Known, Unknown, and Unknowable Uncertainties

In normative decision theory, the weight of an uncertain event in a decision is governed solely by the probability of the event. A large body of empirical research suggests that a single notion of probability does not accurately capture peoples' reactions to uncertainty. As early as the 1920s, Knight made the distinction between cases where probabilities are known and where probabilities are unknown. We distinguish another case –- the unknowable uncertainty –- where the missing information is unavailable to all. We propose that missing information influences the attractiveness of a bet contingent upon an uncertain event, especially when the information is available to someone else. We demonstrate that the unknowable uncertainty –- falls in preference somewhere in between the known and the known uncertainty.     

Choices at various levels of uncertainty: An experimental test of the restated diversification theorem

Our “Restated diversification theorem” (Skogh and Wu, 2005) says that risk-averse agents may pool risks efficiently without assignment of subjective probabilities to outcomes, also at genuine uncertainty. It suffices that the agents presume that they face equal risks. Here, the theorem is tested in an experiment where the probability of loss, and the information about this probability, varies. The result supports our theorem. Moreover, it tentatively supports an evolutionary theory of the insurance industry—starting with mutual pooling at uncertainty, turning into insurance priced ex ante when actuarial information is available.     

Inferring beliefs as subjectively imprecise probabilities

We propose a method for estimating subjective beliefs, viewed as a subjective probability distribution. The key insight is to characterize beliefs as a parameter to be estimated from observed choices in a well-defined experimental task and to estimate that parameter as a random coefficient. The experimental task consists of a series of standard lottery choices in which the subject is assumed to use conventional risk attitudes to select one lottery or the other and then a series of betting choices in which the subject is presented with a range of bookies offering odds on the outcome of some event that the subject has a belief over. Knowledge of the risk attitudes of subjects conditions the inferences about subjective beliefs. Maximum simulated likelihood methods are used to estimate a structural model in which subjects employ subjective beliefs to make bets. We present evidence that some subjective probabilities are indeed best characterized as probability distributions with non-zero variance.     

Behavioral Probabilities

This article introduces the concept of behavioral probabilities, along with an econometric procedure for jointly estimating these probabilities as well as individual utility functions. Behavioral probabilities that guide decisions differ from posterior probabilities that are reported after receiving risk information. The underlying process that generates behavioral probabilities reflects a behavioral anomaly as the new risk information takes on an excessive role. While utility function estimates are consistent with theoretical predictions, considering behavioral probabilities alters their implications. Viscusi's research is supported by the Olin Center for Law, Business, and Economics.     

The Optimality of the Expert and Majority Rules Under Exponentially Distributed Competence

We study the uncertain dichotomous choice model. In this model a set of decision makers is required to select one of two alternatives, say support or reject a certain proposal. Applications of this model are relevant to many areas, such as political science, economics, business and management. The purpose of this paper is to estimate and compare the probabilities that different decision rules may be optimal. We consider the expert rule, the majority rule and a few in-between rules. The information on the decisional skills is incomplete, and these skills arise from an exponential distribution. It turns out that the probability that the expert rule is optimal far exceeds the probability that the majority rule is optimal, especially as the number of the decision makers becomes large.     

Personal probabilities of probabilities

By definition, the subjective probability distribution of a random event is revealed by the (‘rational’) subject's choice between bets — a view expressed by F. Ramsey, B. De Finetti, L. J. Savage and traceable to E. Borel and, it can be argued, to T. Bayes. Since hypotheses are not observable events, no bet can be made, and paid off, on a hypothesis. The subjective probability distribution of hypotheses (or of a parameter, as in the current ‘Bayesian’ statistical literature) is therefore a figure of speech, an ‘as if’, justifiable in the limit. Given a long sequence of previous observations, the subjective posterior probabilities of events still to be observed are derived by using a mathematical expression that would approximate the subjective probability distribution of hypotheses, if these could be bet on. This position was taken by most, but not all, respondents to a ‘Round Robin’ initiated by J. Marschak after M. H. De-Groot's talk on Stopping Rules presented at the UCLA Interdisciplinary Colloquium on Mathematics in Behavioral Sciences. Other participants: K. Borch, H. Chernoif, R. Dorfman, W. Edwards, T. S. Ferguson, G. Graves, K. Miyasawa, P. Randolph, L. J. Savage, R. Schlaifer, R. L. Winkler. Attention is also drawn to K. Borch's article in this issue.     

