) To ascertain your losses, you w i l l again spin the pointer on the disc. You w i l l lose $2 only i f the pointer stops in the shaded area, which represents 807o of the disc. - 238. -Your earnings from each bet w i l l consist of any positive amount of money gained minus the $1 that you have to pay the experimenter for playing the bet. For instance i f you played the gamble above, and spun the pointer and i t stopped in the 60% chance of winning area, you w i l l have won $10. If you then spin the pointer and i t stops in the 80% chance of losing area, you w i l l lose $2; so your net gain is $10 - $2, i.e. $8. Subtracted from this sum is the $1 financial obligation placed upon you by the experimenter. . So you would receive $7. Your total earnings w i l l be calculated from the results of a l l the gambles you decided to play. ARE THERE ANY QUESTIONS? Ref: M.Inf Bet 1: FAIR CHANCE OF HIGH PAYOFF, FAIR CHANCE OF HIGH LOSSES 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness s;ale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? - :239 -Bet 2: POOR CHANCE OF HCGH PAYOFF, GOOD CHANCE OF HIGH LOSSES. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 3: FAIR CHANCE OF MODERATE PAYOFF, FAIR CHANCE OF MODERATE LOSSES. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 4: POOR CHANCE OF MODERATE PAYOFF, GOOD CHANCE OF LOW LOSSES. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 5: POOR CHANCE OF HIGH PAYOFF, GOOD CHANCE OF MODERATE LOSSES. 1. Do you think this i s an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 6: FAIR CHANCE OF LOW PAYOFFS, FAIR CHANCE OF MODERATE LOSSES. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? - 240; -Bet 7: GOOD CHANCE OF MODERATE PAYOFFS, LOW CHANCE OF MODERATE LOSSES. 1. Do you think this is an attractive gambe? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 8: LOW CHANCE OF LOW PAYOFFS, HIGH CHANCE OF LOW LOSSES. 1. Do you think this i s an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 9: POOR CHANCES OF MODERATE PAYOFFS, GOOD CHANCE OF HIGH LOSSES. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 10: FAIR CHANCE OF HIGH PAYOFFS, FAIR CHANCE OF MODERATE LOSSES. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 11: GOOD CHANCE OF LOW PAYOFFS, LOW CHANCE OF LOW LOSSES. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? - 24:1 -Bet 12: FAIR CHANCE OF HIGH PAYOFFS, FAIR CHANCE OF LOW LOSSES. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness seal -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 13: LOW CHANCE OF LOW PAYOFFS, HIGH CHANCE OF HIGH LOSSES. 1. Do you think this is an attractive gamble? 2. How would you rate i t on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 14: GOOD CHANCE OF HIGH PAYOFFS, LOW CHANCE OF LOW LOSSES. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness seal -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 15: GOOD CHANCE OF LOW PAYOFFS, LOW CHANCE OF MODERATE LOSSES 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness seal -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 16: GOOD CHANCE OF MODERATE PAYOFFS, LOW CHANCE OF LOW LOSSES. 1. Do you think this is an attractive gamble? 2. How would you rate i t on a degree of attractiveness scale? -5 0 +5 1 1 1 . 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? - 242 -Bet 17: FAIR CHANCE OF LOW PAYOFFS, FAIR CHANCE OF LOW COSTS 1. Do you think this is an attrative gamble? 2. How would you rate i t on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 18: GOOD CHANCE OF HIGH PAYOFF, LOW CHANCE OF MODERATE LOSSES 1. Do you think this is an attractive gamble? 2. How would you rate i t on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 19: FAIR CHANCE OF LOW PAYOFFS, FAIR CHANCE OF HIGH LOSSES 1. Do you think this is an attractive gamble? '2. How would you rate this on a degree of attractiveness scale -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 '3. Would you bet on this gamble? Bet 20: POOR CHANCE OF HIGH PAYOFFS, GOOD CHANCE OF LOW LOSSES 1. Do you think this is an attractive gamble? ... 2. How would you rate this on a degree of attractiveness scale -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 21: GOOD CHANCE OF LOW PAYOFFS, LOW CHANCE OF HIGH LOSSES. 1. Do you think this i s an attractive gamble? 2. How would you rate this on a degree of attractiveness scale -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? - 243 -Bet 22: FAIR CHANCE OF MODERATE PAYOFFS, FAIR CHANCE OF HIGH LOSSES. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness seal -5 0 +5 1 1 1 1 1 1 . 1 1 . I l l 3. Would you bet on this gamble? Bet 23: FAIR CHANCE OF MODERATE PAYOFFS, FAIR CHANCE OF LOW LOSSES. 1. Do you think this is an attractive gamble? 2. How would you rate i t on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 24: GOOD CHANCE OF MODERATE PAYOFFS, LOW CHANCE OF HIGH LOSSES. 1. Do you think this is an attractive gamble? 2. How would you rate this in a degree of attractiveness seal -5 0 + 5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 25: LOW CHANCE OF MODERATE PAYOFFS, GOOD CHANCE OF M3ERATE LOSSES. 1. Do you think this is an attractive bet? 2. How would you rate i t on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Bet 26: GOOD CHANCE OF HIGH PAYOFFS, LOW CHANCE OF HIGH LOSSES. 1. Do you think this is an attractive gamble? 