UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Dilemmas in decision making : a methodological test case in economic anthropology Prattis, James Ian 1970

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1970_A1 P73.pdf [ 11.42MB ]
Metadata
JSON: 831-1.0103965.json
JSON-LD: 831-1.0103965-ld.json
RDF/XML (Pretty): 831-1.0103965-rdf.xml
RDF/JSON: 831-1.0103965-rdf.json
Turtle: 831-1.0103965-turtle.txt
N-Triples: 831-1.0103965-rdf-ntriples.txt
Original Record: 831-1.0103965-source.json
Full Text
831-1.0103965-fulltext.txt
Citation
831-1.0103965.ris

Full Text

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 <ti CO o fl m cu O 'A u CO CO i-l tx cd K o w co o c U CO Range of Exploitation *Virginia M *Storm Winds Aunna Franton Gulf M2 *Lumarco Laurelin *Miss Christie Miss Mell May S *Annie L Beatrice S *Broken Islands Eddie May Flora M Miss Lorna North Wind *Harco Polly Ann *Ruby Ann Sea View *Sea Kay Seal Spray *Slave Carodic E. McAllister N. Anderson G. J o l l i f f e L. Stoffer G. Hardy G. Huson K. J o l l i f e T. Christiansen M. Stoffer S. Huson P. Bruce J. Bruce A. Johnson A. James G. Bruce N. Bruce N. Johnson A. Dick B i l l Scow G. Mellish A. Bruce A. Bristol L. Bristol L. White S. Commano R. Bianco H H H L H M H H M H M M H L M H M M H M M H M M H M L L L L L L L L L L L L L L L L M H M L H M M H M M H L H L H L H M M H M L H M H L M M H H L L H H H H L M M L L L L L L L M L L M M L L H H H H L L L L L L L L L M H M L H H L L L L L L 3 4 4 4 2 2 1 3 H H 5 3 1 4 4 5 3 4 3 4 3 3 4 4 4 4 5 4 1 4 6 1 5 5 3 5 6 2 4 3 1 6 2 6 1 4 3 3 5 3 4 3 4 3 4 1 1 3 5 5 5 3 1 5 55 40 A.C. Q.C.-Fraser H 56 15 1/2 caste Namu-Pender Harbour M 50 30 A.C. Q.C.-Fraser M 30 10 A.C. Q.C.-Fraser H 55 15 A.C. Low; around A.B. L 40 20 K Rl-Pender Harbour M 28 8 A.C. Q.C.-Fraser H 50 20 A.C. Q.C.-Fraser H 45 20 A.C. Namu-Fraser M 50 30 K Hardy Bay-J.S. L 49 30 K Rl-Pender Harbour M 27 9 K Rl-Pender Harbour M 50 30 K A.B. area L 32 10 K Rl-Pender Harbour M 26 10 K Rl-Pender Harbour M 19 2 K Rl-Pender Harbour M 40 20 A.C. Rl-Pender Harbour M 45 30 K Namu-Fraser M 50 35 K Rl-Pender Harbour M 35 10 A.C. Q.C.-Fraser H 26 4 K Rl-Pender Harbour M 29 11 A.C. RI-Fraser M 27 9 A.C. RI-Fraser M 30 10 A.C. RI-Fraser M 50 35 K RI-Fraser M 32 2 A.C. Rl-Pender Harbour M 12 K 14 AC * Boats that I went out on. Index: Q.C: Queen Charlotte Islands A.B.: Alert Bay Rl: Rivers Inlet J.S.: Johnston Straits A.C: Anglo-Canadian K: Kwakiutl Ul . 4 Test of Propositions My theoretical propositions were: 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. 2'. Individuals w i l l show an increased tolerance for high-frisk stra-tegies, with respect to the outcome to which HU is attached, as wealth increases. 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. Tests on Strategy Risks Table XXIV shows the number of high and low risk strategies taken as information and wealth varied. A strategy is categorised as high or low risk upon analysis of actual decisions of where to fi s h . Given a set of alternatives a strategy was classified as high risk when the fisherman chose that alternative which maximised gain, and as low risk when an alternative was chosen which minimised loss. Table XXIV: Strategy Risks Information Wealth Wealth L M H L M H L 2 23 6 L 52 25 0 H 8 40 1] Information H 52 19 0 n=12 n=12 n=Z Number of high risk strategies Number of low risk strategies taken. taken. - 142 -In Table XXV this raw data has been converted into propor-tions of high and low r i s k s taken. Table XXV: Proportion of High and Low Risks: Taken by A l e r t Bay  G i l l n e t Fishermen Wealth Low Medium High Information Low High Low High Low High Proportion of High Risks Taken 4% 13.4% 48% 67.7% 100% 100% Proportion of Low Risks Taken 90% 86% 52% 32.3% 0% 0% Proposition 1' From the data i n Table XXV one can see that low wealth fishermen took low r i s k s t rategies 96% of the time when information was low and 86.6% of the time when information was high. When con-trasted with the fac t that medium wealth fishermen took low r i s k strategies only 52% of the time under conditions of low information and 32.3% of the time when information was high, one can take the resu l t s as strongly supportive of proposi t i o n 1' which states that low r i s k s t rategies are preferred by LW i n d i v i d u a l s . The low and medium wealth fishermen are d i r e c t l y comparable as N=12 for each category. - 143 -Proposition 2' This proposition related an increased preference for high risk strategies to increases in wealth. Under conditions of low information medium wealth fisher-men took 44% more high risks than low wealth fishermen. There was a 54.3% increase in high risks taken when information was high. Though these categories of fishermen are directly comparable, com-parison is d i f f i c u l t between medium wealth and high wealth cate-gories as there were only two high wealth fishermen. However the fact that these latter individuals took high risks a l l the time irrespective of information lends support to proposition 2', though too much significance should not be attached to a sample of two. Proposition 3' This proposition related increased preference for high risk strategies to increases in information. Table XXV shows that low wealth fishermen took high risks 4% of the time when information was low and 13.4% of the time when information was high. The information increase produced a 9.4% increase in the proportion of high risks taken. Medium wealth fishermen took high risks 48% of the time under conditions of low information whereas 67.7% of the strategies - 144 -opted for under high information were high risk. This represents an increase of 19.7% in the proportion of high risks taken as i n -formation increases. For the high wealth category, information did not make any difference to risk taking as both skippers consistently took high risks irrespective of information, but as I pointed out previously the small sample size is not a sound basis for comparison. Test on Investment Risks When risk-taking with respect to capital investment was tested for in the Alert Bay sample the f i t was less good. Proposition 1' was supported, as a l l 13 low wealth fisher-men took low investment risks. However proposition 2' was not confirmed as some of the logical states (MW L.Inf; HW L.Inf; HW M.Inf, etc.) were not ob-served i n the empirical world. The same can be said for proposition 3'. It was not empirically possible to get the ordinal increments in risk-taking to test the propositions. This means that one or more intervening variables, functional to risk-taking, operate in such a way as to prevent particular configurations or states from appearing. The f i r s t possible intervening variable I would like to con-sider i s education. Recall the information circuits model I discussed - 145 -in Chapter 2 -- in that I visualized access to information circuits about technology as a function of wealth and education, and that risk-taking on technology was a function of information category (Figure 10). Figure 10: Information Circuits Model Of the 26 Alert Bay cases, the model was correct for 18. The eight cases that, did., not conform to the model were: Predicted Inf Category No. 4 - MW MEd HInf MInf 7 - -MW- MEd HInf MInf 8 - MW MEd HInf MInf 9 - MW LEd MInf LInf 17 - LW MEd MInf LInf 18 - MW LEd MInf LInf 20 - MW MEd HInf MInf 25 - MW LEd MInf LInf Of the 7 MW MEd configurations four had medium access to in-formation as predicted, three had high access. This difference can be located between the significance of circuit e with respect to d for par-ticular individuals. For 4 fishermen e'ifj^ d, for three c\~r—P.-- 146 -Similarly of the three MW/LEd combinations, a l l had MInf in-stead of LInf as the modal predicted. Of the five LW/MEd, only one case did not conform to the prediction; for this individuals h*^g, and h ^ f. Thus each one of these six contrary cases involve a differen-t i a l significance attached to only adjacent information c i r c u i t paths. This factor and the fact that almost 70% of the cases ran according to prediction makes i t reasonable to accept the model as holding for the majority of cases, though i t is evident that the relative weightings of wealth and education have to be revised for some parts of the model. The model is less accurate with respect to the next step of predicting risk-taking in terms of the level of access to information circuits. A l l the cases in the H.Inf category and the L.Inf category took investment risks as predicted by the model, but of the eight fi s h -ermen in the M.Inf category two took High Investment Risks, and six took Low Investment Risks. This suggests that in terms of capital in-vestment on new technology H.Inf as measured by access to information circuits is the generally necessary requisite for high risk. Congruence Between Investment and Strategy Risks For 23 of the Alert Bay cases there was a 1:1 congruence between investment risks and strategy risks, but the three exceptions are espe-c i a l l y noteworthy. - 147 -17. North Wind -- The skipper on this boat had not invested in any new technology for his boat, yet was observed to take high stra-tegy risks f a i r l y consistently. Of five low information situations, he twice elected to take high risks, and on five high information situations he took high risks four times. On my shore interview I learned two things: (a) that higher risks than were safe for the boat were being taken; (b) the lack of cap-i t a l investment in the boat was mainly a function of the fact that i t was not worth investing in.' The skipper, Neil Johnson, was looking ahead in terms of getting a new boat, but in order to float a loan from his fishing com-pany he would have to turn in a good performance with regard to the amount of fish he landed. His elected strategy to achieve this end was to take higher than usual exploitative risks. 25. Slave -- Sonny Commano, the skipper of the Slave took low investment risks on his boat, but on both high and low information consistently elected to take high strategy risks. One explanation for this could be attributed to the general consensus amongst Alert Bay fishermen that no one knew the currents and tides quite like Sonny, which would mean that some modern technology would be superfluous to him -- radar, depth sounder, etc. (But there were others like him, who took high strategy risks yet also invested in technology.) - 148 -Another explanation may be attributable to Sonny's "fish sense". But, in general he just seems to be an exception. 26. Carodic -- Roy Bianco, skipper of the Carodic had inves-ted $18,000 in the boat and much of the latest technology, but usually always opted for low risk strategies. He informed me that this was chiefly due to the fact that he had only been a fisherman for two years, and didn't feel he knew enough about tide, current and run conditions to fi s h the hazards. So though he had taken high investment risks on technology, he had decided to take low strategy risks. These exceptions point out something very important -- viz. that risk-taking as a means to increase fish intake can be used in one or both of two ways. Investment risks on technology provide" a fisherman with a better means to exploit the general fish resource. Better equipped boats had a greater range and mobility and could fish areas that poorly equipped boats were denied access to. High risks on fishing strategies served to place the fisher-men in a pas i t i o n where they could exploit the greatest quantity of salmon; though this often placed the fisherman in a hazardous position with respect to his boat's and own physical well-being. - 149 -Relevance of Cultural Factors None of the Kwakiutl fishermen in the sample had a high range of exploitation, and i t could be argued that this has a cultural base viz. Indians prefer to exploit areas close to their home base while whites prefer to go further afield. But I think an economic explanation suffices. Of the twelve Indians in the sample, 11 took low investment risks on technology, which implied that they were prohibited from exploiting the entire inshore fisheries because of deficient technology. Also, white fishermen who similarly took low investment risks had a limited range of exploitation for the same reasons. However, some of the data I collected gave support to the no-tion that certain cultural factors are very relevant^to differences in risk-taking. During the course of fieldwork on the Alert Bay boats, I was constantly struck by the technological ware displayed on two par-ticular categories of boats --Japanese and Yugoslav gillnetters. Time constraints prevented me from doing an adequate survey, but data collected from five Japanese boats and three Yugoslav boats with respect to investment in technology indicated that the individual fishermen concerned took far higher investment risks than one would expect. The table below shows that the wealth and education combina-tions 1produce a higher than expected information level (not just ad-jacent information ci r c u i t paths), resulting in the expected high invest-ment risks. Boat Skipper U t i l i t y W Ed. Inf. Inv. Reference C i r c . Risks Age Experience Group Range H i k a r i Nugget Symphony II. L i b r a Bonnie Lock A l b i o n Ranger M. Kamachi T. Asano S. Muramatsu M. Kamachi T. Oikawa I. Malavoz Al b i o n Commando R. Sekula Tony A. Peradenic H H H H H H H H M M M M M M L L L L L L M M M H H H H H H H H H H H H H H H H 50 45 30 48 35 45 35 55 28 20 10 25 15 18 10 20 Japanese Japanese Japanese Japanese Japanese Yugoslav Yugoslav Yugoslav H H H H H H H H - 151 -Two of the Japanese boats, Hikari and Symphony II, carried radar. The other boats did not, largely as a function of the fact that the five boats always fished together, and information from the boats with radar would be relayed to the other three. The Nugget and Symphony II had f ibreglass planing hulls which meant more speed. A l l had the very latest in electronic equipment — radio telephones, automatic pilots and depth sounders. This group of five represented a considerable investment in fishing technology. A l l five Japanese fishermen lived in the Japanese sector of Steveston, within walking distance of one another. Japanese was spoken at home, and as far as I could gather the majority of their social re-lationships were within the Japanese community. The skippers of the Hikari and Libra had been1 interred dur-ing World War II and had had their seine boats confiscated. They f e l t they had received a very bad deal over this and after twenty years there was s t i l l a note of bitterness about i t . The three Yugoslavs possessed what were perhaps the most beau-t i f u l gillnetters on the entire West coast. Built at a cost of $30,000 each, by the Albion Boat Co. of Haney, B.C., the boats were the very last word in technology — high-speed engines and planing hulls, radar, automatic