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

Predictive and Postdictive Validity

Predictive and postdictive validity concern the relationship of the scale under development to some external criterion separated in time and context (Netemeyer et al., 2003). More specifically, it is the extent to which a scale on a given construct is related to the scores on a criterion measure administrated (a) at a future point of time (i.e., predictive validity), or (b) at a previous point of time (i.e., postidictive validity) (Cronbach & Meehl, 1955). Having said that, the following research question was posed:

RQ 6: Does the eWOM trust measure significantly predict (postdict) eWOM-related consumer behaviors (e.g., opinion seeking, adoption, giving, passing)?

To assess predictive validity, a sample of undergraduate business students from an Austrian university was invited to participate in a short survey three months after the participants had handed in a questionnaire with the eWOM trust scale. A fraction of the participants were contacted in class and were asked to fill out a paper-based questionnaire, while to the second half an email was sent including a link to an online survey. In total, 89 usable questionnaires were obtained by this process. 64% of the sample were females and the average age was 22.5 years (range: 19 to 37 years). In the research instrument, participants were asked to rate a series of statements concerning their online review behavior within the last two months. Specifically, seven questions for measuring review purchase influence (Cheung et al., 2008; Lee & Koo, 2012; Park & Lee, 2009), six items for review seeking (Cheung & Hung, 2010; Chu & Kim, 2011; Park & Lee, 2009), and three items for review giving behavior (Bailey, 2005; Chu & Kim, 2011) were included. Respondents were asked to answer the questions on a 7-Point Likert scale ranging from 0 (strongly disagree) to 6 (strongly agree). The internal consistency measures and correlations are reported in Table 40. The eWOM trust measure and the scores of the measures of consumer review usage were strongly correlated, being .75 for purchase influence, .74 for opinion seeking, and .42 for opinion giving (all p < .001). These correlations represent a high predictive power of the eWT-S that is well above the regularly observed correlation between attitude scales and behaviours. According to Wicker (1969), correlations are seldom as high as .30 (cited in Hogg & Vaughan, 2011). In addition, considerable correlations among the three behavioral constructs (ranging from .44 to .84) - all significant on the .001 level - were revealed. These results indicated the ability of the new scale in helping to explain and predict future consumer behaviors.

Mean

SD

Developed measure of eWOM Trust

Future

purchase-

influence

Future eWOM

opinion

seeking

behavior

Future eWOM opinion giving

behavior

Developed measure of eWOM Trust

3.57

.86

(.95)1

Future purchase influence

3.05

1.49

75***

(.95)1

Future eWOM opinion seeking behavior

3.35

1.64

.74***

.84***

(.95)1

Future eWOM opinion giving behavior

1.74

1.65

.42***

44***

59***

(.83)1

Notes: Sample: 5 (n=89); Internal consistency estimates (Cronbach’s alpha); Pearson correlation, 2-tailed;

»»» = p < .001.

Another sample of university students was used to evaluate the scale’s postdictve validity. Here, a similar data collection procedure was applied. While participating in a marketing course, students were asked to fill out a questionnaire containing a series of 7-Point Likert items emphasizing their behavioral patterns concerning online customer reviews within the last two months. Specifically, the brief questionnaire included six items to measure the influence of online reviews on past purchases (Cheung et al., 2008; Gilly et al., 1998; obermiller & Spangenberg, 1998; Park & Lee, 2009), six items measuring past opinion seeking behavior (Bansal & Voyer, 2000; Chu & Kim, 2011), and two items targeting past opinion giving behavior (Chu & Kim, 2011). All items were measured with a 7-Point Likert scale ranging from 0 (stongly disagree) to 6 (stongly agree). Additionally, demographic variables were also included. After three months, all respondents received an email with a link to the online followup questionnaire containing the eWOM trust scale. On the introductory page, a definition of online customer reviews and some examples were provided. Then, the students were asked to complete the questionnaire for course credit. In total, 50 completed and assignable questionnaires were used for analysis. 70% of the sample were females and the average age was 23.9 years (range: 21 to 33 years). Table 41 reports the observed Pearson correlation coefficients among the items. Internal consistency of the constructs was assumed, as the Cronbach alphas were above the proposed threshold level. Exploratory factor analysis ensured unidimensionality of the constructs in advance. Again, all correlations showed expected patterns, as they were significant, considerable and positive ranging from .58 (for past purchase influence and opinion seeking behavior; both p < .001) to .33 (for past opinion giving behavior, p < .05). The analysis further showed that - as with the predictability sample - all behavioral constructs were significantly related. Taken together, pre- as well as postdictve validity were both supported by empirical data, indicating a high trust-behaviour consistency.

Mean

SD

Developed measure of eWOM Trust

Past

purchase-

influence

Past eWOM

opinion

seeking

behavior

Past e WOM opinion giving

behavior

Developed measure of eWOM Trust

3.50

.73

(.93)1

Past purchase influence

2.82

1.37

.58***

(.96)1

Past e WOM opinion seeking behavior

3.43

1.31

.58***

.80***

(.92)1

Past e WOM opinion giving behavior

1.95

1.35

.33*

57***

37**

(.79)1

Notes: Sample: 5 (n=50); 1= Internal consistency estimates (Cronbach’s alpha); Pearson correlation, 2-tailed; »»» = p < .001; ** = p < .01; * = p < .05.

 
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