|
Ttp+Online Banking
|
bet | 16/17 | Sana | 24.03.2017 | Hajmi | 0.62 Mb. | | #1948 |
IV. EMPIRICAL RESULTS
A. Confirmatory Factor Analysis
The research model has five constructs with interrelated dependence relationships or causal paths, requiring a structural equation model (SEM) analysis [9, 41]. SEM analysis requires constructs to be assessed rigorously by confirmatory factor analysis (CFA) [1, 29, 41, 93], to examine convergent and discriminant validity. CFA results were obtained using software package LISREL 8.30.
(1) Sample Size:
Experts consider the minimum sample size for reliable SEM analysis to be from 100 [9], to 150 [3] to over 200 [10]. Because our research model is relatively simple, with only five constructs, the sample size of 199 is considered adequate.
(2) Item Reliability:
Item reliability denotes the amount of variance in an item due to its underlying construct. Table 2 shows that t-values for all of the standardized factor loadings of items are significant at p<.01, indicating item reliability. Additional evidence was found using average SMC (squared multiple correlation), which denotes the explanatory power of items related to latent variables. SMC figures in Table 2 lie between .73 and .84, indicating item reliability.
|
| |