I am checking the validity of questions in AMOS. Construct Validity and Item Validity of the questions fulfill the criteria. But Convergent and Divergent Validity conditions are not satisfied.
I seek your help regarding this aspect. How do I proceed with my analysis?
Can further analysis be conducted when convergent and divergent validity conditions are not satisfied?
What is the solution for this problem?
Kindly Help. Thank You.
Construct validity generally includes both convergent and divergent validity, since it is based on the idea that the measure fits a broad pattern of correlation, according to what it does and does not predict.
Check if the values of the Average Variance Explained (AVE) for each factor are over .50, ensuring that the explained variance is greater than the residual variance (Bagozzi & Yi, 1988), or, at least, if all exceed the cut-off value of .40 (Diamantopoulos & Siguaw, 2000).
Check also if you are in the presence of discriminant validity, given that the variance extracted from each factor is greater than the values of the squared correlations between each pair of factors (Fornell & Lacker, 1981).
Good luck with your research!
Respected David Sir and Lisete Madam,
Thank You very much for clearly explaining the results.
Dear Srinivasan,
Here is one of my researches which i used Construct validity to evaluate the measurements scales (includes content, convergent, and discriminate validity),
It may help you,
Regards,
Dr.Alhaddad
Article The Effect of Advertising Awareness on Brand Equity in Social Media
Respected Abdullah Sir,
Thank You very much for sending the paper.
Hi Srinivasan,
If the scale (questionnaire) is developed by you as adhoc, you can check for the EFA as there may be higher cross loading by an item in different constructs i.e. the item may have very similar loadings in two constructs. Hence it is not able to discriminate itself from the other constructs. This sometimes occurs in adopted scale as well which may be because of the bias in response given by respondents or poor construction of scale. Better try running EFA for the data to explore how well the items are falling under a construct ( this helps in knowing which item is creating trouble). Second you can delete the items with very less loadings and also items which covary much with the other item. this might help you to improve Convergent and Discriminant validity
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics
- Survey Design
- Questionnaire