Journal of Economic Psychology 31 (2010) 421–434

Contents lists available at ScienceDirect

Journal of Economic Psychology

journal homepage: www.elsevier .com/ locate/ joep

How a psychologist informed economics: The case of Sidney Siegel

Alessandro Innocenti *

Department of Economic Policy, Finance and Development (DEPFID), University of Siena, Piazza S. Francesco 7, 53100 Siena, Italy

a r t i c l e i n f o

Article history:Received 18 November 2008Received in revised form 20 January 2010Accepted 26 January 2010Available online 1 February 2010

JEL classification:B21B31C78C91D03

PsycINFO classification:2260

Keywords:Experimental economicsPsychologyBehavioral economicsBargaining theoryUtility theory

0167-4870/$ - see front matter � 2010 Elsevier B.Vdoi:10.1016/j.joep.2010.01.010

* Tel.: +39 0577 232785; fax: +39 0577 232793.E-mail address: innocenti@unisi.it

1 The first to be mentioned was Daniel Kahneman,Lecture, Smith placed Siegel first on the list: ‘‘I am indfaculty from 1955 to 1967 for warm tolerant supportencounters over the years on institutional and experfamily who have made all these years the best years

a b s t r a c t

In the 1950s before Kahneman and Tversky showed how behavioral economics could bringeconomics and psychology into a unified framework, a social psychologist, Sidney Siegel,entered the realm of economics and laid the foundation of experimental economics. Thispaper gives an assessment of Siegel’s effort to meld psychology and economics and showsthat Siegel was not only a contributor to the methodology of experimental economics butalso a pioneer of behavioral economics. Although his legacy was paramount in the work ofthe Nobel Prize winner Vernon Smith, Siegel endorsed a very different approach to makinginterdisciplinary research effective.

� 2010 Elsevier B.V. All rights reserved.

1. Introduction

Celebrating his 2002 Nobel Prize in Stockholm, Vernon Smith acknowledged the status of five scholarly pioneers of the‘‘intellectual movement” which led to the foundation of experimental economics. The second on the list, and maybe the leastknown,1 was a psychologist who had passed away in 1961, Sidney Siegel, whom Smith had met for the first time 6 weeks beforehis untimely death. In Smith’s words, in this short period Siegel had ‘‘strongly influenced me in becoming committed to exper-imental economics” (Smith, 2003). This tribute might seem surprising since Siegel had conducted his first experiment in 1954

. All rights reserved.

and the others, in order, Amos Tversky, Martin Shubik and Charles Plott (Smith, 2002a). In the Nobel Prizeebted to Syd Siegel for technical and conceptual inspiration; to John Hughes, Stan Reiter, and the Purduebeginning with my first experiment; to Charles Plott, Charles Holt and Martin Shubik for many valuable

imental issues; to students, visitors, the current ICES team, and especially to my growing but toleratingof my life.” (Smith, 2002b, p. 502)

422 A. Innocenti / Journal of Economic Psychology 31 (2010) 421–434

and published only two books and a handful of papers in psychology journals. Nevertheless, his brief acquaintance with eco-nomics had been significant enough to make him one of the scholars to whom Smith was most indebted.

Figuring out the exact reason for Smith’s indebtedness is not in fact straightforward. Ten years before, Smith had attrib-uted Siegel’s influence on his work to the insight that experimental subjects could deviate from utility maximization becausethey can get bored by repeating the same decision over and over. This intuition led Siegel to introduce some sources of var-iability in the laboratory, which restored maximizing behavior among subjects.2 Smith claimed that this, seemingly minor,problem was ‘‘of fundamental methodological importance, and I think it is unfortunate that it was not more widely knownamong experimentalists in both economics and psychology” (Smith, 1991a, p. 5). In this way, Smith addressed the questionof the relation between economic and psychology to which Siegel dedicated most of his brief career. He added, ‘‘Perhaps itwas not possible for this work to be widely known in either of these two cultures, if economists were willing to accept the pre-mise of the paper without evidence and if psychologists were unwilling to accept the premise with evidence” (Smith, 1991a, p.6). Smith’s point was that, before Kahneman and Tversky proved that behavioral economics could bring psychology and eco-nomics into a unified framework, Siegel had already given an example of how economics could be improved by absorbingexperimental methods. Moreover, he had shown how economists and psychologists could collaborate not only in the laboratorybut also in theoretical work. As Smith claimed, the value of his contribution has not been fully appreciated. The aim of this paperis to shed light on Siegel’s effort to meld psychology and economics and to argue that he was not only a contributor to the meth-odology of experimental economics but also a pioneer of behavioral economics. This appraisal will also indicate that Smith’stribute did not fully appreciate the richness of Siegel’s legacy.

Section 2 discusses Vernon Smith’s view on the relationship between economics and psychology. An account of Siegel’slife and career is presented in Section 3, while Section 4 assesses his contribution to experimental and behavioral economicsand contrasts it with Smith’s approach. Section 5 draws some conclusions.

2. Vernon Smith’s contrast between economics and psychology

Today, the relations between economics and psychology have become a popular topic under the heading of ‘‘newer”behavioral economics3 (Camerer, 1999; Rabin, 1998). Nonetheless, the path leading to this outcome was not straightforward.At the time that Siegel made his contribution, other scholars forcefully claimed that economics would benefit from incorpo-rating findings and theories from psychology.4 It is a fact, though, that these efforts to develop an alternative approach tomainstream economics were in practice neglected for a long time. The work of Tibor Scitovksy, Harvey Leibenstein, GeorgeKatona, and Herbert Simon had to wait three decades to be fully appreciated and developed. Only recently has the so-calledcognitive revolution made clear that focusing on the psychological determinants of choices and, specifically, on brain func-tioning as an information-processing device is quintessential to the understanding of economic behavior.5 This delay wasmainly due to the fact that mainstream economics adopts the concept of perfect rationality and utility maximization as abenchmark to which psychological insights should be tailored. In this way, economics would preserve its integrity: ‘‘We pre-dict that mainstream economics will ultimately meet the behavioral challenge by developing a new quasi-rational synthesis.Such a synthesis, for example, will identify when and where the framing of choices dramatically affects what choices aremade and it will study how framing is conducted and countered in the real world. Better predictions, say of consumerchoices, will be the result, but the standard framework of economic maximization will be for the most part preserved” (Laib-son and Zeckhauser, 1998, p. 32).

Judgments like the above take for granted that psychology and economics can successfully complement each other with-out dismissing their distinct methodological assumptions. Psychology can inject greater realism into economic models with-

2 ‘‘I first learned the full significance of reward saliency in experiments from the psychologist, Sidney Siegel (1961). The context was the very simple Bernoullitrials experiment. On each trial the subject’s take was to predict which of two independent events would occur. The more frequent event occurred withprobability p (say, 70%) and the less frequent event with probability (1 � p), where p is usually not known by the subject. For two decades before Siegel’s 1961paper, psychologists had been doing these experiments, always instructing the subject to ‘‘do you best” to predict correctly on each trial. The standard observedoutcome was probability matching in which the proportion of times the more frequent event is chosen, is p = p. Of course if your objective is to maximize thenumber of correct predictions, then as soon as you have determined which is the more frequent event, you should predict it 100% of the time. Consequently, thetypical conclusions from the data was that the model of rational maximizing behavior had to be rejected. Syd Siegel thought this was a curious conclusionbecause subjects had no incentive to maximize the number of correct predictions other than perhaps the homegrown satisfaction of being correct.Siegel arguedthat, from the subject’s point of view, the task was excruciatingly boring, and hypothesized that the subject’s diversification of his prediction strategy was ameans of reducing this boredom. Specifically, he hypothesized that the subject’s choice criterion was composed of two additive subjective utilities: the utility ofa correct prediction and the utility of variability, where the latter was proportional to p(1 � p), which is the variance of outcomes in a Bernoulli process with [p.808] parameter p. This assumption has the property that the utility of variability was greatest when p = 1/2; i.e., when diversification, or boredom relief, wasgreatest” (Smith, 1998, p. 113).

3 I am indebted to one of the referees for pointing out that older behavioural economics aimed at developing an alternative to neoclassical economics, whilenewer behavioural economics limits itself to studying deviations from the neoclassical paradigm yardstick. Herbert Simon can be considered the key interpreterof the former and Daniel Kahneman and Amos Tversky of the latter.

4 See Lewin (1996) and Sent (2004) for a detailed account of the history of behavioural economics.5 See Glimcher (2009) for a brief history of the recent growth of neuroeconomics.

A. Innocenti / Journal of Economic Psychology 31 (2010) 421–434 423

out rejecting its inductive orientation and, at the same time, economics can benefit from psychological insights withoutrenouncing the rationality principle.

