Best readings for SEM http www semsoeasy com

Best readings for SEM http: //www. semsoeasy. com. tw/

參考書 http: //www. semsoeasy. com. tw/ 3

第一天大綱 1. The basic of Structural Equation Modeling 1. The growth of SEM 2. SEM terms and symbols 2. Amos的操作環境與模型建立 1. 2. 3. 4. 5. 6. 7. Amos introduction and operation 路徑分析 (path analysis) 迴歸分析 (regression analysis) 結構模型分析 (SEM analysis) 圖形調整及輸出 SEM model report Goodness of fit (GOFs) http: //www. semsoeasy. com. tw/ 4

第一天大綱 3. SEM identification 1. How many sample size are need? 2. Rules for determining SEM model parameters. 3. Model identification 4. The estimated principle of SEM http: //www. semsoeasy. com. tw/ 5

第二天大綱 4. 驗證式因素分析 (Confirmatory Factor Analysis, CFA) 1. 2. 3. 4. 一階驗證性因素分析(first order) 組成信度 (CR)與平均變異數萃取量 (AVE) 收斂及區別效度 Second order SEM model 5. Using covariance matrix analyze SEM 1. Making covariance matrix 2. Analyze SEM http: //www. semsoeasy. com. tw/ 6

The growth of SEM (Herhberger, 2003) • SEM has become the preeminent multivariate method of data analysis. • SEM has become the preeminent place in which to publish developments in SEM. • As long as SEM continues to respond to the needs of scientific workers, it will continue to flourish, from its present, healthy young adulthood into, perhaps, an immortality that bypasses old age. http: //www. semsoeasy. com. tw/ 7

Publishing Research in Marketing Journals Using SEM ˇ ˇ ˇ ㄨ ㄨ 1. Papers not reporting an SEM model have a greater likelihood of getting rejected. 2. Papers reporting an SEM model are rated more favorably by reviewers. 3. The use of SEM influences reviewer recommendations and paper quality 4. AMOS applications in marketing have a greater chance of being rejected. 5. Goodness (badness) of fi t indices are positively (negatively) related to paper quality and reviewer recommendations. http: //www. semsoeasy. com. tw/ 8

SEM terminology • Independent predictor exogenous (external) variables – Affect other variables in the model. – 本身不具測量誤差(unique, error variables)。 – 模型中沒有被任何箭頭刺到的變數即為外生變數，有 潛在及觀察變數兩種。 – 如下圖的信任，為一外生觀察變數。 http: //www. semsoeasy. com. tw/ 11

SEM terminology • Experimental research • independent variables. • Non-experimental research • predictor criterion variables. • SEM research • Exogenous Endogenous variables. 15 http: //www. semsoeasy. com. tw/

SEM terminology and symbols • Observed (or manifest) • Latent (or factors) • Direct effects • Reciprocal effects • Correlation or covariance 16 http: //www. semsoeasy. com. tw/

Amos操作範例介紹 科技接受模型 Technology of Acceptance Model, TAM http: //www. semsoeasy. com. tw/

Amos操作環境與模式建立 • 所有程式 IBM SPSS Statistics IBM SPSS Amos ? ? Amos Graphics http: //www. semsoeasy. com. tw/

Amos起始畫面 http: //www. semsoeasy. com. tw/ 22

http: //www. semsoeasy. com. tw/ 23

TAM路徑分析 非標準化模型 http: //www. semsoeasy. com. tw/

報表解釋 http: //www. semsoeasy. com. tw/

TAM SEM 非標準化模型 http: //www. semsoeasy. com. tw/

SEM報表輸出 http: //www. semsoeasy. com. tw/

SEM報表輸出 http: //www. semsoeasy. com. tw/

Title 常用輸入語法(一) 配適度 測量量數 統計量 關鍵詞 卡方值 cmin 自由度 df TLI tli CFI cfi GFI gfi AGFI agfi 格式 format p-value p 群組 group 卡方/自由度 cmindf 模式 model RMSEA rmsea http: //www. semsoeasy. com. tw/ 50

Absolutely Model Fit 1. 2. 3. 4. 5. 6. χ2 (CMIN) small is better p值 (樣本大於 200不用報告) p>0. 05 Chi-square/df 3~1 (嚴謹) ~5 (寛鬆) GFI (配適度) >0. 90 AGFI (調整後…) >0. 90 RMR (殘差均方和平方根) <0. 05 (一般由SRMR取代, 可以不用報告) 7. SRMR (標準化…) <0. 05 8. RMSEA(漸近…) <0. 08(可接受) <0. 05(良好) http: //www. semsoeasy. com. tw/ 60

