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

Eliminating all children who showed evidence of consistent phonetic cue reading ability yielded much different results. With 34 prealphabetic readers (Ehri, 1998), the strongest correlations with rudimentary decoding were with letter naming accuracy (r = .58) and knowledge of nursery rhymes (r = .41), variables closely tied to enriched home literacy experience. No phoneme awareness test explained more than 12% of the variability of alphabetic insight in these prealphabetic readers. It is with these prealphabetic readers that prediction is critical; these are the children at risk for reading delay.

Because we expect a phoneme awareness test to predict the ease of decoding acquisition in prealphabetic readers, the correlation of a test with existing code knowledge is not particularly informative. To assess the validity of a phoneme awareness test, we need to examine its power to forecast children's ease or difficulty in achieving alphabetic insight. To measure this potential for learning to decode, we counted the number of teaching trials it required prealphabetic kindergartners to learn to use the letters S and M to distinguish rhyming printed words like SIX and MIX. Phonetic cue reading, decoding initial or boundary letters, is the first step in using the alphabet, a step that had been taken by two-thirds of our kindergarten participants. For these beginning readers, phonetic cue reading manifested alphabetic insight. However, the other third of the kindergarten sample had too little understanding of what letters represent to use phonetic cues.

The TPI (r = -.47) proved most accurate in predicting the number of trials it would take to teach children to distinguish rhyming words by using their first letters, followed by the Yopp-Singer (r = -.44). In a stepwise multiple regression analysis, only the TPi explained a significant amount of variation in the best fitting regression equation for explaining alphabetic insight. Thus the Tri proved superior to two competitive phoneme awareness tests for assessing prealphabetic readers' potential for learning the alphabetic code, explaining 19% of the variance in the number of Trials to alphabetic insight. These data suggest that the Tri may be a better indicator of potential to gain alphabetic insight than either the TOPA or the Yopp-Singer.

This finding should be viewed cautiously. Though we assessed nearly a hundred kindergarten children, only 34 children qualified for the dynamic assessment of learning by scoring at chance levels on the test of phonetic cue reading. These children were prealphabetic in their understanding of written language (Ehri, 1998); they did not consistently use even the beginning letters in printed words to cue phonemes in pronunciations. Given the available population of kindergarten children and the limited resources available to the study, these were all the prealphabetic kindergarten children available for participation. Using categorized data and a nonparametric statistic, the Spearman rank-order correlation, only the TPI (r = .47) was significantly correlated with Trials at a conservative .01 alpha level, though the difference between this correlation and the correlation of the YoppSinger (r = .38) with Trials was not statistically significant.