The problem with this view is that a long-standing debate has compellingly shown that economics and psychology havefollowed different paths to scientific knowledge (Hertwig & Ortmann, 2001). The list of arguments supporting this judg-ment is long and varied. A first aspect is related to the very object of analysis. While economics mostly adheres to thehypothesis of a representative agent to build its abstract models, psychology and, more generally, cognitive sciences ana-lyze individual behavior in all its wide and disparate variety. By assuming heterogeneity, psychologists are prone to think ofpeople as being motivated by different and often conflicting driving forces (Lewin, 1996; Lopes, 1994; Zwick, Erev, & Bude-scu, 1999). This perspective contrasts with economists’ hypothesis that monetary incentives are key determinants of deci-sion making (Camerer & Hogarth, 1999; Smith, 1976). This assumption of selfishness is taken by economists at face value.Not only are people considered to be interested only in utility to themselves, but they are also indifferent to the differencesamong individuals (Rabin, 1998, 2002). On the other hand, psychologists rarely deal with strategic situations without con-sidering the effect of interpersonal comparisons (Croson, 2005; Lewin, 1996). This observation seems to contradict the stan-dard view that economists are interested in social institutions and psychologists in individuals. On the contrary, thepsychological approach to social interaction assumes that behavior is, to a great extent, environment-dependent, while,although economists acknowledge that the decision frame matters, they practically ignore the influence of guessing others’decision frames and status on individual choices (Loewenstein, 1999). At the linguistic level, the gap between psychologyand economics is even more pronounced. Psychologists fail to assimilate words like equilibrium or externality, but theirvocabulary, expanded by more frequent interdisciplinary connections, is richer and more varied than economists’, whichhas been freed from ambiguities by mathematical language (Murningham & Roth, 2006; Zwick et al., 1999). Finally, labo-ratory methods in psychology and economics are markedly different. Economists, unlike psychologists, provide subjectswith written instructions, repeat experimental trials, pay subjects on the basis of their performance and ban the deceptionof subjects (Bonetti, 1998; Hey, 1991; Ortmann & Hertwig, 1997; Ortmann & Hertwig, 2002). These arguments providegrounds enough to explain why psychology and economics have remained so long apart in their efforts to understandand predict human behavior. If a methodological turn could have contributed to bridge the gap between the two disci-plines, it would have been the introduction of laboratory methods to economics. But a closer examination of the historyof experimental economics and, particularly, of the contribution of one of its key actors, Vernon Smith,6 shows that thiswas not exactly the case.

It is well known that Smith’s (1962) first experimental paper refuted Chamberlin’s (1948) imperfect market experimentalfindings. By introducing the double oral-auction procedure and a sequence of trading periods in place of Chamberlin’s unin-terrupted series of exchanges, Smith obtained the convergence of trading price and quantities to equilibrium values and theconfirmation of the so-called Hayek Hypothesis: ‘‘Strict privacy together with the trading rules of a market institution are suf-ficient to produce competitive market outcomes at or near 100% efficiency” (Smith, 1982, p. 167). Some authors have arguedthat this paper marked the beginning of the early history of experimental economics,7 but at the same time it had an ambig-uous effect on the following developments. On the one hand, a new tool confirming what had already been proved theoreticallyshould have been more readily accepted, but on the other hand Smith’s findings weakened the critical impact of experimenta-tion on the validity of the rational approach to economic behavior.8

In effect, Smith’s leading role made his view on the relations between psychology and economics greatly influential. Headdresses it in Smith (1991b), the concluding section of which contains the following passage: ‘‘Why is it that human sub-jects in the laboratory frequently violate the canons of rational choice when tested as isolated individuals, but in the socialcontext of exchange institutions serve up decisions that are consistent (as though by magic) with predictive models based onindividual rationality? Experimental economists have no good answers to this question, although adaptive learning modelssuch as those of Lucas (1987) are suggestive. We need the help of psychologists, undeflected by battles with straw men”(Smith, 1991b, p. 894).

To illustrate what this help should be, Smith concludes the paper with a revealing metaphor: ‘‘Language learning in chil-dren occurs in a social context. Without contact with people, children do not learn to speak. If they have such contact, theylearn to speak in the total absence of formal instruction. But the same can be said of decision making: I could substitute‘‘make market decisions” for ‘‘speak” in the last two sentences and they would apply to what we have learned in the labo-ratory about adults. On the basis of cognition alone, without the language of the market and ongoing social interaction withother agents, rational decision is frustratingly illusive” (Smith, 1991b, p. 894).

Put differently, if laboratory findings contradict the principle of rationality, psychology may help to reconcile the latterwith the former by explaining what tacit and unconscious learning processes enhance optimality when moving from theindividual to the collective level. Psychologists are asked not to elucidate the reasons for contradictions, but rather to provideempirical support to the reference model. The corollary is that the empirical evidence collected by psychologists is placed inan ancillary position with respect to the axioms at the basis of rational decision theory.

6 Charles Plott was the other key actor, who worked with Vernon Smith in the 1970s. See Ortmann (2003) for a thorough account of Plott’s contribution.7 See Lee (2007) and Lee and Mirowski (2008) for a detailed account of the historical importance of Smith’s work.8 See Moscati (2007) for a discussion of this point.

424 A. Innocenti / Journal of Economic Psychology 31 (2010) 421–434

Smith’s view is grounded in his criticism of the psychological approach to experimentation, made explicit not only in the1991 paper, but also in the 2002 Nobel lecture. The argument runs as follows. When conducting experiments, psychologistsusually focus on the individual isolated from the context. In so doing, they ignore the role of institutions in reinforcing, oreven inducing, individual rationality. Thus, it is not surprising that in their experiments subjects systematically deviate fromrational behavior.

Smith’s (1991b) critical target is Tversky and Kahneman’s empirical work. Their statement that ‘‘deviations of actualbehavior from the normative model are too widespread to be ignored, too systematic to be dismissed as random error,and too fundamental to be accommodated by relaxing the normative system” (Tversky and Kahneman, 1987, quoted inSmith, 1991b, p. 890) is critically affected by the fact that they ask individuals isolated from social interaction to chooseamong hypothetical alternatives. It is not surprising that in this framework subjects are not capable of achieving the efficientoutcomes that in real markets and in well-designed experiments they attain unconsciously.

Smith returns to the same issue in the 2002 Nobel lecture. His view is unchanged: ‘‘Thus, if people in certain contextsmake choices that contradict our formal theory of rationality, rather than conclude that they are irrational, some askwhy, re-examine maintained hypotheses including all aspects of the experiments – procedures, payoffs, context, instruc-tions, etc. – and inquire as to what new concepts and experimental designs can help us to better understand the behavior.What is the subjects’ perception of the problem that they are trying to solve?” (Smith, 2002b, p. 510). But now Smith has aclearer view on the rationale justifying this interpretation. After being directly involved in neuroeconomics (McCabe, Houser,Ryan, Smith, & Trouard, 2001), he refers to arguments similar to (Laibson and Zeckhauser’s (1998)) quoted above and toGigerenzer’s (1996) criticism of Kahneman and Tversky’s approach.9 Human activity being ‘‘dominated by unconscious, auto-nomic, neuropsychological systems that enable people to function effectively without always calling upon the brain’s scarcestresource – attentional and reasoning circuitry” (Smith, 2002b, p. 507), learning processes are neither conscious nor recalled.Constructive rationality has to be supplemented by ecological rationality, which is the outcome of cultural and biological evo-lutionary processes. People’s decision processes cannot be thus detected by asking them but they can only be discovered in thelaboratory, where the effect of the market can be artificially created. The expected finding is that people will turn out to be morerational than they are aware of being because the social mind ‘‘solves complex organization problems without conscious cog-nition” (Smith 2002b, p. 553).

Smith’s faith in the Hayek Hypothesis has been attributed by Lee and Mirowski (2008) to his training: ‘‘The present paperargues that the old metaphor dies hard, especially when those trained in physical sciences keep entering the economics pro-fession, and also maintains that the very metaphor that played an important role in the early development of neoclassicaleconomics helps in detecting a unifying intellectual thread running through Smith’s economic orientation and practice, bothof which have been enormously influential in the development of experimental economics.”

This attitude also permeated the reaction that the young Vernon Smith had on his first meeting with Sidney Siegel. In aletter to the editor of the Journal of Political Economy on October 26, 1961, Smith writes, ‘‘Incidentally, I have just today metSidney Siegel for the first time and I have his to-be-published material on experiments in oligopoly. The results are terriblyinteresting. Duopolists, who only know their own profit outcomes (incomplete information), go to the Bertrand competitiveprice solution. Triopolists do the same but faster. This suggests that my competitive price results might be achieved in stillthinner markets. But the real shocker is the effect of complete information in which duopolists know each other’s profit out-comes. As the amount of information increases, duopolies decrease their tendency to the Bertrand competitive price. Theinvisible hand only works when it is invisible?!” (Smith, 1961).

