Combining information: Probability summation and probability averaging in detection and discrimination.

Detection may improve if a stimulus offers 2 cues rather than 1. This is sometimes attributed to probability summation of independent detections, which provides an especially simple model for sensory information combination. However, this model assumes a strong bias toward the positive response, which is not appropriate for discrimination. The probability summation model is here extended to apply to discrimination and to allow different degrees of summation, ranging from complete through partial to probability averaging, and the use of this model is illustrated for the method of constant stimuli. It allows performance based on independent decisions to be distinguished from performance (e.g., integration) that is better than summation can explain. Models for the discrimination of complex stimuli may provide a tool for studying the development of expertise in areas where this involves a perceptual component, such as clinical judgment. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Improving Predictive Accuracy with a Combination of Human Intuition and Mechanical Decision Aids

This study examines the intuitive combination of human judgment and mechanical prediction under varied information conditions. As expected, mechanical prediction outperformed human intuition when based on the same information, but a combined approach was best when judges had access to relevant information not captured by the model (information asymmetry). The model was useful for differentiating between the event outcomes (improved slope), while eliminating the bias caused by base-rate neglect. Human intuition was useful for incorporating relevant information outside the scope of the model, resulting in improved slope and reduced judgment scatter. The addition of irrelevant information was detrimental to judgment accuracy, causing an increase in bias and a reduction in slope. These results provide insight into how and when combining mechanical prediction and human intuition is likely to result in improved accuracy. Copyright 1998 Academic Press.     

Does a general monitoring skill exist?

Two experiments investigated whether monitoring is better characterized as a domain-specific or a domain-general phenomenon. In Exp 1, college students' performance and discrimination accuracy were not correlated across 8 different domains, whereas confidence and judgment bias were. With tests matched on all salient dimensions except content, in Exp 2, students' performance, confidence, discrimination, and bias were correlated across all or most domains. In addition, confidence was correlated even after the effect of performance was removed. These findings lend qualified support to the domain-general hypothesis, which states that monitoring within a specific domain is governed by general metacognitive processes in addition to domain-specific knowledge. Theoretical and educational implications are discussed, with particular attention given to the origin and development of the general monitoring skill. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

An associative framework for probability judgement: An application to biases.

Three experiments show that understanding of biases in probability judgment can be improved by extending the application of the associative-learning framework. In Experiment 1, the authors used M. A. Gluck and G. H. Bower's (1988) diagnostic-learning task to replicate apparent base-rate neglect and to induce the conjunction fallacy in a later judgment phase as a by-product of the conversion bias. In Experiment 2, the authors found stronger evidence of the conversion bias with the same learning task. In Experiment 3, the authors changed the diagnostic-learning task to induce some conjunction fallacies that were not based on the conversion bias. The authors show that the conjunction fallacies obtained in Experiment 3 can be explained by adding an averaging component to M. A. Gluck and G. H. Bower's model. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Driven by power? Probe question and presentation format effects on causal judgment.

It has been proposed that causal power (defined as the probability with which a candidate cause would produce an effect in the absence of any other background causes) can be intuitively computed from cause-effect covariation information. Estimation of power is assumed to require a special type of counterfactual probe question, worded to remove potential sources of ambiguity. The present study analyzes the adequacy of such questions to evoke normative causal power estimation. The authors report that judgments to counterfactual probes do not conform to causal power and that they strongly depend on both the probe question wording and the way that covariation information is presented. The data are parsimoniously accounted for by an alternative model of causal judgment, the Evidence Integration rule. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Integration of contingency information in judgments of cause, covariation, and probability.

Integration of contingency information underlies many cognitive tasks including causal, covariational, and probability judgments. The authors' feature-analytic approach was used to account for the findings that people differentially weight specific types of conjunctive information in causal (Experiment 1) and noncausal (Experiment 2) contingency judgments. These findings were explained in terms of positive-test and sufficiency-test biases, which were found in both judgment domains. The same biases, however, were not observed in normative conditional-probability judgments (Experiment 3). The authors argue that this discrepancy is owing to the differential clarity of normative criteria in these domains. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Motivated sensitivity to preference-inconsistent information.

If preference-inconsistent information initiates more effortful cognitive analysis than does preference-consistent information, then people should be more sensitive processors of information they do not want to believe than of information they do want to believe. Three studies supported this prediction. Study 1 found that inferences drawn from favorable interpersonal feedback revealed a correspondence bias, whereas inferences drawn from unfavorable feedback were sensitive to situational constraint. Study 2 showed this sensitivity to the quality of unfavorable feedback to disappear under cognitive load. Study 3 showed that evaluations of the accuracy of favorable medical diagnoses were insensitive to the probability of alternative explanation, whereas evaluations of unfavorable diagnoses were sensitive to probability information. The importance of adaptive considerations in theories of motivated reasoning is discussed. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Irrelevant sound disrupts order information in free recall as in serial recall

The claim that the sensitivity of free recall to disruption by irrelevant sound is a function of the extent to which rote rehearsal is employed as a mnemonic strategy was investigated in two experiments. The degree of disruption by irrelevant sound in terms of both item and order information was contrasted under serial and free recall instructions. Irrelevant sound was found to disrupt order and item information equally in serial and free recall tasks (Experiment 1). Contrary to previous reports, an effect of irrelevant sound was also demonstrated on free recall of particularly long lists, and the interaction between list length and retention interval in the irrelevant sound effect was examined (Experiment 2). Generally, the results support the view that irrelevant sound disrupts the use of order cues.     

