Menu
Home
Log in / Register
 
Home arrow Engineering arrow Measuring Electronic Word-of-Mouth Effectiveness: Developing and Applying the eWOM Trust Scale
Source

Exploratory Factor Analysis

F or item reduction and testing the dimensionality of the construct, a principal component factor analysis (PCA) was again carried out on the 25 manifest indicators after the factorability of the data was verified. EFA resulted in a definite factor structure similar to this research’s earlier findings. All items - except Ab6 (“relevant”) - loaded considerably on their intended components. During the reliability stage, this item showed a clear relationship with the ability dimension. However, according to data from the new sample, the item is strongly associated with the integrity/honesty dimension. Hence, this item was considered ambiguous and was deleted.

Subsequently, by evaluating the data by the various decision rules described earlier, the number of factors was restricted to five and the model was re-estimated. This resulted in the explanation of 73.70% of the total variance. All remaining 24 items exhibited significant loadings on their appropriate component of the sub-dimensions of eWOM trust (see Table 27). Due to sample size, factor loadings of .40 and higher were regarded as significant (Hair et al., 2010). Two items (In7, Be6) showed significant, but relatively weak loadings (.43 and .44) and also low communalities; however, the rest of the items passed the .60 benchmark. Considerable crossloadings (>.32) were not detected.

Table 27: Results of the EFA (Sample 4)

Sub

dimension

Variance

Explained/

Eigenvalue

Item

Factor Loading

Communality

MS4

Ability

4.80% / 1.15

Ab7

Ab8

Ab9

Ab10

Ab11

.61

.57

.92

.61

.87

.78

.71

.80

.76

.76

.96

.98

.96

.98

.98

Integrity/

Honesty

51.51% / 12.36

In2

In3

In4

In5

In6

In7

In9

In10

.78

.87

.98

.73

.81

.43

.77

.92

.74

.73

.80

.75

.79

.59

.69

.64

.98

.97

.97

.97

.97

.99

.98

.98

Benevolence

3.87% / .93

Be1

Be2

Be3

Be6

.90

.76

.89

.44

.77

.70

.75

.50

.90

.94

.91

.97

Willingness to rely

10.44% / 2.51

Wi1

Wi4

Wi5

Wi8

.94

.90

.89

.84

.83

.79

.79

.80

.93

.95

.95

.96

Willingness to depend

3.07% / .74

Wi2

Wi6

Wi7

.86

.70

.77

.74

.73

.76

.96

.97

.96

Notes: Total variance explained: 73.70%; Extraction method: Principal Component Analysis; Rotation method: Promax with Kaiser Normalization; Rotation converged in 7 iterations; Kaiser-Meyer-Olkin Measure of Sampling Adequacy (MSA) .96 and Bartlett’s test of Sphericity: sig. .001; Effective sample size: n=526; Item to respondent ratio: approx. 1:21.

In addition to the explorative factor analysis, the individual item properties were examined. All items demonstrated desirable qualities advanced by established scale development guidelines (DeVellis, 2012), including means that are close to the center of the theoretical range, relatively high item variances, and the correspondence of the observed item ranges to the theoretical ranges.

 
Source
Found a mistake? Please highlight the word and press Shift + Enter  
< Prev   CONTENTS   Next >
 
Subjects
Accounting
Business & Finance
Communication
Computer Science
Economics
Education
Engineering
Environment
Geography
Health
History
Language & Literature
Law
Management
Marketing
Mathematics
Political science
Philosophy
Psychology
Religion
Sociology
Travel