Why P-value is sig. ? • • In practice, chi-square is not considered to be a very useful fit index by most researchers. It is difficult to get a nonsignificant chisquare (indicative of good fit) when samples sizes are much over 200 or so. Maruyama, G. (1998). Basics of Structural Equation Modeling. Thousand Oaks CA: Sage. Tanaka, J. S. (1993). Multifaceted conceptions of fit in structural equation models. In K. A. Bollen, & J. S. Long (eds. ), Testing structural equation models. Newbury Park, CA: Sage. http: //www. semsoeasy. com. tw/

View interface properties • page layout – 調整繪圖區域的大小 http: //www. semsoeasy. com. tw/ 63

View interface properties • Formats – 可以用來增加 圖形估計參數 的小數位數 http: //www. semsoeasy. com. tw/ 64

View interface properties • Colors – 改變模型中 圖形的顏色 http: //www. semsoeasy. com. tw/ 65

View interface properties • Typefaces – 更改圖形上 顯示的字型 http: //www. semsoeasy. com. tw/ 66

View interface properties • pen width – 改變圖形上箭 頭的大小及框 線的粗細 http: //www. semsoeasy. com. tw/ 67

View interface properties • Misc (雜項) – 圖形是否 顯示標籤 http: //www. semsoeasy. com. tw/ 68

View Analysis Properties • Estimation http: //www. semsoeasy. com. tw/

How many sample size are need? • rule of 10: 10 observations per indicator (Barclay, et al. 1995; Chin 1998; Chin, and Newsted 1999; Kahai and Cooper 2003) • Bentler and Chou (1987) suggest sample size to free parameter are 5: 1 (資料符合常 態，無遺漏值及例外值情形下)，否則要10倍的樣 本數 http: //www. semsoeasy. com. tw/ 72

How many sample size are need ? 小樣本容易導致收斂失敗、不適當的解(違犯估 計) 、低估參數值及錯誤的標準誤 Hair, Joseph F. , William C. Black , Barry J. Babin , Rolph E. Anderson. (2009) Multivariate Data Analysis (7 th Edition). Englewood Cliffs, N. J. Prentice Hall. http: //www. semsoeasy. com. tw/ 73

http: //www. semsoeasy. com. tw/ 74

SEM參數設定原則 (Raycov & Marcoulides, 2006) Rule 1: All variances of independent variables are model parameters. Rule 2 : All covariances between independent variables are model parameters Rule 3. All factor loadings connecting the latent variables with their indicators are model parameters. Rule 4. All regression coefficients between observed or latent variables are model parameters Rule 5. The variances of, and covariances between, dependent variables as well as the covariances between dependent and independent variables are never model parameters. Rule 6. For each latent variable included in a model, the metric of its latent scale needs to be set. http: //www. semsoeasy. com. tw/ 75

All variances of independent variables are model parameters. http: //www. semsoeasy. com. tw/

All covariances between independent variables are model parameters http: //www. semsoeasy. com. tw/ 77

Amos路徑分析與SPSS迴歸比較 SPSS輸出結果 http: //www. semsoeasy. com. tw/ 78

All factor loadings are model parameters. http: //www. semsoeasy. com. tw/ 80

All regression coefficients are model parameters http: //www. semsoeasy. com. tw/ 81

The variances of, and covariances between, dependent variables as well as the covariances between dependent and independent variables are never model parameters http: //www. semsoeasy. com. tw/

For each latent variable included in a model, latent scale needs to be set http: //www. semsoeasy. com. tw/

Model identification • • p：All observed variables in model Degree of freedom (df)= (p)(p+1)/2 T-rule: free parameters T≦ df (必要條件) T<df：過度辨識 (over identified) T=df：恰好辨識 (just identified) T>df：不足辨識 (under identified) http: //www. semsoeasy. com. tw/ 85

TAM parameter estimation • P = 12 • df = 12*(12+1)/2= 78 • free parameters – – Rule 1: Rule 2: Rule 3: Rule 4: 16 0 8 5 • Model df= 78 -16 -8 -5=49 http: //www. semsoeasy. com. tw/

SEM模型中同時包含 一階及二階CFA構面 http: //www. semsoeasy. com. tw/ 89

SEM模型中同時包含 一階及二階CFA構面 http: //www. semsoeasy. com. tw/ 90

一階驗證性因素分析 http: //www. semsoeasy. com. tw/ 92

Modification Indices 用法 • 「View」 「Analysis properties」 「Output」 「Modification indices」 http: //www. semsoeasy. com. tw/ 95

