People read for many different purposes, and they may read the same text differently depending on those purposes. One reading goal is to obtain information relevant to solving problems. We will be concerned with solving mathematical word problems, and the differences between such reading,and reading for """"""""normal' comprehension of gist. A major focus of the research is to study the underlying cognitive mechanisms (a) for reading, (b) for problem solving, (c) for the interactions between these two classes of processors. We propose a class of processing models, and their associated quantitative representations, for these 2 cognitive domains. The reading model has components for coding Lexical Structural and Meaning information. The problem solving model has components for Number Management, Planning Operations and Arithmetic Computation. The Number Management component has subprocessors for encoding numbers and their scale units, for retaining such information in active memory, and for reordering the information for use in planning the sequence of operations needed for solution. The Operations Planner has subprocesses for developing goals and subgoals to find a solution procedure, and for selecting out the critical information from active memory. The Arithmetic component performs numerical computations, is assumed to be independent of language processing during reading, and will not be studied. In the experiments to test these ideas, the subject fixates the center of a computer display, and his successive key-presses extinguish the preceding word, record its reading time, and display the next word. Before or after the main text, subjects view a YES- NO question about either its semantic or quantitative information. The stimuli are varied in terms of (a) linguistic variables (lexical, syntactic and semantic) and (b) quantitative variables (the number & type of operations (+,-,x, divided by), and the format (digits, words) and value of numbers). The reading time patterns over the sequence, and the response times and accuracy for the questions for the various experimental conditions will provide information relevant to the model's components and their interactions. The research provides a theoretical framework in which to understand the cognitive processing involved in reading and problem solving in normals, in people with poor reading or quantitative abilities, and for some subtypes of learning disabled, having developmental dyslexia and/or dyscalculia. Some of the computer display and active learning procedures may be useful for education and remediation.
| Aaronson, D; Watts, B (1987) Extensions of Grier's computational formulas for A' and B'' to below-chance performance. Psychol Bull 102:439-42 |
| Aaronson, D; Ferres, S (1986) Reading strategies for children and adults: a quantitative model. Psychol Rev 93:89-112 |