pi l o t , hydraulic drum, depth sounders with fish finder functions — absolutely everything! It is interesting to note that of the seven boats b u i l t by this firm since 1968 four are owned by Yugoslavs. - 152 The three Yugoslav skippers lived in West Ladner where 60% of the population was Yugoslav. The influence of ethnic ties were quite extensive i n that a special Yugoslav school operated at week-ends, and social relations were more with other Yugoslavs than with English-speaking Canadians. Both groups of fishermen are thus categorised by an ethnic identity and an idea of ethnic separation, and both are intrusive minority groups — i.e. they came to Canada. These considerations may make explicable their preponderance for innovation in terms of a con-sideration of social marginality (Proposition 8, Chapter 2). The implication of social marginality is of individuals at the margin of one given culture, yet occupying a position where they operate in more than that culture. The marginal individual is not bound by wider group norms and in addition has the resources of his own identity group to draw upon.When this is accompanied by a lower evaluation placed upon cost factors than by non-marginal persons, then, in principle, i t would appear that the marginal individual has a much wider scope for ac-t i v i t y . The data I have on the Japanese and Yugoslav fishermen is only suggestive for support of the supplementary proposition relating social marginality to risk. This evidence can only be taken as indica-tive and not conclusive, due to the very small sample size. - 153 -Summary The f i e l d test gave general support to my propositions on risk-taking with respect to strategies preferred by fishermen, but there was less good a f i t when one came to consider investment risks on technology. Intervening variables of education and of social marginality were shown to be relevant in some cases. However the most important find-ing was that low wealth individuals took low risks — which was one of the major contentions I made in my i n i t i a l discussion (Chapter 2), and was also the most strongly supported result from the experiment (Chapter 3). Fishermen exhibited an increased tolerance for high risk stra-tegies as wealth increased which supported proposition 2'. As wealth moved from low to medium this increase was especially marked. Increased high risk taking was also noted as wealth moved from medium to high, but as there were only two HW fishermen in the sample one cannot place too much emphasis on the result, though i t does support the overall contention of proposition 2'. Except for the two high wealth fishermen, everyone else in the Alert Bay gillnet fleet indicated a higher tolerance for high risks as information increased. The former took high risks irrespective of information. This gives partial support to proposition 3'. The experimental data indicated the importance of the proba-b i l i t y consideration of taking risks for LW individuals. It was less easy in the f i e l d to find out exactly what chance factors were operative for a particular fisherman, as his estimated payoffs and losses were I n -u t i l i t y functions ($ .P : and $,.P.) and not straight calculations of J w w 1 1 payoffs and costs. However, I did discover that the chance factor was more a r t i -culated by LW fishermen when "hazard" fishing was envisaged. To foul on any hazard meant at least $200 for net repairs, and the chance of fouling had to be pretty low (tide running away from hazard, etc.) be-fore a LW fisherman would consider setting off a hazard. In these instances the (P^) appeared to be the most significant factor. I would have liked to assess whether the threshold notion vis-a-vis information and risk taking also operated under f i e l d condi-tions . However as I pointed out earlier (page 110 ) high information for strategies was analoguous to the medium information level in the experiment. There was no effective way that information could be increased for the fisherman beyond specifying the range of odds, pay-offs and costs for any alternative. Thus I was unable to see whether increments in information after the threshold point (identified i n the experiment) made any difference to decision strategies employed. Similarly the information circuits model related risk tak-ing to the number of circuits of information a person had access to in terms of his wealth and education. The argument made was that i f a person actively sought out a number of information circuits then one would expect him to idee higher risks on technology. The scaling of information with regard to technology was or-dinal, and i t gave no provision for identifying cut off points such - 155 -as the specification of the range of odds, payoffs and costs. This was because the information circuits model was designed for a dif-ferent problem — that of relating a person's risk-taking behavior to the number of information circuits that he tapped. So the threshold notion seems to occupy the status of a by-product of experimental serendipity as I was unable to adequately assess how i t operated in the f i e l d , due to adverse situational fac-tors vis-a-vis fishing strategies and the lack of an operational defi-nition for the threshold with reference to investment risks on tech-nology. The assumptions of weighting education vis-a-vis wealth which were fundamental to the construction of the information circuits model were shown to be valid in the majority of cases, though a partial revision of the model is necessary because of the exceptions. The proposition with respect to high investment returns at low cost was not tested in the f i e l d , simply because there were not any strategies or technologies that had these features. However, in the next chapter, I shall be marshalling evidence to show that low wealth individuals w i l l respond positively to risks of this nature. Despite the lack of congruence in testing procedures between the experiment and the f i e l d ,in that I was unable to test in the f i e l d a l l the propositions that were tested in the laboratory, both contexts have shown the preference for low risks exhibited by LWHU individuals. - 156 -I now have to show among other things that this is the case for Third World farmers. I w i l l also be looking for cases that seem to refute my assertions, to see whether in fact these contrary examples f a l l within the specified scope conditions which are pertinent to my state-ments . - 157 -CHAPTER 5 FARMERS AS RISK TAKERS Introduction In this chapter I w i l l be examining evidence from develop-ment literature to assess whether or not my assertions about risk-taking behavior are, at a prima facie level, applicable to the do-main of Third World farmers. I should emphasize that the evidence is not definitive. Short of doing fieldwork in a variety of ethnographic contexts I am unable to offer an explicit test of my assertions, as I have to rely on work done by scholars interested i n different sorts of pro-blems. Where possible I infer incentive conditions, information, resources and u t i l i t y dimensions from the existing literature. Ex-cept in those instances where the data is readily amenable to this type of treatment, I should acknowledge that the process of inference is by and large a rule of thumb method. Despite my inability to adequately test specific proposi-tions in this context, I w i l l be subjecting the general model to quite a severe scrutiny, as I w i l l be looking for cases that appear to refute my contentions. This is i n order to set the precise boun-daries for the domain of relevance for my assertions about risk-taking. - 158 -Roger's book "Diffusion of Innovations" (1962) is considered to be a classic sourcebook in this f i e l d and inasmuch as i t reviews and attempts to synthesise available research findings on the diffu-sion of innovations i t certainly i s . But inasmuch as i t continues the Rural Sociology tradition (Beal, Rogers and Bohlen 1957; Gross and Taves 1952; Copp 1958) of studying the characteristics of indi -viduals associated with innovativeness, i t does l i t t l e to develop any theoretical considerations. Rogers does not attempt a theory, though he suggests further research should be concerned with a more s t r i c t l y theoretical endeavour. He is primarily concerned with a taxonomy of adopter cate-gories whereby individuals can be categorised in terms of the speed with which they adopt innovations. The categories he uses — inno-vators, early adopters, early majority, late majority and laggards — are not used in any way to specify why one individual decides to adopt an innovation and another does not. His concern is to esta-blish, after the fact, the patterns of adoption i n any given community. His variables are selected on an a posteriori basis and this is a feature of most of the work conducted on agricultural innovation. Consequently no claims can be made for advancing a theory of innova-tiveness . It is not until one has an a p r i o r i selection of variables, which when accompanied by the specification of their interrelation-ships can one say there is a theory. - 159 -I am also none too sure that Roger's concern with taxonomy i s p a r t i c u l a r l y useful, and wonder whether a more s t r i c t l y t h e o r e t i c a l approach would render h i s categories superfluous. For instance, the difference between "Innovators" and "Late Adopters" could be viewed i n terms of a series of propositions re-l a t i n g a set of antecedent conditions to the types of r i s k s that predictably w i l l be taken. "Innovators" may have a t t r i b u t e s of high access to i n f o r -mation, be capable of e f f e c t i v e l y m o b i l i s i n g resources i n addition to being s o c i a l l y marginal. One would then p r e d i c t that i n d i v i d u a l s with these c h a r a c t e r i s t i c s would take high r i s k s . The "Late Adopter" may be characterised by low wealth and low access to information and predictably w i l l take low r i s k s . Only when a large number of farmers around him have adopted an innovation, such as hybrid corn, can the " l a t e adopter" assess to his s a t i s f a c -t i o n the r e l a t i v e advantages of adopting. His lateness i s a function of the fa c t that he consistently takes low r i s k s and wants to see how everyone else fares before adopting. In these terms adopter categories do not make too much sense as they can be accounted for by reference to d i f f e r e n t i a l r i s k taking. S i m i l a r l y Rogers i s concerned with d i s t i n g u i s h i n g stages i n the adoption process — awareness-interest-evaluation-trial-adoption: (1962: 81). From the point of view of d e c i s i o n theory one i s concerned only with t r i a l and adoption. The preceding categories are merely various - 160 -configurations of "access to information" and " u t i l i t i e s " . The dif-ference between t r i a l and adoption is simply the difference between decisions at time T and time T ,,. If the incentive conditions as n n+1 perceived by the individual change as a function of increased informa-tion or changing u t i l i t i e s , then one can expect some change in decision at T ,,. If additional information does not result i n a change in n+1 incentive conditions, then the decision at T w i l l be the same as n+1 at T , a l l other things being equal, n What I am trying to point out in this brief discussion i s that categories and taxonomies without a theoretical referent remain mere categories and taxonomies, and indicate an interest i n butterfly collecting (Leach 1961) rather than systematic knowledge. A more considerable drawback, for my purposes, i s that Rogers is not explicitly concerned with the actual decision-making process of the individual. The calculations of a particular farmer with respect to yields and payoffs vis-a-vis costs aia not emphasised. Also other researchers (Cancian F. 1967; Ryan and Gross 1943) have tended to regard a l l innovations as equivalent units from the view-point of analysis with the result that the significance of the incen-tive conditions (in terms of both material and social calculation) i s obscured. I have argued earlier that i t is the incentive conditions that define the nature of the risk in adopting an innovation. I have also maintained that incentive conditions, both positive and negative - 161 -can be calculated in social as well as in material terms. In this way decision making can be viewed as an expanded cost-benefit analy-sis (subject to certain constraints) and the preoccupation with cul-tural obstacles to adoption of innovations that characterises the work of Foster (1962), Spicer (1952) and Hogbin (1958) can be put in a more r e a l i s t i c perspective. Cultural norms become obstacles to adoption of innovations when adoption would involve some perceived cultural cost, such as disapproval of peers or loss of status. An innovation may be rejec-ted on the grounds that i t conflicts with group norms and values. A subsistence farmer, dependent on social a f f i l i a t i o n for access to land, labour and means of production, cannot readily flout group norms, as this may invoke sanctions that would cost him his access to means of livelihood. In this sense cultural norms as "obstacles" are merely an additional dimension to the material incentive conditions attached to the innovation. Certain problems exist in the estimation of cul-tural costs and payoffs vis-a-vis a decision to adopt an innovation, but this is a problem of measurement and not of conceptualisation. Thus i t is that my approach takes l i t t l e directly from the Rogers and Rural Sociology tradition, as my discussion of Rogers' adopter categories indicates my concern is with different kinds of questions. Griliches (1957) anticipates much of my approach with his emphasis on economic variables as the major determinants of the pattern of technological innovation. He emphasises the predominance of - 162 -p r o f i t a b i l i t y — the difference between economic returns and the innovation's economic cost — i n explaining the rate of adoption of an innovation. While I endorse his notion of p r o f i t a b i l i t y , I f e e l i t has to be broadened to include c a l c u l a t i o n s of p r o f i t a b i l i t y i n s o c i a l terms. Peasant farmers do react to p r o f i t a b i l i t y of new factors and factor combinations, but they do so as they view p r o f i t a b i l i t y through the constraints of t h e i r own technological (and s o c i a l ) systems. I am not prepared to argue that s o c i o - c u l t u r a l factors cancel one another out leaving s t r i c t l y economic variables as the prime movers of technological change. G r i l i c h e s i s , and to do so overlooks the f a c t that the important thing to assess i s the i n d i v i -dual's perception of p r o f i t a b i l i t y , not only