Files in this item
|(no description provided)|
|Title:||Magnitude Recoding in Variable-Range Memorial Comparisons|
|Author(s):||Cech, Claude Gabriel|
|Department / Program:||Psychology|
|Degree Granting Institution:||University of Illinois at Urbana-Champaign|
|Abstract:||Several experiments examine propositional and analog model explanations of how relative size judgments are made. Analog models generally claim that analog size values are retrieved from memory and compared. Time to judge remembered size thus reflects a memorial pyschophysics. In contrast, one propositional model, the discrete code model, claims that size codes or features are compared, and that time to judge remembered size reflects feature-matching operations. It is argued that efficiency considerations determine how codes are generated from the data base. Consequently, propositional models predict that coded size will vary as a function of magnitude code efficiency in given problem-solving contexts.
Initial experiments require subjects to compare remembered sizes of pairs of animals. Subsequently, subjects compare only the small (FLEA to BEAVER) or large (SHEEP to ELEPHANT) animals. The discrete code model claims that magnitude codes in the full-range block are more efficient than the same codes in the restricted-range blocks (where significantly more comparisons involve animals with codes that do not differ initially, requiring a stage of code disambiguation). The recoding hypothesis predicts that size will be recoded to make comparisons more efficient. In accord with predictions, results indicate that subjects treat the larger small animals and the smaller large animals as large and small animals, respectively.
Further experiments demonstrate that likelihood of recoding depends on how efficient codes are when the opportunity to recode occurs. Moreover, it is shown that recoding depends both on the range of magnitudes encountered and on specific magnitude pairings. When subjects receive intermingled comparisons of just small or just large animals throughout the experiment, separate categories of animals (large versus small) are established, each recoded into large and small sub-halves.
Analog models that might predict such changes in results purportedly indicative of change in coded size are examined and found wanting. Nevertheless, analog accounts could claim that recoding in these experiments reflects knowledge of the small number of experimental stimuli used. Hence, a final experiment demonstrates recoding when no stimulus is presented more than once.
The implications of these results for models of memorial magnitude representation in comparative judgment experiments are discussed briefly.
Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 1981.
|Date Available in IDEALS:||2014-12-13|