ON RELATIONS BETWEEN COGNITION AND READING STYLE: INDIVIDUAL DIFFERENCES AND DEVELOPMENTAL TRENDS
Abstract (summary)
To establish the validity of a novel representation of reading style, eye movements were monitored during detailed reading, proof reading, and skimming. Classifying eye movements as regressive, intra-word, word-to-word, word-skipping, and line-switch movements showed expected shifts in the relative frequencies of these movements as a function of reading condition. The observed frequency distributions of saccade lengths were used as input to a simulation which assumed no local control of eye movements. Computing the relative frequencies of the simulated word-related eye movements and comparing them to the observed frequencies identified instances of local control common to all and specific to one or two of the three reading conditions.
For more detailed analyses of relations between cognitive skills and reading style, the eye movements of disabled and normal readers were monitored during oral and silent reading. The subjects were also tests on word recognition, phonological skill, and WISC-R I.Q.) Individual differences in the cognitive variables within the two groups were related to eye movements through a structural equation model. Verbal comprehension comprising four of the WISC-R subtests and word recognition were uniquely related to two orthogonal dimensions of reading style based on the combination of word-skipping and regressive eye-movements. These relations held within groups of normal and disabled readers and in oral and silent reading.
These results were replicated with a larger sample of subjects who represented a wide age range (7.8 to 16.8 years). The structural relations of the previous model were estimated between latent instead of observed variables. The same structural model explained the normal readers' data. For the disabled readers verbal comprehension showed an influence on both dimensions of reading style in the oral reading condition.
The wide age range permitted an expansion of the model to explain the developmental relation between cognitive variables and reading style. The observed correlations were consistent with a model that postulated only indirect influences of age on reading style via cognitive variables. Thus, developmental changes in reading style could be interpreted as epiphenomena of correlated developmental changes in cognitive variables.