Now that EFA shows uni-dimensionality with only one factor extracted and CFA conducted ;then what shall be next steps for the following in scale development/construction.
i) convergent validity
ii)discriminant validity
Hello Yumna,
Congratulations on finding and confirming a simple structure for your set of indicators in the scale.
In general, how and what kind of evidence for score validity is required depends on the intended use of the scores from your scale. Given that you're seemingly primed to work on avoiding the outcomes of either the "jingle" or "jangle" fallacies, here's the sequence I'd suggest:
1. Look for other measures/scales purporting to measure the same or similar characteristic/attribute, and determine the degree of correspondence across them (e.g., convergent evidence).
2. If you still have the energy and resources to follow that up, then comparing scores from your scale with those from measures purporting to tap different characteristics/attributes would be helpful (discriminant evidence).
Note: Both aims can be accomplished within the same data collection. You do need to be mindful of not asking for so much time and effort on the part of participants that you fail to get full information.
Good luck with your work.
If you have a model that specifies the relationships between your independent and dependent variables and any relevant covariates, then you can use this to assess convergent and discriminant validity. In other words, the validity tests do not need to be a separate step from your overall analysis.
The following sources may help:
(17) (PDF) An Easy Approach to Exploratory Factor Analysis: Marketing Perspective Noor Ul Hadi (researchgate.net)
(17) (PDF) Specifying the Problem of Measurement Models Misspecification in Management Sciences Literature (researchgate.net)
Yumna Ali Hi, as you have said that you have applied CFA after EFA, I think you may need to focus more on the model fit indices in order to prove your construct validity. Convergent and discriminant validity although they come under construct validity however, it may not be necessary since you have only one construct.
- Badges
- Science topic
- Similar topics
- Psychology