Date of Graduation


Document Type


Degree Type



Eberly College of Arts and Sciences



Committee Chair

Philip N. Chase.


Stability in responding to stimuli within equivalence classes has implications for the maintenance of classes and the ease at which they can be reformed. One way of investigating stability is to examine accuracy and speed of responding as a function of nodal number, or the number of nodes between stimuli in a class. Previous research suggests that subjects respond more quickly and more accurately on relations involving fewer nodes (Fields, Adams, & Verhave, 1989 [May]; Fields, Adams, Verhave, & Newman, 1990; Fields, Adams, & Verhave, 1993; Fields, Landon-Jiminez, Buffington, & Adams, 1995; Spencer & Chase, 1996). A second way to investigate stability is to compare accuracy and speed of responding as a function of types of relations: trials that test baseline, symmetry, transitivity, and combined symmetry and transitivity relations. Research has shown that subjects typically respond to baseline and symmetric relations faster than transitive and combined relations in tests for emergent relations. In the current research, tests were conducted after stable responding in accordance with equivalence relations had been established. Within-class preference tests were used to assess the effects of nodality and relation types on stability. A within-class preference test consists of match-to-sample trials with three or more class-consistent comparisons that occur after confirmation of class formation. In the first experiment, subjects more frequently chose comparisons related to the sample via fewer nodes than those related via more nodes. In addition, subjects chose comparisons related to the sample via symmetry as often or more often than those related via trained baseline relations. Subjects also chose both symmetry and baseline more often than transitive and combined relations. Experiments 2 and 3 investigated the possibility that effects observed in the first experiment were due to the order of training and testing. The results of Experiment 2 were consistent with the results of Experiment 1. The order of testing in Experiment 3 revealed some differences. Performance on the nodal tests was more variable. In addition, only one subject demonstrated highly accurate and stable performance on tests for equivalence. In Experiments 2 and 3, the third comparison sometimes appeared to serve as a contextual stimulus for choosing between the other two comparison stimuli. Experiment 4 evaluated effects of a class-specific reinforcer arrangement during training on responding during post-class-formation within-class preference tests. The class-specific reinforcer arrangement increased stability on nodal-test responding. Relational test results were consistent with the previous three experiments. Results are discussed in terms of theoretical implications for the substitutability of stimuli in equivalence classes, and for application to education, particularly in learning languages and other complex curricula involving stimulus classes.