Search

Retrieval Effects on Confidence in General Knowledge Winston R. Sieck (sieck.3@osu.edu) Department of Psychology, The Ohio State University 1827 Neil Ave, Columbus, OH 43210 USA J. Frank Yates (jfyates@umich.edu) Department of Psychology, The University of Michigan 525 E University Ave, Ann Arbor, MI 48109 USA At a recent panel discussion sponsored by the Ohio Board of Education, credential-laden fellows of the Discovery Institute argued with absolute conviction that the Darwinian theory of evolution is flawed, and that we must have been created by intelligent design. What underlies the confidence we have in our beliefs and knowledge? A typical assumption underlying many models of confidence in general world knowledge is that assessments of arguments that favor or oppose chosen answers will primarily, if not exclusively, determine confidence in choice (e.g. Griffin & Tversky, 1992; Koriat, Lichtenstein, & Fischhoff, 1980; also see Allwood & Granhag, 1996). Differences between models largely reflect distinct proposals for how evidence assessment is accomplished, and the extent and manner of bias that is postulated to exist in the process. For example, Koriat et al. (1980) suggested that memory search produces reasons for and against presented alternatives, and that assessment is biased in that choice-consistent reasons are relied upon more heavily than choice-inconsistent reasons. The current study tests the hypothesis that confidence also depends in part on successful retrieval of topical information that is not directly relevant towards arriving at a choice. Successful retrieval of facts about the general topic is used as evidence that one is knowledgeable about the subject area, and thus likely to be reasoning correctly. Method Participants (n=159) were presented with a series of questions following the form: “Which species has a longer gestation period: (a) chimpanzees, or (b) humans?” Participants first chose one of the two alternatives as more likely to be correct, and then specified a probability between 50% and 100% that their choice was, in fact, correct. Subjects in a reasons condition were asked to write all possible reasons for and against each alternative answer, prior to choosing. Participants in a recall condition were asked to recall all of the facts they could about the topic of each question before choosing. Control condition participants simply answered the questions with standard, non-directive instructions. Results The principal results are shown in Table 1. Mean confidence, proportion correct, and overconfidence were virtually equivalent across conditions. Accuracy discrimination (mean confidence given correct – mean confidence given incorrect) was larger for recall than the other two conditions. Also, the correlation between confidence level and choice accuracy was larger for recall than the other conditions. Table 1: Confidence/Correct Indices by Condition. Indices Mean Conf. Prop. Correct Accuracy Discrimination Pearson's r Control Reasons Recall Discussion Writing all possible reasons for and against alternatives had no impact on choice or confidence. Writing all facts that could be recalled about the topic of each question resulted in confidence judgments that better discriminated between correct and incorrect answers, as compared with control and reasons conditions. Preliminary results from coding of the listed reasons and recollections (not shown), indicate that approximately the same amounts of information were generated under the two procedures, but that important qualitative differences in the protocols seem to exist. It appears that successful recall of facts that are relevant to the topic, but that do not constitute reasons for or against presented alternatives, is taken as critical evidence that chosen alternatives are correct. Acknowledgments This research was supported by NSF Grant SES-9911301. References Allwood, C. M., & Granhag, P. A. (1996). The effects of arguments on realism in confidence judgements. Acta Psychologica, 91, 99-119. Griffin, D., & Tversky, A. (1992). The weighing of evidence and the determinants of confidence. Cognitive Psychology, 24, 411-435. Koriat, A., Lichtenstein, S., & Fischhoff, B. (1980). Reasons for confidence. Journal of Experimental Psychology: Human Learning and Memory, 6, 107-118.