i n s t r i c t l y material terms but i n terms of s o c i a l p r o f i t . This i s why I am concerned with states of nature, u t i l i t i e s and subjective evaluations of p a r t i c u l a r outcomes. I t should be apparent that I cannot r e a d i l y use the work done on a g r i c u l t u r a l innovations, as the parameters required for decision theory are r a r e l y f u l l y s p e c i f i e d . Thus I w i l l only draw on those studies where I can i n f e r or estimate the parameters necessary to assess d i f f e r e n t aspects of my concern with the actual process of decision making. To s t a r t o f f with a case that gives a p o s i t i v e colour to my l i n e of enquiry I would l i k e to consider Bauer's examination of Malayan Rubber Smallholdings (Bauer 1946). - 163 -Case No. 1: Malayan Rubber His c r i t i c a l report on the administration of and stipulations with respect to smallholdings includes a detailed documentation of two instances where opportunities for investing in technical change were passed up by the vast majority of smallholders. These instances concerned the processes of replanting and new planting respectively. Replanting involves the digging up of old rubber trees and planting new ones in their stead, while new planting refers to the process of planting new trees on new acreages. New plant-ing was prohibited by the government, although the regulations were relaxed in 1939-40. As part of his survey Bauer distinguished between peasant, medium and estate holdings as a function of scale. Peasant holdings were characterised as having a resident owner, tapped by family l a -bour and comprising an area of 10-12 acres. Medium holdings of 60-100 acres saw the increased use of con-tract labour for tapping, while the large estates relied entirely on wage-labour for tapping. The great bulk of the Malay-owned acreage is in peasant holdings; some of the smaller Chinese holdings and a few Indian holdings also f a l l in this category. The estates were largely controlled by Europeans. Replanting The process of replanting involved the uprooting of existing trees and their replacement by a new stand planted on the same area. - 164 -Replanting enabled the rubber producer to weed out older un-productive trees and to replace them with new clonal types. This was essentially a long-term investment, as i t took several years for the new stand to reach maturity. The cost for labour in the process of digging out the old trees, and the heavy expenditure required for manuring to produce a second stand of mature trees on what is more often than not unsuitable ground, constituted a prohibiting factor for the peasant smallholder. Before World War II there was hardly any replanting on small-holdings, and what there was, was almost entirely on the larger Chinese and Indian holdings, which are hardly smallholdings i n the accepted sense of the word. Replanting was, however, widely adopted on estates (1946: 91). While i t is feasible to replant on large estates i t is impractical to do so on peasant smallholdings for economic and techni-cal reasons. The lack of capital required to pay not only for the heavy expenses involved but also to bridge the loss of incomes from the time of the f e l l i n g of old trees to the maturity of the new stand constrained peasant holders from replanting. Also i t was technically impossible to replant part of a holding of a few acres successfully, as the area replanted would be closely surrounded by mature trees, which would intercept the sunlight and whose roots would compete for food with the undeveloped rootlets of the newly planted trees, resulting in stunted trees. The combined - 165 -result of these various factors has been virtually to confine re-planting to estates (1946: 93). When Bauer questioned the smallholder as to the absence of replanting — the reply was always the same — ina b i l i t y to afford the loss of income from the felled stand and the technical d i f f i c u l t i e s of replanting part of a peasant holding (1946: 108). If one assumes that wealth is coterminous with size of hold-ings, Bauer's account of replanting strongly supports proposition 1' which states that under conditions of LW and HU, individuals w i l l show a preference for low risk strategies with respect to decisions involving the outcome to which HU is attached. This preference for low risk i s compounded of two main fac-tors : — peasant smallholders did not go for marginal returns to inves tment. — also they seemed to be more concerned with cost minimiza-tion:: than gain maximization. That I can assume that U for increased productivity is high is borne out by Bauer's frequent references to the energy and market orientation of the peasant smallholders. In fact the whole thesis of his report is to change rubber planting restrictions in favour of the smallholders mainly on the grounds that they can do more for the export industry than can the larger, more costly estate operations. Also the progression from peasant - medium - estates is direct-ly proportional to an increase in replanting, from none at the peasant - 166 -level, to a number of instances at the medium level to most estates at the upper level. This evidence is supportive of proposition 2'which relates tolerance for high risk to increasing wealth. New Planting New planting to the extent of 5% of the 1938 planted acreage was permitted i n 1939-40. The peasant category took virtually no advantage of the small amount of new planting permitted, and what area was replanted was almost entirely on the larger Chinese and Indian holdings (medium) (1966: 14). "I did not find a single instance of an owner of less than 20 acres undertaking new planting without purchasing planting rights. In other words, every instance of an area planted in 1939-40 I saw i n the course of this extensive tour belonged to a man who either had the capital to buy planting rights, or had sufficient land to be classed as the owner of a medium holding rather than as a peasant smallholder." (1946: 116) New planting was organized in terms of share certificates i n denominations and multiples of 1/20 of an acre which entitled the owner to plant rubber to the extent of 5% of his 1938 reigstered acreage under rubber. Thus an owner of 3 acres would be entitled to plant 1/7 of an acre, while the owner of a 5 acre holding was entitled to plant 1/4 of an acre. - 167 -Any land held by peasant smallholders (which was not under rubber) usually had other crops growing on i t — f r u i t trees, tapioca, etc., and new planting on t h e i r own land could only be undertaken at the cost of destroying another source of income. Owners who had no unplanted reserve land had to apply to the land o f f i c e s f o r the a l i e n a t i o n of land on which to carry out new plan t i n g . For the peasant smallholder the amounts of land were small and usually a f a i r distance from h i s e x i s t i n g holding. A smallholder i s often prepared to operate a holding a mile or two away from h i s house, but i t i s not f e a s i b l e f o r him to go f a r a f i e l d to tap 16, 24 or 32 trees. The f r a c t i o n a l acreage of new pla n t i n g was thus of no value to the i n d i v i d u a l peasant smallholder, unless he had unused reserve land or i f there was unalienated land close to h i s own holding. These l a t t e r two conditions were very r a r e l y met. Bauer concludes that "New planting by the great majority of the smallholders was thus ruled out." (1946: 120) Once again i t was the estates and wealthier medium holdings that c a p i t a l i s e d on the new planting. This instance i s again supportive of my f i r s t two proposi-tio n s . LW producers are observed to take low r i s k s by merit of t h e i r attempts to keep costs to a minimum and not to go for marginal eco-nomic advantages i n terms of better rubber trees. The wealth gradient from peasant-medium-estate holdings was again accompanied by an i n -crement i n r i s k tolerated with respect to new pl a n t i n g . - 168 -Also Bauer makes frequent references (1946: 121, 122, 135) to the fact that the estates and some medium owners were much better informed about the precise nature of the 5% new planting permitted, whereas the peasant holders very often did not know what the whole process was about. While I cannot glean from Bauer's report any information with respect to education, his references give support to my asser-tion that the taking of high risks increases as a function of infor-mation (Proposition 3'). Summary One could view the lack of i n i t i a t i v e i n terms of new and replanting as peasant conservatism, but this would be doing them an injustice. It is clear from Bauer's report that the peasant small-holders made very shrewd economic calculations and rejected the oppor tunities offered in terms of cost considerations and also because i t was not worth their while (economically) to either replant or new plant. Bauer states in paragraph 63 "...the decisions of the small holders seemo quite definitely influenced by f a i r l y rational compari-sons between the advantages of the alternatives i n terms of income, expenditure, effort and risk." (Bauer 1946) - 169 -LW Farmers and Risk From an example of general support I would like to move to the specific and take each proposition i n turn. Proposition 1' referred to LWHU individuals indicating pre-ferences for low risk strategies. This assertion can be examined in terms of three substantive aspects with respect to agricultural inno-vations : (A) 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; (B) Innovations that offer a low return tend not to be adopted; (C) Farmers seek to minimise the cost factor of an innovation, hence they are more concerned with cost reduction than high returns. (A) The high investment return at low cost thesis goes a long way to explain the present agricultural revolution in Asia. The basis of the revolution li e s in a technical breakthrough in the choice and recombination of desirable genetic sources of a small number of high-yielding varieties of rice, wheat and coarse grains at the International Rice Research Institute (IRRI) near Manila and other research centres in Mexico and India. These varieties, being virtually hand-tailored physiologically and morphologically for f e r t i l i z e r responsiveness, showed fantastic results i n widespread t r i a l s and demonstrations with yields double and even treble those of the traditional varieties grown under the same conditions of good irrigation and heavy f e r t i l i z a t i o n . - 170 -This factor accompanied by f e r t i l i z e r subsidies (Punjab) and seed price subsidies (Philippines) produced a low-cost high investment-return innovation. Since 1966 when the hybrids were f i r s t distributed i n the Pun-jab, Indian farmers in two years planted about 5 million acres each of the high-yielding varieties of rice and wheat and nearly 3 million acres of those of corn, sorghum and millets (Hsieh Sam-Chung 1968). Pakistan had planted approximately 2 million acres of Mexican wheat by 1968, and 1 million acres of IRRI rice. In the Philippines the area planted to the "miracle rice" IR-8 increased to 400,000 acres by 1967. Yet in 1965 Schultz sounds like a prophet of doom i n his book "Economic Crisis in World Agriculture". Before the introduction of the hybrids there was a gloomy, almost desperate concern for the ina b i l i t y of Asian farmers to produce enough food for basic nutritional needs; while population growth was accelerating. But in 1967 Dr. D.S. Athwawl of the Punjab Agricultural Uni-versity announced that by 1970 India would be self-sufficient i n grain and rice crops (Economist 1967). In the' 1969 annual report of the United Nations Food and Agriculture Organisation the director-general forecast that the major problem facing some underdeveloped countries is that of handling a surplus of agricultural production (F.A.O. 1970). Extrapolating from this report the international science journal "Nature" predicts that even countries such as India are now faced with the p o l i -t i c a l l y contentious problems of how to manage a surplus (Nature 1970). - 171 -In these few short years I do not think the surplus is a function of any drastic change in the Indian peasant farmer. I would maintain that he is s t i l l operating on a low risk strategy. The dif-ference lies i n the economic opportunities offered by the hybrid grains and rice. The opportunity to invest in these new crops constitutes a low risk strategy for the farmer, as his costs are minimised through subsidies and the incentive conditions of high yields reassures him, at the very worst, of a harvest bigger than that which can be obtained through using traditional varieties. (B) The evidence from Bauer indicates a shrewd assessment of eco-nomic potential with regard to new planting and replanting on the part of the Malayan smallholder. Where the returns were seen to be marginal the smallholder did not risk his investment capital. Schultz' thesis in "Transforming Traditional Agriculture" (1964) is based on the notion that peasant farmers in traditional agri-cultural systems are not going to invest further in traditional produc-tion factors because the return to investment is so low. I think his assertion is generalisable to any production factor, be i t traditional or not, as no farmer in the LWHU state can afford to risk his "sur-vival k i t " for marginal benefits. His discussion of the Guatemalan community of Panajachel shows that the community is not an isolated subsistence economy, but is closely integrated into a larger market economy. - 172 -"Yet hoes, axes and machetes are not replaced by better tools and equipment. There is not even a wheel. Coffee leaves used as fer-t i l i z e r are not replaced or supplemented by chemical f e r t i l i z e r s . Tra-ditional varieties of corn are not replaced by hybrid seed. Traditional breeds of chickens are not replaced by better hens for producing eggs and broilers for producing meat." (1964: 35) This is not because the Guatemalian Indian is not looking for ways to improve his l o t . Tax notes that "He is on the lookout for new and better seeds and ways of planting" (Tax 1953). But such improvements as the Indian could use have an exceed-ingly small effect upon production. Tax calculated that the return on further investment in available tools, domestic animals and growing crops was too low to provide an investment incentive (1953: 118 f f ) . A similar argument is advanced by Schultz to explain the s i t -uation in Senapur, India. In Senapur there is a larger stock of ma-te r i a l capital than in Panajachel. But despite this the two villages have in common a low rate of return to investment in a l l forms of capi-tal used in agricultural production. The similarities of both villages could be considered as a function of the fact that both economies