驗證式因素分析的深度探討 - hwai.edu.tw

1 Amos SPSS Amos 2 200120062009 3 and Gerbing (1988), Structural Equation Modeling in practice: a Review and Reccomended Two-Step Approach, Psychological Bulletin, 103, 3. , (1989). Structural equations with latent variables. New York: Wiley. , , Issues And Opinion on Structural Equation Modeling , MIS Quarterly, (1), , , Adamantiosand Judy A. Siguaw(2000), Introducing LISREL: A Guide for the Uninitiated. London: Sage and Larcker (1981), Evaluating Structural Equation Models with Unobservable Variables and Measurement Error, Journal of Marketing Research, 18 (February), 39-50. , K. O., & Wong, S. P. (1996). Forming inferences about some intraclasscorrelation coefficients. Psychological Methods, 1(1), Rex B.

2 (2005). Principles and Practice of Structural Equation Modeling (2. nd. ed.). New York: Guilford Press. , Koufteros,& Pflughoeft (2003) Confirmatory analysis of computer self-efficacy. Structural Equation Modeling, 10(2): 263-275. 4 vs. Cronbach s ( ) (Intraclasscorrelation coefficient, ICC) (Composite Reliability, CR) (Average Variance Extracted, AVE) (convergence) (discriminant) (cross validity) Q & A 5 Commentary: Issues and Opinion on Structural Equation population from which the data sample was distribution of the data to determine the adequacy of the statistical estimation conceptual model to determine the appropriateness of the statistical models results to corroborate the subsequent interpretation and conclusions.

3 6 1. ( )2. ( )3. ( ) 7 ( ) 8 not everything that can be counted counts, and not everything that counts can be Einstein 9 Cornbach s (intra-class correlation, ICC) ( ) ( ) ( ) ( ) MultiTrait-MultiMethod(MTMM) 10 Cronbach salpha 11 Cronbach s 12J1,J2 & J4,J5 13 Cronbach s 22(1)1iSKKS = __1(1)sKrKr =+ 14 Cronbach s Cronbach sAlpha ~ Item-total correlation (Nancy et al.)

4 P 67)SPSS for Intermediate Statistics Use and Interpretation (2nd Ed.) (2005) 15 (IntraclassCorrelation Coefficient, ICC) ICCs ( ) Inter-rater reliability ICCs two-way ANOVA ICC : (MaGraw& Wong, 1996) 16 (IntraclassCorrelation Coefficient, ICC) 17 ICC 18 19 F1e1F2e2e3e4e5e6x1x2x3x4x5x6 Charles Spearman 20 CFA 1. 2. ( )3. F1e1F2e2e3e4e5e6x1x2x3x4x5x611 Karl Joreskog 21 ( ) F1e1F2e2e3e4e5e6x1x2x3x4x5x611 22 ( ) F1e1F2e2e3e4e5e6x1x2x3x4x5x611F31 23 CFA 111111111111111111111 1111111111111111 11111 24 ?

5 Jobpositione111incomee21educatione31marr iagee41sexe51agee61 25 EFA CFA loading loading ( ) CFA EFA ML PCA (theory-driven) (data-driven)CFAEFA 26 SEM S ( ) ( ) (model implied covariance) S ( ) 27 (ML, ADF, WLS, ULS) CFA ( ) S ( ) 1. 25 2. 28 ? S (ML) 29 SEM 1 1 CFA dd x1x2 y1y2eexx eeyy eeyy eexx 11 30 CFA 1.

6 (Bollen, 1989)2. ~ (Bollen, 1989)3. (Bollen, 1989)4. 31 CFA CFA 32 CFA EFA LOYALTYLOYAL1e111 LOYAL2e21 LOYAL3e31 LOYAL4e41 LOYAL5e51 LOYAL6e61 Total Variance of VarianceCumulative %Total% of VarianceCumulative %Initial EigenvaluesExtraction Sums of Squared LoadingsExtraction Method: Principal Component Analysis. 33 CFA ( )Loyalty CFAchi-square= degree of freedom=9norm chi= agfi=.603rmsea=. Loyalty CFAchi-square=.762 degree of freedom=2norm chi=.381gfi=.999 agfi=.994rmsea=. 35 36 How to deal with non-normality? 37 1 LOYALTYLOYAL1V1e1W11 LOYAL4V2e4W21 LOYAL5V3e5W31 LOYAL6V4e6W41 ( ) F1F2F3 2e211 4e41 5e51 6e61 7e71 1e81 2e91 4e111 5e121 6e131 1e1411 2e151 3e1611 39 => (factor loading) => (CR)> > salpha> 40 (composite reliability.)

7 CR) CR ( Hair,1997) , Fornelland Larcker(1981) ( )2/(( )2+( )) (J reskogand S rbom, 1996) 41 (Average variance extracted AVE) AVE VE Fornelland Larcker(1981) ( ) AVE= ( 2)/(( )2+( )) (J reskogand S rbom, 1996) 42 CFA Chi square = = 62 norm-chi= = .001agfi=.919 gfi=.945rmsea=.051F1F2F3 7e7 1e8 2e9 6e13 2. 13. 44 1. ( ) 2. bootstrap 95% 1 (Torkzadeh, Koufteros, pflughoeft , 2003) 1 reject ( ) (Anderson and Gerbing,1988, Bogozzi et al.

8 , 1991) AVE (Fornell and Larcker, 1981) AIC ( )(Kline, 2005, p151) 45 Chi square = = 62 norm-chi= = .001agfi=.919 gfi=.945rmsea=. 2e21 4e41 5e51 6e61 7e71 1e81 2e91 4e111 5e121 6e131 1e141 2e151 3e161cov12cov13cov23 46 3 2 1 6 5 4 2 1 7 6 5 4 1 2 3 2 1 6 5 4 2 1 7 6 5 4 2 47 ( ,bootstrap) 48 Bootstrap 49 SEM 50 AVE AVE 51 ECVI AIC ECVI AIC 2 1 ECVIAICM odel 23 ReducedFactor 13 ReducedFactor 12 Full model3 factors 52 CFA Chi square = = 62 norm-chi= =.

9 001agfi=.919 gfi=.945rmsea=.051F1F2F3 7e7 1e8 2e9 6e13 square = = 62 norm-chi= = .001agfi=.919 gfi=.945rmsea=. 53 (Browne, 2006)1. CFA 2. 3. 4. ( ) 5. ( )6. 7. ( )8. 54 ( ) Cross validity - SEM validity generalization CFA 55 Diamantoulosand Siguaw(2000) p:130 56 ( SPSS ) Amos ( ) ( )

10 57 SPSS 58 59 vvv1_1F1vvv2_1F2vvv3_1F3 2v1_1e211 4v2_1e4a1_11 5v3_1e5a2_11 6v4_1e6a3_11 7v5_1e71 1v6_1e81 2v7_1e9a4_11 4v8_1e11a5_11 5v9_1e12a6_11 6v10_1e131 1v11_1e1411 2v12_1e15a7_11 3v13_1e16a8_11ccc1_1ccc2_1ccc3_1a9_11a10 _1vvv1_2F1vvv2_2F2vvv3_2F3 2v1_2e211 4v2_2e4a1_21 5v3_2e5a2_21 6v4_2e6a3_21 7v5_2e71 1v6_2e81 2v7_2e9a4_21 4v8_2e11a5_21 5v9_2e12a6_21 6v10_2e131 1v11_2e1411 2v12_2e15a7_21 3v13_2e16a8_21ccc1_2ccc2_2ccc3_2a9_21a10 _2 60 61 62 63 SPSS & Amos 64 http://