To determine optimal and least effective response modes of learning disabled children, the procedure described by Owen, Braggio, and Ellen (1976) and the Illinois Test of Psycholinguistic Abilities were used. On the basis of pretest and posttest comparisons on selected subtests, the performance of these children could be described as individual or multiple combinations of covert-, vocal-, or visual/manual-response modes. This learning style analysis was validated by presenting a paired-associate task to the children using an administration similar to (matched condition) or different from (unmatched condition) the child's optimal response mode. More items were recalled under the matched condition than under the unmatched condition. Also, the children with multiple response modes recalled more items than the learning disabled children with single optimal response modes. These results suggest that disabled children may do poorly on academic tasks because they may not have enough optimal response modes, or are unable to select the most appropriate mode of responding as required by a task. The use of optimal response modes as an aid in the diagnosis and remediation of educational handicaps is discussed.