perform highly effi c i e n t l y i n using the factors at hand, and farmers i n both villages have shown a reluctance to invest in new production factors mainly as a function of the marginal productive advantage of the new factors. - 173 -Before leaving the "no risk for marginal return" hypothesis, recall that the experimental findings indicated that a low payoff was attractive to some LW individuals when they were reasonably sure that the payoff would accrue, and that their costs were low and prospects of failure low. Unfortunately, in the f i e l d and from the literature I was unable to find empirical instances of this situation so I cannot offer an adequate f i e l d assessment of the experimental findings. However, inasmuch as the experiment has had significance for other f i e l d results I think i t would be reasonable to suggest as a development strategy that some people w i l l go for new production factors that offer marginal productive advantages i f (a) the calculated advantage is likely to accrue; (b) the cost of the factor is low; (c) the prospect of the factor f a i l i n g is also low. (C) The implications of C are really the obverse of B. Because farmers are poor they cannot afford marginal investment returns as the prospect of loss is too great a risk to their "survival" k i t . The cases cited by Bauer and Schultz indicate that the main-tenance of the status quo with regard to production factors in Malaya, Panajachel and Senapur is a function of inhibiting costs for investment. Wealth and Risk Proposition 2' related tolerance for risk directly to increases in wealth. E.M. Rogers in his book "Diffusion of Innovations" (1962) - 174 -gives general support to this notion though he does not specifically deal with risk. Wealth is often associated with other predictors of adoption of agricultural practices like simple ab i l i t y to afford the innovation, specialization, mental abi l i t y and access to the best s c i e n t i f i c infor-mation. Though there may be a complex of intervening variables, wealth is a. good gross predictor of a tendency to adopt innovations. In a study of 108 Indian villages Fliegel, Roy, Sen and Kivlin (1968) sought to determine the extent to which villages differed in ut i l i z a t i o n of modern agricultural technology. Taking the village as their unit of analysis they focussed attention on innovations in agricultural technology and constructed an index which measured the rate of adoption of agricultural innovations. Taking this as a de-pendent variable the authors correlated i t with a series of other fac-tors as independent variables to assess the significance of these ex-ternal constraints on adoption rate. As a measure of affluence, the proportions of families pay-ing Rs. 10 or more in village property taxes was found to be positively and significantly related to adoption rate (30 zero order;.24 1st order partial) (1968: 56). The effect of this measure was found to be inde-pendent of other measures significantly related to adoption. This measure served as a rough guide not only to family af-fluence as village resources, but also to the a b i l i t y of villages to tax themselves. With limited state and federal resources, this a b i l i t y may be crucial for successful development. - 175 -That village affluence is related to adoption indicates a reciprocal and reinforcing effect. Modern agricultural inputs require cash investments— villages with relatively high levels of affluence have the economic resources either in the hands of the farmer or avail-able from local people in the form of loans, to invest in modern inputs. The authors state on page 60 "Any increase in levels of l i v -ing t a s measured by tax index] w i l l provide more opportunity to adopt practices which may require i n i t i a l investment or risk taking." Information and Risk Proposition 3' related tolerance for risk to increments i n information. Fliegel et a l i a in their study,while not explicitly concerned with information, correlated a number of factors with adoption rate that could be taken as indicative of different levels of information. Their sample had three types of villages i n terms of time of exposure to development programs. The stage III villages which had had more than ten years of exposure to the community development pro-gram showed more agricultural innovation than the villages with less exposure. The adoption index was correlated with the stages measuring period of development exposure and the data indicated an association which was s t a t i s t i c a l l y significant at the 1 per cent level (0.25 corre-lation coefficient). - 176 -Adoption rate was also found to be a function of the proximity to agricultural and extension services. A proximity scale covering six institutions -- V.L.W. HQ, Co-operative, Panchayat, Veterinary Centre, Godown, Block HQ, — was significantly related to agricultural adoption (correlation coefficient of 0.28; significant at 1 per cent level). It is reasonable to assume that proximity is directly related to the proportion of information that is made available to farmers. The amount of contact each village had with the extension agent was highly correlated with adoption. A combined index of four measures of contact had a correlation coefficient of 0.50 with adoption rate. Again I am operating on the assumption that increased contact with an extension agent is an indicator of increased information being made available to farmers. These three measures (a) length of exposure to development programs, (b) proximity to extension and co-operative headquarters and (c) contacts with the agricultural agent are a l l positively and significantly related to the adoption rate of agricultural innovations. A l l these measures imply a function of information and dis-semination of information through various channels, and the data sup-ports the notion that risk taking on innovations increases with information. In a previous chapter I visualised access to information circuits as a function of wealth and education. From the Indian data i t is not possible to assess adequately the combination of these fac-tors with regard to information, but in a round-about way one can show the significance of education. - 177 -The authors found that villages with a high proportion of literates tended to score high on the measure of agricultural adoption, and show that literacy is part of a complex of "modernization" variables associated with development. The link between literacy and information is a f a i r l y obvious one — I am assuming that by being literate a farmer has access to more information channels than i f he was non-literate. U t i l i t y Considerations The preceding sections indicate empirical support for my assertions; I now propose to examine several contrary cases to see i f in fact my propositions are refuted. Recall that my remarks about risk-taking were bounded by a u t i l i t y parameter. I stated that my concern lay with decision making in a situation where a high u t i l i t y was held for the outcome, whatever that outcome may be. In the empirical world of contrary cases one can expect two things: (a) the high u t i l i t y stipulation may not be present in the decision situation or (b) the high u t i l i t y stipulation may be present in the decision s i t u -ation, but be in conflict with another high u t i l i t y outcome which takes priority. The following case histories are examples of the situations described above. - 178 -Moturiki Community Development Scheme Moturiki Island i n the F i j i group was chosen as the locale for a community development project i n 1948 by the F i j i a n government (Hayden 1954). The scheme saw the introduction of a copra marketing board, training of Moturikans at a local agricultural station, the planting of coconut nurseries and hardwoods, the establishment of an agricultural school for boys with piggeries and poultry. New cash crops such as rice and dalo were introduced and markets for copra, timber and cash crops were assured. A community development team was resident on the island for a year and initiated these projects. Two years after the team's withdrawal from the community the planting of cash crops had ceased, the clearing of coconuts and care of the nursery had been abandoned, the nursery and reforestation schemes no longer functioned and nobody attended the agricultural school. The scheme had had the advantages of ready markets for cash crops and products and of providing Moturiki with considerable i n -creases in agricultural productivity — yet these opportunities were not taken advantage of. The major reason can be found i n Hayden's own words "Life was pleasant and leisured on Moturiki — food was not a problem" (1954: 6). Thus one could not expect the islanders to increase their productivity in food at the expense of leisure when they had enough for their needs. - 179 -Extra food was the way i n which they regarded the extensive a g r i c u l t u r a l schemes. As there was more than they could eat and ex-change, i t was obviously not needed, so the projects were abandoned. The Moturikans exhibited a low u t i l i t y f o r increasing t h e i r a g r i c u l t u r a l p r o d u c t i v i t y , and the whole scheme seemed to be more a function of the community development team's idea of what was required, rather than as a response to the islanders wants or needs. Food-Fishing by Indian G i l l n e t t e r s from A l e r t Bay In the v i l l a g e of A l e r t Bay, where I based my fieldwork, a remarkable difference i n f i s h i n g attitudes occurred when Indian f i s h -ermen went food-fishing. By law they were allowed to f i s h the r i v e r s with nets provided that the f i s h caught were used f o r domestic consump-ti o n . Here was a s i t u a t i o n where no one exhibited a high u t i l i t y for increasing f i s h intake, and nobody ventured to take high r i s k s as i t was not very important to get a large catch. Sonny Commano on the Slave took me along on one food-fishing t r i p to the Nimpkish River, opposite A l e r t Bay on Vancouver Island. Of the s i x sets he made, he never once took a high r i s k . He acknowledged the f a c t that more f i s h could be had close to the r i v e r banks but the shallowness and number of trees i n the water decided him against i t . He fished the middle of the r i v e r , and the whole a f f a i r was very much l i k e a pleasant afternoon's trout f i s h i n g . - 180 -I gathered from Sonny's statements, that food-fishing did not carry the same sense of urgency to get big catches as did commercial fishing. There was no way to s e l l the fish that was caught, and even with canning and salting there was usually more than enough to consume. This surprised me because of Sonny's reputation as a com-mercial fisherman, as he was noted for the high fishing risks he took. But in the food-fishing situation he was operating with a different preference rating for the outcome — increase fish intake^ He stated that i t was not important to him how much fish he caught, as long as he got some. Discussions I had with Arthur Dick and B i l l Scow confirmed this attitude towards food-fishing. Nobody took high risks simply because i t was not that important to pu l l in big catches, as there was more than enough for domestic consumption. The Moturiki community development scheme and the food-fishing situation had the necessary preconditions vis-a-vis resources and information to expect that high risks would be taken. But in ac-tual fact the Moturikans turned down the prospect of ready markets and the Indian fishermen did not attempt to maximise fish intake. This would seem to contradict my assertions about risk taking. However, neither situation met the required scope conditions with ref-erence to the specified u t i l i t y parameter (both instances were charac-terised by a low u t i l i t y for increasing productivity) and are thus outside the domain of my theoretical statements. They do not provide a rebuttal to my assertions. - 181 -Spanish American Farmers and Hybrid Corn Spanish American farmers of E l Cerrito i n the Rio Grande V a l l e y of New Mexico grew corn as t h e i r major crop. Their y i e l d s were low com-pared to the harvests of midwestern farmers, and the general q u a l i t y of the corn was low. hybrid corn which gave three times the y i e l d of that grown t r a d i t i o n a l l y . The farmers recognised the need for b e t t e r production and once the a b i -l i t i e s of the new corn were indicated by means of a demonstration p l o t , h a l f the farmers i n the community planted the hybrid and each at l e a s t doubled h i s production the f i r s t year. In the following year three-quarters of the farmers i n the v i l l a g e had accepted the new seed. I t produced high y i e l d s and was w e l l w i t h i n the f i n a n c i a l means of the i n d i v i d u a l farmers. In 1946 the county extension agent succeeded i n introducing Yet four years a f t e r the f i r s t i n troduction a l l the farmers had ceased to plant the hybrid and were using t h e i r o l d corn. Table XXVII: Adoption of Hybrid Corn by Spanish American Farmers Year No. of Farmers P a r t i c i p a t i n g No. of Farmers Y i e l d 1946 1947 1948 40 60 30 84 84 84 D 0 U B L E D 1949 1950 3 0 84 84 - 182 -The agent's follow up enquiry into the eventual rejection of the hybrid established that the farmers s t i l l f e l t the need for better yields, and that they wanted to increase their productivity in order to market surplus corn (Apodaca 1952: 38). The farmers can be quite definitely categorised as HU for increasing agricultural productivity. But what was at stake was a u t i l i t y for corn quality which had a priority over corn quantity. The agent eventually discovered that the farmers and their wives were unhappy about the way in which the new corn made t o r t i l l a s — which was the basic staple in the community's diet. The texture and taste were considered inferior to the traditional corn. Also t o r t i l l a s had a complex of symbols associated with their manufacture — a woman's housecraft, social offerings and pride were at stake, as well as the necessity of eating t o r t i l l a s at every meal. Thus one can view the rejection of the hybrid corn not as a function of the farmers having a low u t i l i t y for increasing produc-t i v i t y ; but as a function of their having a higher u t i l i t y for corn quality and the t o r t i l l a complex. Some Implications The examples I have examined indicate that a basic pattern of decision making can be expected i n different socio-cultural settings. When I examined material to support the notion that LWHU farmers take low risks, I found this consideration to hold across a variety of social structures — viz . Panajachel - Senapur - Malaya. - 183 -"Senapur, for instance, has a long-established caste