Explicit and implicit processes in multicue judgment

Abstract

Expert judgment is often exercised in situations in which multiple pieces of information are available and relevant in varying degrees. For example, a doctor may be attempting to diagnose the likely illness of a patient on the basis of symptoms, medical history, clinical examination, and results of imperfectly diagnostic tests, as well as relevant demographic information, such as gender, age, and occupation. To take another example, a personnel manager might be judging the suitability of a candidate for a position, taking into account work experience, qualifications, interview performance, and results of psychometric tests. These are examples of what are termed multicue judgments, where the various dimensions of information are the cues, whose values have particular and distinct instantiations for individual cases.What makes multicue judgment of psychological interest, as well as of practical importance, is the potential for bias when such judgments are made intuitively. If experts lack self-insight into the processes underlying these judgments, they may be unconsciously biased. For example, it may be contrary to law or company policy for the personnel manager to take into account age or gender in his or her decision making. If no explicit algorithm is followed, however, we may have nothing better than self-report from the decision maker to establish which cues were actually used and with what relative weighting. For this reason, a tradition of research based on social judgment theory was developed some years ago to capture the tacit policies that decision makers use when making multicue judgments. Social judgment theory has its origins in the psychology of Egon Brunswick and is rooted in the ecological approach to human cognition-specifically, by relating the way in which the available information is modeled by the environment, on the one hand, and by the judge, on the other. For the history of the theory, the reader is referred to Doherty and Kurz (1996) and, for details of the associated lens model methodology, to Cooksey (1996).Social judgment theory uses multiple linear regression to measure policies that are implicit in the judgments that people make-a technique known as policy capture. In such studies, participants-or judges-are asked to make wholistic judgments about a series of cases in which information about the same set of multiple cues is available. The implicit policy can then be inferred by performing a multiple linear regression analysis from the cues onto the judgments. The resulting beta weights tell us which cues influenced the judge and to what relative extent. In some studies, an objective criterion of correct performance is also available. In the case of personnel selection decisions, for example, we may be able to assess job performance at a later date. In a full lens model analysis, a second multiple regression is performed from the cues onto the criterion. A simplified (ignoring nonlinear and nonadditive aspects) version of the lens model equation is then given as follows: achievement = ...

Learning strategies in amnesia

“…Reber et al, 1996), but also for individuals without neurological problems (e.g. Evans, Clibbens, Cattani, Harris, & Dennis, 2003;Gluck et al, 2002). In contrast, Lagnado et al (2006) found no evidence for a lack of task knowledge and self-insight.…”

Easy does it: The role of fluency in cue weighting

“…Although normative models propose that decisionmakers must weight cues by their objective validity, it has been shown time and again that this is a difficult undertaking (Evans, Clibbens, Cattani, Harris, & Dennis, 2003, Goodie & Crooks, 2004Permut, 1973;Peterson & Pitz, 1985;Tversky & Kahneman, 1974). In this paper, we have proposed that people will often weight information according to the ease with which it can be processed.…”

On the resolution of conflict in dual process theories of reasoning

“…It seems highly doubtful that these various accounts can all be mapped onto each other or onto just two underlying cognitive systems (Evans, 2006a). It may be, for example, that parallel-competitive models will turn out to be appropriate in domains such as inductive rule learning (Sloman, 1996), multicue judgement (Evans, Clibbens, Cattani, Harris, & Dennis, 2003), and social cognition (Smith & DeCoster, 2000). There are a number of different kinds of implicit processes, with an important distinction being between those that deliver content for conscious processing (the main function of my proposed heuristic system) and those that can influence behaviour directly, without ever reaching consciousness-for example, through implicit learning or the acquisition of automated cognitive skills.…”