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

To test the validity of the Test of Phoneme Identities (TPI), 97 kindergartners were administered the TPI and two other tests of phoneme awareness, measures of alphabet and nursery rhyme knowledge, and a measure of rudimentary decoding ability. The TPI proved reliable (a = .91) and comparable to the other phoneme awareness measures in predicting decoding ability in the full kindergarten sample. Thirty-four kindergartners identified as prealphabetic readers were taught to use the letters S and M as phonetic cues to distinguish rhyming printed words, an indicator of alphabetic insight. The TPI was shown to be more effective than the nursery rhyme and alphabet measures and the other phoneme awareness tests in predicting the number of lessons required for a student to learn to distinguish phonetic cues.

A RECENT SURVEY FOUND THAT PHONEME AWARENESS is currently the hottest topic in literacy education (Cassidy & Cassidy, 1998/1999), and for good reason. Simple measures of phoneme awareness predict success in reading acquisition more accurately than expensive IQ and reading readiness tests (Stanovich, 1993/1994), and they distinguish successful beginners from stragglers (Adams, 1990). Moreover, efforts to teach children to recognize phonemes have sometimes had dramatic effects in eliminating barriers to reading (e.g., Bradley & Bryant, 1985).

Oddly amid all this success, measuring phoneme awareness has proven surprisingly difficult. Phoneme awareness research has been hampered by researchere use of a wide range of instruments and paradigms (Stahl & Murray, 1994,1998; Yopp, 1988). Not all children understand how to respond to phoneme awareness tests. Whereas some answer effortlessly, relying on their spelling knowledge to finesse the tests, other children may cry bitter tears of frustration.

What is phoneme awareness, and how does it help children learn to read? Phonemes are best described as the vocal gestures from which words are constructed in a language (Liberman & Liberman, 1992). Because these vocal gestures are produced very rapidly (10 to 20 per second) and because they overlap with other phonemes in a syllable, they are difficult to identify. Phoneme awareness means recognition of these vocal gestures in their natural contexts, spoken words. Children who can recognize phonemes can understand the meaning of letters in an alphabetic writing system, for letters represent the phonemes of spoken words, and spellings map out the pronunciations of words at the phoneme level.

But how can we determine if children recognize phonemes in spoken words? Phoneme awareness has typically been operationalized as scores on various measures of phoneme manipulation, usually segmentation or blending tests. For example, we test segmentation by asking children to "break the word apart" and "tell me each sound in the word in order," and we test blending by asking them to "tell what word we would have if these sounds were put together" (Yopp, 1988,1995).

Manipulation tests presume a depth-chart model of phonological awareness (e.g.,Adams,1990) in which children progress from segmenting larger, more obvious speech units (words, syllables) to smaller, less obvious, finergrained units (onsets, rimes) until they are aware of the smallest, most abstract units: phonemes. In the depth-chart model, phonological awareness is viewed as a generalized sensitivity to the structure of spoken language, a skill that extends, in the final analysis, to any phoneme in a language. When we ask children to segment the phonemes in grew, we assume it does not matter that the particular phonemes /g/, /r/, and /ii/ constitute the word; we expect a generalized sensitivity to whatever phonemes are there. In contrast, we do not expect generalized sensitivity to emerge with other critical concepts in early childhood education. For example,we would not expect children to achieve letter recognition by developing sensitivity to initial letters, final letters, and finally to complete spellings. We teach letters one by one.

Recent research suggests the phoneme awareness that is causal in learning to read is knowledge of phoneme identities rather than skill in manipulating phonemes (Byrne & Fielding-Barnsley, 1989, 199 o, 1991; Murray, 1998). Phoneme-to-word matching tests are prototypical tests of phoneme identities. For example, Wallach and Wallach (1979) told children entering first grade, "Some words start with the sound /ml, like ma or mud or me" (p. 200), and asked them to decide which illustrated word matched the isolated sound ("Does man or house start with /m/?"). This simple test discriminated children who were well prepared for first grade from those likely to struggle with learning to read.

Murray (1998) found that instruction in phoneme manipulation improved blending and segmentation ability but made little difference in kindergarten children's alphabetic insight, that is, in their ability to apply the alphabetic principle that letters represent phonemes in words. In this study, all participants were taught to associate eight letters with their sounds, for example, to say "sss" when shown the letter S. A manipulation group was taught to segment and blend any phoneme,while an identity group was taught the identities of the eight phonemes symbolized by the letters one phoneme at a time. A control group listened to stories read aloud and composed language experience stories. The manipulation group surpassed the other groups in segmentation and blending, but these children did not demonstrate alphabetic insight; they could not reliably determine, for example, whether the printed word MAD represented mad or sad. In contrast, children taught the eight phoneme identities were more likely to use correspondences to distinguish rhyming words, that is, to achieve alphabetic insight. Knowledge of phoneme identities helped them understand correspondences well enough to use initial letters in identifying words. Still, results of this study were ambiguous, because the children taught phoneme identities did not demonstrate superior phoneme awareness in posttests. This may have owed to the insensitivity of the experimental measures of phoneme awareness with kindergarten children.