system, whereas Panajachel has singular f l e x i b i l i t y i n the movement of families up and down i t s social status scale. In Senapur the families of the privileged castes, mainly the Thakur, have perpetuated their wealth, privileges and social status for many generations. In the Guatemalan community, Tax found marriages cutting across wealth lines and much mobility on the economic and social scale; moreover the social and eco-nomic gap separating the top and the other families has not been substantial." (Schultz 1964: 44, footnote 8). Both of these communities differ radically from the Malayan example, yet risk-taking was observed to be similar in a l l three cases. More evidence to support the idea of basic patterns of deci-sion making can be obtained from Fliegel et alia's study (1968). They used discrete sets of variables to assess the difference that agrarian structure, occupational diversification, caste structure, severity of village faction disputes, and religious diversity had upon the adoption rate of agricultural practices. Their hypotheses regarding the influence of agrarian struc-ture on adoption of farm practices were not supported by the data. It was found that neither population pressure on available land nor the extent of dominance i n land ownership by a few landlords ware related to the adoption index. With regard to occupational diversification i t was hypothesised that the predominance of agriculture would be associated with relatively high levels of adoption of farm practices. This was not supported. Similarly the hypotheses regarding caste structure were not supported. It was hypothesized that villages dominated by a single caste would rank high in adoption of farm practices. The rationale for - 1 8 4 -this hypothesis rests on the fact that dominant castes tend to be agri-cultural castes in agricultural villages and that this should maximize s k i l l inputs into agriculture (1968: 65). The second variable concerning caste is a ranking of the r i t -ual status of the most numerous caste. The hypothesis was that villages with numerically dominant castes of high status w i l l be high in adoption of farm practices. This contention was not supported by the data collected. The severity of village faction disputes was used to measure horizontal differentiation in village structure, and the hypothesis was that factionalism would be negatively associated with adoption. This was not supported. Neither was the consideration that religious diver-sity would have a negative effect on adoption. These findings lead the authors to conclude that structural variation did not make any significant difference to the adoption rate (1968: 69). That this can be said of India, which has long been typified as having every socio-cultural obstacle to development imaginable, gives some encouragement to the hard-pressed development-planner. These considerations have a number of implications for the way in which one may view social structure and culture from the stand-point of analysing decision making. Firs t of a l l structure is not a given, an entity that exists through time. It is a heuristic device used by an analyst in order to solve different types of problems, (hence the many different definitions of social structure!). - 185 -S o c i a l structure can only be relevant or s i g n i f i c a n t when i t provides access to resources that are valued. Resources can be viewed i n s t r i c t l y material terms — access to land and markets, etc. — but the wider economic^meaning covers a l l behavior. I f prestige as a valued resource can only be obtained through a p a r t i c u l a r structure (kinship, age organisation) then that structure i s important to the actor making decisions about the resource prestige. In terms of farmers adopting a g r i c u l t u r a l p r a c t i c e s , i n the vast majority of cases s o c i a l structure i s not the avenue f o r access to resources — information and technology — concerned with a g r i c u l t u r a l p r o d u c t i v i t y . In India t r a d i t i o n a l structures such as caste and the patron-c l i e n t networks are no longer the exclusive channels for resources as-sociated with a g r i c u l t u r a l p r o d u c t i v i t y . Resources such as technology and information are a v a i l a b l e from sources outside of the t r a d i t i o n a l networks and while the l a t t e r may s t i l l be important for c e r t a i n prestige considerations, they are no longer the structure used to gain access to the resources needed for a g r i c u l t u r a l production. The point I would l i k e to emphasize i s that any structure, be i t t r a d i t i o n a l or modern, only becomes important when the resources i t can provide and d i s t r i b u t e are valued as necessary. T r a d i t i o n a l group norms and structures become s i g n i f i c a n t when group, a f f i l i a t i o n c a r r i e s with i t access to a resource (land, labour,prestige, r i t u a l status, etc.) that an i n d i v i d u a l values. This - 186 -being the case, i t is unlikely that an individual w i l l risk his ac-cess by flouting group norms. Given his high u t i l i t y for a resource he is constrained to play the game according to whatever rules and constraints are implied in a particular access route. As soon as an individual obtains access to the resource he values outside of any group a f f i l i a t i o n s , then he is less constrained to observe group sanctions, as he does not need the group to gain ac-cess to the resource he wants. This perspective i s , I submit, generalisable to a l l resources and a l l structures. Complementary to this view of structure, one can place cul-ture — the general scapegoat for failed development plans — i n a similar perspective. I would submit that cultural values provide the parameters within which individuals make decisions; and cultural norms are established not through a mystical genesis but by regular patterns of aggregate decision making. As the pattern of decision making changes at the aggregate level, so does the norm. This view of social structure and culture should free us from discussing them as "obstacles" or "drawbacks" to innovation and development. They are part of the antecedent conditions to decision making, and once one can specify the implications of these conditions then one can predict the type of decisions that w i l l be made. - 187 -CHAPTER 6 CONCLUSION Introduction My i n i t i a l considerations focussed on the problem of farmers making decisions about agricultural innovations, then I progressed to a higher level of generality wherein I referred to individuals taking risks. The testing procedure I used involved both laboratory and f i e l d settings, while corroborative evidence was supplied from secondary sources. From this progression I intend to draw three types of impli-cations — methodological, theoretical and practical, and in doing so intend to i l l u s t r a t e both the strengths and the weaknesses in the ap-proach I have used. Under "Methodological Implications" I w i l l dis-cuss processes of enquiry and types of models that can be used by eco-nomic anthropologists. In addition I propose a "paradigm" for research enquiry in order to i l l u s t r a t e the progression of the work attempted here and the required revisions. "Theoretical Implications" considers the way in which the basic subject matter of anthropologists — social and cultural behavior — may profitably be viewed, while under "Practical Implications" I w i l l be concerned with possible policy considerations suggested by this research. - 188 -Methodological Implications In his inaugural lecture at the College de France in 1960, Levi-Strauss in addressing the problem of kinship analysis indicated the impossibility of varying at w i l l the complex relationships i n -volved or apparent in the f i e l d situation (Levi-Strauss 1966). He does, however, maintain that i t is possible to construct models of the systems we are interested in, in the laboratory, to deter-mine how they would function when the basic relationships are trans-formed and manipulated to f i l l out every logical possibility. Thus we would obtain two series of phenomena — laboratory results and f i e l d examples. The process then becomes one of succes-sive trips from the laboratory to the f i e l d , f i e l d to the laboratory to find a f i t . But as i n so many of his endeavours Levi-Strauss reasons for a particular procedure are rarely made expl i c i t . He does not expound on the nature of models, nor on this process as a testing de-vice for theory. However, the exercise he proposes seems meaningful mainly in terms of theory construction and a precedence for laboratory work. With reference to decision theory considerations, in the laboratory one can set up and control a series of logical p o s s i b i l i t i e s for individuals taking risks which would allow one to systematically manipulate a set of incentive conditions and a set of external constraints with respect to a given decision task. This arrangement provides the basis for the type of model-building that Levi-Strauss believes is important. - 189 -I maintained (in Chapter 2) that the laboratory and f i e l d are particular research settings in which my assertions about risk-taking behavior could be tested and compared. This assumption was conditional upon the same scope conditions being met in each setting. The scope conditions placed an individual decision-maker within parameters of wealth and subjective u t i l i t y for some outcome, information and incen-tive conditions for any risk; and my propositions predicted the type of decision strategy used. In the laboratory and the f i e l d settings these conditions were met, but where I anticipate objections is with respect to the operationalisation of variables such as wealth and information into low and high categories. Can a $1 wealth category be s t r i c t l y comparable to my mea-sure of low wealth for Alert Bay fishermen? Not directly, but I am not going to argue that exactly the same scope conditions applied in the two situations with regard to measurement of parameters. I w i l l argue that the wealth and information gradients relative to the fi e l d and laboratory situation are comparable as each was calculated from a baseline of the lowest category having the property of not being able to afford a loss in resources. The theoretical propositions are translated into different substantive settings via different operational definitions, because I am not concerned with wealth or information per se but with gradients of these variables relative to particular contexts. - 190 -Where there is a wealth gradient, information gradient and a decision involving risk — relative to a particular situation — these are the scope conditions I specify must be present in a l l cases. The methodology I have espoused provides necessary links between farmers accepting innovations, fishermen deciding where to fish and sophomore students gambling. I am trying to say something general about the sorts of risks people w i l l take when they value an end highly. I am in fact dealing with a means (decision) ends (goal) relationship. The end can vary — maximise productivity, increase r i t u a l cleanliness, or promote family solidarity. What I am submitting i s that any action decided upon towards that end can be analysed in terms of incentive conditions, information processing and u t i l i t y preferences. The calculation of the incentives, both positive and negative can take place on one or more dimensions — but that this calculation is made is the major assumption of my thesis. On Models and Research Cycles Anthropologists have long argued among themselves and with economists regarding the range of applicability of economic theory. I do not propose to go any further into the controversy as I stated earlier (Chapter 1) that I view economic theory as a set of axioms which have standard interpretations in a market context, but which are uninterpreted in a non-market context. - 191 -This implies that the "real" question 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, and is a plea for getting on with the job of constructing and testing models instead of talking about putative domains of relevance. The progression of the work attempted here has demonstrated the usefulness of borrowing analogies from economics. Using a decision theory format and modifying i t to include situational constraints I have shown that a similar type of decision making occurs in different substantive contexts. The model I built was not a perfect f i t , but adequate enough to give one confidence in the methodological assumptions that were fun-damental to i t s construction. The lack of congruence in testing pro-cedures meant that in some cases i t was not possible to test a l l the propositions i n a l l contexts. However this is a function of adverse situational contexts and inadequate operationalisation of variables and does not cast doubt on the methodological assumptions used to construct the model. Perhaps the best way to assess the merits of the approach attempted here is to suggest an ideal research paradigm and measure my work against this. In addition to the implications this may have for enquiry i n economic anthropology i t should suggest necessary re-visions as far as the model attempted here is concerned. - 192 -I think that much of the debate in economic anthropology could be rendered superfluous i f i t were possible to get the proponents in both camps to adhere to a research paradigm that has served science extremely well (Greer: 1969). This paradigm can be characterised as a cycle having five distinct steps: (1) select a particular problem; without a problem no one — no matter how hard he tries — can graduate beyond imprecise generalities. (2) the problem is conceptualised i n such a way as to suggest a t r i a l solution. A model is constructed which places the problem within a larger frame of reference. (3) the t r i a l solution is tested, logically and empirically. (4) the testtresults are evaluated against what was expected, which leads to (5) the extension and editing of the model, and the generation of new problems. The problem I was concerned with was i n terms of individuals making decisions under conditions of uncertainty and loss. I constructed a model of risk-taking which predicted the way individuals would make decisions subject to constraints of information, wealth and subjective u t i l i t y for the outcome the decision is about. The laboratory test indicated that my statements were appli-cable to a domain narrower than I had expected. - 193 -(1) It indicated that increments in information — after the range of odds and payoffs were specified — had no significant influence upon decision-making. (2) It also indicated that some individuals from low wealth categories would take