Barry Anderson

Date of Award


Document Type


Degree Name

Master of Science (M.S.) in Psychology



Physical Description

1 online resource (125 p.)


Psychology, Experiments, Memory, Classification




Memory for crossed and nested classifications was investigated. Two experimental groups were exposed to stimuli which could be organized by both a crossed and nested classification. The stimuli consisted of nine drawings in a 3 x 3 matrix. Each drawing is characterized by attributes on five dimensions. The nested classification requires four dimensions to organize the nine drawings, while the crossed classification requires two dimensions. Of the five dimensions, three are unique to the nested classification, one is unique to the crossed classification, and one is common to both classifications. Subjects were presented the stimuli so that either the crossed or nested aspects were emphasized. This emphasis was accomplished by manipulating both the temporal order of rehearsal and the physical format, which were confounded in this experiment. Both nested and crossed groups first rehearsed the common dimension. The crossed group secondly rehearsed the other crossed dimension and thirdly the three nested dimensions. The nested group secondly rehearsed the three nested dimensions, then thirdly the crossed dimension. Also, the physical format for the two groups differed by having different lines separating the drawings of their sets. Three measures were taken during the recall of the sets. The temporal order of recall of the stimulus attributes was noted. Errors were counted when the attributes of any dimension were recalled in the wrong structural location, or were not recalled at all. Clustering was measured for each recall trial. Clustering is the tendency to recall all of one dimension, then all of another dimension, etc., rather than mix dimensions together. It was hypothesized (1) that subjects in the nested group would make fewer errors in the recall of the stimuli than would the crossed subjects; (2) that the crossed group would cluster in their recall to a greater extent than the nested group, and (3) that the two presentation methods would induce the subjects to record in long-term memory their respective structures better than the structure of the other group. Twenty four Portland State University undergraduates participated in the study. The error hypothesis (1) was not confirmed. For short-term memory, both groups made the same number of errors. The clustering hypothesis (2) was not confirmed. The induction of structure hypothesis (3) was partially confirmed. The nested group recalled the nesting with fewer errors than the cross. The crossed group recalled both structures equally poorly. A learning curve hypothesis (4) was not confirmed. A subject expectation effect hypothesis (5) was partially confirmed. Subjects in both groups increased their errors when they tried to rearrange the material in their memory in certain ways before recall. Several post-hoc analyses were performed. For short-term recall, errors were correlated with absolute rearrangement. Absolute rearrangement measures the difference between dimensional position in the presentation and recall sequences. The position change of the dimensions unique to the nested classification was found to correlate significantly with errors for both groups. The experimental hypotheses and post-hoc analyses are consistent with the hypothesis that the nested classification tends to override the cross classification when the two are in direct competition. Overall, the experimental results provide support for the hypothesis that the experimental procedure can induce the memory of nested classifications far better than the memory of cross classifications.


If you are the rightful copyright holder of this dissertation or thesis and wish to have it removed from the Open Access Collection, please submit a request to and include clear identification of the work, preferably with URL

Persistent Identifier