Processes on the Borderline Between Cognitive Abilities and Personality: Confidence and its Realism

Theoretical and Historical Background of Research on Confidence

There are two traditions in psychological studies of confidence. One tradition treats confidence as a personality trait and employs a typical format for the self-assessment of such traits. For example, ‘assertiveness’ and ‘bold and bashful’ aspects of the extroversion dimension include features of self-confidence (e.g. McCrae and Costa, 1990). Although we have used these and other related scales in our work, confidence-as-personality trait will [Page 546]be treated as a marginal topic in this chapter. This is because empirical evidence suggests that such personality measures do not correlate to any substantial degree with our own procedures for assessing confidence (Kleitman et al., 2003; Pallier et al., 2002).

The second tradition of research on confidence has a long history in psychology that is inextricably linked to well-defined cognitive activities, typically in providing an answer to a test item. There are three distinct streams in this tradition. Psychophysical studies of confidence started with the work of Fullerton and J.M. Cattell (1892). Classical psychophysicists routinely collected three bits of information in their studies of threshold performance: accuracy, speed, and confidence. These three dependent measures provided relevant information for the interpretation of psychophysical functions. More recent research following this stream of work was reviewed by Vickers (1979) and Baranski and Petrusic (1999).

The second stream comes from psychologists in the area of decision-making. Their typical question is whether those who know more also know more about how much they know. ‘Know’ refers to accuracy, and ‘knowing how much they know’ relates to confidence (Lichtenstein and Fischoff, 1977). Two important theoretical approaches have been dominant in the study of confidence: the heuristics and biases approach (Kahneman et al., 1982) and the ecological approach (Gigerenzer et al., 1991). As we shall elaborate later, the heuristics and biases approach attributes the discrepancy between ‘knowing how much they know’ and ‘know’ to systematic personal tendencies. The ecological approach attributes this discrepancy to the characteristics of tasks that may attract the use of wrong cues in choosing the answer to a test question and adapting a wrong normative model.

...