risks for low payoffs when the prospect and amount of loss were low, and the prospect of the payoff was high. (I had asserted that this category of individual would not go for low payoffs under any circumstances.) On the credit side, however i t showed unequivocally that low wealth individuals did have a definite preference for low risk strategies, as the exceptions i n (2) above s t i l l indicated a prefer-ence for low risks. Those individuals'who opted for marginal payoffs only did so when the remaining incentive conditions were strongly in their favour. In other words low payoff bets were taken by some low wealth individuals when i t was highly likely that they would be success-ful (high P ) and when the cost factors were negligible (low P , low $ ). w ± 1 Increments in wealth from low to medium were accompanied by significant increases in high risk taking,though high wealth individuals did not take significantly different types of risk to medium wealth individuals. I think, however this result may be due rather to an in-adequate distinction between upper wealth levels than to a uniform pattern of risk taking across medium and high wealth levels. The f i e l d test suggested that further constraints on my model were necessary, as i t did not adequately predict the investment risks - 194 -taken by a group of Japanese and Yugoslav gillnetters. The insertion of the supplementary proposition with respect to social marginality and risk taking accounted for the higher than expected level of risk taking. In empirical terms, one of the logical p o s s i b i l i t i e s in the laboratory — t h a t of high information — did not occur, so I was forced to deal with only two levels of information i n terms of strategy risks. Also the assumptions of relative weighting applied to the wealth and education parameters, which were fundamental to the infor-mation circuits model, were not supported in some cases. Despite these constraints the sample of gillnetters from Alert Bay indicated general support for my assertions about risk taking behavior in terms of the predicted relation between risk taking, u t i l -i t i e s , information and wealth. The evidence marshalled from secondary sources was of a gen-eral corroborative nature and two of the contrary cases were shown to l i e outside my specified scope conditions. However the third case — hybrid corn in New Mexico — illustrated the necessity of considering the constraints imposed upon decision making by conflicting u t i l i t i e s . The New Mexico example showed that while a high u t i l i t y was exhibited for increasing corn yield, there was a higher u t i l i t y attached to corn quality which eventually led to the rejection of the corn hybrid. This suggests that rankings of u t i l i t i e s associated with particular types of outcomes w i l l have to be made in order to specify the dominance of one u t i l i t y over another. - 195 -Editing Editing of the model can be done in two ways. F i r s t l y , as a result of the tests, one can stipulate more precisely the range of applicability of the model; and secondly one can offer revisions of the basic statements. Proposition 1': "Individuals w i l l show a preference for low risk stra-tegies with respect to decisions involving the outcome to which HU is attached." Conditions: LWHU. The experimental test showed that the LW/H.Inf category did not show the anticipated preference for low risk strategies that pro-position 1' predicted. However the fact that I was unable to find an empirical counterpart to the LW/H.Inf possibility suggests (a) that i t ;is not to be found in the real world and (b) that i t s nonoccurrence is a function of the effects of intervening variables (such as education) which prevents the configuration from occuring. In fact the information circuits model used the parameter of education in such a way that no low wealth individual ever had access to high information. As this part of the model was strongly supported by the f i e l d test, i t seems reasonable to stipulate that the experimental result is not a rebuttal to the proposition, as i t deals with a logical possibility which has no empirical referent. The f i e l d test strongly supported proposition 1'. Under conditions of low information low wealth fishermen took 44% more low - 196 -risks than medium wealth fishermen, while high information produced a 54.3% difference in the proportion of low risks taken by low wealth as compared to medium wealth fishermen. These figures show conclusively that low risk strategies are preferred by low wealth fishermen. The discussion of the Agricultural Revolution i n Asia was in terms of peasant farmers being offered high yield, low cost pro-duction factors. The Director General of F.A.O., A.H. Boerma stated in the 1969 annual report that the food problem facing the world in the near future is more likely to be surpluses than starvation (FAO 1970) . The underdeveloped areas of the world increased their agricul-tural production by 5% i n 1967 and 2% in 1968 with estimates for 1969 showing a continuation of the trend. This startling change from F.A.O.'s prophesies of doom and famine (as recently as 1967)has been effected by breakthroughs in agricultural technology, especially in new strains of rice and grains. As i t would be naive to expect peasant farmers to have changed drastically i n three years, I would submit that they are operating on the same strategy that they have been forced to use for centuries — low risk. Development agencies have generally accepted the agricultural revolution as a boon and a windfall; and very l i t t l e has been done in analysing why i t happened in the late 1960's and not before. I would submit that i t was only in the late 1960's that development planners could offer low risk production factors to peasant farmers. - 197 -In India and the Philippines, hybrid grains and rice have been adopted wholesale simply because a farmer does not have to take a high risk to invest in them. The opportunity to invest in these new crops constitutes a low risk strategy for the farmer, as his costs are minimised through subsidies, and the incentive conditions of high yields reassures him, at the very worst, of a harvest bigger than that which can be obtained by using traditional varieties. Thus proposition 1' can be accepted as valid. Proposition 1': "Individuals w i l l show a preference for low risk stra-tegies with respect to decisions involving the outcome to which HU is attached." Conditions: LWHU. 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". The experiment indicated that s t a t i s t i c a l l y significant i n -creases in risk taking did not occur between medium wealth and high wealth categories, (though I suggested that this could be attributed to an inadequate distinction between the upper wealth levels). Sig-nificant increases in high risk taking were observed between low and medium wealth levels. In the f i e l d , fishermen exhibited an increased tolerance for high risk strategies as wealth increased, though comparison was d i f f i c u l t between medium wealth and high wealth categories as I only had two - 198 -fishermen in the latter category. However the fact that these two fishermen never took low risks while medium wealth operators did, gives some support to the overall contention of the proposition. The evidence from secondary sources indicated that wealth operates as a general predictor for risk taking. Bauer (1946) found that risks taken on replanting and newplanting of rubber trees in Malaya were directly related to the wealth levels of the growers concerned. Fliegel et a l i a (1968) showed that the rate of adoption of agricultural technology i n Indian villages was positively and significantly related to measures of affluence. This general support for proposition 2', however, even when taken in conjunction with the fact that high wealth fishermen took more high risks than medium wealth fishermen does not permit one to say with confidence that risk taking increases beyond the medium level. while the f i e l d test and evidence from secondary sources does not contradict the proposition, too much weight should not be attributed to a sample of two and evidence of a general nature. So, while I tend to think that the experimental results may be a function of inadequate distinctions between the upper wealth levels, there is insufficient evidence from other sources to show that the pat-tern of wealth vis-a-vis risk taking suggested by the experiment could be disregarded. These considerations lead to a revised proposition: 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 from low to medium." Conditions: HU - 199 -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 information increases." The most interesting result from the experiment was the threshold point of information vis-a-vis risk taking. Once the range of odds, payoffs and costs had been specified (Medium Information in the experiment) further increments i n information did not produce sig-nificant changes in patterns of risk taking. This implies that a l -though uncertainty is an important part of risk taking, after a certain point has been reached uncertainty can continue to be lowered with no resulting change in decision making. It could be argued that again the experimental distinctions between the upper information levels were weak, but as I pointed out earlier (page 84 ) there is evidence from other experimental results to indicate that high information is distinctly different to medium information. The difference can be attributed to the fact that high infor-mation affects low wealth subjects so that they show?.-: similar kinds of preferences for low risks,as do MW/H.Inf subjects, while medium information operated on low wealth subjects to produce a significantly different pattern of risk taking to that exhibited by MW/M.Inf subjects If the experimental distinction was inadequate then there would be no difference i n the effect that high information and medium information h on low and medium wealth levels. Thus i t is reasonable to accept the threshold effect of information increments upon risk taking as valid. - 200 -Unfortunately the f i e l d test did not provide me with a high information category analoguous to the experimental one. With regard to decisions on fishing strategies there was no effective way that i n -formation could be increased for the fisherman beyond specifying the range of odds, payoffs and costs for any alternative. Thus I was un-able to see whether increments in information after the threshold point, identified in the laboratory, made any difference to decision strategies employed. Information in terms of investment risks on technology was concerned with measuring the number of information circuits an indi-vidual had access to, and did not control for cut off points in speci-fying the four risk factors. As i n the previous proposition, the evidence from secondary sources was of a general rather than a specific nature, which means that the experimental result with regard to information and risk tak-ing can be used to revise proposition 3' as follows: 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 information increases to the point where the range of risk factors i s specified." Conditions: HU. - 201 -Intervening Variables (1) Education I stated e a r l i e r that there had to be some intervening var-i a b l e which prevented the LW/H.Inf configuration from occurring and pointed out that the information c i r c u i t s model v i a the parameter of education was the offending v a r i a b l e . This model v i s u a l i s e d access to information c i r c u i t s about technology as a function of wealth and education, and that r i s k taking on technology was a function of information category. I assumed a r e l a t i v e weighting of wealth v i s - a - v i s education i n order to get a graded sequence of c i r c u i t s . This assumption was not always:borne out by the f a c t s . For instance i n the f i e l d t e s t the MW/MEd category ends up on c i r c u i t e which predicts medium information. Of the seven fishermen i n this category four had high information while three had medium information as predicted. The three fishermen i n the MW/LEd category a l l had medium information instead of low information as predicted. The only other "deviant" from the model was a LW/MEd f i s h e r -man who had medium information instead of the low that was predicted. But as four other fishermen i n th i s category had the predicted i n f o r -mation, I do not f e e l that a r e v i s i o n i s c a l l e d f o r i n th i s s ection of the model. These considerations suggest a new ordering of c i r c u i t s with regard to information category as below: - 202 -From e we ean stipulate that there i s a 4 i n 7 chance (ap-proximately p=0.6) of the individual having access to high information, and a 3 i n 7 chance (approximately p=.04) of an individual having access to medium information. Using the same type of probability statements, the medium risk taking category can be eliminated (I only measured low and high risks in the field) and we can state that for any individual with med-ium access to information circuits there is a 0.25 chance that he w i l l take high risks and a 0.75 chance that he w i l l take low risks. This is based on the fact that of eight medium information fishermen six took low investment risks while only two took high risks. 