Recent research into the reading process has been suggestive of three surface-level factors involved in proficient reading. Those factors are automatic word identification, contextual word identification, and text phrasing. Previous research into these factors has tended to treat the factors in isolation, thus making the tacit inference that the factor under study was the sole surface-level factor of importance and was not appreciably influenced by other co-occurring factors.
The purpose of the present study was to verify the validity of these factors as contributors to reading proficiency, to test a model of reading that employed the factors in a simultaneous and/or interactive fashion, and to identify any developmental changes in the model from third- to fifth-grades.
The subjects for the study were students from two public schools who were asked to complete a variety of reading tasks. Each task was a measure of one of the three independent factors or of a comprehension factor. Multiple regression techniques were used to determine the contributions of each independent variable to comprehension. A model employing the three independent factors, a comprehension factor, and the various measures of these factors was hypothesized for each grade and tested. In the model the effects of automatic and contextual word identification were mediated through phrasing in affecting comprehension. The model was tested using covariance structure analysis (i.e. LISREL).
The results of the analysis supported the validity of the three factors as important variables of surface-level behavior in the reading process. Moreover, the hypothesized model of reading was found to fit the accumulated data reasonably well. Alternative models, employing the same variables in different combinations were found to be less reasonable on the basis of overall fit, amount of variance in comprehension explained and/or the models' degree of parsimony. Inspection of the models for third- and fifth-grades permitted a developmental perspective. The most significant developmental effect was the relative increase in importance of context versus automaticity as the grade levels increased. Explanations for the effects are offered and implications for instructional practice and future research are discussed.