Highlights • We review the literature on culture and decision making. • We comprehensively analyze ten decision making phases. • Possible causes of cultural differences are reviewed and critiqued. • Big-picture questions for future research are discussed., The study of culture and decision making addresses variations in how and why people from different cultures sometimes tend to decide differently. This review is organized around what is intended to be a comprehensive analysis of the distinct fundamental questions that people must answer in the process of making virtually all real-life decisions. Our emphasis was on recent developments as well as identifying important yet neglected topics (e.g., how decision episodes get started—or not, and why some decisions are never implemented). Early as well as current efforts have focused mainly on East Asian and North American Caucasian cultures, with little treatment of other populations. In such studies, individualism and collectivism have been the dominant explanatory factors although related but distinct concepts such as “tightness” and “looseness” have been welcome additions to recent discussions. Throughout, the review emphasizes practical concerns, such as the challenges of intercultural learning and collaboration.

This article has two aims. The first is to present results that partly explain why some automobile drivers choose to use their seatbelts only part time, thereby exposing themselves to unnecessary risk. The second is to offer and illustrate the "cardinal decision issue perspective"((1)) as a tool for guiding research and development efforts that focus on complex real-life decision behaviors that can entail wide varieties of risk, including but not limited to inconsistent seatbelt use. Each of 24 young male participants drove an instrumented vehicle equipped to record continuously seatbelt use as well as other driving data. After all trips were finished, each participant completed an interview designed to reconstruct how he made randomly selected seatbelt-use decisions under specified conditions. The interview also examined whether and how drivers established "decision policies" regarding seatbelt use. Such policies were good predictors of inconsistent seatbelt use. Drivers who had previously adopted policies calling for consistent seatbelt use were significantly more likely than others to actually drive belted. Meta-decisions about seatbelt policy adoption appeared to rest on factors such as whether the driver had ever been asked to consider selecting a policy. Whether a driver made an ad hoc, on-the-spot seatbelt-use decision was associated with a perceived need to make such a decision. Finally, participants with full-time policies were especially likely to deploy their seatbelts by default, without recognizing the need to decide about belt use on a trip-by-trip basis. We end with recommendations for reducing inconsistencies in seatbelt use in actual practice.

Human judgment often violates normative standards, and virtually no judgment error has received as much attention as the conjunction fallacy. Judgment errors have historically served as evidence for dual-process theories of reasoning, insofar as these errors are assumed to arise from reliance on a fast and intuitive mental process, and are corrected via effortful deliberative reasoning. In the present research, three experiments tested the notion that conjunction errors are reduced by effortful thought. Predictions based on three different dual-process theory perspectives were tested: lax monitoring, override failure, and the Tripartite Model. Results indicated that participants higher in numeracy were less likely to make conjunction errors, but this association only emerged when participants engaged in two-sided reasoning, as opposed to one-sided or no reasoning. Confidence was higher for incorrect as opposed to correct judgments, suggesting that participants were unaware of their errors.

Cognitive engineering and decision making has become central to current human factors and ergonomics research and practice. This trend parallels exciting developments in decision making research in general. Experimental economics and judgment and decision making research boom academically and gain much public attention, while theories and economic models of individual choice are expanding and gaining recognition in applications and practical solutions to real-world problems. This symposium brings together three leading figures in different traditions of research in decision making: Gary Klein represents the naturalistic decision making approach. Frank Yates represents research on behavioral decision making and heuristics and biases. Alvin E. Roth, the 2012 Noble laureate for economics, represents game theory and market design applications in economics. The invited speakers will present their views of the different research traditions in decision making and they will provide some thoughts on the trends and possible future developments of research, which we hope will provoke new ideas and discussions in Cognitive Engineering and Decision Making in HFES.

Three studies examined cultural variations in indecisiveness among Chinese, Japanese, and Americans. In Study 1, validated self-report, comprehensive measures of indecisiveness indicated large cultural differences, with Japanese participants exhibiting substantially more indecisiveness than Chinese or Americans. Study 2 provided evidence that such cultural variations correspond to variations in people’s positive versus negative values for decisive behaviors, suggesting that such values are plausibly an important means for motivating and sustaining cultural differences in indecisiveness. Study 3 provided direct behavioral instances of the differences in indecisiveness implicated in Studies 1 and 2. It also suggested that thoroughness might be an important cognitive mechanism whereby cultural differences in indecision actually occur, with thoroughness being especially prominent among Japanese decision makers. Suggestions for theory concerning the nature and foundations of indecisiveness and its cultural variations are developed and discussed, along with plausible implications for real-life practical issues, for example, in politics and management.

