從評分者間信度與實驗處理忠實度評析日本與臺灣自閉症單一個案研究品質

THE QUALITY OF SINGLE-CASE RESEARCH FOR INDIVIDUALS WITH AUTISM SPECTRUM DISORDERS IN JAPAN AND TAIWAN: AN INVESTIGATION OF INTER-RATER RELIABILITY AND TREATMENT FIDELITY

近年來國內外皆倡導選擇有實證研究支持其介入成效的策略或方案，目前國外已有數篇文獻聚焦於探討自閉症研究品質，然而這些文獻所分析的研究主要發表於英語系期刊中，故結論無法推論至發表於非英語系期刊的研究品質。為了瞭解非英語系期刊中自閉症研究的品質，本研究採系統性文獻回顧，分析日本與臺灣主要之特殊教育學術研究導向期刊，以探討其刊登之所有自閉症單一個案研究的評分者間信度與實驗處理忠實度之現況與趨勢。本研究依（1）從日本與臺灣共七份主要特殊教育研究期刊中，搜尋文獻；（2）依預先設定的文獻選取與排除準則，進行研究評估；（3）針對文獻中描述之評分者間信度與實驗處理忠實度，進行資料編碼與分析等三步驟，最後進行194篇符合選取準則之文獻評析。本研究結果如下：在評分者間信度方面，日本於過去數十年間符合單一個案研究評分者間信度標準的研究數量，整體而言呈現增加的趨勢，臺灣文獻則未呈現同樣的趨勢；在實驗處理忠實度方面，日本尚未有單一個案研究描述研究的實驗處理忠實度，而臺灣則已有數篇研究符合國際間檢視實驗處理忠實度資料的標準。就整體研究品質而論，日本與臺灣的自閉症單一個案研究已逐漸符合評分者信度間的標準，然而對於實驗處理忠實度的敘述仍處於初步發展階段。為了提升自閉症介入方案的研究品質，本研究建議在亞洲地區的相關學術研究論文應依建議的指標與標準說明各研究之評分者間信度與實驗處理忠實度。

Given this era of identification of evidence-based practices, previous reviews have provided a snapshot of the current status of the research quality of autism spectrum disorder (ASD)-focused studies. However, including only studies in English language journals does not represent the quality of the literature published in journals of languages other than English. To evaluate the overall quality of ASD intervention research in non-English-language journals, this systematic review summarizes the current status and trends of interrater reliability (IRR) and treatment fidelity in ASD-focused singlecase research published in major academically oriented Japanese and Taiwanese special education journals. To conduct this review, the following three steps were taken: (a) literature search of the seven prominent special education journals in Japan and Taiwan, (b) assessment of potential studies against pre-set inclusion and exclusion criteria, and (c) study coding and analysis of descriptive study characteristics and measures of IRR and treatment fidelity. A total of 194 articles met the inclusion criteria for the review. The results show that an overall increasing trend in the number of articles that reported IRR data with acceptable levels were observed over time in the Japanese journals while no such trend was found in the Taiwanese journals. In contrast, it was found that no article published in the Japanese journals had reported treatment fidelity data while a small number of articles that reported treatment fidelity data with acceptable quality degrees were observed in the Taiwanese journals. As to the overall quality of ASD research, researchers in Japan and Taiwan are increasingly attending to quality with regard to collecting and reporting acceptable IRR data for outcome variables in their single-case autism-related articles. Yet, the evaluation of treatment fidelity and its IRR is still at the initial stage. To improve the overall quality of ASD intervention research, efforts should be made to report both IRR and treatment fidelity data based on the suggested standards with acceptable quality degrees in Asian-language journals.