5. Results and discussion
5.3. Descriptive analysis
Exploratory factor analysis was applied for the identification of the core factors affecting the
consumer brand engagement. The preliminary analysis was conducted using SPSS version 21 to test for the reliability of scales, constructs’ items and correlation.
5.3.1. Factor analysis
Prior to the main analysis, an exploratory factor analysis, further referred to as EFA, dimension reduction technique was implemented to explain the maximum amount of variance with the fewest number of principal components. As a result of conducting factor analysis, a
Kaiser-Meyer-Olkin (KMO) and Bartlett’s test of sphericity measures were obtained in order to test the sample’s adequacy and appropriateness of factor analysis. The KMO value varies between
Table 5. Constructs’ mean, std. deviation, Skewness, Kurtosis.
Construct Mean Std. Deviation Skewness Kurtosis Satisfaction 4.8398 1.13640 -.360 -.283
Trust 4.5014 1.22808 -.322 .127
Commitment 3.6397 1.31120 .286 -.296 Involvement 4.0478 1.37713 .200 -.419 Participation 4.5178 1.11638 -.178 .064 Self-expressive brand 3.3386 1.27687 .092 -.271 Cognitive processing 4.3671 1.17542 -.131 .144 Activation 5.0715 1.39306 -.624 -.191
Affection 4.4687 1.17233 -.201 .099
Brand loyalty 4.3828 1.36390 -.148 -.373 Marker variable 4.2038 1.13710 .014 .068
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
0 and 1, the higher it is, the higher the sampling adequacy. Normally, the value should be more
than .6 to be suitable for conducting factor analysis, in this case the obtained KMO value was .962 indicating that the sample is satisfactory. Also, Barlett’s test of sphericity lends
confidence the significance of the study by indicating the validity of the data collected by testing
for possible redundancy between the variables and should be less than 0.05 to demonstrate the suitability of factor analysis. Under a 95% level of significance, α = 0.05 the obtained p-value
of .000 is less than 0.05, thus proving the factor analysis valid. In addition, the principal factor analysis revealed that only nine of the components have shown eigenvalues above 1, with the first factor accounting for a 45.5% of total variance accounted by all factors; second 7.5%; third 5.9%; fourth 4.4%; fifth 3.8%; sixth 2.8%; seventh 2.1%; eights 1.9%; and ninth 1.7%. Those nine factors have accounted for 75.57% of total variance. Also, the scree plot was used to present the Eigenvalue against each factor. The graph shows that after factor one the curvature of the scree plot changes sharply, with the visible slight drop continuing to factor ninth.
5.3.2. Reliability of the constructs
Customer brand engagement was measured by 56 items. Table 6 presents the measures, and respective factor loadings of scale’s items that were used for further analysis. Several of the
initial items were dropped during EFA because they did not meet the suggested minimum
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
required value for factor loading of .5 (Hair et al., 2010) or they were found to be correlating with an unwanted factor. The internal consistency of constructs was measured by a reliability score- Cronbach’s alpha coefficient. The analysis resulted in Cronbach’s α values ranging from .82 to .95. The Cronbach’s alpha values are also reported in Table 6. Study by Nunally (1978) indicates that values should be equal to or above .70, in order to provide a reliable and rigorous measure. The presented values indicate that the items are related to their respective constructs.
‧
CSAT1 Overall I am satisfied with this brand .912 .948
CSAT2 I am very happy with this brand .903 .946
CSAT3 I am very satisfied with this brand .894 .946
CSAT4 I believe that using this brand is usually a very satisfying experience .878 .948 CSAT5 I am very satisfied with the service provided by this brand .870 .944 CSAT6 This brand does a good job of satisfying my needs .856 .953 CSAT7 I made the right decision when I decided to use this brand .833 .947 CSAT8 The service-products provided by this brand is very satisfactory .811 .951
Trust TRU1 I trust this brand .895 .832
TRU2 This brand is safe to use .891 .892
TRU3 This brand is an honest brand .880 .839
TRU4 I rely on this brand .781 .832
CCOM1
I would be willing to postpone my purchase if the products of this brand were
temporarily unavailable .834 .873
CCOM2
I am so happy with this brand that I no longer feel the need to watch out for
other alternatives .828 .865
CCOM3
I am willing to make small sacrifices in order to keep using the products of
this brand .827 .861
CCOM4 I would stick with this brand even if it would let me down once or twice .824 .869
CCOM5 I am very loyal to this brand .817 .869
SEBR1 This brand reflects my personality .948 .943
SEBR2 This brand is an extension of my inner self .945 .930
SEBR3 This brand mirrors the real me .937 .930
SEBR4 This brand symbolizes the kind of person I really am inside .895 .932
SEBR5 This brand contributes to my image .841 .955
CINV1 This brand is very important to me personally .929 .927
CINV2 Because of my personal values, I feel that this brand is the brand that ought
to be important to me .918 .925
CINV3 Because of my personal attitudes, I feel that this brand is the brand that
ought to be important to me .912 .922
CINV4 Compared with other brands, this brand is important to me .905 .929
CINV5 I’m interested in this brand .844 .943
Participation PAR1 I make constructive suggestions to this brand on how to improve its offering .902 .757 PAR2
If I have a useful idea on how to improve this brand, I give it to someone at
the firm .877 .734
PAR3 I let this brand know of ways that it can better serve my needs .873 .760 PAR4 I spent a lot of time sharing information with others about this brand .616 .887 CPRO1 I think about this brand a lot when I’m using it. .935 .787 CPRO2 Using this brand gets me to think about this brand .916 .747 CPRO3 Using this brand stimulates my interest to learn more about this brand .827 .913 Activation ACT1 Whenever I’m using telecommunication services I usually use this brand .937 .920
ACT2 I use this brand the most .931 .806
ACT3 I spent a lot of time using this brand compared with other brands .861 .815
Affection AFF1 Using this brand makes me happy .944 .893
AFF2 I feel good when I use this brand .925 .880
AFF3 I feel very positive when I use this brand .924 .891
AFF4 I’m proud to use this brand .821 .941
Brand loyalty BLOY1 In the future, I will be loyal to this brand .932 .881
BLOY2 This brand will be my first choice in the future .927 .894
BLOY3 I will buy this brand again .917 .879
BLOY4 I will not buy other brands if this brand is available for sale .842 .930 0.920
Table 6. Measures, factor loadings and Cronbach’s alpha of contructs.
‧
Table 7 shows the correlations among the constructs of the conceptual model, the means and standard deviations.
Table 7. Correlations, means and standard deviations.
5.5. Common method bias
Common method variance (CMV) is a potential concern in behavioral research that can lead to possibly misleading conclusions. Podsakoff et al. (2003) indicate that it is attributable to the measurement method, rather than the constructs the measures represent. Common sources of method biases include common source/rater effects, item characteristics, and item/measurement context effects (Podsakoff et al., 2003). As this research used a self-administered survey, i.e.
data was gathered via the same source and from the same respondents, the occurrence of CMV
Constructs 1 2 3 4 5 6 7 8 9 10 11
Mean 4.872 3.633 4.502 4.066 4.525 3.343 4.365 4.463 5.051 4.394 4.170
SD 1.119 1.272 1.234 1.360 1.126 1.277 1.179 1.180 1.388 1.332 1.131
**. Correlation is significant at the 0.01 level (two-tailed).