Noting that while the number of errors gives a general indication of spelling ability, it may represent an oversimplification of the spelling process, a study examined the spelling ability of 40 Canadian third grade students of average ability. Subjects were divided into low and high spelling ability groups on the basis of the Edmonton (Alberta) Public Schools Spelling Achievement Test. Subjects completed achievement tests, cognitive processing tasks, and a computerized spelling test, which consisted of 30 words selected on the basis of familiarity and predictability. After hearing each word orally and used in a sentence, subjects spelled the word on the computer, then rated their familiarity with the word and perceived spelling of the word. Scores on the spelling tests were analyzed as a function of spelling ability, general cognitive processing, and more global spelling factors. Low versus high ability group analyses did not demonstrate differences in general cognitive processing. Group analyses did, however, reveal differences in patterns of correct spelling and phonetic errors as a function of word familiarity and predictability. Regression analyses also revealed that correct spelling by high ability spellers was best predicted by variables measuring use of phonetic and sequential segmentation strategies. Correct spelling by low ability spellers was best represented by memory for whole words and for letter sequences. (HTH)