臺大管理論叢 NTU Management Review VOL.29 NO.1

Post-Disaster Grain Supply Chain Management with Supplier Hoarding and Regime Intervention 36 results are shown in Table 2. Additionally, we tested the construct reliability on Cronbach’s α statistics. The results of all Cronbach’s α values beat 0.7 (Cronbach, 1951). Another important test is to calculate measurement errors, i.e., common method bias; this study utilized the Harman’s single-factor model (Sanchez and Brock, 1996) to assess the statistical significance. From the unrotated principal components analysis, it is determined all of the factors have eigenvalues greater than 1, which calculates to 68.2% and 21.3% of the total variance and the first factor. The fit indices, (χ2/df = 11.8, Non- Normed Fit Index (NNFI) = 0.63, Comparative Fit Index (CFI) = 0.69, Root Mean Square Error of Approximation (RMSEA) = 0.15), of Harmon’s one-factor test show poor support for the common method bias upon application of the confirmatory factor analysis. In general, we can ignore the common method bias in this study. Table 2 Results of the Measurement Model Construct Items Factor Loading S.E. SMC CR AVE Supplier belief in future earnings (FE) FE1 1.000 .624 .734 .416 FE2 .873*** .112 .492 FE3 .619*** .095 .270 FE4 .640*** .097 .281 Supplier attitude toward weakened operational shock-wave after disaster (ATT) ATT1 1.000 .308 .736 .493 ATT2 1.249*** .195 .799 ATT3 .740*** .108 .380 Perceived relationship quality in main business partner (RQ) RQ1 1.000 .683 .714 .462 RQ2 .722*** .113 .352 RQ3 .735*** .116 .343 Regime intervention (RI) RI1 1.000 .356 .722 .466 RI2 1.239*** .190 .566 RI3 1.140*** .171 .479 Supplier hoarding intention (SH) SH1 1.000 .358 .722 .466 SH2 1.113*** .170 .535 SH3 1.143*** .170 .511 *** Denotes significance at p < .001 Note: S.E. means standardized error.