Writing to Learn vs. Learning to Write
Mark Stellmack
There is a popular notion that by encouraging students to write, we can improve their thinking. This is not a new idea. Almost 30 years ago, Langer (1986) noted that "'Integrating reading and writing' and teaching 'writing across the curriculum' have become the educational parlance of the 1980s" (p. 400). She suggested that writing was used for very limited purposes in most classrooms and went on to note that:
"Common sense, personal experience, and educational lore all suggest, however, that writing can lead to extensive rethinking, revising, and reformulating of what one knows. It can make a person aware of what is known, what is unknown, and even what needs to be known. And yet, few studies have been undertaken to learn when people learn from writing, what different kinds of learning result from different kinds of writing experiences, or how writing can be used to help students understand and remember the material they read" (p. 400).
Langer then asked six students to read social studies passages and to perform one of three types of study tasks: completing short answer questions (i.e., fill in the blanks), taking notes about what they read, or writing an essay about the passage. Students scored higher on a test of knowledge about the readings after writing an essay in comparison to the other two study activities. The idea is that, when preparing an essay, students processed the information that they read more deeply and integrated the knowledge more fully than in the other two types of studying. This is consistent with what is known about memory and depth of processing in cognitive tasks. Note that Langer did not evaluate the quality of the writing. Instead, the writing was used as a tool for learning. Langer's subjects were "writing to learn".
In contrast, in our Research Methods course, students do a lot of writing, but not of the type described by Langer. Our students prepare a manuscript describing research that they designed and executed. The students are never tested on the specific content of their writing; rather, the writing itself is the end product. Langer's students were "writing to learn," but our students are "learning to write." The effects of this kind of writing are not evaluated with a test. Instead, the writing itself is evaluated.
It seems that an instructor should decide whether course writing assignments are intended to enhance learning of the course material (writing to learn) or if the writing itself is the goal of the activity (learning to write). Clearly, in the latter case, it is important to assess the writing for coherence and good mechanics, so assessment tools like grading rubrics and their reliability and validity are of concern. But with respect to writing to learn, is it necessary to write well in order to reap the benefits? Langer (1986) did not seem to be concerned with the quality of students' writing, merely with its secondary benefits to learning. Perhaps evaluating writing quality is not a concern in the "writing to learn" scenario. Perhaps if the instructor attempts to evaluate writing in that case, he or she is adding unnecessary work that does not support the primary goals of the assignment.
One question is whether "learing to write" produces measurable benefits beyond the assignment itself. As an attempt to answer this, I looked at the performance of 328 students in our Spring, 2014 and Fall, 2014 sections of Research Methods. In that course, students perform many activities for credit, including taking four objective exams and writing four papers as sections of an APA-style manuscript. There was a significant correlation between performance on the objective exams and the writing scores, r(326) = .507, p < .001. Is this evidence that learning to write causes students to do well on objective exams? No, of course not. There is a relationship but we cannot infer causation, in part because the writing took place concurrently with the objective exams.
Is success in "learning to write" activities associated with better performance in other academic contexts? To address this question, I looked at performance of the Research Methods students described above in our Statistics course. The correlation between writing scores in Research Methods and performance on Statistics exams was r(326) = .433, p < .001. However, the Statistics course preceded the Research Methods course. Based on this, one could argue (probably unsuccessfully) that teaching students Statistics leads to better writing in Research Methods! And it turns out that the strongest correlation of all was between performance in the Statistics class and on the objective Research Methods exams, r(326) = .713, p < .001. Maybe "Statistics Across the Curriculum" is the way to go.
A more plausible interpretation of these data is that there is a general characteristic of students that leads to generally good academic performance in many different contexts, including writing. Good students write well. Students may learn to write better in their research methods class in that they acquire knowledge about proper scientific writing and formatting, but students who are generally better students tend to produce better writing. There is no evidence, nor any reason to expect, that learning to write will show generalizable benefits to other contexts.