Date of Award

December 2022

Document Type


Degree Name

Master of Science (MS)



First Advisor

Adam Derenne


Adaptation-level theory has been used to explain how the size and shape of generalization gradients depend on the procedure used to study generalization. However, most of the empirical support for this theory has come from studies using a relatively simple stimulus dimension (e.g., lights that vary in brightness). In addition, much of the research has focused on one specific prediction of adaptation-level theory: that responding during a generalization test should shift towards intermediate values within the tested portion of the stimulus dimension. The present research used a more complex set of stimuli. Specifically, the stimulus dimension was based on bar length, but participants (192 undergraduates) assessed bar length indirectly by estimating the combined length of six parallel bars that were nonidentical in length. The research additionally examined adaptation-level theory predictions for how gradients are affected by the amount of discrimination training that participants receive, and the relative difficulty of the discrimination (S+ and S- were relatively similar in one condition and relatively dissimilar in another). Data collection occurred online.The research yielded orderly generalization gradients, but not the outcomes predicted by adaptation-level theory. Instead, the generalization tests showed such effects as a progressive decrease in the area under the generalization gradient, greater area under the gradient when the range of tested stimuli was wider (with this difference emerging more rapidly than adaptation-level theory would predict), and pronounced individual differences in the basic form of the generalization gradient. These results suggest the need for additional theories of stimulus generalization.