test of phoneme identities: Predicting alphabetic insight in prealphabetic readers, The

Journal of Literacy Research, Sep 2000 by Murray, Bruce A, Smith, Kimberly A, Murray, Geralyn G

Another phoneme awareness measure given during the initial session was the Test of Phonological Awareness (TOPA), a commercially published standardized test with established norms (Torgesen & Bryant, 1994). This test uses rows of illustrated words to help children remember words and make comparisons. The TO PA is a word-to-word matching test; the examiner does not explicitly isolate and pronounce the critical phoneme. In part one, children hear a target word and see its illustration, and then search through three other illustrated words to find a matching word with the same beginning phoneme. For example, given bat, they decide if horn, bed, or cup begins with the same phoneme. In part two, they search a set of four illustrated words (e.g., bed, bus, chair, and ball) for a word that begins with a phoneme different from the others.

As a screening test of decoding ability, we gave a test of phonetic cue reading. Phonetic cue reading is using the initial or boundary letters of a word to generate a partial pronunciation sufficient to identify a word from among activated words (Ehri, 1991). In this test, the examiner presents a word printed in capital letters (e.g., MAD), and asks which of two rhyming words it is (e.g.,"Is-this sad or mad?").

The complete test of phonetic cue reading is shown in Table 2. Children who scored beyond chance on this forced-choice measure, stipulated as lo or better out of 12 items, were excused from instruction during the second session of testing, because their performance indicated they probably had the alphabetic insight we intended to teach. The criterion of 10 of 12 correct on the test of phonetic cue reading is based on the probabilities of chance performance given a forced choice. The probability of getting 9 of 12 correct by chance is .0537, slightly above the liberal alpha standard of .05. The probability of getting lo of 12 correct by chance is about .02, a more conservative criterion. Most children (63 of the 97 participants) achieved this level of performance and were excused from the trial teaching.

We began the second session by giving the posttest version of the TPI to all participants. The posttest is an equivalent form of the TPI. Most items on the posttest version require the student to locate the same phoneme at the opposite end of the word; if the phoneme is presented in the initial position on the pretest item, the phoneme is presented in the word-final position on the posttest, and vice versa. The educational purpose of the posttest is to measure phoneme awareness after instruction. It allows an examiner to test for overall gains in phoneme awareness without reusing the same items, and it also provides a means of checking on whether particular phonemes (e.g., short vowels) have been identified.

Children who scored below 10 on the test of phonetic cue reading continued with a dynamic assessment of learning with testing interspersed among teaching episodes. They participated in a lesson about the phonemes represented by the letters M and S. Instruction followed a script adapted from the phoneme identification program employed in a previous study (Murray, 1998; the teaching script adapted for the present study may be found in the Appendix). The instructional script provided for 17 teaching and testing segments. After each teaching segment, the participants were tested for alphabetic insight. In each testing segment, a participant was shown a word printed in capital letters; each word began with M or S. The child was asked to identify the word given a rhyming pair of spoken words. For example, shown MOUTH, the participant was asked, "Is this south or mouth?" If the question was answered correctly, the next card was tendered, until a criterion of five consecutive correct responses was reached, in which case the lesson ended. If the student answered incorrectly on any item, the examiner resumed teaching with the next planned instructional segment. We assumed each test was equally difficult, because in each case the participant looked at a printed word beginning with M or S (e.g., MAP), in which M or S represented the entire onset, and was asked which of two words it might be ("Is this sap or map?). Each lesson segment was brief (see Appendix), and a test could involve from one to five words. Even so, fatigue was inevitable, and we limited the teaching to 17 instructional segments. When an item was failed on Test 17, the session ended.