Judgeability concerns: The interplay of information, applicability, and accountability in the overattribution bias.

In 3 studies, the authors examined the impact of judgeability concerns in the overattribution bias (OAB; G. A. Quattrone, 1982) by manipulating the presence-absence of a constrained essay, the participants' accountability, and the applicability of the available information. A constrained essay was neither necessary nor sufficient to anchor a judgment. When no essay was circulated, no OAB occurred in the cases of accountability or of inapplicability (Studies 1 and 2). When the essay was provided, however, both accountability and inapplicability were needed to eliminate the OAB (Studies 2 and 3). These findings did not result from conversational rules or demand characteristics. They illustrate that people control the expression of a judgment made under uncertainty; people express the judgment to the extent they feel entitled to do so. The results are discussed in the wider context of current multistage models of the dispositional inference process. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Aging and Conditional Probability Judgments: A Global Matching Approach.

Age differences in bias in conditional probability judgments were investigated based on predictions derived from the Minerva-Decision Making model (M. R. P. Dougherty, C. F. Gettys, & E. E. Ogden, 1999), a global matching model of likelihood judgment. In this study, 248 younger and older adults completed frequency judgment and conditional probability judgment tasks. Age differences in the frequency judgment task are interpreted as an age-related deficit in memory encoding. Older adults' stronger biases in the probability judgment task point to age differences in criterion setting. Age-related biases were eliminated when age groups were equated on memory encoding by means of study time manipulation. The authors conclude that older adults' stronger judgment biases are a function of memory impairment. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Implicit memory bias for threat-relevant information in individuals with generalized social phobia.

The authors used a noise judgment task to investigate implicit memory bias for threat in individuals with generalized social phobia (GSP). Participants first heard neutral sentences (e.g., "The manual tells you how to set up the tent.") and social-threat sentences (e.g., "The classmate asks you to go for drinks."). Implicit memory for these sentences was then tested by asking participants to rate the volume of noise accompanying the presentation of these "old" sentences intermixed with "new" sentences that had not been previously presented. Implicit memory for old sentences is revealed when participants rate the noise accompanying old sentences as less loud than the noise accompanying new sentences. Those with GSP demonstrated an implicit memory bias for social-threat sentences, whereas controls did not. This differential priming effect suggests that information about threat may be automatically accessed in GSP. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Selective information use in the development of search behavior.

Age differences in children's use of various sources of information about object location were examined in a study of search behavior in 3-, 4-, and 7-yr-olds. The 2 principal sources of information were general associative knowledge about the typical locations of objects (location specificity) and explicit verbal statements about object location. Age differences in the results reflected increasing utilization of the 2 sources of information when each was considered separately. More importantly, the design allowed consideration of Ss' ability to combine the information to limit search. In this respect, the findings reveal sophisticated information integration on the part of even the youngest Ss tested. At all ages, Ss were able to combine both sources of information when both were relevant and to ignore irrelevant location-specificity information when the verbal information was logically superordinate. (7 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The role of short-term memory capacity and task experience for overconfidence in judgment under uncertainty.

Research with general knowledge items demonstrates extreme overconfidence when people estimate confidence intervals for unknown quantities, but close to zero overconfidence when the same intervals are assessed by probability judgment. In 3 experiments, the authors investigated if the overconfidence specific to confidence intervals derives from limited task experience or from short-term memory limitations. As predicted by the naive sampling model (P. Juslin, A. Winman, & P. Hansson, 2007), overconfidence with probability judgment is rapidly reduced by additional task experience, whereas overconfidence with intuitive confidence intervals is minimally affected even by extensive task experience. In contrast to the minor bias with probability judgment, the extreme overconfidence bias with intuitive confidence intervals is correlated with short-term memory capacity. The proposed interpretation is that increased task experience is not sufficient to cure the overconfidence with confidence intervals because it stems from short-term memory limitations. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Some more sensitive measures of sensitivity and response bias.

In this article, the author proposes a new pair of sensitivity and response bias indices and compares them to other measures currently available, including d′ and β of signal detection theory. Unlike d′ and β, these new performance measures do not depend on specific distributional assumptions or assumptions about the transformation from stimulus information to a discrimination judgment. With simulated and empirical data, the new sensitivity index is shown to be more accurate than d′ and 16 other indices when these measures are used to compare the sensitivity levels of 2 experimental conditions. Results from a perceptual discrimination experiment demonstrate the feasibility of the new distribution-free bias index and suggest that biases of the type defined within the signal detection theory framework (i.e., the placement of a decision criterion) do not exist, even under an asymmetric payoff manipulation. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Outcome feedback: Hindsight and information.

