COGNITIVE PREDICTORS OF WRITING ABILITY

This paper examines the utility of three measures of cognitive ability in predicting writing ability. A review of research in the area of writing is presented that considers a wide array of topics: (1) the development of writing skills, (2) an analysis of speech and writing modes, (3) behavioral approaches for improving writing, (4) educationally-oriented research, and (5) cognitively-oriented theories of writing. The author juxtaposes the constructive theory of memory with the Flower and Hayes model of writing in presenting a combined problem solving-schema theory framework for research in writing. Forty-three undergraduate educational psychology students volunteered for the study. All subjects wrote at least two essays on assigned topics for a set time period. Two independent raters scored each essay (interrater reliability r = .83) on a six-point scale. In addition, each subject participated in three information processing tasks presented on a TRS 80 Model III micro-computer. Task 1 was an adaptation of a memory scanning procedure developed by Sternberg (1966). Task 2 was an adaptation of the Sunday & Tuesday Task employed by Hunt, Lunneborg, and Lewis (1975). Task 3 was an anagram solving task. Correct number of responses were recorded for each subject on Tasks 1, 2, and 3, and elapsed time measures were recorded for Tasks 2 and 3. A composite score for each subject was also calculated for scores on all essays. The independent variables (measures on Tasks, 1, 2, and 3) and the dependent variable (composite writing score) were then entered into a multiple regression analysis. The results found a significant relationship among the predictor and criterion variables, with the significance resident in number of correct responses on Task 3. The authors referred to the sources of individual differences in cognition described by Hunt (1978) in attempting to explain the relationship between organizational skills in working memory and writing ability