臺灣版兒童溝通能力檢核表的因素結構

Abstract

「兒童溝通能力檢核表（Children’s Communication Checklist）」為國際上普遍的社會溝通障礙兒童的篩檢量表，也適合篩檢「自閉症」兒童。然而，鮮少研究探討其因素結構和區別效度。此外，社會溝通隨年紀而增長；但是，這項檢核表的因素結構，能否呈現這項成長趨勢？因此，本研究主要目標在於檢視臺灣版「兒童溝通能力檢核表」的因素結構。其次，本研究評估社會溝通潛在因素，隨年紀而增長的趨勢。 首先，將量表兒童常模資料（N = 1201）拆成兩組樣本，一組使用探索性因素分析尋找因素結構，並挑選出單向度和高共同性題項來構建新的子量表；另一組則使用驗證性因素分析來檢驗可複製性。因素分析結果顯示「結構語言」、「語用」和「刻板語言」三項因素，穩定解釋量表共同變異。 其次，使用雙因素模型來表徵常模資料，並評估量表指標的單向度程度。結果顯示可將量表視作單向度，支持這項量表以「整體溝通綜合分數」作為社會溝通能力的指標。 再者，使用「測量恆等性」檢驗語用潛在因素，是否在性別和年齡有顯著差異？結果顯示隨年紀增加，潛在因素的平均分數隨之降低（亦即，語用進步）；另外，沒有性別差異。因此，這項量表適合依據年紀，來構建常模。 最後，基於因素分析結果作為選題依據的簡式量表，使用自閉症（N = 120）在各複合常模分數的篩檢能力，作為量表效度的檢驗標準，並檢視各種常模分數在「接受者操作特徵曲線（ROC）」表現。結果顯示基於「語用因素」的分數達到最大的曲線下面積，而且以115分以上作為切截點，其特異度84%和敏感度91%。若使用這項量表的目的在於篩檢語用能力缺損，建議使用「語用」常模分數。 總之，本研究的結果顯示，「結構語言」、「語用」和「刻板語言」為臺灣版兒童溝通能力檢核表的因素。而且，「語用」因素隨年紀增長。最後，本研究也呈現如何選擇量表題項，來建構適合篩選自閉症兒童的簡化版量表。

Children’s Communication Checklist – Second edition (CCC2) is a widely-used scale to screen children with social communication impairment or autistic spectrum disorder (ASD). However, it is unclear what factorial structure would underpin CCC2. Additionally, social communication skills undergo great changes from early childhood to late childhood. The first goal of this study was to explore the factorial structure of the CCC2-Taiwan version, and the second goal was to assess whether the CCC2 factor would indicate the developmental trend of social communication during childhood. First, norm samples (N = 1201) of CCC2-Tawan were split into two halves, one half sample for exploratory factor analysis (EFA) to explore the factorial structure and then selecting unidimensional items with high commonality, the other half sample for validating item selection by confirmatory factor analysis (CFA). The results of factor analysis showed that three factors (structural language, pragmatics, and scripted language) well explain the data. Additionally, the bifactor model was used to represent the data, and results suggested that data of CCC2 were essentially unidimensional, in the support of using the “general communication composite, GCC” score as the major measure to indicate social communication skills in individual children. To achieve the second goal of this study, testing measurement invariances of pragmatic factor in CCC2 across gender and ages were utilized. The result indicated that latent mean invariances between gender were not significantly different, while the latent mean of older children was significantly lower (i.e., with better pragmatic skills) than younger children. Therefore, testing measurement invariances showed that it was appropriate to build norms in CCC2 Taiwan version by age, instead of by gender. Finally, a brief form of CCC2 that item selection criteria were based on results of factor analysis were formed to calculate new norm-referenced composite scores for screening ASD, and receiver operating characteristic (ROC) curve analysis was conducted to evaluate screening performance of the brief form. The result showed that the pragmatics composite score was the best measure among various composite scores, with 84% specificity and 91% sensitivity, if using ≥ 115 as cutoff score. General communication composite (GCC) score was relatively poor, with 83% specificity and 88% sensitivity using the same cutoff score. If the purpose of using the CCC2 is only to screen children with pragmatic impairment, the results of this study suggest that the Pragmatic composite score would be the optimal measure. In conclusion, results of this study revelated that structural language, pragmatics, and scripted language were factors of CCC2-Taiwan version. Additionally, this study depicted the validity of a brief CCC2 form to screen children with ASD.