研究目的：本研究的主要目的，是修訂兩歲期自閉症篩檢工具（Screening Toolfor Autism in Two-Year-Olds, STAT, Stone, Coonrod, & Ousley, 2000; Stone, Coonrod,Turner, & Pozdol, 2004）為台灣版本，簡稱T-STAT，適用於層次二的自閉症篩檢工具。本研究報告為初步的結果。研究方法：本研究的樣本是34名生理年齡在24-36個月之間的自閉症類及發展遲緩幼兒。兩組幼兒皆為17名，在生理、心理年齡、性別、及社經水準上皆加以配對。修訂STAT為T-STAT的工作，評估作業維持測量遊戲、要求、相互注意協調能力及模仿能力等四項領域，但將原STAT中測量相互注意協調能力及模仿能力的題項，各做一題的更動，其餘題項不變，計分方式與STAT相同。研究結果：初步的結果發現，不論是STAT或T-STAT，在遊戲、要求及相互注意協調能力等領域皆能有效區分兩組，但模仿能力領域則未有明顯的區辨力。此外，若以STAT的切截分數來做分類，則不論STAT或T-STAT，與自閉症觀察診斷量表（Autism diagnostic observation scale, ADOS, Lord,Rutter, DiLavore, & Risi, 1999）間所建立的敏感度與精確度達94.1%及82.4%；STAT或T-STAT，與臨床診斷間的敏感度與精確度亦達94.1%及82.4%。研究結論：初步的結果顯示，不論是STAT或T-STAT，在篩檢自閉症類幼兒上尚稱滿意；是否這項結果在較大的樣本，及對非典型自閉症的篩檢效果仍佳，需要進一步研究探究。

Purpose: The purpose of the study was to modify the Screening Tool for Autism in Two-Year-Olds (STAT, Stone et al., 2000; 2004) for Taiwan, called the T-STAT, for level-2 screening of young children with autism spectrum disorder. Methods: Thirty-four young children with autism spectrum disorders and developmental delay were recruited (MEAN AGE: 28.18months old, boy/girl= 28/6). Each group had 17 children aged 24-36 months matched for chronological age, mental age, gender, and socio-economic status. Adapted from the STAT, the T-STAT maintained all of the items in the play and request domains, but substituted one item from joint attention domain and one from the imitation domain. The coding algorithm was the same for the STAT and the T-STAT. Results: The initial data revealed that both the STAT and the T-STAT can appropriately discriminate children with autism spectrum disorders and developmental delay on the domains of play, request, and joint attention, but not on the imitative domain. When using a score of 2 as a cutoff, agreement between the STAT or T-STAT and the ADOS (Autism Diagnostic Observation Scales, Lord, et al., 1999) was acceptable; the sensitivity and specificity were 94.1% and 82.4%, respectively. The results of the agreement between the STAT and T-STAT and clinical diagnosis using the DSM-IV-TR (APA, 2000) were the same as the data shown by the ADOS. Conclusions: Our preliminary data demonstrated that the STAT and T-STAT as level 2 screening tools for autism spectrum disorders is promising. Further studies are needed to test more participants with both typical and atypical autism and to clarify the utility of the STAT and T-STAT in young children aged 2 to 3 with autism spectrum disorders. In addition, longitudinal work is needed to understand the stability of the diagnosis.