7 - 203 -Intervening Variables (2) Social Marginality During fieldwork evidence was collected that supported the supplementary proposition relating considerations of social marginality to risk taking. I observed that Japanese and Yugoslav gillnetters took far higher investment risks then one would expect, and that their i n -vestment in technology was considerably more than the Alert Bay collec-t i v i t y of high risk investors. Although I only collected information on five Japanese boats and three Yugoslav boats, for the skippers concerned I was able to establish that their ethnic separation was socially reinforced by higher frequencies of interaction with people of the same ethnic group then with outsiders. This and other considerations of language and residence seemed to meet the conditions of social marginality specified in the section on supplementary propositions. The size of the sample is too small to draw definite conclusions, but on the results of my observations I would like to offer an additional proposition, though of a tentative nature. Proposition 4': "The proportion of individuals preferring high risk strategies, with respect to an outcome to which HU is attached, w i l l be higher i n socially marginal groups than in non-marginal groups; a l l other things being equal." - 204 -Scope Conditions: The U t i l i t y Parameter The conditions s p e c i f i e d as being relevant to my t h e o r e t i c a l considerations placed an i n d i v i d u a l decision maker within parameters of wealth and subjective u t i l i t y f o r some outcome, information and incentive conditions for any r i s k . In Chapter 5 the examination of contrary cases showed that on the whole they lay outside the scope conditions of my concern, i n p a r t i c u l a r they did not r e f e r to the s p e c i f i e d HU parameter. However, one case (that of hybrid corn i n New Mexico) indi c a t e d that while the HU parameter was met, the farmers eventually rejected a high y i e l d / -low cost production farmer. This was due to a higher u t i l i t y being placed on corn q u a l i t y than on quantity, which implies that the s t i p u l a t i o n of "HU fo r any outcome" as a scope condition has to carry an a d d i t i o n a l q u a l i -f i c a t i o n i n terms of ordering of u t i l i t y p r i o r i t i e s . This can be done by s t a t i n g that the scope of my t h e o r e t i c a l considerations applies to "any outcome to which HU i s attached, and which i s not dominated by u t i l i t i e s attached to other outcomes." New Problems To l i s t the new problems generated by the t e s t i n g procedure i s a major task i n i t s e l f — however the more immediate tasks concern: (a) the testing procedure i t s e l f . The immediate task i s to construct better indices and measures i n order to provide a less clumsy t o o l - 205 -than the one used here. The use of better sampling techniques would allow me access to parametric techniques which could pro-vide weightings for the effect different variables have upon risk taking; the attempt to axiomatise considerations of social marginality, information, resources, u t i l i t i e s and risk taking so that a l l the statements can be derived from a set of basic hypotheses; the inclusion of a wider set of decision making strategies. I only dealt with high risk and low risk strategies (maximise gain and minimise loss respectively), whereas calculations of minimax strategies would provide a measure for medium risk. This would give a more extensive set of statements with regard to risk taking and provide a better f i t for actual decision making s i t u -ations; the testing of the revised hypotheses in other situations involving risk; the generation of a new cycle of laboratory and f i e l d tests. Levi-Strauss' dictum that research endeavour should involve suc-cessive trips from the laboratory to the f i e l d , thence from the f i e l d to the laboratory (Levi-Strauss, 1961) implies that the next stage of the research begun here lies in the laboratory. To this end I intend to run another decision making experiment which modifies to a considerable degree the design used in this research. - 206 -This is partly due to the need for better operational defi-nitions, but also is concerned with the fact that a different set of propositions w i l l be tested. The duplex gamble w i l l be used to manipulate the incentive conditions (payoffs, costs, and probabilities) attached to different bets. The use of experi-mental salaries to play for actual stakes to simulate r e a l i s t i c conditions and the subjective evaluations of each gamble's attrac-tiveness proved to be worthwhile. With regard to the latter aspect I would test to see what kind of association exists between the evaluation of the attractiveness of different bets and the decision made whether or not to play a particular bet. This w i l l offer a check on the relation between evaluation scores and decisions to play bets, and w i l l permit me to assess whether the lack of high risk taking by some subjects is related to different subjective evaluations of the bets considered. The major modifications involve the operationalisation of the wealth and information parameters. I was not sure that the distinction made between upper wealth levels was adequate. The wealth gradient w i l l s t i l l be calculated in terms of the lowest category having the property of being unable to afford losses. By demarcating the low wealth level at $2, medium at $4 and high at $8, I w i l l be able to map the wealth categories into an i n -terval scale. The medium and high wealth figures are exponen-t i a l functions of the low wealth category taken as a base. - 207 -(LW=2 ; MW=2 ; HW=2 ) By assigning a measure of 1 to low wealth, 2 to medium wealth and 3 to high wealth, I can map the wealth gradient into an interval scale. The assumption of additivity and transitivity required for an interval mapping are met by considering each wealth level as an exponential function to the base 2. In theoretical terms I w i l l construct a set of proposi-tions which assert that risk taking is an exponential function of wealth. In order to assess more precisely the interrelation be-tween wealth and information with regard to risk taking, I intend to permit subjects to buy information about the gambles. They w i l l be presented with a vague description of each bet and w i l l be able to purchase two types of information. For 50$ the range of any one of the four incentive conditions can be bought; for $1 the precise specification of the appropriate payoff or probability w i l l be given. The mandatory $1 betting stake w i l l mean that the low wealth subjects ($2) can only spend $1 on information as they require the remaining dollar to bet with. The maximum amount that a subject can spend on information is $4 and the minimum is $0. The financial outlays made on information enables one to con-struct an interval scale for information corresponding directly to the amount paid out by the subject. This transformation rests on the premise that subjective u t i l i t i e s attached to small monetary amounts correspond to the objective value of the amount. (Edwards, Lindman, Phillips 1965) - 208 -The advantage of mapping these parameters into i n t e r v a l scales i s that i t allows me access to parametric techniques which permit a c a l c u l a t i o n of the weights that can be a t t r i b u t e d to the incentive conditions and s i t u a t i o n a l constraints with regard to high and low r i s k taking. This would then have implications f o r the subsequent f i e l d t est as the purpose would not simply be to see i f r i s k taking i s i n the predicted d i r e c t i o n but whether or not s i m i l a r types of weightings are attached to the constraints s t i p u l a t e d as f u n c t i o n a l to r i s k taking. This consideration would involve a much more rigorous o p e r a t i o n a l i s a t i o n of v a r i a b l e s , and the main d i f f i c u l t i e s foreseen i n subsequent f i e l d tests are with respect to this problem of measurement. The next experiment w i l l s t i l l deal with bets as unique, di s c r e t e events, though subsequent experiments would be designed to incorporate the e f f e c t s of p r i o r winnings.and losses on de-c i s i o n making. The choice of subjects could also be varied. One obvious domain f o r t e s t i n g i s fishermen themselves, though other occu-p a t i o n a l groups — policemen, stockholders, bartenders — would be f e r t i l e avenues to explore and compare. One parameter that must remain constant, however, i s the u t i l i t y parameter. I t may appear desirable to examine the e f f e c t s that a low u t i l i t y for an outcome would have upon r i s k taking. But t h i s would be a c o n t r a d i t i o n i n terms as one would no longer be dealing with - 209 -r i s k . Risk taking refers to behaviour i n situations where there i s a desirable goal and a lack of certainty that i t can be at-tained, with attendant p o s s i b i l i t i e s of loss (p. 42). Therefore, i f subsequent experiments are to be concerned with testing the-ories of r i s k taking they necessarily involve a high u t i l i t y for the outcomes examined. From some of my e a r l i e r remarks about the experimental s i t -uation as a testing device (p. 54, 92ff, 198) i t could be inferred that I attach more weight to the significance of laboratory re-sults than I do to f i e l d r e s u l t s . I should state that this i n -ference i s not j u s t i f i a b l e , as I regard the laboratory as an equivalent testing ground to the f i e l d . My concern i s to show i t s potential significance as a tool for enquiry i n anthropology. The value of any test — be i t i n the laboratory or the f i e l d — l i e s i n the construction of the research design and not i n the i n f a l l i b i l i t y of either context. Theoretical Implications The immediate theoretical implication i s that I have shown that individuals take risks i n s i m i l a r sorts of ways i n divergent sub-s tantive ins tances. This implies that effective cross-cultural analysis requires a process of abstraction at a l e v e l that does not have s i t u a t i o n a l boundaries. The operational definitions of the abstract concepts i s the mechanism by which s i t u a t i o n a l boundaries are translated into measurable parameters relevant to the model. - 210 -But I must emphasize that the abstract conceptualisation has to be free from the situational referent. The obvious extension of my assertions about risk-taking is into the exchange theory of Homans (1958), Belshaw (1965), Blau (1964), the communication theory of Levi-Strauss (1967) and the transactional models proposed by Barth (1966). I categorised the individuals in my thesis according to a wealth str a t i f i c a t i o n and was concerned with decisions with respect to s t r i c t l y material outcomes — financial increments. To permit wider generality, instead of referring to wealth, I could refer to the posses-sion of a particular resource (prestige, r i t u a l status, control of sanc-tions, etc.) and analyse decision-making in terms of risking that re-source in order to increase i t . Considerations of information, u t i l i t y and incentive conditions would apply, though they would involve dif-ferent operational definitions. Costs and payoffs would have to be calculated along social as well as economic dimensions, and given that the measurement problems could be overcome, one could construct a predictive model of risk tak-ing with regard to particular resources and particular ends. Further implications are with regard to the way in which the traditional subject matter?of.anthropology may profitably be viewed. I argued earlier (Chapter 5) that social structure is a heuristic device used by the observer and is not a given. I stated - 211 -further that social structure exists for any actor in a social system when i t provides access to resources that he values and needs. Resources can be viewed in terms of material commodities such as land and money but I am also referring to intangible resources such as status, leisure, power and security. If "status" as a valued and needed resource can only be obtained through a particular structure then that structure is important to the actor making decisions about the resource "status". From this standpoint of resources and actors making decisions, social structure can be viewed as the outcome of choice behavior. Regu-lar patterns of choice behavior at the aggregate level define the sig-nificance of particular structures such as caste, or kinship groups in terms of the resources they can control and distribute. Once the desired resources can be obtained by alternative means and sufficient individuals decide to use those means, then the importance of caste or kinship a f f i l i a t i o n wanes and new structures emerge as a function of the change in patterns of choice behavior. From this viewpoint i t would seem that decision making provides a basic building block for the study of social behavior. Supplementary to this view of social structure is the way in which one could regard culture. Instead of the role of culture as a residual category in analysis, one could put i t into a more precise perspective by stating that cultural values provide the parameters within which individuals make decisions. - 212 -Ethnographic descriptions of p a r t i c u l a r value systems and cultures then permit one to designate the c u l t u r a l l y perceived a l t e r -natives that are appropriate to p a r t i c u l a r d ecision 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) for choosing between the a l t e r n a t i v e s . This implies that the major task for the anthropologist i s the i d e n t i f i c a t i o n of these constraints. I would state that men use c u l t u r a l l y appropriate means to cope with c u l t u r a l l y defined problems but the assumption behind a decision theoretic approach i s that the means chosen are chosen r a -t i o n a l l y . Men act i n order to get the things they value and need through the r a t i o n a l use of the rules and resources of t h e i r p h y s i c a l and c u l t u r a l environments. The r u l e s , rewards, resources and man's perception of them can change and vary through time and space, but the argument made here i s that the process of r a t i o n a l c a l c u l a t i o n i s a constant cross c u l t u r a l v a r i a b l e . P r a c t i c a l Implications The possible p r a c t i c a l implications of the work attempted here can perhaps most r e a d i l y be seen i n the area of planning for development i n the Third World. This i s on the grounds that t h i s thesis i s concerned with the micro-process of decision making. For instance Michael Lipton (1969) i n a review of Adelman and Thorbecke (1966) notes that Part I of the book which i s concerned with "Development Theory" bears no r e l a t i o n to Part II which i s t i t l e d "Development Planning". - 213 -This does not surprise him as he states "Can one expect ap-plied macro-economics to be properly linked with theory when we are so dismally lacking in a micro-economics of development or, indeed, in tested propositions about the decision procedures of farmers, who com-prise two-thirds of economic decision-takers i n underdeveloped countries and produce half their output?" (1969: 641) He comes out strongly in favour of concentrating research and f i e l d energy on the analysis of micro-processes and of the authors in this volume states "If even 10% of this talent were devoted to the testing of micro-economic hypotheses — as regards for example, peasant decisions — the next book of this type might be really useful" (1969: 645) . However, Lipton may have overstated his plea for exclusive micro-studies as they are a beginning for systematic enquiry and not an end in themselves. Methodologically micro-theory in terms of l i m i -ted problem areas and a concern with formalisation is a l l that one can effectively handle at the moment and not lose track of rigour. It should be emphasised, however, that micro-analysis is only a building block for macro-theory. Although my specific considerations are in terms of a micro-theory of risk taking, in the context of farmers and agricultural i n -novations i t