The competent academic problem solver : toward an integrated model

Abstract

Clinical and theoretical models based on the assumption that young school-age children's private speech guides and directs ongoing motor behavior have met with growing enthusiasm. The empirical basis for the assumption has not been naturalistically validated, particularly as it is measured by the differential employment of speech and motor behavior by children of varying competence in an academically related task. The present study investigated the interface of two streams of behavior, verbal and motor, with 25 first-grade children (11 female and 14 male). Videotapes of the children individually completing an easy and a hard version of a fill-in-the-blank reading and writing assignment developed specifically for this study were analyzed. Several innovative techniques, including separate coding systems for speech and motor behavior were used. Three criteria were used to distinguish competent children from the less competent: 1. scores on standard tests for I.Q. such as the WISC-R, and reading competence as measured by the Gates-MacGinnitie Reading Test (vocabulary and comprehension subtests); 2. task performance as measured by number correct; 3. personality dimensions as measured by reflection-impulsivity. Initial analyses included group differences and correlational techniques. I.Q., a general competence measure did not discriminate between high and low I.Q. children in task performance, verbal or motor activity. Scores on the Gates-MacGinnitie Reading Test, a measure more specific to the task, discriminated between better and poorer readers, with better readers completing more trials and answering more trials correctly. During the easy task, the more competent readers worked more quickly with a larger percentage of time actually on-task. During the hard task, both groups worked equally as long and the more competent readers completed more trials correctly. Reading aloud was not significantly correlated with reading competence while task-irrelevant speech and task-directed private speech were significantly, negatively correlated with reading competence. Multivariate analyses indicated that motor behavior accounted for the greatest variance in both hard and easy tasks in predicting outcome. On the hard task, a measure of prior success, Gates-MacGinnitie Reading Vocabulary subscores, was also highly predictive. No significant differences were found in differential use of speech or motor behavior with respect to the reflective-impulsive dimension though the "fast-accurate" children were preponderantly Gates-MacGinnitie competent readers. Differential patterns of verbal and motor behavior were analyzed through the use of a conditional probability model for the joint occurrence of the two behavior streams; thus, showing patterns of behavior co-occurring at given points in time or sequences of events in a manner not possible with summary parametric statistics. Task-directed private speech was most likely when the child was off-task. Sequential conditional analyses, viewing behavior "X" as a cue for behavior "Y", indicated that few behaviors acted as consistent cues with the exception of reading aloud and looking-at-the-board which cued each other. These results imply: 1. At a clinical level, intervention programs in academic settings whose primary emphasis is the modification of private speech to strengthen deficits should concentrate on the behavior most predictive of success, namely, motor activity. Such motor activity includes necessary conditions such as looking-at-the-board, remaining on-task, not engaging in task-irrelevant behavior, and keeping quiet; 2. At a theoretical level, the role of private speech as a major element in the construct of impulsivity-reflection was questioned. Overt task-related private speech (other than reading aloud), by this age group, does not necessarily exercise functional, facilitory control of motor behavior; 3. At a methodological level, joint use of a parametric and conditional probability analyses increased the amount of information and power of statements generated from the data.