OH MY GOD. I've been spending my night looking for way to solve my discriminant validity problems until I found this. Thank you very much for helping me. May God bless you always, Sir!
Sir.. thank you for your brilliant suggestions! One question - what if we are left with only two questions after deleting some of the problematic questions? Would that not be a concern? Is there a minimum number of questions a construct should have?
Great Dr. Thank you for this. Can you kindly suggest a solution for the Fornell and Larcker Criterion DV issue in which the first or the top value (the square root of AVE) has to be higher than the values below if I'm not wrong, but in my case, the top diagonal AVE value is lower than the values below. What am I supposed to do to resolve it?
Hello Sir, thank you for your video. Do you think checking the correlation between indicators of the problematic variables will help identify the problematic indicators? I found that useful. And do you have any references we can cite for the procedure (first check the STDEV.P, then check cross-loadings) ? Thank you.
Yes, that may help. As for reference, please refer to Collier, J. (2020). Applied structural equation modeling using AMOS: Basic to advanced techniques. Routledge.
Thank you so much for the video. I have some issues with my Cronbach's Alpha values. Some of them are less than 0.7, the values in my paper are around 0.67 - 0.68. Could you please tell me what should I do to solve the problem? Thanks again!
Dear Dr. Fawad if the cross-loading between items is less than 0.1, that item should be deleted to improve discriminate validity. Is there any reference for this? actually, in my PhD thesis, I cant put anything without reference. Your help in this regard is highly appreciated.
Please refer to Sarstedt, M., Hair, J. F., Pick, M., Liengaard, B. D., Radomir, L., & Ringle, C. M. (2022, May). An updated assessment of model evaluation practices in PLS-SEM: an abstract. In Academy of Marketing Science Annual Conference (pp. 85-86). Cham: Springer Nature Switzerland.
Hi Professor, thank you for your videos and explanations! Please, I have a question regarding this.. Can I use (traditional/original) PLS algorithm for evaluate the measurement model and consistent PLS algorithm to evaluate the structural model at the same time? Due to I have an issue with the Fornel and Larcker DV using consistent PLS algorithm.. But when I checked it using the original/traditional ones, the results were good.. And for other analyses, the results were also good using consistent PLS algorithm.. Thank you for your response!
@@researchwithfawad Thank you so much for your responses, Professor. That means it's not possible to use Normal PLS algorithm (for measurement model) and Consistent PLS algorithm (for structural model) at the same time? 🙏
Thank you so much for your videos and for sharing your knowledge! Your help couldn't be more timely for me! I have a question: Can I get the paper or the reference where you got that SDs for the responses should be greater or equal to 0.25, to consider a valid response? I would like to cite them in my research.
Thanks for watching. Here is the reference Collier, J. (2020). Applied structural equation modeling using AMOS: Basic to advanced techniques. Routledge.
Thank you so much for such a wonderful explanation. I have one query. My HTMT ratios are perfect but in Fornell-Larcker Criterian some values are negative which probably results in negative betas which are not supported by literature. Can you please explain how to change negative betas into positive?
Dear Sir, I would like ask you, if we need to eliminate any particular item from a construct to solve issues related to Convergent Validity, do we use the same eliminated model or rather use the original model for Discriminant Validity test?
Thank you so much for your helpful videos professor. Is it possible to teach us about best worst modeling (or max diff) with SPSS. Thank you in advance
Yes but my question is in smart pls 4 fornel larker criterion table dependant variable loadings are showing on upper row and independant is on lower row does it have issues
I still have a question about you removing LC1 and LC2, why do you consider it problematic because i have the same problem, where the value is all over 0,10 but i still have 1 htmt over 0.90 so im confused what other variables considered problematic
I followed this but the second I deleted the responses with relationship of two problematic variables, on PLS I got an issue of latent variables using different scales. Is there a solution?
Halo Sir. I had tried to solve the discriminant validity, but until kept two indicators, HTMT criterion still not reach. Can I ignore HTMT validity ? and any reference like my case?
Good Question and suggestion. Well this bunch of information is not in one single paper but i accumulation of knowledge over the past 5-6 years of working on SEM tools like SmartPLS/SEMinR/AMOS. Primarily you can use the following references Collier, J. E. (2020). Applied structural equation modeling using AMOS: Basic to advanced techniques. Routledge. Hair JF, Hult GTM, Ringle CM, Sarstedt M. Thousand Oaks, USA: Sage Publications; 2014. A Primer on Partial Least Squares Structural Equation Modeling (PLS-SEM) Hair Jr, J. F., Hult, G. T. M., Ringle, C. M., Sarstedt, M., Danks, N. P., & Ray, S. (2021). Partial least squares structural equation modeling (PLS-SEM) using R: A workbook.
