Abstract
The effects of proofreading on the spelling accuracy of fifth-grade students were studied by having them read and detect errors in text containing misspelled words. The treatment involved an error detection task requiring a decision as to whether an underlined stimulus word embedded in a sentence was correctly spelled or misspelled. Serving as their own controls, the students were exposed to three spelling exposure frequency conditions during the proofreading treatment: (a) four exposures to a misspelling, (b) two exposures to a misspelling and two exposures to the correct spelling, and (c) four exposures to the correct spelling. The students proofread two types of spelling words in terms of sound-spelling correspondence; half were predictable and half were unpredictable. The major finding was that exposure to correctly spelled words did improve spelling accuracy for immediate and delayed posttests. Exposure to misspelled words did not significantly affect the spelling accuracy of the sample as a whole, but the accuracy of a few outliers was substantially impaired. Unpredictable words were found more difficult to spell than predictable words. No interaction was found between spelling ability and spelling accuracy improvement as the result of proofreading correctly spelled words; poor spellers improved as much as average and good spellers.