Eliminative and Enumerative Behaviour in a Conceptual Task

Abstract

P. C. Wason's paper “On the Failure to Eliminate Hypotheses in a Conceptual Task” is criticized on the grounds that the task set is in important respects untypical of problem solving situations in general. It is suggested that few or no subjects are “Eliminators” in Wason's sense and that although many are “Enumerators” in his sense, enumerative behaviour is a function of the situation. An experiment is reported in which it appears that modification to the situation can sharply reduce the incidence of enumerative behaviour and increase the subject's chance of success at an early stage. The frequency with which hypotheses are directly eliminated is shown to be very low in all groups.

Hypothesis generation, sparse categories, and the positive test strategy.

“…This issue is reminiscent of early discussions of the rule-discovery task (Wetherick, 1962;Wason, 1962) in which a distinction was drawn between falsification of a hypothesis explicitly considered by a participant (falsification in H E ), and the kind of implicit falsification that occurs because the oracle's responses are inconsistent with a great many rules, regardless of whether the learner has explicitly considered them (falsification in H I ). The analysis in this paper treats both of these events as genuine falsifications, even though the learner would be unaware of implicit falsifications.…”

Exploring the determinants of dual goal facilitation in a rule discovery task

“…In particular, such coding is insensitive to participants' expectations about test outcomes (Poletiek, 1996); for example, a triple receiving positive feedback may have been generated with falsification in mind (Wetherick, 1962; see also Caverni & Rossi, 1997, for corroboration of Wetherick's proposal). In this paper we avoid labelling triples as either confirmatory or disconfirmatory, and instead follow others (e.g., Oaksford & Chater, 1994;Valle´e-Tourangeau et al, 1995) who simply code triples as either ''positive'' or ''negative'' (i.e., conforming or not conforming to the experimenter's rule).…”

“…Notwithstanding concerns that have been voiced in the literature over Wason's confirmation-bias account of the 2-4-6 task (Evans, 1989;Klayman & Ha, 1987), there is little doubt that the initial 2-4-6 triple provides a compelling lure that encourages formulation of overly narrow initial hypotheses (Wetherick, 1962). In a recent study, for example, Cherubini et al (2005) showed that participants derive initial hypotheses that retain the relational regularities that are salient in the 2-4-6 triple, since such regularities have high ''information relevance'' (see also Van der Henst et al, 2002, for further demonstrations of ''relevance effects'' arising with the 2-4-6 triple).…”

Contrast class cues and performance facilitation in a hypothesis-testing task: evidence for an iterative counterfactual model

“…Early criticism of Wason's (1960) "confirmation bias" account came from Wetherick (1962), who claimed that it was simply the experimenter's essentially devious provision of the 2-4-6 exemplar triple that lured participants into the production of the overly restricted hypotheses that are commonly announced. More recent support for Wetherick's proposal has come from research by Van der Henst, Rossi, and Schroyens (2002), who invoked "relevance theory" (e.g., Sperber, Cara & Girotto, 1995) to argue that participants expect that the information provided by the presented seed triple will be of value for solving the task.…”