魏氏兒童智力測驗第三版的預測效度：三年追蹤研究

Predictive Validity of the WISC-Ⅲ (Chinese Version): A Three Year Follow-Up Study

魏氏兒童智力測驗第三版（WISC-Ⅲ）（中文版）基本上是英文版的中文翻譯版，已於1997年於台灣發行，目前可以說是台灣最受歡迎的兒童智力測驗之一。本研究的目的係檢驗其在不同年級和不同種族間的預測效度（predictive validity）。參與研究的受試者來自兩個不同的小學，其中一所為平地小學，另外一所則為山地小學。研究樣本是由階層隨機抽樣產生，包括一年級30位、三年級30位和五年級30位，共計90位學生參與。受試者由合格之心理師依據指導手冊所定的標準程序進行魏氏兒童智力測驗第三版（中文版）10個分測驗的施測。本研究中魏氏兒童智力測驗第三版（中文版）的效標變項（criterion variable）為受試者連續三年的學校學業成就（academic achievement），最後共有84位學生進入結果分析。對於三個不同年級的受試者和兩種不同族別的受試者而言，學業成就和語文智商（VIQ）、作業智商（PIQ）和全量尺智商（FSIQ）間的相關係數全部達到統計上的顯著水準。其中除了山地小學的受試者之外，全量尺智商對於學業成就的預測較語文智商和作業智商為佳。進一步比較語文智商和作業智商對於不同年級受試者的學業成就的預測：對於一年級受試者而言，作業智商較語文智商有較佳的預測力；對於三年級受試者而言，作業智商和語文智商有相同的預測力；對於五年級受試者而言，語文智商則較作業智商有較佳的預測力。最後比較兩個不同小學，發現全量尺智商對於學業成就的預測貢獻，平地小學大於山地小學；相對地，人口統計學變項（包含年級）對於學業成就的預測貢獻，則是山地小學大於平地小學。對於不同年級和不同族別的受試者，魏氏兒童智力測驗第三版（中文版）即使預測三年後的學業成就，依然具有預測效度。此外，本研究的臨床意涵及限制也一併在內文討論。

The Wechsler Intelligence Scale for Children-Third Edition (WISC-Ⅲ) (Chinese version), which is basically the translation of the English version, has been published in 1997. It is at present the most popular children's intelligence test in Taiwan. This study examined the predictive validity of the WISC-Ⅲ (Chinese Version) across grade, race and follow-up year. The participants enrolled in this study were students recruited from two different primary schools. Using stratified random sampling, a total of 90 students, including 30 1st graders, 30 3rd graders and 30 5th graders, were enrolled in this study. They were administered the WISC-Ⅲ (Chinese Version) with the 10 subtests in standard procedure described in the test manual. The criterion variable of the WISC-Ⅲ (Chinese Version) was the participant's school academic achievement in the following three years. The results for 84 students were analyzed. Full Scale IQ was a better predictor of academic achievement than Verbal IQ or Performance IQ across grade, race and follow-up year, except for aboriginal children. Furthermore, examination of the Verbal IQs' and Performance IQs' correlations with academic achievement across grades revealed three different results. In grade 1, Performance IQ was a better predictor than Verbal IQ. In grade 3, Performance IQ and Verbal IQ had equal predictive power. In grade 5, Verbal IQ was a better predictor than Performance IQ. Finally, WISC-Ⅲ IQs had larger contributions to academic achievement in non-aboriginal children than in aboriginal children. Moreover, demographic variables, especially the grade variable, had larger contributions to academic achievement in aboriginal children than in non-aboriginal children. Results suggests that the WISC-Ⅲ (Chinese Version) has predictive validity across grade, race and follow-up year, even as much as three years later. Implications and limitations of the current findings have also been discussed.