This proposal is seeking continued support for a program of research involving relational learning in mentally retarded and nonretarded individuals, which has been funded by NICHD for 3 years. The specific aim of the present investigators is to analyze and enhance the relational characteristics of stimulus arrays in order to facilitate performance on tasks such as oddity and match-to-sample. Both mentally retarded and young nonretarded children are particularly prone to failure on these relational tasks (Green et al., 1990; Soraci, Deckner, Baumeister, & Carlin, 1990). Converging evidence from various studies suggests that a critical factor in the performance discrepancies between retarded and nonretarded children is a differential sensitivity to relational information. We are in the formative stages in the development of a general theory of relational learning that would have implications for an understanding of the locus of performance differences across varying levels of intellectual functioning (e.g., Soraci, Baumeister, & Carlin, in press). In the proposed studies, emphasis will be placed on the role of the structures of visual arrays in facilitating detection of relevant stimulus relations as a "front-end" (Brown, Collins, & Deguid, 1989) intervention strategy. The proposed studies are divided into two sections: (a) Global Stimulus Array Organization: The Oddity Paradigm, and (b) Stimulus-Specific Organization: The Detection Paradigm.