2. How would you rate this on a degree of attractiveness seal -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? - 244 -Bet 27: LOW CHANCE OF LOW PAYOFFS, HIGH CHANCE OF MODERATE LOSSES. 1. Do you think this i s an attractive gamble? 2. How would you rate this on a degree of attractiveness scale? -5 .0 +5 1 1 1 1 1 1 1 1 1 1 1 3. Would you bet on this gamble? Ref/H.Inf INSTRUCTIONS TO SUBJECTS I would like to thank you for helping me in this experiment. The experiment you are about to participate in calls for you to make evaluations about different gambles that w i l l be presented to you. The gamble has four components, an amount to win ($ w); the chance of winning ( C W ) ; an amount to lose ($L) ; and a chance of losing ( C L ) . It can be represented by visualising two discs below: The l e f t hand disc represents the chances of winning a certain amount (607o chance of winning $4) ; the right hand disc represents the chances of losing a certain amount (407> chance of losing $5). - 245 -Each gamble has these four components. The experimental task is for you to decide which of the gam-bles in the questionnaire you would play for real stakes. In order to do this you w i l l be given a salary at the beginning of each session. Your salary may be small, $1; moderate $3; or high $5. By a random assignment your salary w i l l be $1, $3, $5, i n this session. You w i l l need your salary to bet with, with the result that you may end up making a lot of money, none at a l l , or somewhere in be-tween. There is a financial obligation placed upon you in that for any bet you decide to play, you w i l l have to pay the experimenter $1 out of any possible winnings. This means that i f you decided to play a particular bet and in fact won $10 you would receive $10 - $1, i.e. $9. This $1 obliga-tion remains constant for every bet you decide to play. To make i t clear what you have to do, examine the bet below. You w i l l see a verbal description of a gamble, and a series of questions. Take special note of the f i r s t question as this enables you to get more information about the gamble. Bet A: GOOD CHANCE of HIGH WINNINGS, FAIR CHANCE of MODERATE COSTS. (V 1. Which two of the four components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? - 246 -3. How would you rate i t on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? The f i r s t question permits you to gain more information about the gamble. For instance, i f you want to know what the chances of lo-sing (C^) and the amount to- win ($ w) are, ask the experimenter and he w i l l give you the precise figure, e.g. 60% chance of losing; 4$ to win. The second question concerns your evaluation of the gamble. The third question refers to your rating of the gamble. The indifference point at zero represents a situation where the gamble ap-pears neither attractive nor unattractive. The fourth question calls for your decision. The verbal description of each component covers a range of values as follows: A GOOD chance of winning or losing has a range of 707. to 100% A FAIR chance of winning or losing has a range of 40% to 707. A POOR chance of winning or losing has a range of 0% to 407> HIGH winnings or losses range from $5 to $10 MODERATE winnings or losses range from $1 to $5 LOW winnings or losses range from $0 to $1. (These ranges are written on the blackboard for your reference). This means that Bet A: GOOD CHANCE of HIGH WINNINGS, FAIR CHANCE of MODERATE COSTS could be either (1) 1007, chance of winning $10, 407. chance of losing $1.50 OR (2) 717, chance of winning $5 , 697. chance of losing $4.50. You w i l l note that one verbal description covers a wide range of possi-b i l i t i e s . (2) is very different to (1), yet they are both covered by the same verbal description. - 247 -As you go through the questionnaire, regard each gamble as discrete in that you w i l l be considering each one in terms of having your salary in your hand and being obligated to pay the experimenter $1 for every bet that you play. To determine your earnings for taking part in the experiment, those gambles that you decide to bet on in the questionnaire, you w i l l in actual fact play for real stakes. Say for instance, you decided (in the questionnaire) to bet on the gamble: FAIR CHANCE of HIGH WINNINGS; HIGH CHANCE of MODERATE LOSSES and that this verbal description meant: 60% chance of winning $10; 80% chance of losing $2. To determine how you would do on this gamble, you w i l l spin a pointer on the disc to determine your winnings. Only i f the pointer stops in the shaded area w i l l you win $10. (The shaded area covers 607o of the total disc and accurately portrays the chance element.) To ascertain your losses, you w i l l again spin the pointer on the disc. You w i l l lose $2 only i f the pointer stops in the shaded area, which represents 807> of the disc. '248-$ L = ?2 Your earnings from each bet w i l l consist of any positive amount of money gained minus the $1 that you have to pay the experimenter for playing the bet. For instance, i f you played the gamble above, and spun the pointer and i t stopped in the 60% chance of winning area, you w i l l have won $10. If you then spin the pointer and i t stops in the 80% chance of losing area, you w i l l lose $2; so your net gain is $10 - $2, i.e. $8. Subtracted from this sum is the $1 stake necessary to play the bet. So you would receive $7. Your total earnings w i l l consist of any positive amount you may have accumulated. To stay in the \"game\" however, you must always have at least a $1 stake before you can play the next bet. This means that i f you won $7 on the f i r s t gamble played, but lost $8 on the next one you played -- then you would be out of the game. This is because you would not have the $1 stake that is necessary to enable you to play any subsequent bets. ARE THERE ANY QUESTIONS? - 249 -Ref: H.Inf Bet 1: FAIR CHANGE OF HIGH PAYOFF, FAIR CHANCE OF HIGH LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gambles? 3. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? Bet 2: POOR CHANCE OF HIGH PAYOFF, GOOD CHANCE OF HIGH LOSSES 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? 3. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? Bet 3: FAIR CHANCE OF MODERATE PAYOFF, FAIR CHANCE OF MODERATE LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this i s an attractive gamble? 3. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this, gamble? Bet 4: POOR CHANCE OF MODERATE PAYOFF, GOOD CHANCE OF LOW LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? 3. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? 250 Bet 5: POOR CHANCE OF HIGH PAYOFF, GOOD CHANCE OF MODERATE LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? 3. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? Bet 6: FAIR CHANCE OF LOW PAYOFFS, FAIR CHANCE OF MODERATE LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? 3. How would you rate this on a degree of attractiveness scale? -5 0 + 5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? Bet 7: GOOD CHANCE OF MODERATE PAYOFFS, LOW CHANCE OF MODERATE LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this i s an attractive gamble? 3. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1' 1 1 1 1 4. Would you bet on this-gamble? Bet 8: LOW CHANCE OF LOW PAYOFFS, HIGH CHANCE OF LOW LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? 3. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? - 251 -Bet 9: POOR CHANCE OF MODERATE PAYOFFS, GOOD CHANCE OF HIGH LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? 3. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? Bet 10: FAIR CHANCE OF HIGH PAYOFFS, FAIR CHANCE OF MODERATE LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? 3. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 T 1 1 1 1 1 1 4. Would you bet on this gamble? Bet 11: GOOD CHANCE OF LOW PAYOFFS, LOW CHANCE OF LOW LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this i s an attractive gamble? 3. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? Bet 12: FAIR CHANCE OF HIGH PAYOFFS, FAIR CHANCE OF LOW LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this i s an attractive gamble? 3. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? - 252' Bet 13: LOW CHANCE OF LOW PAYOFFS, HIGH CHANCE OF HIGH LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this i s an attractive gamble? 3. How would you rate i t on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? Bet 14: GOOD CHANCE OF HIGH PAYOFFS, LOW CHANCE OF LOW LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? • 3. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? Bet 15: GOOD CHANCE OF LOW PAYOFFS, LOW CHANCE OF MODERATE LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this i s an attractive gamble? 3. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? Bet 16: GOOD CHANCE OF MODERATE PAYOFFS, LOW CHANCE OF LOW LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? 3. How would you rate this on a degree of attractiveness scale? -5 0 + 5 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? Bet 17: FAIR CHANCE OF LOW PAYOFFS, FAIR CHANCE OF LOW COSTS. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? 3. How would you rate i t on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? - ?53 -Bet 18: GOOD CHANCE OF HIGH PAYOFFS, LOW CHANCE OF MODERATE LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? 3. How would you rate i t on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? Bet 19: FAIR CHANCE CF LOW PAYOFFS, FAIR CHANCE OF HIGH LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? 3. How would you rate this on a degree of attractiveness seal -5 0 +5 l l l l l l l l l l ' l 4. Would you bet on this gamble? Bet 20: POOR CHANCE OF HIGH PAYOFFS, GOOD CHANCE OF LOW LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? 3. How would you rate this on a degree of attractiveness seal -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? Bet 21: GOOD CHANCE OF LOW PAYOFFS, LOW CHANCE OF HIGH LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? 3. How would you rate this on a degree of attractiveness seal -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? - 25,4 -Bet 22: FAIR CHANCE OF MODERATE PAYOFFS, FAIR CHANCE OF HIGH LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? 3. How would you rate this on a degree of attractiveness scale -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? Bet 23: FAIR CHANCE OF MODERATE PAYOFFS, FAIR CHANCE OF LOW LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? 3. How would you rate i t on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? Bet 24: GOOD CHANCE OF MODERATE PAYOFFS, LOW CHANCE OF HIGH LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? 