Dear SmartPLS-users,
I have a question.
I'm writing my master's thesis about loyalty creation in the news media business. One of my research objectives is to measure the impact of several news brandpersonality dimensions (Sincerity, Ruggedness, Trustworthiness etc.) on reader loyalty by measuring their congruity with their perception of the news media along those dimensions.
Intuitively I think it makes sense to define Brandpersonality as a Second-order construct in a reflective-formative model, where the underlying dimensions are measured reflectively as First-order constructs. But the most valid way of setting up that kind of model in SmartPLS is by use of the repeated indicator approach (Mode B as described in Becker et al, 2012).
My suspicion is, whether the weights from each of the First-order constructs on the Second-order construct (Brandpersonality) becomes skewed in some kind of way when using the repeated indicator approach.
Thus, my question is: How do I measure the total effect of each First-order construct on Loyalty when using the repeated indicator approach?
Is it as simple as calculating indirect effects? Eg. just multiplying the path coefficient from the First-order construct on the Second-order construct with the path coefficient from the Second-order construct to the endogenoues construct (loyalty), in the following manner:
Rugged -> Brandpersonality -> Loyalty
Sincerity -> Brandpersonality -> Loyalty
etc.
Hope you can help!
Best,
Jacob