自閉症兒童對事件記憶之研究:電腦輔助訪談

Abstract

自閉症兒童可能因為不懂性騷擾的本質而容易成為性罪犯的攻擊目標。在性犯罪中，定罪的關鍵與受害者的供詞有極大的關係，因受害者通常是唯一的目擊證人，然而，自閉症兒童通常有缺乏社交情緒互動的特徵，造成訪談極度困難。一般來說，使用現有的訪談方法由自閉症兒童身上所獲得的正確資訊比從正常發展兒童身上所獲得的要少得多。本研究目的在於使用新的訪談輔助 - 電腦動畫訪談人員，在法庭訪談中給予自閉症兒童更多的協助。本研究的研究對象包含來自全台各縣市共30名國小一、二年級學童 (18:12) ，其中一半為確診自閉症案例(13名男童與2名女童) (M=7.52, SD=0.71)，另一半為正常發展兒童(13名男童與2名女童) (M=7.67, SD=0.66)。學童們在參與尋寶活動一週後接受訪談，各組其中一半的學童由電腦訪談，而另一半則由人員訪談，訪談問題中包含了六題誤導性題目，用以測試研究對象的暗示感受能力。研究結果顯示，無論是由電腦或人員執行訪談，兩組研究對象所展現的記憶表現(例如回憶正確度，回憶數量，誤導性問題反應的正確度等) 結果相似，此結果對未來相關研究與應用方面提供了寶貴的資訊貢獻。

Children with autism spectrum disorder (ASD) can be vulnerable to sexual predators as they may not understand the nature of the abuse. In sexual crimes, there is a heavy reliance on the victim’s testimony for conviction because the victim is often the only eyewitness. However, children with ASD typically show characterized deficits in social-emotional reciprocity, which makes them extremely difficult to interview. In general, the amount and accuracy of information extracted from interviews using current methods are often less for ASD individuals than for typically developing children. In this study, we aimed to provide more support for children with ASD in forensic interviews by way of a new interviewing aid: a computer-animated interviewer. A sample of 30 children studying in grades one and two (18:12) were recruited throughout Taiwan, with half of the children diagnosed with ASD (13 boys; 2 girls) (M=7.52, SD=0.71) and the other half being typically developing (TD) (13 boys; 2 girls) (M=7.67, SD=0.66). The children participated in a “scavenger hunting” event and were interviewed a week later regarding the target event, with half of each group being interviewed by the computer and the other half interviewed by a live human. The participants were also asked six misleading questions aimed to examine their suggestibility. Results revealed that both groups of children showed comparable abilities in memory performance (i.e., recall accuracy, recall amount, and response accuracy to misleading questions) when interviewed by the avatar interviewer and the live human interviewer. The results show encouraging implications for future applications.