Query regarding discriminant validity

[Mohammad Moddassir Alam]

Dear Sir and Group members 

For checking the discriminant validity, there are approaches such as: through Gakingston MS-Excel Validity Master Output or through Fornell and Larcker(1981) criteria of AVE>interconstruct co-relation. 

Now, suppose if I want to check the discriminant validity by following the Fornell and Larcker criteria, then do I need to check the correlation of estimate in the Analysis properties while running the measurement model? Should these values be compared with the AVE.

My second query is what do the values after ASV column under different construct signify in the Gakingston MS-Excel Validity Master Output

Sincerely,

MD.MODDASSIR ALAM

[Neeraj Kaushik]

Dear Alam

I understand you've 2 queries:

1. Which method of Discriminant validity to use?

2. What do ASV & others indicate?

Well, you may use any method of checking Discriminant validity. Some of the times, reviewers have any preference and may ask for that very method. 

Secondly, these values (AVE, MSv & ASV etc) are not validity rather 

AVE>MSV and AVE>ASV indicate that Discriminant validity holds true.

So in itself any value have no meaning, but fulfilling the criterion is important.

Best wishes 

--
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 dataanalysistraining@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 dataanalysistraining+unsubscrib...@googlegroups.com.
To post to this group, send email to dataanalysistraining@googlegroups.com.
Visit this group at https://groups.google.com/group/dataanalysistraining.
For more options, visit https://groups.google.com/d/optout.

[Mohammad Moddassir Alam]

Dear Sir, 

Thank you for valuable insights.In addition, I have few queries:

1) To calculate the inter-construct squared correlation, should I check mark the correlation of estimate in the output option of Analysis properties in AMOSS.

2) What do the values after ASV column under different construct headings denote in  Gakingston MS-Excel Validity Master Output

3) While giving different indices value under model fit, we quote p-value. What does this p-value denote? Is it 

  the p-value of the measurement model.

Warm Regards 

Md.Moddassir Alam 

 

To unsubscribe from this group and stop receiving emails from it, send an email to dataanalysistraining+unsub...@googlegroups.com.

To post to this group, send email to dataanalysistraining@googlegroups.com.
Visit this group at https://groups.google.com/group/dataanalysistraining.
For more options, visit https://groups.google.com/d/optout.

--

MD.MODDASSIR ALAM

[Neeraj Kaushik]

Answers:

1. Analysis Properties->Output->

Tick on Squared Multiple Correlation

2. After ASV means on the Right-hand side of ASV?

3. p-value always refer to hypothesis testing. Plz remember, In whatever technique, you see the p-value, it always refer to some hypothesis.

Now my question to you is: What does p-value refers to in ANOVA table, Regression Coefficient table of Regression output?

[Mohammad Moddassir Alam]

Respected Sir,

Thank you for giving me time for replying to my queries. 

Yes, I meant to the right of ASV column.I still have few doubts, which are :

1.Is there any relation between shared variance (MSV ,ASV)  and squared multiple correlation.

2.Do we check hypothesis in the measurement model which is performed to check the validity of measures as SEM is performed to test the hypothetical relationship?

3.Do we check hypothesis in zero order CFA for eg. to show the relationship of manifest variables to its latent construct? 

[Neeraj Kaushik]

Dear Alam

You're not only late but careless too.

You've not bothered to answer my questions.

Best wishes

Neeraj

To unsubscribe from this group and stop receiving emails from it, send an email to dataanalysistraining+unsub...@googlegroups.com.

[Mohammad Moddassir Alam]

Dear Sir,

I sincerely apologise for the delay in responding to your question. 

Ans: p- value tests the null hypothesis that the regression coefficient is equal to zero (no effect). A low p-value (< 0.05) indicates that null hypothesis can be rejected.

Regards 

Moddassir 

--

MD.MODDASSIR ALAM

[Neeraj Kaushik]

Dear Alam

You're correct.

The p-value (mentioned for t-test) in Regression Coefficient checks the Null hypothesis that Reg coeff=0 so p<0.05 rejects this hypothesis and we conclude that reg coeff is robust.

In ANOVA output, null hypothesis is that r2=0. Again p<0.05 rejects this hypothesis and we conclude that coeff of determination is not zero rather a robust value.

Now coming to ur questions:

1. Plz remember in CFA we deal with 'r' only but we call it by several names e.g. Alpha, Beta, Correlation.

Average of (Square of avg beta) is called as AVE.

Square of relation between 2 constructs constitutes r2 called by ASV & MSV.

That's why ascertaining Discriminant Validity is checking whether relation fo a construct with its statement is higher than its relation with other constructs or not?

AVE > MSV

AVE > ASV

2 & 3. Validity is not a thing or statistical value rather a combination of several conditions e.g. For convergent validity we see:

i. Alpha> 0.7

ii. AVE > 0.5

iii. Alpha > AVE

If all three are ok, we conclude Construct Validity hold true.

But in the output we still get p-values which checks the same null hypothesis as is given in Reg Coeff output i.e. whether Reg coeff=0

It indicates here that whether a construct is able to explain the statement or not? (Remember MLE Regression works opposite to OLS)

So do remember whenever there is any p-value we're checking any hypothesis.

Plz feel free to ask for any/every doubt.

Best wishes

 

[Mohammad Moddassir Alam]

Respected Neeraj Sir,

Thank you very much, for the answers to my queries.

Warms Regards,

Md.Moddassir Alam

--

MD.MODDASSIR ALAM