Cognitive dimensions of numerical rule induction.

Abstract

The cognitive determinants of number analogy performance were studied by systematically manipulating the processing demands imposed by the items. To explore sources of developmental differences in analogical reasoning, subjects were included from two age levels, upper elementary school and college. To allow the investigation of individual differences in reasoning ability, the elementary school children were selected to represent two general aptitude levels, average IQ and high 1Q. In each group a combination of process and content knowledge factors accounted for more than 70% of the variance in item solution difficulty. The most, critical processing demand affecting successful performance was the amount of solution-related information to be assembled and managed in working memory. Process and content-knowledge factors differentiated adults from children, whereas only specific, factual knowledge competencies seemed to differentiate between IQ levels in children.Efforts to understand the nature of academic intelligence (Neisser, 1976) have increasingly shifted to analyses of the cognitive determinants of intelligence-test performance (e.g., Hunt, 1978;Pellegrino & Glaser, 1979;Sternberg, 1979). Attempts have been made to specify models of performance on intelligence-test tasks and to use such models as the basis of analyses of age and skill differences. The present study is an attempt to understand academic intelligence more completely by examining process and content-knowledge requirements of perfor-This article is based on the first author's doctoral dissertation submitted to the University of Pittsburgh.

A cognitive analysis of number-series problems: Sources of individual differences in performance

“…Similarly, striking individual differences exist in manipulating numerical information (e.g., Hunter, 1977). Recognition of memorized series and calculation may well be domain-specific forms of knowledge, but they probably are not task-specific: both procedures may be applicable to a wide variety of tasks that require numerical skills (Holzman et al, 1982;Muth, 1984). In contrast, checking and self-termination are related to efficiency and may well be both task-and domain-independent, but neither characteristic appeared to be related to skill in Experiment 2.…”

Strategy Use in Reasoning Training With Older Adults

“…These pattern descriptions have been studied extensively in prior research on inductive reasoning (Holzman, Pelligrino, & Glaser, 1982;Kotovsky & Simon, 1973). The participants learned through modeling, feedback and practice procedures to identify the pattern and solve letter series problems involving the pattern.…”

Computer models solving intelligence test problems: Progress and implications

“…The latter involves variations with respect to reference (last number, second last number, third last number), linear vs. tree recursion (as for factorial vs. Fibonacci), and usage of one or more operators. These dimensions were identified from a psychological study of the difficulty of number series problems by Holzmann et al [145]. Although IGOR2 is an AI system, the authors propose that IGOR2 can also be viewed as a plausible cognitive approach to learning complex rules.…”