This study was designed to (a) assess the cognitive modifiability of preschool deaf and hearing children using dynamic and static-conventional instruments, and (b) compare the factorial pattern of the cognitive measures in both groups. The sample was composed of 26 deaf and 26 pair-matched hearing preschool children from 4·1 to 6·3 years of age. All of the children were given, individually, Raven's Coloured Progressive Matrices (RCPM), Children's Analogical Thinking Modifiability (CATM), the Picture Analogy Test (PAT), and the Analogies subtest from the Snijders-Oomen Nonverbal Intelligence Test (SON), and were initially matched on a motor integration scale. No significant differences were found between the two groups on RCPM, Analogies (SON), or PAT. On the CATM, a dynamic assessment instrument, the deaf children scored lower than the hearing children at pretest but improved their scores more than the hearing children when taught; both groups scored similarly after teaching. Results are interpreted according to the Structural Cognitive Modifiability and Mediated Learning Experience theory (Feuerstein, Rand, & Hoffman, 1979) and the unique communication barriers of deaf children.