3. How would you rate this in a degree of attractiveness scale -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? Bet 25: LOW CHANCE OF MODERATE PAYOFFS, GOOD CHANCE OF MODERATE LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? 3. How would you rate i t on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? - '255. -Bet 26: GOOD CHANCE OF HIGH PAYOFFS, LOW CHANCE OF HIGH LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? 3. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? Bet 27: LOW CHANCE OF LOW PAYOFFS, HIGH CHANCE OF MODERATE LOSSES. 1. Which 2 of the 4 components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? 3. How would you rate this on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1 4. Would you bet on this gamble? C 0 M M E N T S: Table IV: Bets Categorised Per Subject Subjects Cell 1 SR FR VR Cell 2 SR FR VR Cell 13 SR FR VR Cell 4 SR FR VR Cell 5 SR FR VR Cell 6 SR FR VR Cell 7 SR FR VR Cell 8 SR FR VR Cell 9 SR FR VR 1 2 4 1 — 5 6 7 1 — 8 2 — 7 4 — 8 4 1 6 2 — 2 2 3 4 6 6 4 — 7 4 — 8 5 — 8 2 — 8 4 — 3 4 5 4 7 2 — 7 2 — 8 4 — 5 8 2 — 8 5 2 4 4 5 7 7 3 — 8 3 — 7 2 — 5 8 4 — 7 3 — 5 2 5 3 6 1 — 7 8 4 — 5 1 — 7 4 — 8 4 1 6 2 3 — — 6 1 — 7.3 — 8 3 — 5 1 — 8 3 — 6 5 1 8 4 — 7 3 7 1 — 4 7 3 — 8 2 — 7 1-— 8 1 — 8 2 — 8 2 1 8 4 4 4 1 — 7 2 — 8 4 — 8 4 — - 7 1 — 7 6 — 8 3 — 9 2 4 1 — 6 1 — 6 1 — 8 1 — 8 2 — 7 4 — 7 1 — 8 5 1 10 0 3 5 0 — 6 1 — 7 3 — 8 3 — 8 4 — 8 3 — 8 5 — Total 25 41 3 — 46 3 — 65 16 — 74 23 — 74 27 — 68 23 — 75 33 2 77 3.7 5 M O I—I X H o CO l-tl o o 3 o Hi w X! T3 fD l-i H-B fD 0 fT fa t f Co ro U i ON SR = slightly risky FR = f a i r l y risky VR = very risky H CO cr M fD co I < M M - 257 -Table V: Frequency of Bets Played Per Cell Cell Cell Cell Cell Cell Cell Cell Cell Cell Bet Type Bet 1 2 3 4 5 6 7 8 9 7 4 10 8 9 8 10 7 10 10 10 1 2 2 3 10 9 8 9 9 S.R. 12 2 4 8 10 10 10 8 10 10 14 8 10 10 10 10 10 10 10 10 Bets 16 2 1 5. 8 8 8 7 9 10 18 2 7 7 10 10 10 10 10 10 23 6 6 5 7 8 10 8 8 8 26 .- 3 2 8 10 7 10 9 10 1 ,1 1 4 3 4 6 3 - 2 . 1 1 3 4 2 3 6 F.R. 4 - -- 2 -- 1 -- 1 4 11 - 1 -- 5 7 8 5 •6 7 Bets 15 - -- -- 1 3 4 3 6 5 17 - 1 2 1 2 1 20 - -- -- 4 1 3 4 2 21 - -- 1 1 1 1 2 24 - 3 5 .4 4 5 6 2 5 - 1 4 V.R. 6 -8 - 1 1 Bets 9 -13 -19 -22 -25 -27 -- 258 -Table VI: Raw Data Matrices Wealth Information LW MW HW L l 25 41 46 Ml 65 74 74 HI 68 75 77 LW MW HW L l 0 3 3 Ml 16 23 27 HI 23 33 37 LW MW HW L l 0 0 0 Ml 0 0 0 HI 0 2 5 (a) SR Bets (Slightly risky) (b) FR Bets (Fairly risky) (c) VR Bets (Very risky) Using the formula X.. = i j x 100 the raw data 1 J • (NT).. x (N , ).. v I i j rb i j are corrected for unequal c e l l frequency giving three new data matrices to which a fourth has been added which combines the two top levels of riskiness into one category. (X is the c e l l entry i n the i t * 1 row and j ^ column; B is ; the number of possible bets taken in that c e l l ; i s the number of possible bets that could have been taken in that c e l l where N^ . is the number of individuals in the c e l l and is the number of risky bets i n that cell.) - 259 -Table VII: Corrected Data Matrices LW MW HW L l 31.3 51.3 57.5 Ml 81.3 91.3 91.3 (a) SR HI 85 93.8 96.3 LW L l 0 MW 3.3 HW 3.3 Ml 17.8 25.6 30 HI 25.6 36.4 41 (c) FR LW L l 0 Ml 0 HI 0 MW 0 0 2 HW 0 0 5 (b) VR LW MW HW L l 0 3.3 3.3 Ml 17.8 25.6 30 HI 25.6 38.4 46 FR (d) and VR - 26.0 -APPENDIX IY . Chi Square. Median Tests on Propositions 1, 2 and 3. Proposition One Cells 1 and 2 The null hypothesis to be tested i s that the median amount of losses w i l l be the same for each c e l l . Table la Cell 1: Losses: Cell 2: Losses: Subject No. Bets Played Cell 1 Bets Played Cell 2 1 14,23 2 7,11,14,18,23 10 2 14,23 2 7,12,14 5.5 3 7,14,18,23 9 7,14,18,23,26 16.5 4 7,12,14,23 6.5 7,12,14,18,23 11 5 10,16 4.5 7,10,14,18,23 12.5 6 14,23 2 7,14,23 5.5 7 7,14,16 5.5 3,7,12,14,16,18,23,26 22 8 7,12,14,23 6.5 7,12,14,18 9 9 14,18 4.5 3,7,10,14,26 19 10 0 7,14,18 8 These results give a median of 6.5. When the median i s found, one sets up a table of observed frequencies for each c e l l . This is done by determining the number of losses above and below the over-all median, for each c e l l . Observed Frequencies Cell 1 Cell 2 Above Median 1 8 9 Below Median 9 2 11 10 10 20 - 261 -The expected frequencies have 507., of the cases above the med-ian and 507o of the cases below the median for both c e l l s ; viz. 5. Expected Frequencies Cell 1 Cell 2 Above Median 5 5 Below Median 5 5 ? k 2 2 X'' = (o-e) where X = chi square; k = no. of categories; o = 1 e observed frequency in a category; and e = expected frequency in a cate-gory. In this the number of df = K - 1 = 1 therefore a correction has to be made for continuity. This simply involves subtracting .5 from each absolute difference between an e and an o. The formula used i s X 2 = k /(o-e) - . 5 7 2 i e — 2 \"2. 2 X 2 = /(1-5) - . 