Many, if not most, real-life decisions depend heavily on judgments, which are opinions as to what was, is, or will be the state of some relevant aspect of the world. Thus, for instance, physicians’ medical treatment decisions are often predicated on judgments as to the true cause of the patient’s symptoms, and jurors’ verdicts rest on their beliefs about whether the defendant actually committed the crime in question. Studies have shown that the kinds of judgments that inform people’s decisions can differ substantially and surprisingly according to the cultural backgrounds (e.g., Chinese versus Japanese versus American) of those rendering the judgments, particularly judgments expressed as probabilities (e.g., ‘‘There’s a 75% chance that that will happen’’). This article reviews major cultural variations in key characteristics of probability judgments (e.g., their overconfidence) and why those differences exist. It also discusses the theoretical and practical implications of the findings, including for collaborations among decision makers from distinct cultures.

Almost all real-life decisions entail attribute conflict; every serious choice alternative is better than its competitors on some attribute dimensions but worse on others. In pre-decisional “coherence shifting,” the decision maker gradually softens that conflict psychologically to the point where one alternative is seen as dominant over its competitors, or nearly so. Specifically, weaknesses of the eventually chosen alternative come to be perceived as less severe and less important while its strengths seem more desirable and significant. The research described here demonstrates that difficult multiattribute decision problems are aversive and that pre-decisional coherence shifting aids individuals in regulating that emotional discomfort. Across three studies, attribute conflict was confirmed to be aversive (Study 1), and skin conductance responses and ratings of decision difficulty both decreased in participants who coherence shifted (Study 2). Coherence shifting was also diminished among decision makers who were depleted of regulatory resources, known to be required for common emotion regulation mechanisms. Further, coherence shifting was shown to be relatively common among people who reported strong suppression tendencies in everyday emotion regulation (Study 3). Overall, the data suggest that, at least in part, coherence shifting serves as a tool that helps decision makers manage the pre-decisional discomfort generated by attribute conflict. Theoretical and practical implications are discussed.

This paper considers the relationship among the inferential style characterizing a society, its litigation system, and other aspects of its culture. Differences between common law and Continental systems of litigation are associated with a stronger individualistic tendency in Anglo-American society than in Continental societies. William Twining’s model of the ‘rationalist’ tradition in Anglo-American adjudication and evidence scholarship, as modified by an overlay of individual rights, tracks important differences between the common law system, on the one hand, and Asian litigation systems, particularly East Asian systems, on the other, and between Western and Asian styles of thinking. The paper speculates that one reason the jury system has taken root in the Anglo-American system is that Anglo-American culture is more likely to generate among its people habits and customs of deliberation that will make them willing and effective jurors.

The purpose of this research was to determine the mechanisms underlying the graphical effect identified by Stone, Yates, and Parker (1997) , in which graphical formats for conveying risk information are more effective than numerical formats for increasing risk-avoidant behavior. Two experiments tested whether this graphical effect occurred because the graphical formats used by Stone et al. highlighted the number of people harmed by the focal hazard, causing the decisions to be based mainly on the number of people harmed (which we label the “foreground”) at the expense of the total number of people at risk of harm (which we call the “background”). Specifically, two graphical formats were developed that displayed pictorially both the number of people harmed and the total number at risk, and use of these display formats eliminated the graphical effect. We thus propose that the previously discussed graphical effect was in fact a manifestation of a more general foreground:background salience effect, whereby displays that highlight the number of people harmed at the expense of the total number of people at risk of harm lead to greater risk avoidance. Theoretical and practical implications are discussed.

