臺灣一般兒童與臨床兒童之魏氏兒童智力量表第三版(WISC-III)因素恆等性研究

WISC-III Factorial Invariance Across Normal and Clinical Samples

本研究主要目的在檢驗魏氏兒童智力量表第三版(WISC-III)在一般兒童與特殊兒童組群間之因素恆等性。文中根據1,100 名WISC-III 標準化樣本及1150 名臨床兒童樣本資料,運用結構方程模式進行多樣本驗證性因素分析。本研究以WISC-III 12個主要分測驗與四個潛在因素架構為假設基準模型，分二階段完成檢驗。階段一先分就二組進行單樣本驗證性因素分析，以建立適當之基準因素模式。階段二則建立階段性逐步設限的巢套模型，逐一檢驗因素架構、因素負荷量、與殘差變異數之跨組別恆等假設。研究發現除了少處細微之殘差變異或共變性差異外,絕大多數之檢驗參數均具跨樣本恆等特性，WISC-III 在一般兒童與特殊兒童組群間具有相同之四因素架構與因素負荷量。整體而言，研究結果支持部份因素恆等性，台灣臨床兒童與一般兒童之WISC-III 分數是具有相同之解釋意義。

Taking as its sample a normal group of 1,100 children and a clinical group of 1,150 children in Taiwan, the purpose of this study was to test for invariance in the WISC-III factorial structure across normal and clinical samples. Results of our two-stage multi-sample confirmatory factor analysis supported partial factor invariance. Among all examined parameters, only a few discrepancies involving unique error variances and subtly correlated residuals were found across groups. Evidence of partial measurement invariance indicated that the hypothesized four-factor model generally fit the real data fairly well. It was found that while measurement accuracy for a few subtests differs between and among groups, the overall main structure and factor loadings were generally invariant across normal and clinical samples. Therefore we concluded that WISC-III scores can be interpreted as having the same meaning across groups.