Several studies employing the "20 questions" test reported that reflective children asked proportionally more constraint-seeking (CS) questions than did impulsive children. The finding was interpreted as supporting the generally accepted view that reflective children are more mature and more efficient problem solvers than impulsive children. Inexplicably, the reflectives' higher proportion of CS questions was not associated with fewer questions to solution, that is, with greater efficiency. The present study examined an alternative hypothesis that differences in performance of impulsive and reflective children on the "20 questions" test are due to individual differences in preferred perceptual processing strategy rather than in cognitive maturity of problem-solving strategy. Efficiency of performance of reflective and impulsive children was shown to be related to type of stimuli and experimental conditions employed, supporting a perceptual rather than cognitive-maturational interpretation.