Escalation effects are instances in which a decision maker continues to commit resources to a losing course of action, solely because prior resource allocations have been made. The research described in this article examined the potential influence of financial budgets on such escalation effects. Determining such influences is important for a variety of reasons, including the prevalence of budgets in almost all real-world organizations where escalation effects would, in principle, significantly compromise the organizations' welfare. The results of five studies employing undergraduates and part-time master's degree students with working experience demonstrated the following: (a) the mere existence of financial budgets does not in and of itself affect the incidence of escalation effects; (b) escalation effects occur among experienced participants even when explicit future costs and benefits are provided; (c) the prospect that making additional investments to escalate commitment to a project would overspend a budget tends to reduce significantly decision makers' tendencies to exhibit escalation effects; (d) even when investment expenditures would not exceed overall budget limits, the prospect of overspending the budget for a substage of a multistage investment project has the same suppressive influence on escalation effects as the prospect of overspending an entire budget; and (e) a side effect of the use of budgets is that they can induce decision makers to eschew investment opportunities that generate net benefits. There is also evidence that the prospect of exceeding a budget introduces two mechanisms into participants' decision-making processes—more marginal cost-benefit reasoning and greater sensitivity to proscriptive consequences of exceeding a budget. Implications are discussed.

Exemplar and connectionist models were compared on their ability to predict overconfidence effects in category learning data. In the standard task, participants learned to classify hypothetical patients with particular symptom patterns into disease categories and reported confidence judgments in the form of probabilities. The connectionist model asserts that classifications and confidence are based on the strength of learned associations between symptoms and diseases. The exemplar retrieval model (ERM) proposes that people learn by storing examples and that their judgments are often based on the first example they happen to retrieve. Experiments 1 and 2 established that overconfidence increases when the classification step of the process is bypassed. Experiments 2 and 3 showed that a direct instruction to retrieve many exemplars reduces overconfidence. Only the ERM predicted the major qualitative phenomena exhibited in these experiments.

Previous studies have revealed surprising and persistent cross-cultural variations in overconfidence, whereby respondents in some Asian cultures (e.g., Chinese) exhibit markedly higher degrees of overconfidence than respondents in other cultures (e.g., in the United States and Japan). Most of those demonstrations have entailed general knowledge tasks (e.g., answering questions such as whether Europe is larger than Australia). The present studies sought to determine whether such cross-cultural variations extend to judgments about the kinds of events that bear upon more common practical decisions and to aspects of accuracy other than overconfidence. Subjects in Taiwan, Japan, and the United States made probabilistic differential diagnoses of fictional diseases in a stochastic artificial ecology. Results revealed that previously observed cross-cultural variations do indeed generalize. The data were also informative about several potential accounts for such variations, e.g., arguing against a proposal that they rest on different emphases on discrimination rather than calibration, but consistent with the influences of culture-specific cognitive customs, including responsiveness to explicitly displayed information, regardless of its presumed validity. Copyright 1998 Academic Press.

Many of us believe that, after writing about a subject, we understand it more deeply. Studies in education indicate that writing does indeed enhance comprehension. Three experiments examined whether similar “exposition effects” exist for decision making. In these experiments, subjects were confronted with standard framing problems. Positive exposition effects would require that the influence of alternative frames on subjects' choices be diminished by exposition demands. Control subjects made choices under customary, non-exposition instructions. Others chose after writing rationales for their selections (exposition), after explicitly planning to write such rationales, or merely in anticipation of writing the rationales. Exposition reduced framing effects in each of the experiments. The magnitudes of the effects were greatest for subjects who wrote before choosing. Also, exposition markedly increased subjects' confidence that their choices were appropriate. Theoretical and practical implications are discussed.

Overconfidence in general knowledge is typically stronger among Asian than among Western subject groups. The research described here examined the possibility that such differences might be a manifestation of previously reported extreme response styles on the part of Asian respondents. This hypothesis was evaluated by comparing overconfidence implicit in directly reported judgments and judgments inferred from decisions Chinese and American subjects made about wagers in which they could earn actual, material goods. Contrary to the response style hypothesis, indications of extreme Chinese overconfidence were unaffected by whether judgments were direct or inferred from decisions. However, American subjects' inferred judgments were even more overconfident than their direct judgments. The bias in all subjects' inferred judgments indicates that, in disagreement with some interpretations of recent developments in the literature, overconfidence is indeed a “real,” consequential phenomenon, not a data-analytic artifact. An additional, serendipitous finding was that the inferred judgments of both Chinese and American subjects were far less variable than their direct judgments.

