부산대학교 도서관

Purpose: In a feasibility study for a randomized controlled trial of treatments for phonological disorders conducted over a period of 8 months, we examined 6 clinically relevant outcome measures. We took steps to reduce error variance and to maximize systematic variance. Method: Six children received traditional treatment (Van Riper, 1939) and 7 received expansion points (Smit, 2000), a treatment program with both phonological and traditional elements. Outcome measures, which were applied to both word list and conversational samples, included percentage of consonants correct (PCC; Shriberg & Kwiatkowski, 1982), PCC for late and/or difficult (L /D) consonants and number of L/D consonants acquired. Results: In repeated-measures analyses of variance, all measures showed significant differences from pretreatment to posttreatment, and the word list measures were associated with very high power values. In analyses of covariance for between-groups contrasts, the adjusted expansion points mean exceeded the adjusted traditional treatment mean for every measure; however, no differences reached significance. For the L /D PCC (conversation) measure, the contrast between groups was associated with a large effect size. Conclusion: We recommend that practitioners use outcome measures related to a word list. We recommend that researchers consider using L/D PCC on the basis of conversational samples to detect differences among treatment groups. [ABSTRACT FROM AUTHOR]