Siegel’s findings seemed to be fully supportive of Smith’s beliefs. But when they met again six weeks later, their agree-ment turned out to be less than expected: ‘‘A really important event at Stanford was my meeting Sidney Siegel, who wasa fellow that year at the Center for Advanced Study in the Behavioral Sciences. I knew Syd only six weeks before, very incon-veniently, he died. (I have never forgiven him. What a great experimental scientist!) I showed him my work. He was skep-tical, too, but it was different; his was the skepticism of a scientist, not a wise guy. He had ideas, suggestions, and challengesfor me that emanated from a deep commitment to the science of behavior. Through his cutting criticism came excitementand implicit encouragement” (Smith, 1991a, p. 156). What Smith and Siegel disagreed on was exactly the way to integratepsychology and economics, as I intend to show in the remainder of the paper.

3. The case of Sidney Siegel

Siegel’s life led him to acquire an openness of mind that helped him to become a pioneer of experimental economics.10

Siegel’s parents emigrated to the United States from Romania at the turn of the century and settled in New York to run first a

9 Smith makes reference to Gigerenzer in the opening footnote of his Nobel Lecture: ‘‘The title [Constructivist and Ecological Rationality in Economics] wassuggested to me in the paper by Joel Norman, ‘Two Visual Systems and Two Theories of Perception: An Attempt to Reconcile the Constructivist and EcologicalApproaches,’ Behavioral and Brain Sciences, 2002. After finishing this paper I found that my use of the term below had been used by Gigerenzer, Tood, and ABCResearch Group (1999), for application to ‘fast and frugal decision making,’ by individuals: ‘A heuristic is ecologically rational to the degree that it is adapted tothe structure of an environment’(p. 13)” (Smith, 2002b, p. 503). On the debate between Gigerenzer and Kahneman–Tversky, see Gigerenzer (1996) andKahneman and Tversky (1996).

10 These biographical notes rely on materials and articles provided by Sidney’s son, Jay Siegel, and on the memoir written by Sidney’s second wife, AlbertaEngvall Siegel (1964), who was professor of psychology at Stanford University and played a very active role in Siegel’s professional life.

A. Innocenti / Journal of Economic Psychology 31 (2010) 421–434 425

bakery and then a restaurant. Sidney was born on 1916 as the last of five children. His childhood was spent helping his parentsin their businesses and attending primary schools. With the Great Depression, his family’s financial situation deteriorated and,in the early thirties, Sidney lived on the road, roaming back and forth across the United States, finding occasional jobs in thesummer and coming back to New York during the winter. In those years he became a sort of juvenile delinquent, who playedpool for cash, stole objects and extorted money. He failed to finish high school and served some time in jail. In his second wife’swords, this behavior was tempered by his innate aversion to any kind of violence: ‘‘Throughout his life, Sid was strongly sym-pathetic with ‘underprivileged’ and ‘delinquent’ kids, but he felt no kinship with the hoodlums. His own delinquencies had beenstrictly a way of getting by in the life of a city which offered no jobs and few channels to success to its slum children” (EngvallSiegel, 1964, p. 4).

This period came to an end in 1939 after his first marriage and his father’s death, when he moved to Los Angeles.There he attended a school for radio repairers and was then hired as a technician in a radio shop. Two years laterhis first son Jay was born. To earn a living for the family, Siegel applied to be enrolled in the Army, but his attemptwas frustrated by his record of having been affected by tuberculosis some years before. He therefore started workingas a civilian employee in the Army Signal Corps, where he learned the principles of electronic engineering. In 1942he first entered an academic environment by attending courses on engineering and mathematics at Stanford Univer-sity. Then he began teaching science and mathematics in a secondary school in San Jose, although he was not a col-lege graduate and therefore not eligible to teach in a high school. To solve this problem, he enrolled at San Jose StateCollege in 1945, while continuing school teaching. He obtained a bachelor’s degree in 1951, at the age of thirty-five.Then he enrolled in graduate school at Stanford University, where his research career began, to last less than adecade.

He soon decided to focus his attention on social psychology. His first project was entitled ‘‘Cognitive Ambiguityand Ethnocentrism,” the results of which led to his Ph.D. thesis on ‘‘Certain Determinants and Correlates ofAuthoritarianism.” The thesis included Siegel’s first experiment, in which, to detect a measure of authoritarianism,students were asked to match some photographs of faces with an equal number of generic sentences randomly cho-sen from different and unrelated sources. The number of associations arbitrarily made by the students was consid-ered as an indicator of overconfidence and, more specifically, of ‘‘intolerance for cognitive ambiguity” (Siegel, 1954,p. 251).

Siegel submitted his Ph.D. thesis in the fall of 1953. In the same year, he came into contact with two young philosophers,Donald Davidson and Patrick Suppes, at Stanford. Davidson had received his education at Harvard, where he studied math-ematical logic with Willard Quine, and had moved to Stanford in 1951, where he was to become one of the most distin-guished analytical philosophers of the past century, whose discussions of the concepts of action, truth andcommunicative interaction generated considerable debate in philosophical circles around the world. Patrick Suppes had ob-tained his Ph.D. from Columbia University in 1950 and soon became acquainted with the logician McKinsey, who was writ-ing an introductory book on game theory (McKinsey, 1952). His influence led Suppes to study utility theory and to write apaper on the foundations of probability (Suppes, 1951), the first of a series which would make him a key contributor to thefield. Davidson also became involved in this topic when Mc Kinsey moved from RAND Corporation in Santa Monica to Stan-ford in 1952.11 The result of their common interest was the project of experimentally measuring utility. To perform this task, itwas necessary to adapt the laboratory procedures used in psychology. The promising, albeit not so young, student Sidney Siegelwas invited to join the project.

The experiments were carried out in the spring of 1954 and the results reported in a Stanford technical paper in August1955 (Davidson, Siegel, & Suppes, 1955). The preliminary version was revised and enlarged into the book ‘‘Decision Making:An Experimental Approach” (Davidson, Suppes, & Siegel, 1957).

Suppes’ memory of the period deserves to be quoted: ‘‘This was my first experimental work and consequently in a gen-uine sense my first real introduction to psychology. The earlier papers on the foundations of decision theory concerned withformal problems of measurement were a natural and simple extension of my work in the axiomatic foundations of physics.Undertaking experimental work was quite another matter. I can still remember our many quandaries in deciding how tobegin, and seeking the advice of several people, especially our colleagues in the Department of Psychology at Stanford” (Sup-pes, 1979, p. 8). Siegel played a key role in the project, as proven by the experimental design that anticipated most of his laterwork. The main motivation of the study was to replicate Mosteller and Nogee’s (1951) test of utility maximization, which,according to the authors, was affected by a number of methodological limitations.12 In the experiment, a series of choicesbetween playing and not playing were submitted to fourteen students. When they exhibited indifference between two options,a number was associated to the bets until the underlying utility function was traced. Davidson, Suppes and Siegel claimed thatMosteller and Nogee’s method did not allow the inference of any consequence of the interval scales separating the points of theutility function, unless it was proved that the numbers assigned were unique up to a linear transformation. Moreover, their

11 In an interview in which Davidson acknowledges this debt, he also remembers the dramatic circumstances of McKinsey’s moving to Stanford: ‘‘But in factMcKinsey was the guy who was teaching both of us [Davidson and Suppes]. He was one of the inventors of quantified modal logic, though he didn’t publishmuch of his stuff. We hired him because he was with the RAND Corporation in Santa Monica, and there was all this stuff about his being a bad security riskbecause he was a homosexual. So they took away his security clearance and Stanford hired him. Then McKinsey committed suicide” (Lepore, 2004, p. 236).

12 The drawbacks of Mosteller and Nogee’s design are discussed in Camerer (1995, p. 620).

426 A. Innocenti / Journal of Economic Psychology 31 (2010) 421–434

experiment was criticized for two other methodological features. First, every choice was expressed as the acceptance or therejection of a gamble, and therefore unbalanced in favour of the second option for risk-prone subjects. Secondly, objective prob-abilities were implicitly assumed to be equal to subjective probabilities.13 Davidson, Suppes and Siegel amended these flawsand found that 15 subjects out of 19 chose as if they were maximizing expected utility. Thus, their preferences could be rep-resented by a utility curve unique up to a linear transformation. This evidence led the authors to claim that their experimentwas the first ‘‘to measure subjective probability behavioristically on the basis of empirically determined utilities” (Davidson,Suppes, & Siegel, 1957, p. 25).

In summer 1954, Siegel moved to Pennsylvania State University to join the faculty in the Department of Psychology. Thenext step in his career was twofold. On the one hand, he consolidated his abilities as a statistician by teaching a graduatecourse in statistics and, on the other hand, he further improved experimental techniques for measuring the utility function.The first activity led him to write Nonparametric Statistics for the Behavioral Sciences (Siegel, 1956), which rapidly became thestandard non-technical handbook for researchers in all the social sciences. The second effort aimed to provide an estimationof the intervals of the utility function. In two papers published in psychological journals (Hurst & Siegel, 1956; Siegel, 1956),Siegel devised an ingenious procedure for determining an ordered metric scale of preferences in the laboratory, which im-proved the method proposed by Coombs (1950).

