CHAPTER 4 RESEARCH RESULTS
4.4 The Results of Objective 3: The Results Examine the Positive Influences of
4.4.3 The Observed Variation Relationship Analysis of Measurement
Chi-Square = 6.145; df = 2; Relative Chi-Square = 3.073; p-value = .050 GFI = .997; NFI = .997; TLI = .975; CFI = .998
RMSEA = .061; RMR = .004
Figure 4.3 The Results Verify the Construct Validity of Causal Variables on Model of Tourists’ Satisfaction (SATI) Confirmatory Factor Analysis
Techniques
4.4.3 The Observed Variation Relationship Analysis of Measurement
significance level of 0.01, the size of a relationship of 105 pairs of variables was at a moderate level (r < 0.4), a relationship of 14 pairs variables was at a moderate level (0.4 ≤ r < 0.6), and that of 26 pairs was rather high (0.6 ≤ r < 0.8). The variables that were correlated at the highest were perception (Informational Service: PERC (G)) and perception (Festival Product: PERC (B)) (r = 0.808) while satisfaction (Outcome Quality: SATI (Q)) and perception (Festival Product: PERC (B)) (r = 0.122) were correlated at the lowest, and satisfaction (Outcome Quality: SATI (Q)) and perception (Convenient Facility: PERC (C)) were not correlated.
Considered the statistics of Barlet’s Test of Sphericity, it was shown that the value was 7671.497, df = 171, P = 0.000. It revealed that the correlation matrix differed from the identity matrix with a statistical significance level of 0.01 which corresponded to the analysis result of Kaiser-Mayer-Olkin measure that was equal to 0.898. KMO close to 1 indicated that the observed variables were highly correlated and suitable for measuring a congruence between the research model and the empirical data in the future. The reason for testing the mentioned statistics as if the variables were identity matrix and not correlated, they were not suitably proposed for the factor analysis, as shown in table 4.16
Table 4.17 The Correlation Coefficient, the Mean and Standard Deviation of the Observed Variables (n = 550)
4.4.3.1 Detecting Multicollinearity
Multicollinearity occurs when two or more explanatory variables are highly linearly connected in a study and get an effect on regression analysis. Each independent variable’s tolerance and VIF (Variance Inflation Factors) are two collinearity diagnostic criteria that aid in the detection of multicollinearity. When the tolerance is more than 0.1, these variables would not be highly significantly proportional or correlated with each other and the value of VIF is not greater than 10 (Belsley, 1991).
Table 4.18 The Regression Coefficient of Explanatory Variables for Tourists’
Perception, Experience, Satisfaction and Tourists’ Loyalty Results
(n = 550) B SEB Beta t-value P Tolerance VIF
(Constant) .056 .155 - .364 .716 - -
PERC (A) .150 .033 .170 4.615 .000** .522 1.917
PERC (B) .040 .039 .050 1.025 .306 .297 3.371
PERC (C) .031 .029 .048 1.050 .294 .338 2.957
PERC (D) -.116 .039 -.147 -2.947 .003** .285 3.504 PERC (E) -.099 .053 -.109 -1.876 .061 .208 4.810
PERC (F) .030 .037 .036 .813 .417 .356 2.811
PERC (G) -.058 .043 -.069 -1.328 .185 .259 3.862
EXPE (H) .056 .047 .064 1.191 .234 .242 4.134
EXPE (I) .143 .033 .196 4.291 .000** .340 2.945
EXPE (J) .007 .035 .009 .208 .835 .395 2.530
EXPE (K) .324 .043 .332 7.576 .000** .368 2.718 SATI (L) .058 .040 .068 1.465 .144 .331 3.018 SATI (M) -.015 .037 -.018 -.401 .689 .370 2.703 SATI (N) -.009 .031 -.011 -.306 .760 .506 1.976 SATI (O) .104 .032 .129 3.282 .001** .458 2.185 SATI (P) .049 .038 .060 1.280 .201 .322 3.107 SATI (Q) .285 .034 .353 8.308 .000** .390 2.562
** P < 0.01; R = 0.790, R2 = 0.624, Adjusted R Square = 0.612, F = 52.039**
According to table 4.18, revealed that a relationship between independent variables and tourists’ loyalty categorized at a high level (R=0.790) and all independent variables could jointly predict the quality of life as high as 62.4 percent (R2 = 0.624) which proposed great enough relative to a statistical significance level of 0.01 (F = 52.039). It meant that the studied factors can be used for estimation.
