Date of Award
Doctor of Philosophy (PhD)
Dr. Linda S. Siegel
Classification, the recognition, construction, and ordering of classes, is a basic part of thought process. One aspect of classification, class inclusion, requires the quantitative comparison of a superordinate class and its included subclass. According to Piagetian theory, class inclusion is the definitive test of classification competence because inclusion ability demonstrates understanding of the relationships between different hierarchical levels of classes.
Consideration of the relationship of language and thought leads to critical examination of the role of language in assessment of classification skills. To the extent that classification ability precedes or is independent of language development, procedures which rely on linguistic comprehension and production will yield results which are incomplete and possibly invalid. If we vary language in assessment, we should obtain a better understanding of early classification strategies and ability.
This research on class inclusion treats language as an independent variable in two days, by examining alternate verbal forms of the inclusion question, and by reducing language in that question. Within the Piagetian research tradition there is only one classification study with reduced language, and there is one recent experiment which compared an alternate verbal form to the standard Piagetian question for class inclusion. Several experiments are reported here which examined the effects of language variation on class inclusion performance.
In the first alternate form of the inclusion question the word "all" modified the superordinate class, and children gave more correct answers in this case than with the standard unmodified question. Further examination indicated that children performed better with the "all" question because they understood that question better than the standard question.
With the second alternate inclusion form, Siegal's question, children were asked whether they wanted the superordinate class or the subclass. This was an attempt to replicate a previously reported finding that young children gave more correct answers to this form than to the Piagetian question. In the work reported here, after the children answered they were told to take their choice, and when analysis included both this behavioural measure and verbal responses, there were no differences in performance on the standard versions of Siegel's and Piaget's questions. This discrepancey with the earlier finding is discussed in terms of methodological differences between the studies. In contrast to the standard version results, when the word "all" was used in both questions the performance differences obtained were consistent with the previous finding: younger children gave more correct answers to Siegel's than to Piaget's question.
Language in the inclusion question was reduced by using visual symbols to identify the subclasses and the superordinate class. Four-year-olds gave more correct answers to the reduced language question than they did to the standard question in which words identify the classes.
A mathematical model is presented which estimates underlying components of correct and error response to the inclusion question, allowing anlaysis of the strategies that children use with different forms of the inclusion question. For example, the four-year-olds did not do better with the reduced language question because they were demonstrating more inclusion logic with that form. The difference in results was based on use of different non-inclusion strategies: with the reduced language question the children were more likely to guess, and with the standard question they were more likely to compare the subclasses to each other. Use of the mathematical model was critical in preventing false interpretation of the results.
The methodological and analytic implications for class inclusion research are discussed. In general, these studies demonstrate the usefulness of alternate approaches to inquiry, analysis, and interpretation of the existing data concerning the development of thought processes.
Hodkin, Barbara, "Language Effects in Assessment of Class Inclusion Ability" (1978). Open Access Dissertations and Theses. Paper 688.