Conducting Process Research
If we could actually hear ourselves think, studying reading and writing processes would be an easy and straightforward task. Of course we can't -- we can only listen for individual notes and sequences and try to pick up the melody. The process-tracing methodologies described in this volume share two important features: they aim to describe processes rather than products, and, in doing so, they do not tell the whole story or catch the whole tune. Working with process data, as with any other type of data, is an interpretive act: we make inferences about writers' goals, strategies, and understandings based on clues gathered from their notes, drafts, process logs, interview responses, protocol comments. To use the Hayes and Flower ( 1980) visual metaphor, we see the porpoise only when it breaks the surface of the water; from these glimpses we infer the path it follows below.
In the classroom and in other research settings, we enhance our chances of gathering rich and telling data by carefully controlling the circumstances under which we conduct our research and by being aware of the nature of those circumstances. We are careful not to generalize too far, for example, from studies in the lab or in the classroom, where particular sets of constraints (on time, audience, motivation) apply. And we are careful to clearly articulate the nature of our interest to the writers and readers we study. For example, when we collect "think-aloud" protocols, we want writers to verbalize their thoughts, plans, and strategies as they work, but we don't want them to step back and narrate those processes for us or to reflect on or interpret their actions. We