A readable writing approach to text revision was evaluated in 5 experiments with 1,169 17–20 yr old males. Revision involved use of word lists and the application of a syntax rule. It resulted in a reduction of 6 grade levels in the readability score. The revisions failed to produce practical improvements in comprehension on several reading-to-do tasks. Meaningful improvements in comprehension were achieved only for low-ability readers on a reading-to-learn task when vocabulary was simplified. Results are interpreted in terms of cognitive-processing requirements. The inadequacies of readability formulas and guidelines as a means of controlling the production of text are discussed, and an alternative, transformer approach is advocated. (36 ref)