5 7 2 + /(8-5) - .57 + /(9-5) - .5/ + /(2-5) - .57 = 3.5 2 + 2.5 2 + 3.5 2 + 2.5 2 5 = 7.46 which is significant at the (.01) level and enables us to reject the null hypothesis, i.e. there are very significant differences between the medians of both cells w.r.t. losses. - 262 -Chi Square Median Test on Cells 4 and 5 Table lb Subject No. Cell 4 Bets Played Cell4: Losses Cell 5 Bets Played Cell 5: Losses 1 7,12,14,16,18.26 17.5 7,10,11,12,14,16, 22 18,26 2 7,12,14,16,18,23 11 3,4,10,12,14,18,21, 37 23,24,26 3 7,11,12,14,16,18, 20.5 7,10,11,12,14,15,16, 25.5 20,23,26 18,26 4 7,11,12,14,15,16, 24 7,10,11,12,14,15,16, 34 18,20,23,26 18,23,24,26 5 10,11,12,14,16,18, 18.5 7,10,12,14,18,23,26 21 26 6 1,3,7,12,14,16,18, 30.5 1,7,10,11,12,14,16, 38 20,23,26 18,23,24,26 7 7,10,11,12,14,16, 30.5 5,7,10,11,12,14,16, 26.5 18,21,23,24 18,23,26 8 7,11,12,14,16,18, 20.5 7,10,11,12,14,15,16, 33 20,23,26 18,20,23,24,26 9 7,12,14,18,23,24, 25 7,10,11,12,14,16,18, 25.5 26 23,26 10 7,10,12,14,18,24, 22.5 3,10,12,14,16,17,18, 30.5 26 23,24,26 These results have a median of 25. Observed Frequencies Expected Frequencies Cell 4 Cell 5 Cell 4 Cell 5 Above Med. 2 8 10 Above Med. 5 5 Below Med. _8 _2 10 Below Med. 5 5 10 10 20 X 2 = k /(o-3) - .5/2 i = =_ e = /(2-5) - .5_72 + 7(8-5) - .572 + /(8-5) - .5_72 + /(2-5) - .5_7 , _ - - . -=5.00 which is sufficiently high to reject the null hypothesis at - 263 \" Cells 7 and 8 Table lc Subj ect Cell 7 Cell 7: Cell 8 Cell 8 No. Bets Played Losses Bets Played Losses 1 5,7,10,11,12,14,16, 27.5 1,5,7,10,12,14,15, 38.5 . 17,18,20,26 16,17,18,20,23,26 2 1,7,10,11,12,14,15, 42.5 1,7,10,12,14,16,18, 30.5 16,18,20,23,24,26 20,23,26 3 12,14,18,23,26 16.5 3,7,10,12,14,16,18, 37 23,24,26 4 10,12,14,18,26 16.5 1,3,7,10,12,14,16, 41.5 18,20,23,24,26 5 14,16,18,23,24,26 22.5 7,11,12,14,15,16, 31.5 18,20,21,23,26 6 1,7,10,11,12,14,16, 38 3,4,7,8,10,11,12, 31 18,23,24,26 14,15,18,23,24 7 7,10,11,12,14,16,18, 23 7,10,11,12,14,15, 26.5 23,26 16,18,23,26 8 7,10,12,14,15,18,23, 24.5 1,7,10,11,12,14,15, 49 26 16,17,18,21,24,26 9 7,10,11,14,16,18,21, 44.5 7,10,11,12,14,16,18, 22 23,24,26 26 10 3,7,10,12,14,16,18, 37.5 7,10,11,12,14,15,16, 34 20,23,24,26 18,23,24,26 These results give a median of 31.25. Observed Frequencies Expected Frequen< Cell 7 Cell 8 Cell 7 Cell 8 Above M 4 4 10 Above M 5 5 Below M _6 _j4 10 Below M 5 5 10 10 20 X 2 = k /(-o-e) . .5/2 = /(4-5) -.bj1 + 7(6-5) - .572 + /(6-5) - .5/2 + 7(4-5) - .5/ 5 5 5 5 5 5 5 5 - 264 -= .2 which means the null hypothesis cannot be rejected. This im-plies that at high levels of information there is no significant dif-ference between the way LW and MW operators take risks. Proposition 2 The null hypothesis being tested is that the median amount of gains w i l l be the same for each i c e l l . Cells 1, 2, 3 Table 2a Subj ect Cell 1: Cell 2: Cell 3: Cell 3 No. Gains Gains Bets Played Gains 1 11 23 12,14,16,18,23 29.5 2 11 18.5 7,12,14,18 26 3 22 28.5 7,12,14,18 26 4 22 29.5 7,10,12,14,18,23,26 44.5 5 11 29.5 14,18,23 18.5 6 11 14.5 3,7,12,14,16,23,26 36.5 7 14.5 44 7,14,16,23 18 8 22 26 4,7,12,14,18 29.5 9 15 29.5 4,7,10,12,14,16,18 40.5 10 0 18.5 7,12,14,16,18 29.5 These results give a median of 22. Observed Frequencies Expected Frequencies Cell 1 Cell 2 Cell 3 Cell 1 Cell 2 Cell 3 Above M 0 7 8 15 Above M 5 5 5 Below M 10 3 _2 15 Below M 5 5 5 10 10 10 30 df =3-1=2 - 265 -X 2 = \\ (o-e) 2 e = (o-5) 2 + (7-5) 2 + (8-5) 2 + (10-5) 2 + (3-5) 2 + (2-5) 2 5 5 5 5 5 5 = 5 + 4 + 9 + 5 + 4 + 9 5 5 5 5 = 15.2 which is high enough to reject the null hypothesis at the p = .01 level, i.e. there are significant differences between the cells as wealth varies with information constantly low. Test between 2 and 3 median = 29 ObS. Freq. Cell 2 Cell 3 Above M 4 6 10 Below M _6 _4 10 10 10 20 Exp. Freq. 5 5 5 5 X 2 = k /(o-e) - .572 i 5 2 = 1.5 + 4 5 = 1.8 - not significant. Test between 1 and 2 median = 21.3 - 266 -Obs. Freq. Cell 1 Cell 2 Above M 3 8 11 Below M _7 __2 _9 10 10 20 Exp. Freq. 5 5 5 5 X 2 = k i /(o-e) -e • l l2 = 1.5 5 .2 + 2.52 + 5 1.52 + 2.52 5 5 - 3.4 must reject the null hypothesis, though this is close enough to 3.8 to point to a significant difference. Chi Square Median Test on Cells 4, 5, 6 Table 2b Subject Cell 4: Cell 5: Cell 6: Cell 6: No. \" Gains Gains Bets Played Gains 1 37 44.5 7,10,12,14,15,16,18,23,24,26 49 2 33 62.5 1,3,7,10,12,14,17,18,23,24,26 60 3 40 46.5 1,3,7,10,11,12,14,15,16,18,23,26 61 4 41 53.5 7,10,11,12,14,16,18,23 49 5 42 44.5 1,7,10,11,12,14,16,17,18,23,24,26 60 6 63 60 7,11,12,14,18,23 30.5 7 46 52.5 7,10,11,12,14,16,18,23 41.5 8 49 61 3,7,10,11,12,14,15,16,18,21,23,26 54.5 9 40.5 49 7,10,11,12,14,15,16,18,23,26 50 10 44.5 51.5 3,7,10,11,12,14,16,18,23,24,26 56 These results give a median of 49. _ 267 Observed Frequencies Expected Frequencies Cell 4 Cell 5 Cell 6 Cell 4 Cell 5 Cell 6 Above M 1 5 6 12 Above M 5 5 5 Below M _9 _5 _4 18 Below M 5 5 5 10 10 10 30 df =3-1=2 e = (1-5) 2 + (5-5) 2 + (6-5) 2 + (9-5) 2 + (5-5) 2 + (4-5) 2 5 5 5 5 5 5 = lji + 0 + l + l_6 + 0 + l_ 5 5 5 5 = 6.8 which i s sufficiently high to reject the null hypothesis at p = .05. /N.B. Firther tests between 4 and 5, and then between 5 and 6 show that for 4 and 5 the null hypothesis is rejected at p = 0.01, but be-tween 5 and 6 the null hypothesis i s not rejected. This means that there are s t a t i s t i c a l l y significant differences in tolerance for risk between individuals in cells 4 and 5, but not between individuals In cells 5 and;^/ Cells 4 and 5 median = 46 - 26.8 -Above Med. Below Med, Obs. Freq. Cell 4 5 2 8 8 2 /(2-5) - .5) 2 + 7(8-5) ;72 .51 = 2.5Z x 4 5 = 5.00 Exp. Freq. Cell 4 5 5 5 5 5 Cells 5 and 6 median = 52 Obs. Freq. Cell 5 6 4 5 Above Med. Below Med. X 2 = ((4-5) = .1 .5) •+ 0 + ((4-5) ,5) 2 + 0 Exp. Freq. Cell 5 6 5 5 5 5 Chi Squ. Median Test on cells 7, 8, 9 Table 2c Subject No. 1 2 3 4 5 6 7 8 9 10 Cell 7: Gains 56.5 68.5 33.5 37.5 33 60 49 45.5 53.5 67.5 Cell 8: Gains 76.5 63 59 70 51 50.5 50 59.5 45.5 53.5 Cell 9: Bets Played 7,11,12,14,16,18,20,26 1,4,7,10,12,14,16,18,23,24,26 