Effects of training directionality and class size on equivalence class formation by adults

Equivalence classes are more likely to emerge after comparison-as-node (CaN) instead of sample-as-node (SaN) training for pigeons, typically developing children, and individuals with mental retardation, but not for typically functioning older children or adults. Thus, differential effects of CaN and SaN training were directly correlated with a subject's biological/developmental status, and they were inversely correlated with a subject's behavioral repertoire at the start of training. By implication, CaN and SaN training should differentially influence the establishment of larger equivalence classes by adults if training is conducted with an inefficient procedure such as the simultaneous protocol. Two 5- or 7-member equivalence classes were established. For each class size, the baseline was formed by CaN or SaN training. A larger proportion of subjects showed the emergence of 7-member classes after CaN rather than after SaN training. No differences were found with 5-member classes. For adults, then, likelihood of equivalence class formation was an interactive function of CaN or SaN training and class size. In addition, likelihood of learning the baselines was an inverse function of class size but was not influenced by CaN or SaN training. The speed of learning baselines for the 5-member classes did not differ during CaN and SaN training. In contrast, the baselines for 7-member classes were learned more slowly during CaN than SaN training. Many of these results support the Saunders and Green (1999) discrimination analysis of equivalence class formation.