A previous study of recall of letter strings by good and poor beginning readers (Shankweiler, Liberman, Mark, Fowler, & Fischer, 1979) revealed that the performance of good readers was more severely penalized than that of poor readers when the letter names rhymed. To determine whether the differences in susceptibility to phonetic interference extend to materials that more closely resemble actual text, we designed an experiment to test recall of phonetically controlled sentences and word strings. As in the case of letter recall, we found that, although good readers made fewer errors than poor readers when sentences or word strings contained no rhyming words, they did not excel when the materials contained many rhyming words. In contrast to manipulations of phonetic content, systematic manipulations of meaningfulness and variations in syntactic structure did not differentially affect the two reading groups. We conclude that the poor readers' inferior recall of phonetically nonconfusable sentences, word strings, and letter strings reflects failure to make full use of phonetic coding in working memory.Much evidence suggests that adult subjects employ a phonetic representation during comprehension of both spoken and written material (see, for example, Baddeley, 1978;Kleiman, 1975;Levy, 1977;Liberman, Mattingly and Turvey, 1972;Tzeng, Hung, and Wang, 1977). In several studies of beginning readers, we and other investigators (Byrne & Shea, 1979;Mark, Shankweiler, & Liberman, 1977;Shankweiler, Liberman, Mark, Fowler, & Fischer, 1979) have found new support for the involvement of phonetic representation in the reading process: The ability to make effective use of phonetic representation appears to be correlated with success at learning to read.The possibility of associating children's reading ability with their use of phonetic representation was explored first by Liberman, Shankweiler, and their colleagues (Liberman, Shankweiler, Liberman, Fowler, & Fischer, 1977), who assessed the role of phonetic representation in letter string memory. Using a modification of Conrad's (1964) procedure, they asked good and poor readers in the second grade to recall a string of consonants in which the letter names eitherThe second and third authors are also at the University of Connecticut, Storrs. This work was partially supported by a grant from the National Institute of Child Health and Human Development (Grant HD-01994) to Haskins Laboratories and by a NICHD postdoctoral fellowship (Fellowship HD05677) to Virginia Mann. Thanks are due Robert Katz. who also aided in testing the reading ability and IQ. Bruno Repp and Leonard Katz provided helpful assistance with the statistical treatment of the data. We also wish to express our appreciation for the generous cooperation of the staff of the Parker Memorial School in Tolland, Connecticut. Special thanks are due the principal, Robert Lincoln; the three second-grade teachers, Judy Lombard, Diane Iadoresta, and Jean Medick; and the school secretary, Florence Peacock. We are also grateful to ...