Shanahan, T. & Lomax, R. (1986). An Analysis and
Comparison of Theoretical Models of the Reading-Writing Relationship. Journal
of
Educational Psychology, 78 (2),116-123.
Introduction
This study compared and evaluated three theoretical models of the
relationship of learning to read and write.
The models were evaluated on the basis of their ability to account for
the relationships found in reading and writing data collected from second and
fifth graders. The three theoretical models were identical in the components of
reading and writing; however, how information was transferred across reading and
writing varies. The first model was
the interactive model where reading can influence writing and writing can
influence reading; the second model was the reading-to-writing model
which states that reading knowledge can influence writing but no writing can
influence reading; and the third model was the writing-to-reading model
where writing influences reading, but reading doesn't influence writing.
Method
This study involved 256 second-graders and 251 fifth-graders.
The measures used in this study were selected to increase the ability to
identify relationships between reading and writing at the two grade levels.
These tests provided equivalent measures of analogous reading and writing
skills. The measures were selected
to be similar, but appropriate, across grade levels. The reading components included word analysis, vocabulary,
and sentence and passage comprehension; the writing components included
spelling, vocabulary, sentence structure, and story organization. The second
graders completed the Phonetic Analysis Test of the Stanford Diagnostic Reading
Tests (SDRT); the Reading Comprehension Test of the Gates-MacGinitie Reading
Test (GMRT); the Vocabulary Test of the SDRT, and a cloze test which had
grade-level appropriate passages. The
fifth graders completed the Phonetic Analysis Test of the SDRT; the Reading
Comprehension Test of the SDRT; the Vocabulary Test of the GMRT; and a cloze
test. In addition, each
student was asked to complete a grade level appropriate spelling test that was
analyzed for accuracy in spelling, phonemic ability and visual ability. Also,
each student wrote two stories that were analyzed for the mean t-unit length,
vocabulary diversity, and grammatical features.
Results
The reading and writing scores were analyzed using the linear structural
relationship (LISREL) model, which consisted of a structural equation model that
described the theoretical relationships among unobservable, hypothetical latent
variables using linear equations and a measurement model that considered the
latent variables as observable indicator variables to measure error.
Students were tested on all of the variables for each grade.
The LISREL was used to evaluate the component models of the
reading-writing relationship.
In comparison of the three theoretical models of the reading-writing relationship, the interactive model was found to be most superior of all three models t the second and fifth grade levels. At both grade levels the reading-to-writing model, was in greater agreement with the data than was the writing-to-reading model. In all three models, the fifth graders had a greater correlation than the second graders. Although the interactive model was superior to the reading-to-writing model at the second grade, the two models fit the data approximately equally well at the fifth grade.
The results within the reading components of the interactive model, revealed that spelling did not influence reading vocabulary significantly at the second-grade level; syntax did not influence reading comprehension at the fifth-grade level; and reading comprehension did not influence story structure at the fifth-grade level. Word analysis knowledge influenced reading comprehension ability of the second graders compared to the fifth graders. Vocabulary knowledge in reading comprehension was more prevalent in the fifth graders compared to the second graders. All the results of the reading components of the interactive model were similar to the reading-to-writing model and writing-to-reading model; however the reading-to-writing model's reading component of word analysis had a greater influence on the vocabulary knowledge of fifth graders compared to the second graders.
Within the writing
components of all three models, spelling was related to the vocabulary diversity
of the fifth graders compared to the second graders. Vocabulary diversity had a
greater influence on the syntax of second graders compared to fifth graders.
Vocabulary diversity had a greater influence on story structure of the
fifth graders compared to the second graders.
Syntax had a greater influence on story structure at both grade levels;
however, this influence was only significantly different for the second graders
compared to fifth graders in the reading-to-writing model. The
reading-vocabulary/writing-vocabulary relationships were higher when reading was
learned before writing, which implies that the importance of reading is to
writing development. Also,
the influence of comprehension on syntax, suggests the importance of reading in
the writing process. The reading and writing relationship presented some
differences between the models. Within
the interactive models and reading-to-writing model, the word-analysis /spelling
relationship was strong at both grade levels.
Discussion
In conclusion, the interactive model provided the best description of the
reading-writing relationship with second and fifth graders.
The reading-to-writing model was superior to the writing-to-reading model
suggesting a greater amount of knowledge transfer from reading to writing.
From this finding, it was reasonable to infer that the traditional
approach, that is learning reading before writing was unnecessarily inefficient.
A disconnected approach to reading and writing fails to take advantage of
the knowledge sharing opportunities a student can experience when reading and
writing are integrated. Overall, the combination of reading and writing benefits
were apparent in this finding.
The limitations of this study was that it only examined the product or information variable. Future research into the overall reading-writing relationship needs to consider the process variables simultaneously with the product or knowledge measures. To maximize the measurements, the reading and writing components should be measured during the performance of identical or related task. If a writing model was established that included components that were not parallel to reading, the relationship may have been smaller than what was presented in this study. There is a need to develop more appropriate measurement methods for examining the relationship between reading and writing.
[Return to Additional References]