Moreover, Table 4.10 indicated that the lowest tolerance value was 0.208 and the greatest value 0.552 whereas the lowest value was higher than the tolerance limit; Tolerance > 0.1. The value of VIF was 1.917 the minimum and 4.810 the largest whereas the largest value was lower than the VIF limit. It could be seen that each variable had variance that did not overlap with the other variables.
Therefore, there were no problems with multicollinearity or high correlation and the variables can be used for the analysis of the structural equation model.
4.4.3.2 The Reliability and Validity of the Study
The reliability coefficient, known as Cronbach’s alpha coefficient, refers to the degree of dependability, consistency, or scale stability. The indicators of highly reliable constructs are highly interrelated and indicate that they all seem to measure the same (Hair et al., 2010). The general acceptance of the lower limit for Cronbach’s alpha is 0.7, although it may decrease to 0.60 in exploratory research (Hair et al., 2010). In this study, the reliability coefficients of 21 factors tested with 550 samples (see Table 4.18)
The average variance extracted (AVE) is a measure of convergence among a set of items representing a construct. The average percentage of variation is explained among a construct’s items (Hair et al., 2010). Fornell & Larcker (1981) indicated that the AVE value should exceed 0.5 for a construct, and the reliability composite reliability (CR) value should exceed 0.60 for a construct. (Diamantopoulos
& Siguaw, 2000) of all constructs.
Table 4.19 Reliability and Validity
(n = 550)
Latent variable
Reliability Validity
Alpha
Coefficient (CA) CR AVE
Tourists’ perception (PERC) .911 .597
PERC (A) .920
PERC (B) .918
PERC (C) .920
PERC (D) .916
PERC (E) .916
PERC (F) .917
PERC (G) .917
Tourists’ experience (EXPE) .818 .534
EXPE (H) .916
EXPE (I) .917
EXPE (J) .920
EXPE (K) .920
Tourists’ satisfaction (SATI) .899 .573
SATI (L) .919
SATI (M) .921
SATI (N) .920
SATI (O) .921
SATI (P) .921
SATI (Q) .922
Tourists’ loyalty (LOYA) .732 .578
LOYA (R) .920
LOYA (S) .919
According to table 4.19, revealed that with regards to the composite reliability of latent variables (CR), Tourists’ perception (PERC), Tourists’ experience (EXPE), Tourists’ satisfaction (SATI), and Tourists’ loyalty (LOYA) had the composite reliability (CR) ranging from 0.732 to 0.911 which considered quite high as it was greater than 0.70.
Tourists’ perception (PERC), Tourists’ experience (EXPE), Tourists’
satisfaction (SATI), and Tourists’ loyalty (LOYA) had the average variance extracted (AVE) of factors and latent variables ranging from 0.534 to 0.597. It implied that all observed variables can describe quite significantly the variance of factors and latent variables which is greater than 0.70 in each factor.
With reference to the mentioned statement, it can be summarized that the composite reliability (CR) is quite high meaning greater than 0.50 and the observed variables can describe quite significantly the variance of latent variables (AVE) in each factor which is greater than 0.70. Regarding analysis result of the reliability of observed variables (Internal consistency reliability), it can be seen that the observed variables had high level of reliability to measure each factor. It indicates that from the assessment of the measurement model, there is an outstanding evidence showing that defining all factors and latent variables are all correct and reliable.