PA tasks used in O’Conner et al (1995).
Volume 87(2) June 1995 p 202C217
O´Connor, Rollanda E.1,4; Jenkins, Joseph R.2; Slocum, Timothy A.3
Assessing phonological manipulation skills.
The pretest and posttest tasks were based on tests used in studies of phonological awareness and kindergarten prediction of reading skill in first grade (Ball & Blachman, 1991; Cunningham, 1990; Fox & Routh, 1984; O´Connor et al., 1993; Perfetti, Beck, Bell, & Hughes, 1987; Rosner & Simon, 1971; Yopp, 1988). Each task included three unscored practice items with feedback and an opportunity for the child to repeat the correct response. Following each testing item, the examiner wrote the child´s response and then provided the correct response. We wanted to select children genuinely low-skilled in phonological manipulations and expected some children to have difficulty with phonological tasks because of inexperience with the directions and format. By providing feedback on all items, we were able to eliminate children who quickly learned the new tasks. None of the actual testing items (except items on the mastery tests) was used in training. These measures are shown in the Appendix.
Although one study reported blending of phonemes to be too easy for kindergarten children (Bentin & Leshem, 1991), others reported blending ability in kindergarten to be a good predictor of later reading ability (Lundberg, Olofsson, & Wall, 1980; Perfetti et al., 1987; Shankweiler & Liberman, 1989; Uhry, 1992; Yopp, 1992). The first 10 items were three-phoneme words presented in onset-rime format. The examiner said, M (pause) ake. What word is that? The next five words were presented as three separate phonemes (top).
Children segmented the first 10 words into an onset-rime format (e.g., The examiner said, Make. Tell me two sounds in make.), and 5 additional words into three phonemes. Children received 1 point for each correctly segmented portion of the word (1C2 points for onset-rime segmentation and up to 3 points for segmentation into three phonemes). We did not penalize children who completely segmented words during the onset-rime phase of the task (e.g., for the first 10 words, we awarded 2 points for a response of make or make).
Syllable deletion was based on Berninger´s (1986) modification of the Rosner Test of Auditory Analysis (Rosner, 1979), which she expanded to 10 items for the kindergarten level. We included the syllable deletion task because it is an easier phonological task than deletion or segmentation of phonemes and has been used for early identification of children at risk for reading failure. Examiners state a word (baseball) and ask the child to say the word minus one syllable (Say it again, but don´t say base).
Rhyme was included because it may be a precursor to later developing phonological skills (MacLean, Bryant, & Bradley, 1988), or it may represent a source of individual differences in phonological awareness (Stanovich et al., 1984). Although the ability to rhyme develops early for many children (Bryant et al., 1990), our work with low-skilled and developmentally delayed youngsters has found rhyme production to be challenging for them in kindergarten. Following an explanation and examples, the examiner gave the instruction for the first of five items: Say a word that rhymes with make. Correct responses included real or nonsense words. For incorrect responses, the examiner modeled correct alternatives.