Academic journal article The Psychological Record

Eye-Movements, Training Structures, and Stimulus Equivalence Class Formation

Academic journal article The Psychological Record

Eye-Movements, Training Structures, and Stimulus Equivalence Class Formation

Article excerpt

After human participants learn conditional discriminations, they usually form equivalence classes, which are defined by the properties of reflexivity, symmetry, and transitivity (Sidman & Tailby, 1982). These conditional discriminations could be arranged in an MTS task where A stimuli are presented as sample stimuli and comparisons as B stimuli (AB relations), and sample as B stimuli and comparisons as C stimuli (BC relations). Reflexivity is a conditional relation which entails that a stimulus has a relation to itself. This is tested by presenting stimulus Al as a sample, and then presenting stimuli Al, A2, and A3 as comparisons. The reflexive condition is true if each A stimulus is matched to itself without any programmed consequences. The symmetry relation entails that the conditional relation is bidirectional. After training to match stimulus A to stimulus B (AB) and stimulus B to stimulus C (BC), symmetry is evidence if B is matched to A (BA) and C is matched to B (CB) without any direct reinforcement. Following training of the AB and AC relations, transitivity is evident if A is matched to C (AC) without reinforcement. Finally, a global equivalence test, which is a symmetry and transitivity test, assess whether C is matched to A (CA) (Sidman, 1992, 1994).

The term training structure has been used to describe how different conditional discriminations are sequenced during training (e.g., R. R. Saunders & Green, 1999). Three training structures (linear series [LS], many-to-one [MTO], and one-to-many [OTM], have been used for establishing the prerequisites for testing the emergence of equivalence classes (see Fig. 1 with an example of training of 12 conditional discriminations). In the present study, participants were trained on 12 conditional discriminations and tested for the formation of three five-member classes. For LS, participants were trained each relation in a sequence: A [right arrow] B [right arrow] C [right arrow] D [right arrow] E. For MTO, participants were trained such that many samples were trained to one comparison: A [right arrow] E, B [right arrow] E, C [right arrow] E, and D [right arrow] E. For OTM, participants were trained such that one sample was trained to many comparisons: A [right arrow] B, A [right arrow] C, A [right arrow] D, and A [right arrow] E (see Fig. 1). In LS, the A and E stimuli serve only as samples or comparisons, respectively, whereas the B, C, and D stimuli serve as both samples and comparisons. During training in both MTO and OTM, no stimulus serves more than one function (i.e., each stimulus is only seen as a sample or as a comparison). The number of conditional discriminations will increase for all three training structures as a function of class size. Hence, by increasing the class size (number of members) the number of nodes will increase for the LS while it will always be one node-classes only for MTO and OTM (see Fig. 1). As a series of experiments have been done to study equivalence class formation as function of the different training structures (e.g., Arntzen, Grondahl, & Eilifsen, 2010; Arntzen & Hansen, 2011; Arntzen & Holth, 1997), it has been assumed that outcomes on equivalence tests should be the same for the three training structures. However, inconsistencies have been reported, and some of the differences could be related to procedural variables (see Arntzen, 2012, for a discussion).

R. R. Saunders and Green (1999) discrimination analysis provides a parsimonious account of why different training structures can result in different outcomes on tests for stimulus equivalence. This account is based upon the number of simultaneous and successive simple discriminations required during training. The discrimination analysis holds that each training structure requires a different number of simultaneous and successive simple discriminations during training. One basic assumption in the discrimination analysis is that simultaneous discrimination is easier than successive discrimination. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.