The purpose of this paper is to compare topical structure analysis (TSA) and a checklist approach　in terms of the effects on EFL writers' macro level revision. Two classes of Basic English Composition, which consisted of forty-eight Japanese university students, were asked to revise their written texts. One class was taught to how to perform TSA and the other class was given a checklist of macro level revision and taught how to use the list.
The revised texts were analyzed in two respects: (1) holistic evaluation and (2) increase or decrease of T-units and T-unit movements. The result of holistic evaluation showed that the class using TSA performed better than the class with the checklist approach. In particular, TSA produced better effects for the lower proficiency group. Analysis concerning T-units clarified that the TSA group n10ved T-units more vigorously than the checklist group, which led to better organization of the revised texts.
These results lead to the conclusion that the checklist approach doesn't show the writers systematic connections an10ng detection, diagnosis and fixing concerning text problems. In contrast, drawing TSA diagrams guides the writers from detecting to diagnosing and to fixing of the problems.