One-shot decisions under Linear Partial Information

This paper purports to make a contribution to the analysis of a class of decisions that has received little attention in the literature, although it appears to be of considerable importance. Certain decisions cannot be repeated but must be made under fuzziness in the sense that state probabilities are not exactly known (LPI-fuzziness). The analysis of Linear Partial Information is applied to the principle of neglecting small probabilities found by Allais (1953), enabling the decision maker to break away from the maxmin criterion. By systematic exploitation of the fuzzy information available, strategies are shown to exist that provide payoffs whose lower bound exceeds the maxmin benchmark with sufficiently high probability. The same methodology is shown to be useful for dealing with the case of only ordinal preference orderings that are so typical of those crucial decisions that may be made only once in a lifetime.     

Subjective expected utility theory revisited: A reductio ad absurdum paradox

A paradox is posed and analyzed in which people reverse their preferences for information on probabilities versus prizes once the range of the unknown probabilities is sufficiently narrowed. This reversal is shown to be incompatible with both objective expected utility (EU) as well as subjective versions in which the same probability transformation applies to all random variables. Experimental data are presented showing that the reversals occur with small, medium and large payoffs.The present paradox is compared with those of Allais and Ellsberg, and found to differ in substantive ways. It raises further questions about the normative status of expected utility theory, especially its treatment of probability and value. The paradox specifically calls into question EU's substitution and compound probability axioms.     

Preference and belief: Ambiguity and competence in choice under uncertainty

We investigate the relation between judgments of probability and preferences between bets. A series of experiments provides support for the competence hypothesis that people prefer betting on their own judgment over an equiprobable chance event when they consider themselves knowledgeable, but not otherwise. They even pay a significant premium to bet on their judgments. These data connot be explained by aversion to ambiguity, because judgmental probabilities are more ambiguous than chance events. We interpret the results in terms of the attribution of credit and blame. The possibility of inferring beliefs from preferences is questioned.¹     

The market for (Ir)reproducible econometrics

Context: As the volume and complexity of research have increased, the amount of time spent on Institutional Review Board (IRB) review has decreased. The complexity of research has expanded, requiring increasingly specialized knowledge to review it. Dilemma: Under the current system, increasing numbers of research studies requiring expertise in ethics, new technologies or diverse study designs place a substantial burden upon local IRBs and often result in substantial variability among their reviews. This lack of uniformity in the review process creates uneven human subjects’ protection thus undermining the intent of the Common Rule. Objectives: To outline a scenario for expert centralized IRB review via implementation of a national virtual IRB review system overseen by the Office for Human Research Protections (OHRP). Conclusions: The complicated ethical issues and science involved in much of current research warrant an expert review panel. Centralized review would enable expert review specific to the research at hand, ensure consistency in human subjects protection, reduce the burden on local IRBs, and may reduce time spent obtaining approval. A centralized virtual system would allow IRB members to remain at their institutions while providing unprecedented expert review through currently available technology, and make information regarding monitoring and adverse event reporting available online in real-time.     

How to consult an expert? Opinion versus evidence

In this article, two modes of non-binding communication between an expert and a decision-maker are compared. They are distinguished mainly by the nature of the information transmitted by the expert. In the first one, the expert reports only his opinion (soft information) concerning the desirability of a certain action, whereas in the second one, he is consulted to provide evidence (hard information) to convince the decision-maker. The expert’s ability to provide evidence increases with the precision of his information. The article shows that requiring evidence is always beneficial to the decision-maker whereas it is beneficial to the expert, if and only if the preferences of both agents are different enough.     

The influence of probabilities on the response mode bias in utility elicitation

The response mode bias, in which subjects exhibit different risk attitudes when assessing certainty equivalents versus indifference probabilities, is a well-known phenomenon in the assessment of utility functions. In this empirical study, we develop and apply a cardinal measure of risk attitudes to analyze not only the existence, but also the strength of this phenomenon. Since probability levels involved in decision problems are already known to have a strong impact on behavior, we use this approach to study the impact of probabilities on the extent of the response mode bias. We find that the direction in which probabilities influence measured risk aversion is the opposite in the certainty equivalence (CE) method versus in the probability equivalence (PE) method. Utilizing the CE elicitation approach leads to an increase of risk seeking for gambles involving high probabilities. For the PE method, subjects tend to behave risk averse with gambles of high probabilities. This behavior is reversed in the gain domain. This “tailwhip” effect is consistently replicated in several experiments, involving both loss and gain domains of lotteries.     

Partial information on decisional competences and the desirability of the expert rule in uncertain dichotomous choice situations

The paper analyzes the expert resolution problem by employing extended versions of the uncertain dichotomous choice model. The main purpose of this study is to illustrate how the analysis of optimal decision rules can be carried out while dispensing with the common restrictive assumption of full information regarding individual decisional competences. In contrast to most previous studies in this field we here evaluate the expert rule under alternative assumptions regarding the available partial information on judgmental competences rather than compare it in an ad hoc manner to some common alternative rules, such as simple majority rule. A fuller optimality analysis allowing the evaluation of all relevant decision rules, and not merely the expert rule, is attempted for a five-member panel of experts assuming a uniform distribution of individual competences. For three-member groups the optimality issue is resolved by assuming no information on individual competences and interpreting the expert rule as an even-chance lottery on skills.     

