4.4.1.2 The Discriminant Validity Analysis
In order to ensure further, it is important to establish construct validity and the
discriminant validity of outer model. Hence, prior to hypotheses testing, discriminant
validity was ensured. Discriminant validity refers to the level to which items can
differentiate among different constructs in which it shows that the items of different
constructs are not overlapping. Additionally, discriminant validity of measures share
variance between each individual construct and hence, it should be higher than the
variance shared among specific constructs (Compeau, Higgins, & Huff, 1999).
In this study, the discriminant validity of measures was established through Fornell and
Larcker (1981) method, where the square root of AVE for the entire constructs was
replaced at the diagonal elements of the correlation matrix as demonstrated in Table 4.6
Seventeen items were dropped to enhance the result and those items are no more in this
study. All the other values are less than 0.90 and hence, the outer model‘s discriminant
validity was established as suggested by Hair, Black, Babin, and Anderson (2010). The
above results of the outer model‘s construct validity ensure that it is appropriate to test
the proposed hypotheses.
4.4.1.3 The Construct Validity
The construct validity and content validity is defined as the level to which the proposed
items suitably measure the concept of the constructs that they are designed to measure
(Hair, Black, Babin, & Anderson, 2010). Stated differently, items measuring a construct
should load higher on their respective constructs. Hence, items are considered through a
You've reached the end of your free preview.
Want to read all 4 pages?
- Summer '17
- Iking
