Unable to improve discriminant validity

[vibhu johar]

Dear gp members

While testing convergent and discriminant validity of the scale in sem in amos, i found AVE values of all constructs and in two constructs AVE values were less than 5. Then i found CR which came out to be> 7. The convergent validity was fine.  Then testing for discriminat validity root AVE was found to be greater than correlation values between respective constructs but in case of two constructs the Root AVE value was less than correlation values. Then i checked cross loadings in EFA and deleted two items of satisfaction which were loaded in loyalty. Then i again calculated AVE. But still the divergent validity issue was there. The root AVE  of two constructs were still not greater than correlation value. I saw neeraj sirs video on improving discriminant validity. Please guide. 

 Thank you

Vibhu johar

[Garisha Arora]

AVE above  0.5 is very important.  Improvement through modification indices is possible only for minute deviation using modification indices that too a certain limit,  say one or two in a model. If the value is very less from 0.5 e.g. 0.46 / 0.44  then there is data issue. If by deleting items you can improve the AVE then great (without loosing content validity)  otherwise...  You need to check the data discrepancy. Respondents may not have filled the responses correctly or check the scale and critically analyze it.  Is it really a reflective scale?

You need to look upon these aspects critically. As you said C. R.  was OK which should be above 0.7.  But AVE is a compulsive parameter, which should be established. Look on these aspects.

For discriminant validity, please go n check your variables (scale's items)  in which issue is appearing. There might be a possibility at scale level.  If not,  then indeed it's a case of data.

I hope it helps. 

🙏🙏

--
Protocols of this Group:
 
1. Plz search previous post in group before posing the question.
2. Don't write query in someone's post. Always use the option of New topic for the new question. You can do this by writing to dataanalys...@googlegroups.com
3. Its better to give a proper subject to your post/query. It'll help others while searching.
4. Never write Open ended queries. This group intend to help research scholars NOT FOR WORK THEM.
5. Never write words like URGENT in ur posts. People will help them when they are free.
6. Never upload any info about National Seminars/Conferences. Send such info on personal emails. And feel free to share any RESEARCH related info.
7. No Happy New Year, Happy Diwali, Happy Holi, Happy B'day, Happy Anniversary etc. allowed on this group.
8. Few months back there was a facility for asking & sharing the Research Papers. Now there is no provision of asking for the research paper here.
 
Let’s make a better research environment.
---
You received this message because you are subscribed to the Google Groups "DataAnalysis" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataanalysistrai...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dataanalysistraining/CALPc%3Danb1oBTMfTGd-rmVaLKTY-uUM%2BoQMmCRhuGT-MrjkTzOQ%40mail.gmail.com.

[vibhu johar]

Thanks mam for your guidance. 

To view this discussion on the web visit https://groups.google.com/d/msgid/dataanalysistraining/CAEmQ%2B4spJ5jC2nEyA4dZtbF79WSipx-9uLKNaT8DJ1Y_tsNQfw%40mail.gmail.com.

[Garisha Arora]

M sorry there is a slight correction. I wrote it by mistake.  Modification indices is used for model fitness.  Not for validity. So don't consider that particular point as answer for your question. Rest is same as I mentioned. 

Apologies! 

To view this discussion on the web visit https://groups.google.com/d/msgid/dataanalysistraining/CALPc%3Da%3DBgBfwdLAbPYb3azA1vHd14AER8wkGUVtLsdCSDZOQtg%40mail.gmail.com.

[vibhu johar]

Ok mam. no

To view this discussion on the web visit https://groups.google.com/d/msgid/dataanalysistraining/CAEmQ%2B4sCzdYeGMhnneg1b%2B09EG8_cXk5nmrdzPMYVu2RWaGhBA%40mail.gmail.com.