配适度检定.PDFVIP

量表發展工作坊(3) 配配適指標適指標 修正指標修正指標 (Goodness-of-fit Index and Modification Index) ByBy 劉介宇 1 配適度檢定配適度檢定((test of ggoodness-of-fit)) • 配適度檢定原理的虛無假設配適度檢定原理的虛無假設 – 理論上 H 0 : () ^ – 實務上實務上 H : S  0 • Ｓ：樣本共變數矩陣 •• ΣΣ與與ΣΣ((θθ)) ：：未知母體共變數矩陣未知母體共變數矩陣 ^ •  ：根據母體建構參數所複製出(reproduced)的估計共變數矩陣 • CFACFA的基本檢定的基本檢定 ，是期望獲得是期望獲得 「不要拒絕虛無假設不要拒絕虛無假設」」的的 檢定結果，也就是H0被接受的結果，或獲得不顯著的 檢定結果 。 • 也就是: 你設定的model所跑出來的相關係數矩陣=原始 資料的相關係數矩陣資料的相關係數矩陣 2 AssessingAssessing ModelModel FitFit • How well does the model fit the data? • ThisThis cancan bebe answeredanswered byby thethe ChiChi-squaresquare statistic but this test has many problems – ItIt iis samplle siize ddependdentt, so withith llarge samplle sizes trivial differences will be significant – There are basic underlying assumptions are violated the probabilities are inaccurate 3 Example:Example: ReproducedReproduced CorrelationCorrelation MatrixMatrix 4 5 AssessingAssessing ModelModel FitFit • Fiit iindidices – Read through text books and you’ll find that there are tons of fit – Which do you choose? – Different researchers have different preferences and different cutoff criterion for each index – Examples: • CFI-Chi-square based calculation