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

Reliability and Generalizability

In general, reliability describes the consistency of a measure. That is, reliability is concerned with that portion of measurement that is due to permanent effects that persist from sample to sample (Netemeyer et al., 2003). The procedures regularly discussed in literature can be grouped into two general types: (1) test-retest (temporal stability) - which targets the correlation between the same person’s score on the same set of items at two points of time; and (2) internal consistency - which reflects the interrelatedness among items or set of items in the scale. Both issues are addressed in this thesis. While the latter is assessed through multiple criteria (individual corrected item-to-total correlations, the average inter-item correlation among scale items, and a number of reliability coefficients like Cronbach’s alpha and AVE) (Churchill, 1979; Cronbach, 1951; MacKenzie et al., 2011; Nunnally & Berstein, 1994) along the stages of the scale development process, the former is discussed at this point in greater detail.

As mentioned earlier, test-retest reliability cares about the stability of a respondent’s items responses over time (Netemeyer et al., 2003). Typically, a test-retest or “stability” coefficient - which is the correlation between the same measures (and sample) on two different assessment occasions - is used to assess the scale’s consistency. In cases where the magnitude of the stability coefficient is low (with no change in the construct over time), then the reliability of the measure is questionable. Assessing the test-retest reliability is valuable since it is able to provide confidence that the measure reflects the construct and the scale is generalizable to other assessment occasions or is stable over time (DeVellis, 1991; Haynes et al., 1999; Nunnally & Berstein, 1994). Having conceptualized eWOM trust as a stable attitude, this is a critical issue. Recognizing the importance of such an assessment, the following research question is proposed:

RQ 7: Is there a significant and considerable correlation between the measurements of eWOM trust on two different assessment occasions?

Generalizability is another critical issue for scale development. Research ventures that strive for a better understanding of this issue aim to clarify the extent to which a scale can be generalized to other situations, conditions or slightly other contexts under which it is assumed to be applicable (Netemeyer et al., 2003). The following three guiding questions could be raised with respect to the developed eWOM trust measure. First, eWOM trust per se is defined as a relatively stable global attitude concerning the reliance on customer reviews as an institution. Is the scale then also applicable to other contexts like measuring generalized trust in specific eWOM platforms? Second, eWOM trust in general is theorized to influence consumer attitudes and reactions towards individual review messages to a certain degree. Hence, one may ask whether such a relationship truly exists or not. Third, the new measure is developed in Germanspeaking countries (Germany and Austria); therefore, a valid question is: does the scale generalize and replicate across other cultures and languages? In the following, these critical issues will be discussed.

Found a mistake? Please highlight the word and press Shift + Enter  
< Prev   CONTENTS   Next >
Business & Finance
Computer Science
Language & Literature
Political science