1,3,5,7,8,10,11,12,14,16,18,20,23,24> 26 1,4,7,10,11,12,14,16,18,26 3,5,7,10,11,12,14,15,16,13,23,24,26 1,3,7,10,11,12,14,15,16,18,23,26 1,5,7,10,12,14,15,16,18,23,26 1,3,6,10,12,14,16,18,23,24,26 3,4,5,7,10,11,12,14,15,16,18,23,24,26 4,7,10,11,12,14,15,16,17,18,23,24,26 Cell 9: Gains 44.5 62.5 78.5 56.5 64.5 61 64 62.5 68 58 - 2.6.9 -These numbers give a median of 58.5. Observed Frequencies Expected Frequencies C e l l 7 C e l l 8 C e l l 9 C e l l 7 C e l l 8 C e l l 9 Above M 3 5 7 15 Above M 5 5 5 Below M _7 _5 _3 15 Below M 5 5 5 10 10 10 30 df = 3 - 1 = 2 use X 2 = J ( o - e ) 2 e = (3 - 5 ) 2 + ( 5 - 5 ) 2 + ( 7 - 5 ) 2 + (7- 5 ) 2 + (5- 5 ) 2 + ( 3 - 5 ) 2 5 5 5 5 5 5 = 4 + 0 + 4 + 0 + 4 + 0 + 4 5 5 5 5 = 3.2 which i s not high enough for us to r e j e c t the n u l l hypothesis, i . e . the diff e r e n c e s i n tolerance f o r r i s k i n the 3 c e l l s i s not s t a t i s t i c a l l y s i g n i f i c a n t . C e l l s 7 and 8 median =53.5 Obs. Freq. Exp. Freq. C e l l 7 8 C e l l 7 8 Above M 4 5 5 5 Below M 6 5 5 5 X 2 = ((4-5) - . 5 ) 2 + 0 + (4-5) - .5 2 + 0 =0.1 C e l l s 8 and 9 median =59.5 270 Obs. Freq Cell 8 9 Above M 3 Below M 7 X 2 = ((3-5) - .5) 2 = 1.5Z + 4 5 = 1.8 7 Above M 3 Below M Exp. Freq. Cell 8 9 5 5 5 5 Proposition 3 . The null hypothesis being tested i s that the median amount of gains is the same for each case. Table 3a Subject No. Gains: Cell 1 Gains: Cell 4 Gains: Cell 1 11 37 56.5 2 11 33 68.5 3 22 40 33.5 4 22 41 37.5 5 11 42 33 6 11 63 60 7 14.5 46 49 8 22 49 45.5 9 15 40'. ,5 53.5 10 0 40. .5 67.5 These results give a median of 37.5 Observed Frequencies Expected Frequencies Cell 1 Cell 4 Cell 7 Cell 1 Cell 4 Cell 7 Above M 0 7 7 14 Above M 5 5 5 Below M 10 _3 _3 10 Below M 5 5 5 10 10 10 30 - 271 df =3-1=2 2 2 XZ = (c-e) Z (0-5) 2 + (7-5) 2 + (7-5) 2 + (10-5) 2 + (3-5) 2 + (3-5) 2 = 5+ _4 + 4. + 5+ _4+_4 5 5 5 5 = 13.5 which is high enough to reject the null hypothesis at p = .01. 2 [N.B. A chi squ. test on cells 4 and 7 (median = 44.5) gives an X of 1.8 which means we cannot reject the null hypothesis.] Cell 1 cf Cell 4 median = 20.5 Obs. Freq. Exp. Freq• 1 4 1 4 Above M 0 10 5 5 Below M 10 0 5 5 2 2 X = 4.5 x 4 5 = 16.1 Cell 4 cf Cell 7 median = 44.5 Obs. Freq. Exp. Freq• Above M 3 7 5 5 Below M 7 3 5 5 X 2 = ((3-5) - ,5) 2 x 4 5 = .45 x 4 = 1.80 • - 272 -Chi Square Median Test on cells 8, 5, 2 The null hypothesis being tested is that the median amount gains w i l l be the same for each c e l l . Table 3b Subject No. Cell 2 Cell 5 Cell 8 1 10 64.5 76.5 2 5.5 62.5 63 3 16.5 46.5 59 4 11 53.5 70 5 12.5 44.5 51 6 5.5 60 50.5 7 22 52.5 50 8 9 61 59.5 9 19 49 45.5 10 8 51.5 53.5 A cursory glance at the entries in cells 2 and 5 indicates the null hypothesis w i l l be rejected at p = 0.01 as the largest entry in c e l l 2 is smaller than the smallest entry i n c e l l 5. Cell 2 cf Cell 5 . median =22 Obs. Freq; Exp. Freq. Above M 0 10 5 5 Below M 10 0 5 5 X 2 = 4.52 x 4 = 16.1 5 Cell 5 cf Cell 8 median = 49 - 273 -Obs. Freq. Exp. Freq. 2 5 8 Above M 0 6 9 5 5 5 Below M 10 4 1 5 5 5 X 2 = 5 ^ + 1 + 1 6 + 5 ^ + 1 + 1 6 =10+6.8=16.8. 5 5 5 5 5 5 Between cells 5 and 8, the median i s located at 53. Observed Frequencies Expected Frequencies Cell 5 Cell 8 Cell 5 Cell 8 Above M 4 6 10 Above M 5 5 Below M _6 _4 10 Below M 5 5 10 10 20 df = 2-1 = 1 X 2 = k /(o-3) - .572 x ~ ~ 5 = 2(4-5) - .5T2 + 7(6-5) - .5_72 + /T6-5) - .5_72 + 7(4-5) - .572 5 5 5 5 = .2 which is too low to reject the null hypothesis, i.e. there are no s t a t i s t i c a l l y significant differences between cells 5 and 8. Chi Square Median Test on cells 3, 6, 9 The null hypothesis being tested is that the median amount of gains w i l l be the same for each c e l l . - 274 -Table 3b Subject No. C e l l 3 C e l l 6 C e l l 9 1 29.5 49 44.5 2 26 60 62.5 3 26 61 78.5 4 44.5 49 56.5 5 18.5 60 64.5 6 36.5 30.5 61 7 18 41.5 64 8 29.5 54.5 62.5 9 40.5 50 68 10 29.5 56 58 These scores give a median at 49.5 Observed Frequencies Expected Frequencies C e l l 3 C e l l 6 C e l l 9 C e l l 3 C e l l 6 C e l l 9 Above M 0 6 9 15 Above M 5 5 5 Below M IP. _4 _1 15 Below M 5 5 5 10 10 10 30 df = 3 - 1 = 2 X = ± (c-e) e = (0-5) 2 + (6-5) 2 + (9-5) 2 + (10-5) 2 + (4-5) 2 + (1-5) 2 5 5 5 5 5 5 = 5 + 1 + 1 6 + 5 + 1 + 1 6 5 5 5 5 = 16.8 which i s high enough to r e j e c t the n u l l hypothesis at p = .01 l e v e l . [The chi square median test was done on c e l l s 6 and 9 and gave 1.8 which i s not s i g n i f i c a n t enough to r e j e c t the n u l l hypothesis.] C e l l 3 cf C e l l 6 median = 41 - 275 -Obs. Freq. Exp. Freq. Cell 3 6 Above M 1 9 5 5 Below M 9 1 5 5 X 2 = ((1-5) -5) 2 x 4 = 3.52 x 4 = 9.7 5 5 Cell 6 cf Cell 9 median = 58 Obs. Freq. Above M 3 7 Below M 7 3 X 2 = ((3-5) - .5) 2 5 = 1.52 x 4 5 = 1.8. Exp. Freq 5 5 5 5 - 276 -APPENDIX V: Scale for Wealth Stratification for Alert Bay Gillnetters Unit LW MW HW 1 ingle Man 2,500* +3,000 +6,000 + 1 dependent 3,000 +3,000 +6,000 + 2 dependents 3,500 +3,000 +6,000 + 3 dependents 4,000 +3,000 +6,000 +4 dependents 4,500 +3,000 +6,000 + 5 dependents 5,000 +3,000 +6,000 + 6 dependents 5,500 +3,000 +6,000 + 7 dependents 6,000 +3,000 +6,000 + 8 dependents 6,500 +3,000 +6,000 [* Annual Income from a l l sources, not just fishing.] C D ADDENDUM The concern with methodological issues in this thesis is largely due to the view I have of the potential development of economic anthropology. A viewpoint