These results strengthened Siegel’s confidence in experiments and led him to deal with more challenging issues, such aslearning processes. There were two different inspirations for this project. The first was Estes’ model of learning theory; thesecond was Herbert Simon’s 1955 paper on bounded rationality.

Estes’ model (1950), which triggered the debate on the informative and computational capacities of economic agents,aimed to show that learning could be represented as a converging stochastic process. To support his theory, Estes (1954)surveyed some experimental evidence, which attracted Siegel’s interest. The experimental task consisted of betting ontwo different events, to which two quite different probability values were associated. The finding supporting Estes’ modelwas that choice repetition induced subjects to match their predictions to the actual proportions of occurrence of the twoevents, rather than to bet rationally only on the most probable one. Siegel replicated the same experiment and showed thatthe convergence to the more probable outcome could be improved by rewarding subjects with real money and by makingthe experimental task less repetitive (Siegel, 1959; Siegel & Goldstein, 1959; Siegel, Siegel, & Andrews, 1964). Siegel’s designincluded three different treatments (no payoff, reward, and risk), the comparison of which provided unambiguous evidenceof the importance of monetary incentives, and a set of ingenious techniques to relieve the tediousness of the long sequencesof choices.14

The next step in Siegel’s study of learning theory was triggered by Simon’s celebrated paper on behavioral choice, the aimof which was ‘‘to construct definitions of ‘rational choice’ that are modelled more closely upon the actual decision processesin the behavior of organisms than definitions heretofore proposed” (Simon, 1955, p. 114). Simon’s definition of satisficingbehavior relied upon the psychological concept of aspiration level, which defined an alternative as satisfactory. Siegel wasimmediately supportive of this proposal and took it into account in devising an experiment to prove that a behavioral modelof decision making should include an explicit assumption on how individuals define their level of aspiration (Becker & Siegel,1958; Siegel, 1957).

In 1959 Siegel was named Professor of Psychology at Penn State University, where he planned what was to be the lastresearch project of his career. In his continual effort to integrate psychology with economics, he started to collaborate withLawrence E. Fouraker, who taught economics at Penn State.15 Their joint work resulted in two books, ‘‘Bargaining and GroupDecision Making: Experiments in Bilateral Monopoly” (Siegel & Fouraker, 1960) and the posthumous ‘‘Bargaining Behavior”(Fouraker & Siegel, 1963).

The main finding of the 1960 book was that in bilateral monopoly bargainers were inclined to maximize payoffs byselecting the Pareto optimal solution and by dividing the surplus equally. The convergence became more likely whengreater amounts of relevant information were available to the bargainers. Siegel and Fouraker stressed how this out-come contrasted with Schelling’s (1957) point that in a bargaining situation less informed bargainers are in a moreadvantageous position with respect to the more informed ones. The 1963 book extended experimental analysis to oli-gopoly by providing further confirmation of the role of complete information in implementing a Pareto optimal marketsolution.

In the subsequent history of experimental economics, Siegel did not play an active role. He died suddenly of a heart attackon November 29, 1961, while he was working at the Center for Advanced Study in the Behavioral Sciences of Stanford Uni-

13 In support of this point, Davidson, Suppes, and Siegel (1957) quoted Edwards (1954), who showed that, independently of utility considerations, peoplecould prefer some probabilities to others.

14 See Section 4.15 Lawrence Edward Fouraker was born in 1923 and obtained a Ph.D. in economics from the University of Colorado in 1951. In the same year he accepted a

teaching position at Penn State, where he was nominated Assistant Dean for research at the College of Business Administration. After Siegel’s death he joinedthe Harvard Business School (HBS) in 1961, where he was promoted full professor in 1963. Before his death in 1998 he served in many executives roles,including over 10 years as Dean of the HBS.

A. Innocenti / Journal of Economic Psychology 31 (2010) 421–434 427

versity. The legacy he left behind is impressive for a man who had graduated only ten years before at the age of 35. But whatstruck Vernon Smith most is what he could have done to change that history: ‘‘I am persuaded that if Sid had lived he wouldnot only have been the deserving Nobel Laureate who was well out in front of the rest of us, but also the timetable for therecognition of experimental economics would have been moved up, perhaps several years. It is important to be long-lived ifyou are to obtain such recognition” (Smith, 2008, p. 198).

4. Why are experimental economists so indebted to Siegel?

Most historical literature identifies three phases in the evolution of experimental economics: the early years, datingfrom 1948 to the early 1960s, the middle years, almost the whole of the 1980s; and the subsequent maturity.16 Cham-berlin’s (1948) attempt to test an imperfect experimental market took the lead, while the next breakthrough was achievedby Siegel and Fouraker’s (1960) methodological contribution. In particular, their book was crucial in establishing, first, thatthe conversion of subjects’ payoffs into cash rewards is a necessary requirement for the validity of experimental findingsand, secondly, that subjects’ information conditions are key variables in the laboratory. If it is evident that these advancesare attributable to Siegel, it has to be pointed out that he also made important theoretical contributions to the economicissues he dealt with. This appraisal also includes bargaining theory, to which Fouraker’s role was to provide only a verygeneral introduction. Siegel’s continual effort to pursue interdisciplinary collaborations had not yet driven him to searchfor help in other directions.17 Whether these planned collaborations would have improved his theoretical reputation, canhardly be said. But the focus on Siegel’s methodological legacy was certainly due to Vernon Smith, who put in the fore-ground his attempt to meld economics and psychology. In order to fully assess Siegel’s legacy, it will be useful to discussfirst his methodological achievements, then the theoretical ones, and finally to contrast them with Smith’s view, as pre-sented in section 2.

What Siegel did in ten years of research was to propose a set of guidelines for conducting experiments in economics.Later, Smith systematized these procedures but, as he acknowledged himself, he relied heavily on Siegel’s contribution.That Siegel was aware of his pioneering role is revealed by the fact that he did not miss any opportunity to make hismethodological vision explicit. From the 1955 book on utility maximization co-authored with Davidson and Suppes on-wards, he drew up and constantly updated the list of requisites to be met in economic experiments (Davidson et al.,1957; Siegel, 1956; Siegel, 1961; Siegel & Andrews, 1962; Siegel & Fouraker, 1960; Siegel & Goldstein, 1959; Siegelet al., 1964). The current view, as found in many handbooks,18 identifies four principles by which experimental economicsshould abide: procedural regularity, motivation, unbiasedness and calibration. On all these requisites, Siegel’s mark wasindelible.

The principle of procedural regularity is met when experimental design permits replications that will be accepted asbeing valid by other researchers. It requires that instructions are fully detailed, the methods of recruiting subjects areunbiased and all the features of the laboratory environment are as much as possible under the control of the experi-menter. Siegel was perfectly aware that the laboratory is not a socially neutral context, but that it is rather itself an insti-tution with its own formal or informal, explicit or tacit, rules. For this reason he did not disdain to perform experiments inthe field. His first experiment on the measure of the intervals of the utility function was conducted in a penitentiary,where inmates served as subjects and cigarettes as rewards (Hurst & Siegel, 1956). But when conducting experimentsin the laboratory, he took care of every detail with exceptional rigour. Some examples can illustrate this point better thanan exhaustive review.

For his first experiments on utility theory, Siegel devised a special dice on which the numbers were substituted with non-sense syllable, such as ‘‘ZOJ”, ‘‘WUH” or ‘‘QUJ”, that subjects did not associate with any meaning. This expedient aimed toavoid subjects’ betting depending on ‘‘prejudices and superstitions which many people hold concerning familiar events,e.g. heads on coins, evens on dice, etc.” (Siegel, 1956, p. 210). Notwithstanding this, Siegel conducted preliminary trials totest whether subjects preferred any nonsense syllable to another.

In the 1959 test of Estes’ learning model (Siegel et al., 1964), to relieve the boredom of 240 trials of the two-choice uncer-tain-outcome decision, subjects were seated in a swivel chair facing a signal light with two arrows pointing right or left, asillustrated below (Siegel et al., 1964, p. 68).

16 Smith (1992), Roth (1993), Roth (1995), and Leonard (1994) are the first authors to give an outline of the history of experimental economics. More recently,Ortmann (2003), Lee (2007), Moscati (2007), Guala (2008), Lee and Mirowski (2008), Sadrieh and Weimann (2008), and the collection of essays edited byFontaine and Leonard (2005) offer a pluralistic account of it.

