The present study sought to extend a recent study by Savage, Carless and Stuart, by looking at the pre-test phonological skills that predicted improvements in letter-sound knowledge and nonword reading. Results showed overall that phoneme manipulation predicted improvements in nonword reading and letter-sound knowledge even when pre-test scores on the respective dependent variables were controlled. Pre-test letter-sound knowledge was an independent predictor of nonsense word reading. Overall, onset-rime manipulation did not add to the prediction of outcome. Onset-rime manipulation predicted decoding improvements in a subgroup where children were exposed to both rime and phoneme-based teaching.