I am developing a new construct. I am keen to understand conceptually, what should the new construct include? Should it have elements of cause and effects? In my view, this will make the definition of the new construct very comprehensive.
For a new construct, we should run the factor analysis for data reduction and for investigating whether there are underlying latent variables, or factors, that can explain the patterned correlations within a set of observed variables.
After that, construct reliability and validity should be tested through structural equation modeling.
In my opinion, the most important question we should ponder at great length when developing a new construct is should we develop a new construct in the first place? Before designing a new one, we first must answer how are existing constructs insufficient so that it justifies developing an entirely new one. If the causes and effects can be explained sufficiently by existing constructs, then what's the difference? The researcher should identify similar constructs and explain one by one why their new construct is different from those and how that difference matter.
Developing a new construct for the sake of itself would not be a significant contribution to science. The researcher must make a convincing argument to justify his new construct. I believe that the measurement validation phase is secondary to this phase. As for the validity testing of new constructs, I believe the discriminant validity should be emphasized more than the convergent validity.
Thank you Mahfuz and Imam for your responses. I was keen to understand how a new construct should be defined. Let's take the example of Employee Engagement or Job Satisfaction. Should one indicate in the definition what causes job satisfaction and what outcomes it leads to or should one limit just to the subjective feeling? Where should one draw the boundary? Job satisfaction cannot happen in isolation... there is a causative factor and the definition will be clear if one includes the cause too. Of course, in this example, there can be many causes and hence might be difficult to isolate it. If the cause and effects are very clear, should one include them in the definition of a new construct?
Just look at the existing overlapping constructs, identify the gaps and introduce yours. I have just investigated Social norm and its underlying internal/external dimensions. They both cannot serve as underlying dimensions yet they exist in the extant literature only to realise after long hustling and bustling.
- Badges
- Science topic