There is an established literature describing how probabilistic forecasts, and hence forecasters, should be evaluated. The present paper takes a different and heretofore neglected perspective on evaluation. It addresses how those who receive and use probabilistic predictions—forecast ‘consumers’—appraise these assessments. Results indicate that there are reliable and important differences between subjective and formal evaluation principles. Among the distinctive features of common subjective appraisal strategies are: (a) an emphasis on judgments being categorically ‘correct’; (b) special attention to forecast extremeness; (c) the desire for good explanations of forecasts; and (d) the sensitivity of appraisals to how pertinent information is displayed to the evaluator. Theoretical and practical implications are discussed.

A commonly used judgment research task requires the subject to choose among alternative answers to a general knowledge question and then state a probability that the selected response is correct. People often exhibit overconfidence in this task. Such overconfidence is stronger in most Asian countries than in Western countries. This article addresses lay expectations and accounts for this difference, as well as for such overconfidence more generally. An empirical study demonstrated that subjects in both Taiwan and the United States typically expect cross-national variations to be the opposite of those that actually occur. It also showed that dominant lay theories for overconfidence rest exclusively on factors such as inflated self-appraisals. A review revealed that such accounts are consistent with scholarly self-esteem explanations for a variety of self-appraisals. However, existing data imply that overconfidence in general knowledge rests on other mechanisms. The extensive theoretical and practical implications of the observed overconfidence misattributions and cross-national variations are discussed.

Individuals are said to be accountable whenever their performance is monitored and there are consequences (either tangible or intangible) associated with that evaluation. We propose that there are at least two distinct types of accountability. One focuses on justification of the procedure used to arrive at an action (procedural accountability, or PA); the other focuses on the quality of the outcomes of that action (outcome accountability, or OA). In three experiments, subjects judged the likelihood that each of a set of individuals held a particular attitude on the basis of background information about those individuals. Results from Experiments 1 and 2 suggested that PA encourages people to take more of the available information into account. Whether or not this effect was beneficial, however, depended on how relevant that information was to the target judgment. OA had only detrimental effects, increasing the amount of noise (or “scatter”) in subjects' judgments and thus leading to lower accuracy overall. Experiment 3 extended the work on PA, focusing on its relationship to the effects of outcome feedback. Results indicated that PA significantly reduced the tendency to be overly responsive to such feedback by reducing the variability in judgment unrelated to the target event. Practical and theoretical implications are discussed.

Decision making is the lifeblood of every organization and the central focus in the practice of evidence-based management (EBMgt). To help guide the practice of EBMgt, this chapter describes fundamental concepts of decision making and decision management. It begins with a discussion of the features that distinguish decision making from other, related concepts, such as more general problem solving and judgment. It then describes how crucial aspects of decision management are approached especially effectively when decision processes are broken down into essential elements. It then describes the theory known as the “cardinal decision issue perspective,” and illustrates its use as a decision management tool.

According to most prescriptive decision rules, formally equivalent methods of communicating risk information should have identical effects on risk-taking behavior, even if the pertinent displays are different. The present work takes two methods commonly employed in epidemiology, incidence rates and relative risks, and examines their comparative effects on risk-avoidant behavior. In Experiment 1, we presented 108 undergraduates with information about risks associated with different brands of tires and toothpaste and displayed that information either as incidence rates or as a relative risk ratio. For the tires product, subjects given the relative risk format were willing to pay more money for a safer product than were subjects given the incidence rate format. There were, however, no differences between the two conditions for the toothpaste product. Experiment 2 evaluated two potential explanations for the difference in findings between the two products. The majority of the data supported an "editing" hypothesis, which suggests that extreme low-probability risks, such as those associated with tire blowouts, are edited to "essentially nil risk," while more moderate risks, such as those associated with periodontal disease, are considered to be small but significant. These findings are discussed in the context of fuzzy trace theory and related models, which suggest that people reason on the basis of simplified representations rather than on the literal information available.

