The National Assessment of Educational Progress showed that many elementary and secondary students are incompetent writers (Applebee, Langer, & Mullins, 1986). If research is to have a positive impact on writing instruction and achievement, which to date it has not, it will be by increasing our understanding of the processes that underlie writing skill and their development. Three promising theoretical models (Hayes & Flower, 1980; Bereiter, Burtis, & Scardamalia, 1986) specify global writing processes and their development, but the models lack procedural specificity and have been tested almost exclusively with verbal protocol analysis and naturalistic observation, which are scientifically weak and instructionally uninformative. The research proposed here will [1] unpack and test the psychological reality of procedural details hypothesized to underlie the global processes outlined in cognitive theories of writing and [2] test for individual and age-related differences in those processes. Eleven proposed experiments focus on specifics of three global determinants of writing skill-[1] planning, [2] translating, and [3] reviewing-and on [4] the effects of theory-based instruction on writing. The research plan, which will be executed by a team of five collaborating scientists, is theory driven, and it uses converging operations (Garner, Hake, & Ericksen, 1956), individual differences tests (Underwood, 1976), and instructional methods (Belmont & Butterfield, 1977) to maximize the scientific and educational returns of the research. Likely returns are more detailed and valid theory of processes underlying writing, and more effective methods for increasing children's writing skill.