which is not as yet widely accepted. However, I am alive to the fact that my concern with \"new directions\" has a number of d i f f i c u l t i e s , as the methodology of moving from one situation to another creates a number of problems with regard to conceptualization and data collection. In this addendum I intend to discuss b r i e f l y a number of problems dealing with level of analysis, comparability of data collected in different contexts, and level of data analysis, respectively. From an i n i t i a l substantive concern with Third World farmers making decisions on agricultural innovations I move to a higher level of generality and make statements about the way in which particular individuals w i l l take risks. The constraints upon risk taking specified at this second level of generality then require operational definition in specific substantive contexts to assess whether particular categorized individuals take risks in the manner predicted. As part of the cat-egorization process a state of nature conceptualization was used whereby every decision maker could be classified according to their location on two axes -- (a) subjective u t i l i t y and, (b) wealth (with regard to a stated outcome.) This typology placed a decision maker in one of nine c e l l s , and was a necessary prerequisite to ordering the phenomena dealt with in the dissertation. Only three of these cells were considered (LW/HU, MW/HU, HW/HU) as the definition of risk used precluded consideration ( i i ) of any c e l l other than High U t i l i t y . (Risk taking refers to behavior in situations where there is a desirable goal and a lack of certainty that i t can be attained, with attendent poss i b i l i t i e s of loss.) Thus the entire typology was not used, and in a sense could not be used, as the task of explicating the characteristics of every c e l l would require more than one thesis. This does not, however, deter from the value of a state of nature classification, as this typology could be used for a variety of problems in addition to those of risk taking. Data collection was in terms of operationally defining the scope conditions in different contexts. The scope conditions considered here place an individual decision maker within parameters of resources and subjective u t i l i t y with regard to some outcome, information and incentive conditions for any risk. The propositions predict the type of decision strategies that would be employed for given values of the above parameters. Certain problems existed in the measurement of incentive conditions. In the laboratory a l l four incentive conditions were measured in each decision situation, whereas in the f i e l d I measured u t i l i t y functions rather than discrete incentive conditions. (A u t i l i t y function is the probability times the value attached to the outcome). However, this was not a drawback as the propositions'could s t i l l be adequately tested. I was simply concerned with two things. Given that the u t i l i t y function for gains (i.e. Ps ,Us) attached to fishing in area A is greater than that for fishing in area B, what does the fisherman do -- f i s h A or f i s h B. The second aspect i s , is his ( i i i ) decision to f i s h A or B a function of the relative tolerance for losses (the negative u t i l i t y functions). From these two considerations I could infer whether a fisherman took a High or a Low Risk in each decision situation. Comparability between f i e l d and laboratory contexts was effected by establishing that the objective nature of the duplex gambles corresponded closely to the subjective evaluation of them (in ranking the 27 duplex gambles into categories of Low, Medium and High Risk there was 80% agreement among a sample of subjects). This was an important point as the incentive conditions recorded in the f i e l d corresponded to each fishermans unique payoff matrix. On a more general level', comparability is valid provided i t can be shown that the relevant scope conditions are present. Risk is defined in abstract terms as a decision situation characterized by desirable goals, uncertainty as to their achievement and a prospect of loss. To explain variation in risk I use the parameters of information, wealth and incentive conditions. Any situation (whether i f be an investment in a larger boat, a decision to put out a net or to take a particular bet) as long as i t has the above characteristics can be defined as risk. The relevance of the explanatory parameters is assessed'in terms of the operational definition of the parameter with regard to context. This means that the operational definition of information with respect to investment risks w i l l be different to the operational definition of information with respect to strategy risks. I emphasize that this is a substantive difference and not an analytic one. (iv) In terms of data analysis, the \"new directions\" for economic anthropology implied in this work suggest a battery of formal analysis that I do not possess. My own mathematical expertise is limited. Perhaps my methodological ambitions outstripped my command of the requisite technologies. I would submit, however, that I have proceeded in a manner consistent with a concern with \"new directions\" in economic anthropology. In methodological terms I have demonstrated the usefulness of borrowing analogies from economics. Also the