Empirical Findings on the Accuracy of Intellectual Self-Perceptions Depend on the Method Used to Assess Self-Perceived Ability

Research on people’s tendencies to over-, under-, or accurately estimate their abilities relies on assessments of self-perceived ability. Here, we investigated how the choice of an assessment method influences the conclusions drawn in research on two perspectives on accuracy. First, studies using the rank-order perspective are concerned with the correlation between self-perceived ability and an objective measure of “real” ability across individuals. Second, studies using the level perspective view the discrepancy between a person’s levels of self-perceived and real ability as an indicator of the degree of bias (vs. accuracy) and investigate the average discrepancy across people, individual differences in people’s discrepancies, or the association between the individual discrepancies and some outcome variable. In a cross-sectional survey of 281 participants, we assessed self-perceived intelligence in two domains (vocabulary knowledge, reasoning ability) with three measures each (percentile rank, IQ score, number of correct items). “Real” ability was measured with objective tests and transformed to match the scales of the self-assessments. We found that the measure for self-perceived intelligence impacted the conclusions drawn in analyses within the level but not the rank-order perspective. We discuss implications for the robustness of past findings and for the assessment of self-perceived intelligence in future studies.