We investigated a phenomenon called judgmentalovershadowing. Subjects predicted whether each of several patients had a disease on the basis of whether or not the patient had each of two symptoms. For all the subjects, the presence of the disease was moderately contingent on the presence of one of the symptoms (Sl.).In Condition 1 of our first experiment, the presence ofthe dis­ ease was highly contingent on the presence of the other symptom (S2). In Condition 2, the pres­ ence of the disease was independent of S2. Judgmental overshadowing occurred in that the Sl­ disease contingency was judged to be stronger in Condition 2 than in Condition 1. Subsequent experiments showed that judgmental overshadowing depends little on the form of the judgment, is not due to a response bias or contrast effect, and does not depend on subjects' actively diagnos­ ing each patient. These results are consistent with, and are generally predicted by, an associative­ learning model of contingency judgment.

Objective This study examined determinants of self-monitoring of blood glucose (SMBG) in patients with diabetes based on a theoretical model. The model focuses on two equally important components of the decision process within a single framework: (1) making a decision and (2) enacting the decision. Methods Diabetes patients with HbA1c >7% and requiring insulin therapy were recruited from a southeastern Michigan healthcare system. Responses were elicited using a mail-in questionnaire. SMBG was measured using a two-week diary and two recall measures. The applicability of the theoretical model as it relates to SMBG wastested using structural equation modeling (SEM). Results The SEM analysis demonstrated that goal desire is an antecedent to goal intentions and that implementation desire and implementation intentions mediate the path between goal intentions and SMBG. Further, implementation intentions are a significant predictor of SMBG. Conclusion Consistent with the theoretical model, results indicated that implementation desire and implementation intentions mediate the relationship between strength of goal intention and performing specific behaviors to achieve the goal. Practice implications It is evident that in order to be more effective, educational programs or interventions should consider incorporating implementation and action plans to help patients translate decisions and intentions into behavior.

People frequently must make judgments of quantities. This article concerns how increasing the number of cues supporting such judgments affects accuracy as well as the judgment process itself. Research on these issues is sparse. This seems partly due to ambiguity about what it means to say that quantity judgments are accurate. Thus, a major focus of the article is on integrating 2 disparate approaches to analyzing quantity judgment, lens model analysis from psychology and mean squared error analysis from economic forecasting

People's ability to assess probabilities of various events has been the topic of much interest in the areas of judgment, prediction, decision making, and memory. The evaluation of probabilistic judgments, however, raises interesting logical questions as to what it means to be a"good" judge. In this article, we focus on a normative concept of probabilistic accuracy that we call discrimination and present a measure of a judge's discrimination skill. This measure builds on an earlier index (Murphy, 1973) and has the advantages that (a) it can be interpreted as the percentage of variance accounted for by the judge and (b) it is unbiased. By way of deriving this new discrimination measure, we also show that it is related to Pearson's chi-square statistic, a result which may be useful in the future development of hypothesis testing and estimation procedures.

Undergraduate and graduate students in finance courses made probabilistic forecasts of the quarterly changes in the stock prices and earnings of publicly traded companies. Consistent with previous findings ( Stael von Holstein, 1972 ), the overall accuracy of both price and earnings forecasts was very modest; subjects would have been more accurate had they predicted that price changes were equally likely to fall into any of the specified ranges. Also consistent with earlier suggestions of “inverted” expertise effects, undergraduate subjects were more accurate than graduate subjects. Decompositional analyses of subjects' judgments were consistent with the hypothesis that graduate students' relatively poor accuracy was affected by their greater tendency to report forecasts that varied from one stock to the next instead of the same forecast for every one. It is argued that the most plausible explanation is that the graduate subjects responded to cues they thought were predictive, but which actually were not. However, it cannot be ruled out completely that the graduate subjects attended to truly predictive cues, but were simply unable to use them appropriately.

