頁籤選單縮合
題名 | 認知分析與心理計量分析對解平衡桿問題認知發展層次與解題運作成份測量之比較=A Comparison of Cognitive Analysis and Psychometric on Saving Balance-Scale Problems |
---|---|
作者 | 丁振豐; Ting, Chen-feng; |
期刊 | 初等教育學報 |
出版日期 | 19970600 |
卷期 | 10 1997.06[民86.06] |
頁次 | 頁81-125 |
分類號 | 179.5 |
語文 | chi |
關鍵詞 | 構念效度; 認知測量模式; 認知分析; 心理計量枌析; 天平平衡; 線性邏輯斯諦測驗模式; |
中文摘要 | 研究者提出兩種整合心理計量與認知分析的構念效度分析模式,並以本研究來驗 證。研究工具包括研究者所設計的「蹺蹺板測驗」和「平衡桿測驗」等兩套天平平衡問題測 驗。研究對象為國民中小學一至九年級學生,共 4677 位受試者。模式一以題目間的關係來 分析測驗的構念效度。這模式是以認知歷程的作業分析為理論基礎設計題型,建構理論上的 難度順序和反應組型,以心理計量模式分析各題型的難度、不同發展層次的受試者的反應組 型及能力估計值,再比較兩種分析結果,以驗證所設計測驗的構念效度。研究一結果顯示認 知分析與心理計量分析結果接近,研究結果可以說明「翹翹板測驗」的構念效度,也顯示模 式一可以作為分析構念效度分析的方式。模式二以題目內成分的關係來分析測驗的構念效度 。這模式是以認知心理學為理論基礎設計題目內容成分,以引起特定解題運作,以線性邏輯 斯諦測驗模式為心理計量模式分析各成分難度,以驗證所設計測驗的構念效度。研究二結果 顯示題目內容成分可以解釋題目難度變異來源超過 60%。表示本研究所提出的九個成分能分 解「平衡桿測驗」題目的變異,成分可以作為瞭解題目難度的變異來源。研究結果可以說明 「平衡桿測驗」的構念效度,也顯示模式二可以作為分析構念效度分析的方式。 |
英文摘要 | By integrating psychometric and cognitive analysis, two construct validity discoursing models were proposed, implemented, and substandtially confirmed in this study. The construct representation was investigated by cognitive task analysis and psychometric parameter estimation. The target tasks included were the Seesaw-Scale and the Balance-Scales developed by author. The subjects included were 4,677 first to ninth graders. Model I focussed on the items relations. The item type were designed based by cognitive task analysis. The results suggested that the consistancy of cognitive and psychometric analysis was well. The analysis included item difficulty order and the subjects response patrerns in development levels. The power of the validity evidence is substantially increased by this model emprirical research. Model II focussed on the content components relation. The components were designed based on cognitive analysis. The Linear Logistic Test Model was adopted to estimate the component parameters. The results suggested that facets and components parameters could well explain the item difficulty variations (accounted for above 60% difficulty variance). In other words, the content components would servee as the important resource to explain the trait variance. The power of the validity evidence was substantially increased by the emprirical content component analysis data. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。