17 In an interview with the author, Martin Shubik, who met Siegel at Stanford in 1959, gives this version of the story: ‘‘Siegel needed an economist involved ingame theory to plan new experiments and he thought that Fouraker was not expert enough. He was really interested at collaborating with me. When Siegelcame back to Penn State University, I used to fly there once each two or three weeks to work with him, but we did not have time enough to obtain concreteresults” (Shubik, 2008). Their collaboration is also documented in a series of letters, dating from July 5, 1960 to December 14, 1960 (Shubik, 1960).

18 Davis and Holt (1992), Friedman and Sunder (1994), and Friedman and Cassar (2004).

428 A. Innocenti / Journal of Economic Psychology 31 (2010) 421–434

On the right of the chair, there was a table with a platform with two push buttons and a vertical panel with two electriclamps; on the left-hand side, there was another table on which a platform with two buttons and a large mirror reflectingthe wooden board were mounted. The procedure started when one of the arrows on the signal light illuminated. Depend-ing on the randomly pointed direction, subjects had to rotate the swivel chair and face either the panel or the mirror topredict which lamp would illuminate by pressing the appropriate button. This arrangement obliged subjects willing topredict the same light in all the trials to change push button. In this way, ‘‘The boredom of repeatedly making the samecognitive response (e.g. right, right, right, right, right, right,. . ., right) is relieved, as is the kinesthetic boredom of remain-ing seated in the same position for an extended period and pressing the same button over and over” (Siegel et al., 1964,p. 68).

Siegel’s meticulousness could sometimes appear pedantic, if not fussy. Repeating the same task with four and five yearold boys (Siegel et al., 1964, p. 55), they were asked to choose one of two opaque bottles, each containing an object, 200times, of which 100 times were without any reward and 100 times with rewards, represented by a variety of dolls, toy cars,sweets, and trinkets, as shown below.19

19 Neither did Siegel neglect the chance of the boys being satiated: ‘‘It is to be expected that a variety of such prizes is more appropriate for youngsters than acollection of a single kind of item. That a 4-year-old might be quickly satiated if more and more of the same kind of reward began to pile up before him issuggested by the usual notions about diminishing marginal utility and by the fact that young children – and perhaps some other types of subjects as well – arenot likely to make discriminations implied by counting; they may observe that they acquire, say, one, two, three, and then ‘‘many” toy cars, but once they havemany, the acquisition of more may become meaningless. The use of diversity of rewards circumvents this difficulty” (Siegel et al., 1964, p. 148).

A. Innocenti / Journal of Economic Psychology 31 (2010) 421–434 429

In order to avoid order effects, for half of the boys the choice sequence was reversed. Unsurprisingly, this treatment wasbound to fail: ‘‘Among the children who were observed initially in 100 trials of the payoff condition, however, it wouldnot have been possible to maintain cooperation through subsequent trials under the no-payoff condition. Not unexpectedly,these children, who had been playing the game for a prize, were reluctant to continue with the game once the possibility ofreceiving a prize was eliminated” (Siegel et al., 1964, p. 57).

Despite some redundancies, Siegel’s careful attention to detail led him to focus on the motivation issue. Induced valuetheory, as codified later by Smith (1976), imposes the use of a reward medium to induce pre-specified features in experimen-tal subjects and makes their innate characteristics largely irrelevant. The results of the reinforcement treatments (Siegel &Andrews, 1962; Siegel & Goldstein, 1959) convinced Siegel that hypothetical choices were unreliable and that experimentalsubjects had to be rewarded to motivate them adequately. On this issue, he departed from the dominant approach amongpsychologists: ‘‘Because of our belief in the central importance of employing payoffs which are meaningful to subjects, re-wards which in fact they covet, we have little confidence in experiments in which the ‘payoffs’ are points, credits, or tokens.Or perhaps it would be more accurate to say that we have little confidence in the use of the term payoff to label such trivia.The relevance of such experiments to any theoretical notions about reward, payoff, or utility seems to be dubious” (Siegelet al., 1964, p. 148). For Siegel, real payoffs – be they cash, students’ grades, cigarettes or trinkets – made subjects’ choicesresponsive to the variables under investigation, which were generally underrated in the ‘‘no payoff” condition.

What Vernon Smith did some years later was just to develop this insight: ‘‘In thinking through the implications of ‘otherthings in the utility function’, I found Sid Siegel’s paper on two-choice uncertain outcome situation particularly helpful (Sie-gel, 1961). In this binary choice situation, the interpretation of over 20 years of psychology literature had been that peoplewere not rational; specifically, they failed to maximize. Since monetary payoffs had not been used, Siegel hypothesized thatsubjects did not maximize because there was nothing of value worth maximizing, and that the observed matching behaviorof subjects was due to ‘monotony, both kinesthetic and cognitive’ (Siegel, 1961, p. 768). Accordingly, he developed an addi-

430 A. Innocenti / Journal of Economic Psychology 31 (2010) 421–434

tive model of utility with two terms: the first was the utility of reward, the second the utility of variability, diversification, ormonotony relief. The model predicted that subjects would be drawn away from matching toward maximizing by introducingmonetary payoffs, and the greater the payoff levels the nearer would be the response to the maximizing response. The dataconfirmed the prediction. Then Siegel’s ingenuity was turned to a procedure for raising the utility of variability as a treat-ment” (Smith, 1991a, p. 5).

Siegel’s emphasis on the motivation issue did not impair his sensitivity to the experimental requisite of unbiasedness.According to this, experiments should be conducted in a way that does not lead participants to perceive any behavior asbeing expected or correct. It is well known that this argument is so compelling as to require the systematic use of placeboand treatment groups in medical experiments. In some cases, this procedure involves deceiving subjects about what theexperimenter is investigating. On the contrary, in experimental economics it is currently assumed that cheating shouldbe banned. Siegel endorsed this principle because he believed that deceptive trials would created an atmosphere of suspicionand skepticism towards the discipline: ‘‘Anyone who has worked with the repetitive-choice situation under study hereknows that at least some subjects formulate the suspicion that the experimenter is altering the sequence of events as theexperiment proceeds, and some think that he is altering it in response to the specific patterning of their choices. Moreover,many subjects doubt that the sequence of events is random, or they do not understand randomness; they watch for a pat-terning and sometimes make their choices under the assumption that a patterning exists. Our procedures were generallydesigned to demonstrate to the subject that the experimenter does not alter the event sequence as the experiment goesalong” (Siegel et al., 1964, pp. 149–150).

The last principle embodied by Siegel in the toolbox of experimental economists was calibration. To obtain a full andunambiguous correspondence between laboratory findings and theoretical predictions, Siegel’s experiments were as simpleas possible and did not ever involve more than one treatment variable, in order to discern accurately the implications for thetested model.20 In recalling Siegel’s statistical expertise, his wife Alberta Engvall summed up his view on calibration as follows:‘‘He averred that the best-designed experiment is one requiring no statistical analysis at all. Where statistics are needed, thesimpler the better. A major argument for nonparametric tests is their simplicity: their basis is easily grasped, the computationsare straightforward, and no distorting transformations are imposed on the raw scores. Preferring clean and simple designs, Sidhad little use for the analysis of variance and typically voiced his suspicion by proclaiming his inability to understand the mean-ing of any interaction. A simple two-group experiment usually sufficed to test the hypothesis of interest” (Engvall Siegel, 1964,pp. 18–19).

This set of methodological guidelines were applied by Siegel to well-defined theoretical targets, but his view of economicswas far from the confidence in the validity of the Hayek Hypothesis driving Smith’s efforts. It was perhaps the belief that themathematical language of economics made it a perfect field for laboratory testing which initially motivated Siegel in hischoice to enter the field. But this orientation was joined with a critical view of utility theory. Siegel was strongly committedto the idea that the maximization principle had to be further specified in order to be empirically relevant. As a matter of fact,individuals could be considered maximizing agents only as a very rough approximation. In order to explain real behavior, itwas necessary to take account of congeries of other factors, which Siegel gradually brought into focus. In Davidson, Suppes,and Siegel, 1957, the list of factors influencing choice included risk attitude and subjective probabilities; in the experimentson utility intervals (Guthrie, Becker, & Siegel, 1961; Hurst & Siegel, 1956; Siegel, 1956; Siegel & Shepherd, 1959) the socio-economic features of the individuals, being prison inmates, men or women, socially close or distant people; in two-choiceuncertain-outcome decisions (Siegel, 1959; Siegel, 1961; Siegel & Andrews, 1962; Siegel & Goldstein, 1959; Siegel et al.,1964) the saliency of rewards and the removal of boredom; in tests on learning (Becker & Siegel, 1958; Becker & Siegel,1962; Siegel, 1957) the level of aspiration. In this way, the utility function became a heuristic device for performing exper-iments: ‘‘The notion of utility is useful in providing a basis for experimental operations. By conceptualizing the experimentalsituation already described in terms of utility and by specifying relevant components of a utility function, we may identifythose aspects of the experimental situation which are related to behavior in that setting. From this, we may prescribechanges in the situation which will lead to predictable changes in choice behavior. The construct of utility is useful here,then, in the degree to which it provides a rationale for determining the systematic environmental changes which lead to sys-tematic and thus predictable, changes in behavior. The overall utility of any possible outcome may depend on the subjectivevalue of each of several conceptually distinct aspects of that outcome. To predict choices, one must identify the various as-pects (or components, or factors) of the situation to which positive or negative utility is associated. With information aboutthe utility of each component of each outcome, it is possible to assess the utility associated with any particular strategy”(Siegel et al., 1964, pp. 9–10).

