Effect of the number of between-classes reject baseline relations on equivalence class formation

In this study, different training conditions involving six combinations of standard (STA) and altered (ALT) baseline trials were compared in regard to the probability of emergence of three 3-member stimulus classes. The STA and ALT baseline trials established the same within-class select relations, but the STA baseline trials established between-classes reject relations as well, while the ALT trials did not. The number of STA trials included in the baseline was related to higher likelihood of equivalence class formation; this relation, however, was not linear and simple, and rather it depended on the distribution of the STA trials, according with two dimensions in baseline training structure: the number of orthogonal relations (AB, AC) and the stimulus classes involved (Class 1, Class 2, and Class 3). A high probability of equivalence class formation was observed when between-classes reject relations were included in at least two stimulus classes, and in one of these for both orthogonal relations. It is concluded that for a six trials baseline between-classes reject relations must be involved in at least the two orthogonal relations and two classes, with a link between these dimensions, for a high probability of equivalence class formation to be ensured. Results are discussed in terms of an account emphasizing the role of sorting behavior on equivalence relations formation.