I am comparing against a previously validated questionnaire which is similar i some respects.
In my opinion, the most comprehensive method to determine covergent and discriminant construct validity is to compute a multitrait-multimethod-matrix.
Dear Ankur,
If I understand your question correctly it sounds like you could use Exploratory Factor Analysis and compare the correlations within and between your factors (the former should be higher than the latter) - please see my study guide.
Easiest way to accomplish this use pearson correlation, Campell and Fiske (1959) approach as indicated by first answer, frequently used for this purpose. If you are developed a depression scale, you have a convergent validity ıf your scale correlations significant, positive and strong with a previously validated depression scale, and your scale has disriminant/divergent validity ıf your depression scale significant, strongly and negative with a previosly validated happiness scale. However, establishing this values generally require the examine previous studies. ıf your scale moderately correlated with an self-esteem scale, and previous studies also shownt such relationship, still you have convergent and divergent validity. Thus, you have to know what previosly stuies tell you about relationship between constructs is most importsnt matter for to support convergent and disriminant/divergent validity.
Convergent and divergent validity also may examine within a scale using Fornell and Locker (1981) approach in confirmatory and exploratory factor analysis framework.
What you are talking about is criterion validity. If the constructs are same you can do Pearson product moment correlation. More important for your questionnaire would be face validity, content validity and construct validity (determined by CFA or SEM)
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics