Confusion errors in serial learning
Robert Marc Schwartz · cIRcle (University of British Columbia) · 2011
Two experiments examined the relative ability of different probes in eliciting items occupying various positions in a serial list, the shape of the serial position curve, and the distribution of errors in serial learning. In Experiment I, position, sequential, and backward probes were employed after one presentation of a serial list. The results indicated that all three probes had equal eliciting strengths and that all three probes produced similar serial position curves. Analysis of errors revealed a gradient of generalization around the target word. In Experiment II error distribution and the serial position curve were examined in a learned serial list using the missing scan as a probe. Analysis of correct responding showed flat serial position curves, and the distribution of errors around the target word seemed essentially random. The results were discussed in terms of current theoretical representations of serial learning.