I want to know is there any method of establishing construct validity of scales other than EFA and CFA? I see that many researchers use Pearson's Product Moment Correlation method with the help of SPSS. They did bivariate correlations based on scale dimensions with total scores of scale. Is this procedure valid? If yes, please provide references. The example is below:
correlation coefficients of all dimensions of TQLS
S.No Dimensions r value
1 PWB .828
2 PYH .808
3 IPD .715
4 PH .718
5 FS .568
Total .727
Hello Barkat,
actually, validation is something that is not bound to any analytical approach and a factor model CAN be one approach if the set of measures concurs with the factor structure.
More generally, validation means that you choose an approach that is critical--which means that it will show some result if the measure is valid and will not show the result if it is invalid or measures something else. For instance, a correlation matrix with a set of validation constructs can help. This matrix should consist of variables with which the tested measure should correlate (and beyond mere significance to a substantial amount) and it should contain variables with which the measure should NOT correlate. And finally, a third and strong test would be to include measures of similar constructs to infer whether your measure isn't not only new wine in old bottles. It is no accident that "validation papers" routinely skip this part :)
A different strategy could be to apply some experimental design. For instance, study 1 in the paper by Antonakis et al. used a simpel scenario study to analyse the validity:
Antonakis, J., & House, R. J. (2014). Instrumental leadership: Measurement and extension of transformational–transactional leadership theory. The Leadership Quarterly, 25(4), 746-771. https://doi.org/https://doi.org/10.1016/j.leaqua.2014.04.005
That being said, I would emphasize the immense role of qualitative/cognitive probing. When you have a survey, and you discuss items with respondents, you learn so much about the validity of the measure.
Alaimo, K., Olson, C. M., & Frongillo, E. A. (1999). Importance of cognitive testing for survey items: An example from food security questionnaires. Journal of Nutrition Education, 31(5), 269.-275. https://doi.org/10.1016/S0022-3182(99)70463-2
Hope that helps
Holger
For convergent validity you can assess inter-item correlation and item- total scale correlation. For divergent validity, you can assess the correlation between the totals for the different scales(domains) of your questionnaire. This could be the first step in construct validity
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publication