To examine the role of phonetic code in memory, a study investigated the use of phonetic recoding strategies in fifth and sixth graders identified as good and poor readers. Subjects--120 students with IQ scores in the average range--were divided evenly between both grades. Thirty subjects from each grade who scored above the 50th percentile on the California Achievement Test were selected as good readers, and 30 subjects from each grade who scored below the 50th percentile were designated as poor readers. Students were individually tested in two separate sessions separated by one week. For the visual input condition, students were shown 28 monosyllabic words one at a time, asked to say each word aloud, and told to remember each word shown. A second set of words was presented one at a time; students replied "yes" if they thought the word was on the previous list, or "no" if they thought the word was not on the previous list. For auditory presentation the initial set of words was presented verbally and the recognition list presented visually. Word pairs were classified as phonetically similar if they shared the same vowel sound and differed by no more than three consonantal phonetic features. Results indicated that good and poor readers made equivalent numbers of recognition errors on phonetically similar and dissimilar items. Groups could not be distinguished by their use of a phonetic code in working memory. (One table is included, and 31 references are appended.) (MM)