PLS Vid7 Discriminant Validity Measurement and Reporting Cross Loading, Fornell and Larcker Criteria

#DerekOng #SMARTPLS #Research

[@tariqshah1887]

your video is so good, but you should also explain every value, what does it means etc. etc. ... I mean interpretation. I am facing much problem with this and i am unable to find a suitable video

[@georgealexander2272]

hello i want to ask something, if i use cross loading criterion, there is no need for me to use fornell larcker ? thankyou

[@psrajeswarisundaram9724]

Is it convergent reliability or convergent validity? Because i could not find definition or explanation for convergent reliability?

[@Shravan1231000]

I have got HTMT value as 1.025 which loes is between the confidence intervals (0.652, 1.059). So can I consider it to be ok..?

[@derekong1609]

HTMT values close to 1 indicates a lack of discriminant validity. Using the HTMT as a criterion involves comparing it to a predefined threshold. If the value of the HTMT is higher than this threshold, one can conclude that there is a lack of discriminant validity. Some authors suggest a threshold of 0.85

[@Shravan1231000]

@@derekong1609 sir can you further clarify on HTMT inference method?

[@Shravan1231000]

Sir can you share your thesis part cover8ng pls sem reporting for reference?

[@derekong1609]

You may refer to A Primer on Partial Least Squares Structural Equation Modeling (PLS-SEM), by Hair, Hult, Ringle, and Sarstedt

[@Alhamzah_F_Abbas]

brilliant, thank you very much

[@kristinegwen8280]

Hi Derek, what should i do if my HTMT is higher than 0.9?

[@derekong1609]

Hi this means that you need to relook at your measurements...some of the items are probably cross loaded ...and you will need to re specify you measurements

[@kristinegwen8280]

@@derekong1609 thank you so much for your kind feedback

[@maheshsingh3667]

If some of the VIF values are >3 and due to that discriminant validity is missing then what to do?

[@mortizmortiz]

How could I test discriminant validity if I have a 2nd order variable?

[@derekong1609]

I have a video on this it is on higher order constructs please have a watch :)

[@adamkahraman3122]

Dear Derek, Right after measurement model criteria of CR, AVE, rho, and Cronbach Alpha test results are within acceptable limits; does Consistent PLS Bootstrapping also required for factors? The reasoning for that says: To see whether all the indicators are statistically significantly related to their own constructs or not. Is that correct procedure? The bootstrapping as we know applies to path analysis. However, one source mentions the apply bootstrapping for factors as well. Sincere thanks...

[@Ztrygwyrs]

You will bootstrapping factors for LV with formative indicators to determine their outer weights

[@cucutajuandiegohernandezla1735]

Hello. I have a question. Is it correct to use cross-loading method to assess for discriminant validity when the CFA is run under CB-SEM (e.g., ML, WLS, WLSMV)?

[@derekong1609]

CB SEM uses a different method is I am not mistaken

[@amalanwary2710]

Thank you Derek, I do not seem to meet the Fornell-Larcker criterion as my diagonals are not the greatest. What can i do to remedy this situation?

[@derekong1609]

Greeting try using the HTMT method for discriminant validity or remove some of the cross loaded items .....

[@prasan_ale]

Dear sir, could you please explain to me how to do a cross-loading indicator in spss? Also, I couldn't understand Discriminant validity by Fornell and Larcker criterion in your video? could you please explain the process in SPSS?

[@derekong1609]

Hi send me a ln email if you can

[@prasan_ale]

How to do item cross-loading in in SPSS? anyone please recommend me any link?

[@사브리나-u5u]

For HTMT, you report results of PLS algo or PLS bootstrapping?

[@derekong1609]

PLS algo is good enough

[@ahmedoasis]

For Discriminant Validity table, it would be a good idea to show a mean, standard deviations, and correlation table for all the variables ! how can we do that via SmartPLS ?

[@derekong1609]

When you click on your data (green database) you will get a summary

[@ehiremenosunde3187]

Great one

[@ronelsimon]

Hello Derek! What is the acceptable difference between loading and cross loading? For example, is it valid to have a cross loading of 0.709 against a loading of 0.809?

[@derekong1609]

That seems high for cross loading it should be below 0.4 if possible

[@ronelsimon]

@@derekong1609 Thank you! : )

[@rushidix]

Dear Dr, regarding of Disriminant Validity (HTMT) value, you have mentioned in the video that the threshold value is 1 or close to 1. My question is what value should be acceptable or cut off point when the HTMT value is close to 1 ? is it 0.95 ? 0.90 ? 0.85 ? 0.75 ? 0.70 ?

[@derekong1609]

nothing above 0.90

[@rushidix]

@@derekong1609 Noted Dr.. thank you very much

[@teohuipeng8949]

Hi Dr, what can i do if one of my iv HTMT is above 0.9?

[@derekong1609]

You have to check the correlation of your variables and decide what to do with them...it may indicate that your variables are highly correlated so either drop one or combine them both

[@ronelsimon]

Thanks, Derek!

[@Girlintherocket]

Thank you Derek. Great lecture and very clear.

[@derekong1609]

Your welcome

[@zurinapatrick1326]

Dr. is it okay to have only one item in the indicator? It happens that one of my indicator (maturity level) have the value above 0.9 for HTMT. If I drop one item from the maturity level, all value in HTMT turn to green and no more discriminant validity problem detected. But my maturity level will only have 1 item now. Is it still possible to proceed to the next step? Or I need to drop maturity level and proceed to the next step with 3 indicators that have more than 3 items? Thank you

[@derekong1609]

Greeting than that item will be treated as a single item construct