Stankov, L., The University of Sydney, Australia
The analyses of performance on tests of cognitive abilities typically rely on measures of accuracy (and sometimes on speed) of performance. Frequently, however, in reaching a decision regarding a particular course of action individuals must act under conditions of uncertainty. In such situations, the choice depends on one's self-confidence that the choice is likely to be correct. In this paper, I focus on psychometric properties of measures of self-confidence that are derived from confidence ratings of participants' answers on typical cognitive ability test items. I will also report on recent findings with a related construct of self-monitoring. The results of from a dozen of studies to date indicate that self-confidence is a general dimension, independent of personality and ability traits. Reliable assessment of self-confidence can be achieved easily within the typical testing environment. Recent work also indicates that self- confidence may have good predictive validity for low-level jobs that involve dealing with the public and self-monitoring is a good predictor of speeding errors during performance in driving simulators. There are also recent indications that this trait may mediate between personality, vocational interests, and measures of the propensity towards cognitive engagements on one hand and performance on cognitive tests on the other. Given these psychometric properties, it may be useful to routinely use confidence ratings in large-scale testing programs and examine their predictive validity in a broader range of situations.