組成信度(composite reliability; CR) • CR值是所有測量變項信度的組合，表示構面指 標的內部一致性，相當於cronbach’s α，CR愈 高表示構面的內部一致性愈高，0. 7是可接受的 門檻( Hair, 1997, Fornell and Larcker, 1981)。 (Jöreskog and Sörbom, 1996) http: //www. semsoeasy. com. tw/ 96

收斂效度 　 企業形象 　 　 轉換成本 　 　 　 滿意度 　 　 忠誠度 　 　 服務品質 　 　 EI 1 EI 2 EI 3 SC 1 SC 3 SC 4 SC 5 CS 6 CS 5 CS 4 CS 3 CS 1 AL 2 AL 3 tangible reliable response assurance empathy 參數顯著性估計 Unstd. S. E. 1. 000 1. 069. 882 1. 000. 990 1. 110. 920 1. 000 1. 184 1. 081 1. 107. 955 1. 000 1. 211 1. 168 1. 000 1. 340 1. 472 1. 374 1. 089 . 049. 055. 077. 082. 084. 118. 115. 116. 107. 074. 109. 122. 116. 108 t-value 21. 842 16. 178 12. 783 13. 515 10. 905 9. 993 9. 368 9. 569 8. 904 16. 449 15. 677 12. 341 12. 084 11. 841 10. 063 收斂效度 P *** *** *** *** Std. SMC 1 -SMC CR AVE . 866. 955. 769. 770. 848. 658. 638. 713. 683. 718. 623. 752. 933. 875. 668. 831. 819. 800. 657 . 750. 912. 591. 593. 719. 433. 407. 508. 466. 516. 388. 566. 870. 766. 446. 691. 671. 640. 432 . 250. 088. 409. 407. 281. 567. 593. 492. 534. 484. 612. 434. 130. 234. 554. 309. 329. 360. 568 . 900 . 751 . 848 . 584 . 808 . 457 . 891 . 734 . 870 . 576 http: //www. semsoeasy. com. tw/

區別效度 　 AVE 服務品質 . 576 . 759 轉換成本 . 584 . 431 . 764 企業形象 . 751 . 487 . 194 . 867 滿意度 . 457 . 737 . 318 . 359 . 676 忠誠度 . 734 . 510 . 380 . 501 . 536 服務品質 轉換成本 企業形象 滿意度 忠誠度 . 857 Square root of AVE in bold on diagonals Off diagonals are Pearson correlation of constructs http: //www. semsoeasy. com. tw/

二階模型設定 • Big-five personality 的高階模型 http: //www. semsoeasy. com. tw/

二階模型設定 • Big-five personality 的高階模型 101 http: //www. semsoeasy. com. tw/

二階模型設定 http: //www. semsoeasy. com. tw/ 102

二(高)階驗證性因素分析 http: //www. semsoeasy. com. tw/ 107

完整SEM模型分析 http: //www. semsoeasy. com. tw/ 109

共變異數矩陣的製作 • SEM模型分析中 Analysis Properties Output Sample moments打勾 • 執行分析 http: //www. semsoeasy. com. tw/

共變異數矩陣的製作 • View Text Sample Moments Sample Covariances • Sample Covariances按右鍵複製 http: //www. semsoeasy. com. tw/

分析報表重點提示 • Estimate – Scalar • • Regression weights Standardized Regression weights Variance Square Multiple Correlations • Model fit – CMIN, df, CMIN/df, GFI, AGFI, CFI, NNFI, RMSEA, SRMR http: //www. semsoeasy. com. tw/

報表解讀 • Regression Weights(非標準化迴歸係數) • 重點放在顯著性的報告 – C. R. (t-value) – P http: //www. semsoeasy. com. tw/ 116

報表解讀 • Standardized Regression Weights (標準化迴歸係數) • 值愈大表示自變數的影響力愈大 http: //www. semsoeasy. com. tw/ 117

報表解讀 • Square Multiple Correlations (SMC, R 2)： 多元相關平方 • 0. 19(small), 0. 33(medium), 0. 67(large) http: //www. semsoeasy. com. tw/

Questions? http: //www. semsoeasy. com. tw/ 120

三星統計服務有限公司 http: //www. semsoeasy. com. tw/ 121 http: //www. semsoeasy. com. tw/