本研究根據魏氏成人智力量表第三版(WAIS-Ⅲ)臺灣標準化樣本,全體888位年齡介於16至84歲間之成人測驗分數,分析兩個分測驗(語文及作業分測驗各一)組合簡式版本之建立方法,以供實務工作者估計全量表智商之使用。文中除考量心理計量特性、臨床意義性,及施測計分之省時便易性等多項因素外,亦比較文獻中常見三種全量表智商估計方法(比例分配法、線性等化法與迴歸預測法)之估計確度。主要研究發現如下:(1)線性等化法能提供相對較為精確之智商估計;(2)「詞彙,矩陣推理」組合提供最佳估計確度,施測時間約在21分鐘上下。徜若施測時間極為有限,則可考慮「類同,矩陣推理」或「常識,矩陣推理」組合,此二版本也具相當之估計正確性,且僅需12分鐘左右的施測時間。本研究發現與國外文獻建議甚為一致,可作為跨文化效度檢驗之證據。文內同時提供不同組合之智商估計公式供實務界使用,以期達到最大效用。此外,亦對簡式版本之性質、研究重要限制,及未來研究建議提出說明與討論。
Two-subtest short forms of the WATS-Ⅲ were investigated with the Taiwan WATS-Ⅲ standardization sample of 888 normal adults, ages from 16 to 84 years old. Various psychometric characteristics were compared among nine from dyads, as well as qualities of estimation among prorating, linear equating, and regression procedures. The results supported the use of a linear equating procedure to generate two-subtest short- form estimates. Vocabulary-Matrix Reasoning, Similarities-Matrix Reasoning, and Information-Matrix Reasoning were proposed as the more appropriate two-subtest combinations for gross FSIQ estimate. Estimation formulas for all possible nine dyads were also presented for references. However, it is recommended that the examiner first evaluate the pros and cons of short forms and be aware of potential estimation errors. This study provides psychometric information for selecting two-subtest abbreviations of FSIQ that match the needs of the clinicians. The results also provide cross-cultural validation evidence for WAIS-Ⅲ dyads.