testing procedure employed has implications for the manner in which anthropologists may conduct enquiry, as i t suggests that laboratory contexts are as legitimate a source of verification for anthropological problems as f i e l d contexts. It is from these two considerations that I offer a test case for methodology in economic anthropology. - 14 -Though the d i s c u s s i o n has moved from choice to exchange i t r e s t s on the same assumption of c a l c u l u s of maximization and r a t i o n -a l i t y i n choice. I t a l s o assumes that the t o o l s of economics — d e c i -s i o n theory and maximizing models — can a l s o be the t o o l s of the a n t h r o p o l o g i s t . These are the \" i n p r i n c i p l e \" p o s s i b i l i t i e s , and a number of a n t h r o p o l o g i s t s have endorsed the idea of using formal d e c i s i o n theory to s o l v e a n t h r o p o l o g i c a l problems (Buchler and N u t i n i 1969; Davenport 1960). But the r e s u l t has u s u a l l y e i t h e r been a d i s c u s s i o n of what ought to be done by a n t h r o p o l o g i s t s i n using formal models or a s u r f e i t ' i of mathematics at the expense of relevance to anthropology. A balance between a n t h r o p o l o g i c a l data and formal models has to be kept c l e a r i n that more s t r i c t l y a n t h r o p o l o g i c a l c o n s i d e r a t i o n s serve to i d e n t i f y the c o n s t r a i n t s and parameters w i t h i n which d e c i s i o n models can be a p p l i e d . This means that ethnographic d e s c r i p t i o n of p a r t i c u l a r value systems and c u l t u r e s permits one to designate the c u l t u r a l l y perceived a l t e r n a t i v e s that are appropriate to p a r t i c u l a r d e c i s i o n s i t u a t i o n s , and to demarcate the p r i n c i p l e s which are determinate (or seem to be) f o r choosing between the a l t e r n a t i v e s . And i t i s i n t h i s way that the a n t h r o p o l o g i s t , w i t h h i s t r a -d i t i o n a l t o o l k i t of a n a l y s i n g value systems and c u l t u r a l l y appropriate ways of doing t h i n g s , can make an enormous c o n t r i b u t i o n to the cross-c u l t u r a l study of d e c i s i o n making. - 19 -CHAPTER 2 THE PROBLEM AND ITS CONCEPTUALIZATION Introduction From my i n i t i a l substantive concern with Third World farmers I generate a number of propositions which link the decisions made by a subsistence farmer about agricultural innovations to his state of re-sources, incentive conditions attached to an innovation, and informa-tion and u t i l i t i e s attached to the outcome \"increase in productivity.\" From these statements specific to peasant farmers, I generalise to three statements about individuals and risk taking. In the interests of parsimony i t would perhaps be better to start with the general state-ments and proceed immediately to testing procedures. However, given the lack;;of precedents for the type of methodology implemented here and the explicational nature of my research, I think i t is necessary to show the step-by-step progression from one level of generality to another. It is also important to il l u s t r a t e the source from which my statements originate. Farmers and Innovations The i n i t i a l substantive problem I am dealing with is the ex-plication of the factors that influence farmers' decisions on agricul-tural innovations. The conceptual approach I intend to use is that of decision theory, go both my substantive and conceptual concerns are within limited boundaries. - 33 -Social Mobility By social mobility I am referring to an index based on role recruitment principles, i.e. ascription and achievement. This rests on a conceptualisation of roles as slots or positions in a social structure for which individuals may compete. My concern with this index is with the rules of competition which define the flow or circu-lation of persons between positions. The index measures the degree to which 'achievement role slots' are present in a particular society. Thus a society, with a high index of social mobility indicates that a greater number of people can compete for roles than in a society with a low social mobility index. The implications of social mobility for innovation are as follows. I f we visualise the competition for a particular role-as re-quiring the attainment of characteristics (a n) ; i t follows that persons aspiring to that role have to acquire those characteristics. Societies low in social mobility are characterised by exclusion rules which prevent everyone in the population from acquiring the necessary prerequisites for particular roles. Societies high in social mobility do not have these barriers. The characteristics (a n) may consist of prestige symbols, expected behavior patterns and, for want of a better word, 'entry fees'; and can be acquired in a variety of ways. If prestige symbols have to be attained, this means that an individual has to involve himself in "@en ; edm:hasType "Thesis/Dissertation"@en ; edm:isShownAt "10.14288/1.0103965"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Anthropology"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en ; ns0:scholarLevel "Graduate"@en ; dcterms:title "Dilemmas in decision making : a methodological test case in economic anthropology"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/35286"@en .