Siegel’s inclination to behaviorism is well exemplified by the way in which he moulded the psychological concept of aspi-ration level by pioneering the concept of reinforcement learning (Camerer & Teck-Hua, 1999; Erev & Roth, 1998). In Siegeland Fouraker (1960), he claimed first that psychological concepts like success and failure were, respectively, equivalent topositive and negative values of the utility function. Then, he assumed that the level of aspiration could be identified bythe utility point associated to the greatest interval between that point and the next below. In this way, decisions giving util-ities below this point caused a psychological feeling of dissatisfaction and repeated experiences of success (or failure) led to a

20 ‘‘In planning the experiments reported in this book, our effort was to employ the simplest possible experimental design. We have deliberately restricted ourstudies to the use of a single-variable design. In any given experiment, only a single independent variable was manipulated, and our interest centered on asingle dependent variable” (Siegel et al., 1964, p. 153).

A. Innocenti / Journal of Economic Psychology 31 (2010) 421–434 431

cumulative process of increasing (or decreasing) the aspiration level. Finally, Siegel claimed that this mechanism was in clearcontradiction with the maximization postulate of neoclassical economics. But this conclusion also sharply contrasted withVernon Smith’s (1962) approach. As shown by Lee’s (2007) historical reconstruction, Smith’s (1962) paper was indeed theobject of harsh criticism by the behavioral scientist William Starbuck, who was Smith’s colleague at Purdue University inthe early 1960s. Starbuck (1963) critically addressed the postulate of an ‘‘unrealistic fixity of mind” underlying Smith’s inter-pretation of laboratory results. In contrast, Simon’s concepts of level of aspiration (LOA) and bounded rationality would havepermitted the reconciliation of economics with psychology on the basis of its ‘‘introspective appeal”. By assessing the his-torical implications of that debate, Lee (2007) concludes that ‘‘cognitive models in behavioral sciences, such as the LOA-based models, had difficulty entering the American economics discipline not simply because they were foreign to econo-mists, but also because they could easily be mobilized to bring into question a certain, well-entrenched idea in economics”(Lee, 2007, p. 27). On the contrary, it was exactly Siegel’s unbiased attitude which drove him far from his earlier belief in theprinciple of utility maximization and to provide the first experimental proof of Simon’s concept of satisficing behavior. Sie-gel’s (1957) experiment showed that, given that decisions are taken by means of sequential procedures, preferences exhib-ited strictly depended on the presentation order. Individuals were inclined to choose the first available alternative abovetheir level of aspiration, but this decision was not necessarily confirmed if choices were repeated because the level of aspi-ration constantly shifted upward or downward. These findings had harmful implications for utility theory: ‘‘In conclusion, itwould seem that a useful behavioral model of decision making should include not only the concepts of utility and subjectiveprobability, as do the present models, but should also include a formulation of the effects of level of aspiration and reinforce-ment on utility. That is, the model should include recognition that utility has a model in its own right, in which the mainconcepts are level of aspiration (LA) and reinforcement effects (R)” (Siegel, 1957, p. 124).

It is revealing that Vernon Smith gave a different interpretation of Siegel’s criticism of the concept of rationality, whichdeserves to be quoted at length: ‘‘Siegel interpreted Simon’s argument as suggesting that the rational model is essentiallycorrect, but more or less incomplete. To make it complete, it was necessary to examine decision problems carefully fromthe utilitarian point of view of the decision maker (not just from the point of view of the experiment/theorist). Note that thisinterpretation differs from the ‘satisficing’ and ‘bounded rationality’ constructions that were later put on Simon’s originalidea, constructions that were critical of the very foundation of rational behavior as conventionally defined. In Siegel’s imple-mentation, actions differ from the predictions of the standard model because of the decision cost. Since the latter is neces-sarily part of the problem of realizing rational outcomes, the result is not just a better descriptive/predictive model. It is abetter normative model of action as experienced by the individual. Thus, the distinction between the descriptive and thenormative model of behavior becomes clouded; neither is cast in objective reality independent of experience. I believe thisis the right way, although certainly not the easiest way, to approach the problem of modelling rational behavior” (Smith,1991a, p. 807).

That Smith’s assessment did not do justice to Siegel’s view is made clear by the way Fouraker and Siegel (1963) ap-proached bargaining theory, itself a particularly suitable area for testing the effects of aspiration level. In a previous paper,Fouraker (1957) had shown himself to be in agreement with Siegel’s idea that bargaining modelling would benefit from mak-ing the psychological determinants of agreements explicit. In the laboratory, this view entailed determining what factorswere more significant for bargainers’ behavior. Siegel and Fouraker’s choice fell on information conditions. The debate raisedby Nash’s (1950, 1953) bargaining model, in which Harsanyi (1956) and Schelling (1960) were key players, led in this direc-tion (Innocenti, 2008). Nash and Harsanyi had proved axiomatically that in conditions of complete information it is rationalto agree on the outcome maximizing the product of bargainers’ utility functions and to distribute it equally in relation to thedisagreement point. Schelling had contended that the process of finding an agreement in bargaining has a component that isinherently empirical, because it concerns a process of intellectual coordination among heterogeneous agents in which thecontext is decisive. Thus, according to Schelling, Nash and Harsanyi’s equilibrium solution fails to solve the bargaining prob-lem because it neglects the learning process through which bargainers define their expectations of others’ behavior. By tak-ing it into account, less information may be better than more information in order to obtain better outcomes. The moreinformed party, knowing that the other party knows only his own payoff, is more forgiving when his opponent makes largedemands.

Siegel and Fouraker’s (1960) experimental findings were generally supportive of Nash–Harsanyi’s model. Bargainers wereinclined to maximize joint payoffs by negotiating the Pareto optimal agreement on the contract curve. But this tendency wasshown to be strictly related to the amount of information available to the bargainers. More information on the other’s payoffsreinforced the convergence to the efficient agreement. Still, prices exhibited the tendency to take the value of the 50-50 splitof the joint payoff, but this finding was dependent on subjects’ level of aspirations, as measured by the recorded successionsof bids and asks. In contrast, evidence for Schelling’s knowledge handicap was ‘‘equivocal”: half of the cases supported it,while the others opposed or failed to support it. The authors concluded that on Schelling’s theory, ‘‘further research, perhapswith larger sample sizes, will be necessary before any strong conclusions may be drawn.” (Siegel & Fouraker, 1960, p. 58).

The posthumous Fouraker and Siegel (1963) did not carry out this research program. The book contained a set of exper-iments which applied the 1960 design to the case of oligopoly. Convergence to Pareto optimal agreements was still con-firmed, but clear evidence was provided that an appropriate manipulation of the environmental conditions (not only thestate of information, but also the details of the bargaining protocols or of the contracts) could implement individual, ratherthan joint, profit maximization. Outcomes were so varied that they were classified according to a list of different subjecttypes, ranging from rivalistic to cooperative. For each group, a psychologically-based justification was also given. For exam-

432 A. Innocenti / Journal of Economic Psychology 31 (2010) 421–434

ple, ‘‘A rivalistic decision could reflect two distinct motivations: (1) that the player derives satisfaction from reducing theprofits of his opponent; typically, he wants to beat his rival, and (2) that the player wishes to send a punitive bid, as a signalthat he is dissatisfied with the past responses of his opponent and wants him to alter his decisions in a manner that benefitsthe signaler” (Fouraker & Siegel, 1963, pp. 204–205).

By focusing on the various psychological determinants of bargaining behavior rather than on the golden rule of utilitymaximization, Siegel demonstrated ‘‘that experimental operations based on a consideration of a psychological construct,utility, lead to predictable choice behavior. The reader is asked to note also the appearance of the word behaves in the corehypothesis. Analyses are directed to the ways in which people actually behave, not to how they say they behave or wouldbehave, nor to how they might expect others to behave. In our judgment, the hypothesis of maximization of expected utilitycan be given a fair test only by research in the behaviorist tradition. Choice behavior must be observed in realistic and sig-nificant choice situations” (Siegel et al., 1964, p. 19).

