Substitution tests have a long history in psychology because of their simplicity of administration and their sensitivity to individual differences related to complex cognitive performance. Despite their widespread use there is no agreement on what the substitution test measures. The present study approached this question by applying a method of componential analysis introduced by Royer [Journal of Experimental Psychology 87 (1971) 335]. Symbols were altered on 20 symbol-digit substitution forms to permit the independent evaluation of three information-processing components: feature encoding, memory, and visual search. A test-retest paradigm demonstrated that the symbol manipulation reliably influenced performance in the predicted manner. A subsequent tachistoscopic visual search experiment demonstrated the validity of the characterization of participants as good or poor searchers on the basis of their symbol-digit performance. The study illustrates the utility of creating a multiform substitution test for evaluating individual differences in feature encoding, memory, and visual search capabilities.