Two experiments investigated the effectiveness of reading backwards as a proofreading technique. Experiment 1 examined the general usefulness of backwards vs. forwards proofreading strategies and the effects of these strategies for two types of typographical errors: misspellings and contextual errors. Subjects, 36 undergraduate students at a California university, were asked to read four passages of 250 words containing 10 typographical errors distributed evenly throughout the passage. Results indicated that the backwards reading strategy was not shown to be an effective technique for proofreading, and, in fact, was actually inferior for detection of contextual errors. Experiment 2, using 18 undergraduates from another California university, examined the proofreading effectiveness of combining forwards and backwards reading by having subjects read each passage twice, sometimes using the backwards, sometimes the forwards strategy. Results indicated that the number of readings had a significant effect on error detection, but the type of strategy used did not. Major findings show that the backwards strategy was not shown by either experiment to be an effective proofreading strategy, and was actually inferior to forwards reading for detecting contextual errors, with no measurable benefits for detecting misspellings. A combination of reading forwards and backwards was found to be no more effective than reading forwards twice, and often required more time. (NKA)