Unlike Smith, Siegel’s attitude led him to adapt theoretical models to experimental results and not the reverse. When lab-oratory findings convinced him that the validity of expected utility theory depended on psychological factors such as riskattitudes, motivation, boredom or level of aspiration, he openly integrated these elements in his models and took full accountof all the theoretical consequences. What moved Siegel on a different path with respect to Smith was just his psychologicallygrounded sensibility. He constantly kept in mind that subjects in the laboratory are different from each other and can bemotivated by a variety of reasons. This implies that economic choices are environment dependent and that experimentsdo not simply aim at building behavioral regularities but also to investigate how the manipulation of external conditionscould make maximization behavior more or less feasible. Only in this perspective can his immediate consonance with Si-mon’s behavioral approach be adequately appreciated. What Siegel claimed is that an alternative to rational choice ormechanical adaptation in learning theory existed. Around the same time, this argument motivated another future NobelPrize winner, Reinhard Selten, who independently conducted a path-breaking experiment on aspirations and adaptationin the theory of firms (Sauermann & Selten, 1962).21 It is thus not surprising that today behavioral economists assign to Seltenthe role of pioneer in the field (Camerer, 2003). On the same grounds, especially if one takes into account how little time Siegelhad to promote and develop his ideas, he surely deserves the credit of being the first to show that psychology could be not anancillary, but a constituent part of the introduction of experimental methods in economics.

5. Conclusions

This paper has tried to show that Sidney Siegel’s view of economic experiments was different from Vernon Smith’s. Siegelwas constantly in search of new empirical regularities and his approach to experimentation was deeply heuristic, in that heaimed at discovering new stylized facts to increase the empirical significance of economic models. Such an orientation didnot conflict with the mathematical nature of economics. As a matter of fact, Siegel considered the ambiguities of the psycho-logical language its main deficiency, the removal of which would improve its effectiveness. But this formal upgrading shouldnot necessarily make psychology an auxiliary of mainstream economics as proposed by Smith. For this reason, Siegel was notonly a pioneer of the methodology of experimental economics but also of behavioral economics, because he endorsed an ori-ginal approach to making interdisciplinary research effective. When Siegel died, his way of integrating psychology and eco-nomics lost its main advocate. There are many reasons to think that only another psychologist could have taken Siegel’sinsights forward and one might wonder why this did not happen. The answer is probably that what he did in nearly 10 yearsneeded much more time to be fully understood and developed: ‘‘We cannot know what Siegel might have done. But thisbook is a deeply impressive record of what he did do. Even with 20 more years than Siegel had, how many of us can aspireto do so much?” (Edwards, 1967, p. 293).

Acknowledgments

I am particularly grateful to the J. William Fulbright Commission in the United States and the Fulbright Commission inItaly for their financial support. I am also indebted to Ernie Lepore and to Rutgers Center for Cognitive Sciences (RuCCS),where I wrote this paper during a sabbatical year as Fulbright Research Scholar.

I am very thankful to Jay Siegel for his kind help and encouragement, to Martin Shubik for fruitful discussions on Siegel’swork, to Duke University Library for access to Martin Shubik and Vernon Smith Papers and to Vernon Smith for allowingaccess to his unpublished materials. Comments by Kyu Sang Lee, Ivan Moscati, Andreas Ortmann and referees are also grate-fully acknowledged.

References

Becker, S. W., & Siegel, S. (1958). Utility of grades: Level of aspiration in a decision theory-context. Journal of Experimental Psychology, 55, 81–85.Becker, S. W., & Siegel, S. (1962). Utility and level of aspiration. American Journal of Psychology, 75, 115–120.Bonetti, S. (1998). Experimental economics and deception. Journal of Economic Psychology, 19, 377–395.

21 Selten’s work was ‘‘unknown to both of us [Siegel and Smith] at the time” (Smith, 2008, p. 197).

A. Innocenti / Journal of Economic Psychology 31 (2010) 421–434 433

Camerer, C. (1995). Individual decision making. In J. H. Kagel & A. E. Roth (Eds.), The handbook of experimental economics (pp. 587–703). Princeton: PrincetonUniversity Press.

Camerer, C. (1999). Behavioral economics: Reunifying psychology and economics. Proceedings of the National Academy of Sciences of the United States ofAmerica, 96, 10575–15577.

Camerer, C. (2003). Behavioral game theory: Experiments in strategic interaction. Princeton: Princeton University Press.Camerer, C., & Hogarth, R. (1999). The effects of financial incentives in experiments: A review and capital-labor production theory. Journal of Risk and

Uncertainty, 19(1–3), 7–42.Camerer, C. F., & Teck-Hua, H. (1999). Experience weighted attraction learning in normal-form games. Econometrica, 67, 827–874.Chamberlin, E. H. (1948). An experimental imperfect market. Journal of Political Economy, 56, 95–108.Coombs, C. H. (1950). Psychological scaling without a unit of measurement. Psychological Review, 57, 145–158.Croson, R. (2005). The method of experimental economics. International Negotiation, 10, 131–148.Davidson, D., Siegel, S., & Suppes, P. (1955). Some experiments and related theory in the measurement of utility and subjective probability. Report No. 4. Stanford

Value Theory Project.Davidson, D., Suppes, P., & Siegel, S. (1957). Decision making: An experimental approach. Stanford: Stanford University Press.Davis, D. D., & Holt, C. A. (1992). Experimental economics. Princeton: Princeton University Press.Edwards, W. (1954). The theory of decision making. Psychological Bulletin, 51, 380–417.Edwards, W. (1967). Review. Decision and choice. Contributions of Sidney Siegel by S. Siegel, S. Messick, & A.H. Brayfield. Journal of the American Statistical

Association, 62, 291–293.Engvall Siegel, A. (1964). Sidney Siegel: A memoir. In S. Messick & A. H. Brayfield (Eds.), Decision and choice. Contribution of Sidney Siegel (pp. 1–24). New

York: McGraw-Hill.Erev, I., & Roth, A. E. (1998). Predicting how people play games: Reinforcement learning in experimental games with unique, mixed strategy equilibria.

American Economic Review, 88, 848–881.Estes, W. K. (1950). Toward a statistical theory of learning. Psychological Review, 57, 94–107.Estes, W. K. (1954). Individual behavior in uncertain situations: An interpretation in terms of statistical association theory. In R. M. Thrall, C. H. Coombs, & R.

L. Davis (Eds.), Decision processes (pp. 127–138). New York: John Wiley.Fontaine, P., & Leonard, R. (Eds.). (2005). The experiment in the history of economics. London: Routledge.Fouraker, L. E. (1957). Professor Fellner’s bilateral monopoly theory. Southern Economic Journal, 24, 182–189.Fouraker, L. E., & Siegel, S. (1963). Bargaining behavior. New York: McGraw-Hill.Friedman, D., & Cassar, A. (2004). Economics lab. An intensive course in experimental economics. London and New York: Routledge.Friedman, D., & Sunder, S. (1994). Experimental methods: A primer for economists. Cambridge: Cambridge University Press.Gigerenzer, G. (1996). On narrow norms and vague heuristics: A reply to Tversky and Kahneman. Psychological Review, 102, 592–596.Gigerenzer, G., Tood, P. M., & ABC Research Group. (1999). Simple Heuristics that make us smart. New York: Oxford University Press.Glimcher, P. W. (2009). Introduction: A brief history of neuroeconomics. In P. W. Glimcher, C. F. Camerer, E. Fehr, & R. A. Poldrack (Eds.), Neuroeconomics:

Decision making and the brain (pp. 1–12). New York: Academic Press.Guala, F. (2008). Experimental Economics, History of. In S. Durlauf & L. Blume (Eds.). The new Palgrave dictionary of economics (Vol. 3, pp. 152–156). London:

Palgrave-MacMillan.Guthrie, G. M., Becker, S. W., & Siegel, S. (1961). Preferences and differences in preference for political candidates. The Journal of Social Psychology, 53, 25–32.Harsanyi, J. (1956). Approaches to the bargaining problem before and after the theory of games: A critical discussion of Zeuthen’s, Hicks’, and Nash’s

theories. Econometrica, 24, 144–157.Hertwig, R., & Ortmann, A. (2001). Experimental practices in economics: A methodological challenge for psychologists? Behavioral and Brain Sciences, 24,

383–451.Hey, J. D. (1991). Experiments in economics. Oxford: Basil Blackwell.Hurst, P. M., & Siegel, S. (1956). Prediction of decisions from a higher ordered metric scale of utility. Journal of Experimental Psychology, 52, 138–144.Innocenti, A. (2008). Linking strategic interaction and bargaining theory. The Harsanyi–Schelling debate on the axiom of symmetry. History of Political

Economy, 40, 111–132.Kahneman, D., & Tversky, A. (1996). On the reality of cognitive illusions. Psychological Review, 103, 582–591.Laibson, D., & Zeckhauser, R. (1998). Amos Tversky and the ascent of behavioral economics. Journal of Risk and Uncertainty, 16, 7–47.Lee, K. S. (2007). Early development of experimental economics at the interdisciplinary crossroads. College of Social Sciences, Ajou University. Working paper.Lee, K. S., & Mirowski, P. (2008). The energy behind Vernon Smith’s experimental economics. Cambridge Journal of Economics, 32, 257–271.Leonard, R. (1994). Laboratory strife: Higgling as experimental science in economics and social psychology. In N. B. De Marchi & M. S. Morgan (Eds.),

Higgling. Transactors and their markets in the history of economics (pp. 343–369). Durham: Duke University Press (History of political economy annualsupplement).