Decomposing Hindsight Bias

Traditional accounts of hindsight bias inadequately distinguish primary hindsight bias from both secondary and tertiary hindsight bias. A subject exhibits primary bias when she assigns a higher ex ante probability estimate to actual outcomes, secondary bias when she believes that she herself would have made the same estimate of the prior probability of an event before receiving outcome information as she made after receiving it, and tertiary bias when she believes that third parties lacking outcome information were unreasonable if they did not make the same prior probability judgments that subjects now possessing such information make.In our experiments, we find that when people can readily calculate the actual ex ante probability of an outcome, they don't reassess that probability when told what outcomes actually occurred. They reassess only in situations in which they are unable to assess prior probabilities or when given information that the outcome was not simply a result of sampling or chance but the result of an imperceptible feature of the initial situation. Observed primary bias may therefore often be rational.     

Prudence and extensionality in theories of preference and value

Luce's axiom governing probabilities of choice is formulated as a principle governing metalinguistic probabilities. IfX, Y, W are sets of options, and δ(X), δ(Y), δ(W) are sentences asserting that choice is made from these sets, then the axiom is $$\begin{gathered} If \pi [\delta (X)] \ne 0 and \pi [\delta (X \cap Y)] \ne 0, then \hfill \\ \pi _{\delta (X)} [\delta (Y \cap W)] = \pi _{\delta (X \cap Y)} [\delta (W)]\pi _{\delta (X)} [\delta (Y)] \hfill \\ \end{gathered} $$ where π is a probability on sentences. The axiom is then entailed by extensionality of the probability π in company with a simple condition on probabilities of truth-functions. Conditions are also given under which the probability π is uniquely represented by a probability on the sets of options. What look to be logical constraints on the metalanguage entail a normative or prudential constraint. Debreu's well-known counterinstance to the axiom as a principle governing probability of choice is examined and a novel and consistent interpretation of the axiom is proposed.     

Estimating subjective probabilities

Subjective probabilities play a central role in many economic decisions and act as an immediate confound of inferences about behavior, unless controlled for. Several procedures to recover subjective probabilities have been proposed, but in order to recover the correct latent probability one must either construct elicitation mechanisms that control for risk aversion, or construct elicitation mechanisms which undertake “calibrating adjustments” to elicited reports. We illustrate how the joint estimation of risk attitudes and subjective probabilities can provide the calibration adjustments that theory calls for. We illustrate this approach using data from a controlled experiment with real monetary consequences to the subjects. This allows the observer to make inferences about the latent subjective probability, under virtually any well-specified model of choice under subjective risk, while still employing relatively simple elicitation mechanisms.     

Evaluating second-order probability judgments with strictly proper scoring rules

Empirical studies have demonstrated that uncertainty about event probabilities, also known as ambiguity or second-order uncertainty, can affect decision makers choice preferences. Despite the importance of second-order uncertainty in decision making, almost no effort has been directed towards the development of methods that evaluate the accuracy of second-order probabilities. In this paper, we describe conditions under which strictly proper scoring rules can be used to assess the accuracy of second-order probability judgments. We investigate the effectiveness of using a particular strictly proper scoring rule the ranked probability score - to discourage biased assessments of second-order uncertainty.     

Quasi-Bayesian Analysis Using Imprecise Probability Assessments And The Generalized Bayes’ Rule

The generalized Bayes’ rule (GBR) can be used to conduct ‘quasi-Bayesian’ analyses when prior beliefs are represented by imprecise probability models. We describe a procedure for deriving coherent imprecise probability models when the event space consists of a finite set of mutually exclusive and exhaustive events. The procedure is based on Walley’s theory of upper and lower prevision and employs simple linear programming models. We then describe how these models can be updated using Cozman’s linear programming formulation of the GBR. Examples are provided to demonstrate how the GBR can be applied in practice. These examples also illustrate the effects of prior imprecision and prior-data conflict on the precision of the posterior probability distribution.     

On subjective probability and related problems

Of late, probability subjectivism was resuscitated by the development of statistical decision theory. In the decision model, which is briefly described in the paper, the knowledge of a probability distribution over the states of nature plays a decisive role. What sources of probability knowledge are legitimate, or at all possible, is the main point at issue. Different definitions, evaluations, and foundations of probability are narrated, discussed, and weighed against each other. The typical research strategy of the statistician is set against axiomatics of subjective or mathematical probability. Finally, the epistemological roots of the probability concept are located by the author in what he calls the etiality principle.