3～6歲兒童漢語語言發育評估量表：語法篇的信效度研究

目的 對3～6歲兒童漢語語言發育評估量表-語法篇進行信效度檢驗。方法 2018年12月—2019年6月期間，採用整群隨機抽樣的方法，抽取成都地區幼兒園1 704名2歲11個月—6歲11個月幼兒，其中男女各852名。採用3～6歲兒童漢語語言發育評估量表-語法篇對幼兒進行測評，並對量表進行信效度分析。結果 共回收有效數據1701份。語法篇包含40個聽句選圖項目。內部一致性信度Cronbach’sα系數為0.834，折半信度為0.795。項目數據整體質量檢驗，被試及項目Infit、Outfit MNSQ值分別為1.00、0.93及0.99、0.93。被試區分度為1.73，被試信度為0.75。40個項目的Infit MNSQ值均介於0.86和1.2 logits之間，與Rasch模型擬合好。項目難度值在-2.55至1.82 logits之間，被試個體能力值在-2.78至5.0 logits之間。結論 兒童漢語語言發育評估量表-語法篇信效度良好，可用於對中國3～6歲兒童語法發展水準進行測評，本量表能簡便、有效地測量兒童語法發展情況，為中國兒童語言障礙的診斷和治療提供參考。
Objective To evaluate the reliability and validity of grammar subscale of Children's Language Development Scale(CLDS). Methods A stratified cluster sampling method was adopted to randomly select 1 704 preschool children(including 852 boys and 852 girls) in Chengdu from December 2018 to June 2019. Children's language ability was evaluated with CLDS, then the reliability and validity of the grammar subscale of CLDS were tested. Results There were 1 701 completed Results in sum. The grammar scale contained 40 items of picture selection after listening to a sentence. Internal consistency reliability Cronbach's α coefficient was 0.834, and the half-fold reliability was 0.795. The infit and outfit MNSQ values of the subjects and the items were 1.00, 0.93 and 0.99, 0.93 respectively. The discrimination validity was 1.73, and the reliability of the scale was 0.75. The infit MNSQ values of 40 items were between 0.86 and 1.2 logits, which fitted well with the Rasch model. The difficulty value of items ranged from -2.55 to 1.82 logits, and the individual ability value was between -2.78 and 5.0 logits. Conclusions The grammar subscale of the CLDS has good reliability and validity. It can be applied to simply and effectively measure 3- to 6-year-old children's grammar development, so as to provide reference for diagnosing and treating speech and language disorders in Chinese children. Copyright © 2022 西安交通大學；中華預防醫學會.