There have long been speculations that graphical and numerical presentations of risk statistics differ in their impact on people’s willingness to pursue actions that could harm or even kill them. But research has been unclear about the processes whereby the pictorial character of graphical displays per se might affect those risky decisions or even whether such effects actually occur. In two studies, we demonstrate that the pictorial nature of a graphical risk display can, indeed, increase risk avoidance. This increase is associated with a heightened impression of the riskiness of less safe alternatives. The results suggest that this picture-driven, intensified sense of riskiness, in turn, rests on two kinds of mechanisms: one cognitive, the other affective. Cognitively, pictorial presentations impose weaker upper bounds on people’s internal representations of the chances that riskier alternatives will bring about actual harm. Affectively, pictures ignite stronger, more aversive negative associations with riskier options and their outcomes.

Cataloged from PDF version of article. Highly knowledgeable people often fail to achieve highly accurate judgments, a phenomenon sometimes called the ‘‘processperformance paradox.’’ The present research tested for this paradox in foreign exchange (FX) rate forecasting. Forty professional and 57 sophisticated amateur forecasters made one-day and one-week-ahead FX predictions in deterministic and probabilistic formats. Among the conclusions indicated by the results are: (a) professional accuracy usually surpasses amateur accuracy, although many amateurs outperform many professionals; (b) professionals appear to achieve high proficiency via heavy reliance on predictive information (unlike what has been observed before, e.g., for stock prices); (c) forecast format strongly affects judgment accuracy and processes; and (d) apparent overconfidence can transform itself into underconfidence depending on when and how forecasters must articulate their confidence. 2003 Elsevier Science (USA). All rights reserved.

Many studies have shown that people have difficulty judging the diagnostic value of conditional probability information with respect to one or more hypotheses. The present research addressed two aspects of performing the diagnostic task in a health care decision: (a) recognition of the information's importance, and (b) correct usage of that information. In experiment 1, health care providers, who are trained in, and regularly exposed, to conditional probabilities imparting diagnostic information, exhibited at least a rudimentary recognition of the need for this information in assessing diagnosticity. Experiment 2 indicated that health care and layperson subjects had difficulty in actually applying the information, however. This difficulty prompts a need for judgment aids and caution in using diagnostic information.

The question of whether aversive outcomes are generally preferred at sooner or at later times in the future is examined. Theories predicting choices of advanced and deferred aversive outcomes are summarized. A previous result suggesting overwhelming preference for more immediate aversive outcomes is challenged on the basis of problems endemic to experimental studies of decision making. An experiment that attempted to circumvent those problems is described. Subjects chose between sums of money that could be lost at various times in the future. The subjects were evenly divided in their preferences for advanced and deferred losses. Subjects in a control condition duplicated the results of the previous experiment that had demonstrated the predominance of preferences for advanced aversive outcomes. The implications of the conclusions for a representation of choice over deferred outcomes are discussed.

Subjects made judgments concerning the strength and direction of the contingency between two dichotomous variables in a situation in which no contingency actually existed. The judgments exhibited a significant primacy effect. The effects of warning and not warning the subjects that they would be required to recall the frequencies of observed event co-occurrences implied that this primacy effect was due to ‘attention decrement’ (Anderson 1981). According to this hypothesis, attention to contingency-relevant information diminishes after the subject is exposed to only a small portion of the available information.

Probability theory requires that the disjunction of two events be judged at least as likely as either of those events. Two studies are described which imply that violations of this requirement are very common. The data also show that subjects do not commit disjunction errors because they misinterpret the disjunctive statement “A or B” to mean “A or B, but not both.” Instead, the pattern of errors is consistent with a judgment process describable as averaging. These results provide support for a “signed combination model of qualitative likelihood judgment.” This model generalizes a previous one which applied to judgments concerning conjunctive events (J. F. Yates and B. W. Carlson, 1986