Investigation of the effect of a task-oriented learning group approach to writing on attitude and skill improvement, as compared to a traditional one-teacher, whole-class design, was the purpose of this study. Fifty high school English students were administered baseline tests of writing skill. An IQ score was also obtained, and the subjects were randomly divided into two groups. The experimental group received instruction including assessment by the writer at the prewriting, writing, and evaluation stages, as well as tasks designed to encourage pooling of information, division of responsibilities, and observation of others' response. The control group received traditional instruction, and at the end of the term, all subjects were retested with alternate evaluation forms and trained rater judgments of actual writing samples. Although the experimental group showed larger mean gains in achievement and attitude improvement than the control group, and a greater mean achievement score than present in normative baselines, no differential effect on writing skills or student attitudes could be proven by the data. The task-oriented design was recommended as a viable alternative to the traditional classroom approach. (Author/KS)