Although "hindsight bias" research has demonstrated that outcome feedback leads people to exaggerate the odds they would have placed on known outcomes, learning theories view feedback as the key to effective adaptation. Our model and data show that outcome feedback has multiple effects. People can extract diagnostic information from feedback despite overestimating what they would have known in foresight. Ss reliably discriminate easy tasks where they "knew it all along" from difficult tasks where they "never would have known it"; differential reactions to feedback provide information useful in other judgment tasks such as assessment of population base-rates and personal knowledge calibration. In Experiments 1–4, feedback increased judgmental accuracy by over 150%. However, a final experiment suggests that certain tasks (insight problems) produce such strong "I knew it all along" reactions that hindsight can overwhelm the information contained in feedback and reduce predictive accuracy. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Use of job information as a criterion in employment decisions of interviewers.

Conducted 2 factorially designed experiments to investigate the extent to which job information was used by interviewers as a criterion for their decisions. In Exp I 48 experienced and 48 unexperienced job interviewers received identical job information. Applicant information was either relevant or irrelevant to the job information. Ss' judgments were based on composites of relevant and irrelevant information segments. Each of the segments was either favorable or unfavorable. Exp II replicated Exp I, except that the 48 Ss did not receive job information. Findings indicate extensive use of job information. Availability of such information reduced the effect of irrelevant attributes on decisions but did not eliminate it. Use of job information was not enhanced by experience in interviewing. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Integration of acoustical information in the perception of impacted sound sources: The role of information accuracy and exploitability.

Sound sources are perceived by integrating information from multiple acoustical features. The factors influencing the integration of information are largely unknown. We measured how the perceptual weighting of different features varies with the accuracy of information and with a listener’s ability to exploit it. Participants judged the hardness of two objects whose interaction generates an impact sound: a hammer and a sounding object. In a first discrimination experiment, trained listeners focused on the most accurate information, although with greater difficulty when perceiving the hammer. We inferred a limited exploitability for the most accurate hammer-hardness information. In a second rating experiment, listeners focused on the most accurate information only when estimating sounding-object hardness. In a third rating experiment, we synthesized sounds by independently manipulating source properties that covaried in Experiments 1 and 2: sounding-object hardness and impact properties. Sounding-object hardness perception relied on the most accurate acoustical information, whereas impact-properties influenced more strongly hammer hardness perception. Overall, perceptual weight increased with the accuracy of acoustical information, although information that was not easily exploited was perceptually secondary, even if accurate. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Automatic memory search and the effects of information load and irrelevant information.

If automaticity is identified with both light capacity usage and little attentional selectivity, then an automatic process should be symptomatized by both small effects of concurrent information load and large susceptibility to interference by similar but irrelevant information. Several experiments are reported that test this prediction for memory search among sets of words. For well-learned memory sets, a small effect of information load co-occurs with a large effect of irrelevant information. By contrast, for arbitrary sets, a large effect of information load co-occurs with a small effect of irrelevant information. These results do confirm the correlation between effects of information load and irrelevant information as a hallmark for identifying automatic processes. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Ubiquitous halo.

Users of rating data have long been concerned about the "halo effect" (high intercategory correlations or low intercategory variance). Prominent sources of this ubiquitous effect are the true intercorrelation of categories and the illusory theories that raters hold about the extent to which categories covary. Cognitive distortions that increase the survival of illusory covariance theories are identified, including failure to adequately attend to hit rates, the differential ease of making "same" or "similar" vs "different" judgments, confirmatory bias in hypothesis testing, and the discounting of impression-inconsistent information. Nine methods (e.g., rating irrelevant categories) currently used to reduce illusory halo are reviewed, and it is concluded that increasing the sample of the ratees' current behavior is the most effective method in use. However, all methods probably leave residual illusory halo. A model of the rating process is described that attempts to direct research toward increasing understanding of the observation, encoding, storage, retrieval, and evaluation sequence, and to illuminate the paradoxical low positive correlation between halo and accuracy reported in 4 recent studies. (169 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Sources of bias in the Goodman–Kruskal gamma coefficient measure of association: Implications for studies of metacognitive processes.

In many cognitive, metacognitive, and perceptual tasks, measurement of performance or prediction accuracy may be influenced by response bias. Signal detection theory provides a means of assessing discrimination accuracy independent of such bias, but its application crucially depends on distributional assumptions. The Goodman–Kruskal gamma coefficient, G, has been proposed as an alternative means of measuring accuracy that is free of distributional assumptions. This measure is widely used with tasks that assess metamemory or metacognition performance. The authors demonstrate that the empirically determined value of G systematically deviates from its actual value under realistic conditions. A distribution-specific variant of G, called Gc, is introduced to show why this bias arises. The findings imply that caution is needed when using G as a measure of accuracy, and alternative measures are recommended. (PsycINFO Database Record (c) 2010 APA, all rights reserved)