Lepore E. (2004). Interview with Donald Davidson. In D. Davidson, Problems of Rationality (pp. 231–266). Oxford: Oxford University Press. <http://ruccs.rutgers.edu/faculty/lepore/Davidson_interview.pdf>.

Lewin, S. B. (1996). Economics and psychology: Lessons for our own day from the early twentieth century. Journal of Economic Literature, 34, 1293–1323.Lopes, L. L. (1994). Psychology and economics: Perspectives on risk, cooperation, and the marketplace. Annual Review of Psychology, 45, 197–227.Loewenstein, G. (1999). Experimental economics from the Vantage-Point of Behavioral Economics. Economic Journal, 109, F25–F34.Lucas, R. E. (1987). Models of business cycles. New York: Basil Blackwell.McCabe, K., Houser, D., Ryan, L., Smith, V., & Trouard, T. (2001). A functional imaging study of cooperation in two-person reciprocal exchange. Proceedings of

the National Academy of Sciences (USA), 98, 11832–11835.McKinsey, J. J. (1952). Introduction to the theory of games. New York: McGraw-Hill.Moscati, I. (2007). Early experiments in consumer demand theory: 1930–1970. History of Political Economy, 39, 359–401.Mosteller, F., & Nogee, P. (1951). An experimental measurement of utility. Journal of Political Economy, 59, 371–404.Murningham, J. K., & Roth, A. E. (2006). Some of the ancient history of experimental economics and social psychology: Reminiscences and analysis of a

fruitful collaboration. In D. De Cremer, M. Zeelenberg, & J. K. Murningham (Eds.), Social psychology and economics (pp. 206–234). Mahwah: LawrenceErlbaum.

Nash, J. F. Jr., (1950). The bargaining problem. Econometrica, 18, 155–162.Nash, J. (1953). Two-person cooperative games. Econometrica, 21, 128–140.Ortmann, A. (2003). Charles R. Plott’s collected papers on the experimental foundations of economic and political science. Journal of Economic Psychology, 24,

555–575.Ortmann, A., & Hertwig, R. (1997). Is deception acceptable. American Psychologist, 52, 746–747.Ortmann, A., & Hertwig, R. (2002). The costs of deception: Evidence from psychology. Experimental Economics, 5, 111–131.Rabin, M. (1998). Psychology and economics. Journal of Economic Literature, 36, 11–46.Rabin, M. (2002). A perspective on psychology and economics. European Economic Review, 46, 657–685.Roth, A. E. (1993). On the early history of experimental economics. Journal of the History of Economic Thought, 15, 184–209.Roth, A. E. (1995). Introduction to experimental economics. In J. Kagel & A. E. Roth (Eds.), Handbook of experimental economics (pp. 3–109). Princeton:

Princeton University Press.Sadrieh, A., & Weimann, J. (Eds.). (2008). Experimental economics in Germany, Austria, and Switzerland. A collection of papers in honor of Reinhard Tietz.

Marburg: Metropolis-Verlag.

434 A. Innocenti / Journal of Economic Psychology 31 (2010) 421–434

Sauermann, H., & Selten, R. (1962). Anspruchsanpassungstheorie der Unternehmung. Zeitschrift für die gesamte Staatswissenschaft, 118, 577–597.Schelling, T. C. (1957). Bargaining, communication, and limited war. Journal of Conflict Resolution, 1, 19–36.Schelling, T. C. (1960). The theory of conflict. London: Oxford University Press.Sent, E. M. (2004). Behavioral economics: How psychology made its (limited) way back into economics. History of Political Economy, 36, 735–760.Shubik, M. (1960). Letters to Sid Siegel, 5 July 1960, 14 December 1960, Martin Shubik papers. Accession 1999-0409, L. Perkins Library, Duke University,

Durham, Box 1.Shubik, M. (2008). Interview to A. Innocenti, April 3.Siegel, S. (1954). Certain determinants and correlates of authoritarianism. Genetic Psychology Monograph, 49, 187–254.Siegel, S. (1956). Nonparametric statistics for the behavioral sciences. New York: McGraw-Hill.Siegel, S. (1957). Level of aspiration and decision making. Psychological Review, 64, 253–262.Siegel, S. (1959). Theoretical models of choice and strategy behaviour: Stable-state behaviour in the two-choice uncertain outcome situation. Psychometrika,

24, 303–316.Siegel, S. (1961). Decision making and learning under varying conditions of reinforcement. Annals of the New York Academy of Sciences, 89, 766–783.Siegel, S., & Andrews, J. M. (1962). Magnitude of reinforcement and choice behaviour in children. Journal of Experimental Psychology, 63, 337–341.Siegel, S., & Fouraker, L. E. (1960). Bargaining and group decision making. New York: McGraw-Hill.Siegel, S., & Goldstein, D. A. (1959). Decision-making behaviour in a two-choice uncertain-outcome situation. Journal of Experimental Psychology, 57, 37–42.Siegel, S., & Shepherd, I. L. (1959). An ordered metric measure of social distance. Sociometry, 22, 336–342.Siegel, S., Siegel, A. E., & Andrews, J. M. (1964). Choice, strategy, and utility. New York: McGraw-Hill.Simon, H. A. (1955). A behavioral model of rational choice. Quarterly Journal of Economics, 69, 99–118.Smith, V. L. (1961), Letter to Professor Harry G. Johnson, Editor Journal of Political Economy. Vernon Smith papers. Accessions 1996-106, Box 14, L. Perkins

Library, Duke University, Durham.Smith, V. L. (1962). An experimental study of competitive market behavior. Journal of Political Economy, 70, 111–137.Smith, V. L. (1976). Experimental economics: Induced value theory. American Economic Review, 66, 274–279.Smith, V. L. (1982). Markets as economizers of information: Experimental examination of the ‘Hayek Hypothesis’”. Economic Inquiry, 20, 165–179.Smith, V. L. (1991a). Papers in experimental economics. Cambridge: Cambridge University Press.Smith, V. L. (1991b). Rational choice. The contrast between economics and psychology. The Journal of Political Economy, 99, 877–897.Smith, V. L. (1992). Game theory and experimental economics: Beginnings and early influences. In E. R. Weintraub (Ed.), Towards a history of game theory

(pp. 241–282). Durham: Duke University Press (History of political economy annual supplement).Smith, V. L. (1998). Experimental economics. Behavioral lessons for microeconomic theory and policy. In D. P. Jacobs, E. Kalai, & M. I. Kamien (Eds.), Frontiers

of Research in economic theory: the Nancy L. Schwartz Memorial Lectures, 1983–1997 (pp. 104–122). Cambridge: Cambridge University Press.Smith, V. L. (2002a). Vernon L. Smith’s speech at the Nobel Banquet. December 10, 2002. <http://nobelprize.org/nobel_prizes/economics/laureates/2002/

smith-speech.html>.Smith, V. L. (2002b). Constructivist and ecological rationality in economics. Prize Lecture, December 8, 2002. <http://nobelprize.org/nobel_prizes/economics/

laureates/2002/smith-lecture.html>.Smith, V. L. (2003). Les Prix Nobel. The Nobel Prizes 2002. Stockholm: Editor Tore Frängsmyr, Nobel Foundation. <http://nobelprize.org/nobel_prizes/

economics/laureates/2002/smith-autobio.html>.Smith, V. L. (2008). Discovery – A memoir. Bloomington: Author House.Starbuck, W. H. (1963). Level of aspiration theory and economic behavior. Behavioral Science, 8, 128–136.Suppes, P. (1951). A set of independent axioms for extensive quantities. Portugaliae Mathematica, 10, 163–172.Suppes, P. (1979). Intellectual Autobiography. In R. J. Bogdan (Ed.), Patrick Suppes. Dordrecht: D. Reidel Pub. <http://www.stanford.edu/~psuppes/

autobio1.html>.Tversky, A., & Kahneman, D. (1987). Rational Choice and the Framing of Decisions. In R. M. Hogarth & M. W. Reder (Eds.), Rational choice. The contrast

between economics and psychology (pp. 67–94). Chicago: The University of Chicago Press.Zwick, R., Erev, I., & Budescu, D. V. (1999). The psychological and economical perspectives on human decisions in social and interactive contexts. In D. V.

Budescu, I. Erev, & R. Zwick (Eds.), Games and human behaviour. Essays in honor of Amnon Rapoport (pp. 3–20). Mahwah, NJ, London: Erlbaum.