could be linked to a model of macro-eaonomic development in terms of the following assumptions: (1) an increase in risk-taking with respect to new factors of produc-tion by peasant farmers w i l l lead to an increase in productivity; - 214 -(2) increased productivity leading to surplus conditions is one of the keys to sustained economic growth because (a) with an agricultural surplus the agricultural production sector of an economy provides productive inputs for other sectors of the economy, (b) this increases incomes and rural purchasing power and i n turn produces higher demands for consumer goods from the i n -dustrial sector, with attendant increases in investment and savings; ultimately this leads to the creation of a situation of self-sustaining growth; (3) the other key to the development problem is that surplus agricul-tural labour be relocated i n a productive capacity in some other sector of the economy — industrial or commercial. In this way my considerations could be tied in with macro-development models of the nature of Ranis and Fei's "Development of the Labour Surplus Economy" (1965). In addition to this tie in with macro-models, there is an urgent strategic reason for concentrating on micro-processes with re-gard to implications for development policy. This is because, by and large, development planners have very l i t t l e information and knowledge to go on when they come to formulate plans. The information from the tests I have conducted should indi -cate to the planner certain constraints and likely outcomes for policy. With the knowledge that poor farmers cannot be induced to take high - 215 -risks on new agricultural practices, and that the present agricultural revolution in Asia is a function of providing low-risk production fac-tors, the planner knows (more than he did) the parameters he has to work with. In a strategic sense my considerations could be regarded as a link between theory and the practical world of development planning. Conclusion What I have attempted to do in this thesis does not coincide with what I would have liked to have done. Measurement problems denied me the level of rigour I thought desirable, but I consider this to be at least a step in the right direction. The procedure of using laboratory tests as a guideline for fieldwork may seem somewhat novel, but I must confess that in other disciplines, the procedures I have used are long since standardized. However, i f this thesis is to be a contribution to the grow-ing formal tradition in economic anthropology, i t s explication of the process of decision making should enable one to have confidence in the methodological assumptions that were fundamental to both model construction and testing procedures. - 216 -BIBLIOGRAPHY A Adelman I. and Thorbecke E. 1966 Theory and Design of Economic Development. Johns Hopkins Press. Apodaca A. 1952 Corn and Custom — The Introduction of Hybrid Corn to Spanish American Farmers in New Mexico. In Spicer E.H. (ed.): Human Problems in Technological Change. N.Y.: Russell Sage Foundation. B Barth F. 1966 Models of Social Organisation. R.A.I. Occasional Papers, No. 23. Bauer, P.T. 1946 Report on the Rubber Smallholdings in Malaya. H.M.S.O. Colonial Research Publications No. 1. Beal, G.M., Rogers E.M., Bohlen, J. 1957 Validity of the Concept of Stages i n the Adoption Pro-cess. Rural Sociology 22: 166-168. Belshaw, C.S. 1965 Traditional Exchange and Modern Markets. Englewood C l i f f s , N.J.: Prentice Hall; Blau, P. 1964 Exchange and Power in Social Life. N.Y.: Wiley. Buchler, I.R. and Nutini H.G. (ed.) 1969 Game Theory in the Behavioral Sciences. University of Pittsburgh Press. Burling, R. 1962 Maximisation Theories and the Study of Economic Anthro-pology. American Anthropologist 64:802-821. - 217 -C Cancian, F. 1967 Stratification and Risk Taking: A Theory Tested on Agri-cultural Innovation. American Sociological Review 32: 912-927. Cook, S. 1966 The Obsolete 'Anti-Market' Mentality: A Critique of the Substantive Approach to Economic Anthropology. American Anthropologist, 68: 323-345. Copp, J.H. 1958 Toward Generalisation in Farm Practise Research. Rural Sociology 23: 103-111. D Dalton, G. 1961 Economic Theory and Primitive Society. American Anthro-pologist. 63: 1-25. 1969 Theoretical Issues in Economic Anthropology. Current Anthropology, February 1969. Davenport, W. 1960 Jamaican Fishing: A Game Theory Analysis. Yale Uni-versity Publications in Anthropology, 59. Department of Fisheries of Canada 1969 Fisheries Statistics of British Columbia 1968. E Economist 1967 Food for India. Economist, May 17. Edwards, W. 1954 The Theory of Decision Making. Psychological Bulletin 51: 380-417. Edwards, W., Lindman, H., and Phi l l i p s , L.D. 1965 Emerging Technologies for Making Decisions. In New Di-rections in Psychology II. T. Newcomb (ed.), N.Y.: Holt, Rinehart and Winston. - 218 -F.A.O. 1967 Annual Report of the Food and Agriculture Organisation of the United Nations. 1965-1966 H.M.S.O. 1970 Annual Report of the Food and Agriculture Organisation of the United Nations. 1968-1969-H.M.S.0. Firth, R. 1939 Primitive Polynesian Economy. London: Routledge and Kegan Paul. 1964a Essays on Social Organisation and Values. London'. Athlone Press, 1964b Social Prerequisites of Economic Growth. UNESCO/55/38. Fliegel, Roy, Sen and Kivlin 1968 Agricultural Innovations in Indian Villages. ;Hyderabad-Institute for Community Development Studies. Foster, G.M. 1962 Traditional Cultures and the Impact of Technological Change. N.Y.: Harper. Goodfellow, D.M. 1939 Principles of Economic Sociology. London: Routledge. Gouldner, A.W. 1960 The Norm of Reciprocity. American Sociological Review 25: 161^178. Greer, S. 1969 Logic of Social Enquiry. Aldine. Griliches, Z. 1957 Hybrid Corn: An Exploration i n the Economics of Tech-nological Change. Econometrica 25: 501-522. Gross, N.C. 1942 The Diffusion of a Culture Trait i n Two Iowa Townships. Ames, Iowa State College. - 219 -Gross, N.C. and Taves, M.J. 1952 Characteristics Associated with Acceptance of Recommended Farm Practices. Rural Sociology 17: 321-327. H Hayden, H. 1954 Moturiki: A Pilot Project in Community Development. London, Oxford University Press. Herskovits, M. 1940 The Economic Life of Primitive Peoples. New York: Knop f. 1952 Economic Anthropology. New York: Knopf. Hogbin, I. 1958 Social Change. Watts: London. Homans, G.C. 1958 Social Behavior as Exchange. American Journal of Sociol-ogy 63: 597-606. Hsieh, Sam-Chung 1968 New Outlook for Asian Agriculture. International Devel-opment Review September 1968. Hoselitz, B.F. and Moore, W. 1966 Industrialisation and Society. UNESCO: Mouton. K Kogan, N. and Wallich, M. 1967 Risk Taking as a Function of Situation, Personality and Group. In New Directions in Psychology III. Newcomb, T. (ed.). N.Y.: Holt, Rinehart and Winston. L Leach, E.R. 1961 Rethinking Anthropology. London, Athlone Press. - 220 -LeClair, E.E. 1962 Economic Theory and Economic Anthropology. American Anthropologist 64: 1179-1203. LeClair,E.E. and Schneider, H.K. 1968 Economic Anthropology: Readings in Theory and Analysis. Holt, Rinehart and Winston. Levi-Strauss,. C. 1966 Scope of Anthropology. Current Anthropology 7,2: 112-123. 1967 Principle of Reciprocity. In Elementary Structures of Kinship. Aldon: Second Edition. Lipton, M. 1969 Review of Theory and Design of Economic Development, by Adelman I. and Thorbecke E. Economic Journal LXXIX: 641-645. M Marsh, CP. and Coleman, A. 1955 The Relation of Farmer Characteristics to the Adoption of Recommended Farm Practices. Rural Sociology 20: 289-296. Moerman, M. 1968 Agricultural Change and Peasant Choice in a Thai Village. Berkeley: University of California Press. Nature 1970 Food Production-Shortage to Surplus? 225; 5227: 9-10. von Neumann,J. and Morgernstern, 0. 1953 Theory of Games and Economic Behavior. Princeton Uni-versity Press. Orans, M. 1968 Maximising in Jajmaniland. American Anthropologist 70: 875-895. o - 221 -P Polanyi, K. 1957 The Economy as Instituted Process. In Trade and Market in the Early Empires. Polanyi, K. Arensberg, CW. and Pearson, H.W. (ed.), New York: Free Press. Prattis, J.I. 1967 The Implications of Technological Change in a Theory of Community Development. Institute of Social Anthro-pology, Oxford University. Unpublished B.Litt. dissertation. R Ranis, G. and Fei, J. 1965 Development of a Labour Surplus Economy. Homewood, I l l i n o i s : Irwin. Robbins, L. 1935 An Essay on the Nature and Significance of Economic Sci-ence. London: Macmillan. Rogers, E.M. 1962 The Diffusion of Innovations. New York: Free Press of Glencoe. Ryan, B. and Gross, N. 1943 Diffusion of Hybrid Seed Corn i n Two Iowa Communities. Rural Sociology 8: 15-24. S' Schultz, T.W. 1963 The Economic Value of Education. N.Y.: Columbia University Press. 1964 Transforming Traditional Agriculture. Yale University Press, 1965 Economic Crisis in World Agriculture. University of Michigan Press. Slovic, P. 1968 Relative Importance of Probabilities and Payoffs i n Risk Taking. Research Bulletin, Vol. 7, No. 7. Oregon Re-search Institute, Eugene. - 222 -Slovic, P. 1969 Differential Effects of Real vs. Hypothetical Payoffs upon Choices Among Gambles. Oregon Research Institute, Eugene. Spicer, E.H. 1952 Human Problems in Technological Change. Russell Sage Foundation. Tax, S. 1953 Penny Capitalism. Smithsonian Institution, Publication No. 16. - 223 -APPENDIX I Duplex Gambles Master Sheet :t No. EV $ P P, w w 1 1 1 0 7.5 .55 7.5 .55 2 -4.875 7.5 .2 7.5 .85 3 0 3.5 .55 3.5 .55 4 -0.15 3.5 .2 1 .85 5 -1.475 7.5 .2 3.5 .85 6 -1.375 1 .55 3.5 .55 7 2.275 3.5 .85 3.5 .2 8 -0.65 1 .2 1 .85 9 -5.675 3.5 .2 7.5 .85 10 2.2 7.5 .55 3.5 .55 11 0.65 1 .85 1 .2 12 3.575 7.5 .55 1 .55 13 -6.175 1 .2 7.5 .85 14 6.175 7.5 .85 1 .2 15 0.15 1 .85 3.5 .2 16 2.775 3.5 .85 1 .2 17 0 1 .55 1 .55 18 5.675 7.5 .85 3.5 :2 19 -3.575 1 .55 7.5 .55 20 0.65 7.5 .2 1 .85 21 -0.65 1 .85 7.5 .2 22 -2.2 3.5 .55 7.5 .55 23 1.375 3.5 .55 .1 .55 24 1.475 3.5 .85 7.5 .2 25 -2.275 3.5 .2 3.5 .85 26 4.875 7.5 .85 7.5 .2 27 -2.775 1 .2 3.5 .85 The expected values of the 27 duplex gambles range from a high of +$6,175 to a low of -$6,175. Only three gambles are " f a i r " — 1,3 and 17 — i n that their expected values are zero. This minimal con-cern with " f a i r " bets reflects the fact that my purpose i n using the duplex gambling design is not to have the two u t i l i t y functions sum to zero, but is to manipulate and vary the four incentive conditions of the bets. This is i n order to assess the significance of each component for risk taking and also to present discrete alternatives for the subjects to choose between. 224 APPENDIX II Ref/L.Inf INSTRUCTIONS TO SUBJECTS* I would l i k e to thank you for helping me i n t h i s experiment. The experiment you are about to p a r t i c i p a t e i n c a l l s f o r you to make evaluations aboud d i f f e r e n t gambles that w i l l be presented to you. Although the information given to you w i l l not be s p e c i f i c , each gamble can be v i s u a l i s e d as having four components, an amount to win ($ ); the chance of winning (C ); an amount to lose ($,); and a w w L chance of l o s i n g (C^). I t can be represented by v i s u a l i s i n g two discs as below: The L e f t hand disc represents the chances of winning a cer-t a i n amount. (60% chance of winning $4); the r i g h t hand dis c represents the chances of l o s i n g a c e r t a i n amount (40% chance of l o s i n g $5). Each gamble has these four components. * This appendix contains three sets of i n s t r u c t i o n s to subjects — low information, medium information and high information categories res-p e c t i v e l y . The only difference i n the i n s t r u c t i o n sheets i s i n terms of the information about the bets which i s dispensed to subjects. The process of playing, and i n s t r u c t i o n s with reference to evaluating and choosing bets to be played remains the same i n each case. - 225 -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 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, and you can regard i t as the stake that is required for each bet. 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. Bet A: Some people have tried this bet. 1. Do you think this is an attractive gamble? •ft 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  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 indif -- 226 -ference point at zero represents a situation where the gamble appears neither attractive nor unattractive. The third question calls for your decision. The description of Bet A does not specify any of the four components of the gamble. However for your information the range of the four components i s as follows: ( $ w ) : Amount to win has a range from $0 - $10. ( C W ) : chance to win has a range from 07, - 1007°. ($•£_): amount to lose has a range from $0 - $10. ( CL ) : chance of losing has a range from 07. - 1007.. 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 to pay $1 for every bet played. To determine your earnings for taking part in the experiment, those gambles that you decided to bet on in the questionnaire, you w i l l in actual fact play for real stakes. Say for instance you decided to play Bet A. Bet A: Some people have tried this bet, and that this corresponded to a gamble (on the master sheet that the experimenter has) that had a 607o chance of winning $6; 807o chance of losing $4. (You w i l l realise that the description "Some people have tried this bet", could allude to a variety of combinations such as 207. chance of winning $10; 507, chance of losing $4. OR 507o chance of winning $5; 507, chance of losing $5. OR 407. chance of winning $3; 607. chance of losing $2, - 227 -and so on. However there is only one such combination on the master sheet.) To come back to Bet A "Some people have tried this bet" which corresponds on the master sheet to 607. chance of winning $4; 807> 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 $1 obliga-tion remains constant for every bet you decide to play. To make i t clear what you have to do, examine the bet below. You w i l l see a verbal description of a gamble, and a series of questions. Take special note of the f i r s t question as this enables you to get more information about the gamble. Bet A: GOOD CHANCE of HIGH WINNINGS, FAIR CHANCE of MODERATE COSTS. <Cw> <V <CL> (V 1. Which two of the four components of the bet do you want precise information about? 2. Do you think this is an attractive gamble? - 246 -3. How would you rate i t on a degree of attractiveness scale? -5 0 +5 1 1 1 1 1 1 1 1 1 1 1  4. Would you bet on this gamble? The f i r s t question permits you to gain more information about the gamble. For instance, i f you want to know what the chances of lo-sing (C^) and the amount to- win ($ 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 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0103965/manifest

Comment

Related Items