@@researchwithfawad dear sir, sorry for follow up question, but I'm using 87 samples for my reasearch. is it still too low? because i have deleted the items and left 3 indicators for each construct but still got 1 htmt higher than 0.9 although fornell lacker criterion test has passed the ideal criteria
I am not able to solve. Anyone can help? AR CB EHM IS IW PEOU PU SUIT TRU AR CB 0.607 EHM 0.607 1.083 IS 0.551 0.659 0.701 IW 0.555 0.846 1.01 0.751 PEOU 0.566 0.57 0.582 0.741 0.621 PU 0.576 0.43 0.422 0.681 0.498 0.918 SUIT 0.577 1.278 0.891 0.6 0.721 0.541 0.424 TRU 0.543 0.499 0.495 0.678 0.582 0.705 0.797 0.47
Thank you Dr I completed my masters because of you at the University of Cape Town. Thank you.
Pleasure. I am glad that the videos helped.
OH MY GOD. I've been spending my night looking for way to solve my discriminant validity problems until I found this. Thank you very much for helping me. May God bless you always, Sir!
Thanks for watching. I am glad you liked the video.
Respected Sir. excellently explained. i learnt a lot from this video. May God bless you
Glad to hear that, Thanks.
Excellent explanation Professor. May you be blessed abundantly!
Thanks. I am glad you liked it.
Respected Sir, your videos has helped me a lot in quantitative methodology for my dissertation. Thank you so much for your efforts. 🙏🙏💐💐
Thank you for watching.
Always wondered if there is a solution for Discriminant validity issues. Thanks a bunch Prof.
Thanks. I am glad you liked it.
Wonderful demonstration! Thanks for this amazing video. It was very much like watching a suspense movie.
Glad you enjoyed it!
Hello, Prof! Thank you so much for the insight! I will try for my data 🤞👍
Thanks for watching
Excellent video. Thank you very much for sharing your knowledge.
Thanks. I am glad you liked it.
Sir.. thank you for your brilliant suggestions! One question - what if we are left with only two questions after deleting some of the problematic questions? Would that not be a concern? Is there a minimum number of questions a construct should have?
Thanks. I am glad you liked it.
Although 2 is low but does the two items cover how you have conceptualized the variable.
@@researchwithfawadNot really. Better to have three at least. I feel.
Great Dr. Thank you for this. Can you kindly suggest a solution for the Fornell and Larcker Criterion DV issue in which the first or the top value (the square root of AVE) has to be higher than the values below if I'm not wrong, but in my case, the top diagonal AVE value is lower than the values below. What am I supposed to do to resolve it?
Please follow the steps in this video
@@researchwithfawad this one, as I can see it, is a resolution for HTMT not F&L criterion. I stand corrected.
Hello Sir, thank you for your video. Do you think checking the correlation between indicators of the problematic variables will help identify the problematic indicators? I found that useful. And do you have any references we can cite for the procedure (first check the STDEV.P, then check cross-loadings) ? Thank you.
Yes, that may help. As for reference, please refer to
Collier, J. (2020). Applied structural equation modeling using AMOS: Basic to advanced techniques. Routledge.
Thank you so much sir. You have taken lot of efforts to resolve the issue practically. Can you please explain why you get chi square infinite issue.
Pleasure. I am glad you liked it. Where are you getting infinite chi-square.
Thank you so much for the video. I have some issues with my Cronbach's Alpha values. Some of them are less than 0.7, the values in my paper are around 0.67 - 0.68. Could you please tell me what should I do to solve the problem? Thanks again!
Pleasure. This is fair reliability. You may have low alpha because of no.of items.
Dear Dr. Fawad
if the cross-loading between items is less than 0.1, that item should be deleted to improve discriminate validity. Is there any reference for this? actually, in my PhD thesis, I cant put anything without reference. Your help in this regard is highly appreciated.
Please refer to
Sarstedt, M., Hair, J. F., Pick, M., Liengaard, B. D., Radomir, L., & Ringle, C. M. (2022, May). An updated assessment of model evaluation practices in PLS-SEM: an abstract. In Academy of Marketing Science Annual Conference (pp. 85-86). Cham: Springer Nature Switzerland.
Hi Professor, thank you for your videos and explanations!
Please, I have a question regarding this.. Can I use (traditional/original) PLS algorithm for evaluate the measurement model and consistent PLS algorithm to evaluate the structural model at the same time?
Due to I have an issue with the Fornel and Larcker DV using consistent PLS algorithm.. But when I checked it using the original/traditional ones, the results were good..
And for other analyses, the results were also good using consistent PLS algorithm..
Thank you for your response!
Use Normal PLS Algorithm.
@@researchwithfawad Thank you so much for your responses, Professor.
That means it's not possible to use Normal PLS algorithm (for measurement model) and Consistent PLS algorithm (for structural model) at the same time? 🙏
tqvm Dr.! always a great help indeed!
Pleasure. I am glad you liked it.
Thank you so much for your videos and for sharing your knowledge! Your help couldn't be more timely for me! I have a question: Can I get the paper or the reference where you got that SDs for the responses should be greater or equal to 0.25, to consider a valid response? I would like to cite them in my research.
Thanks for watching. Here is the reference
Collier, J. (2020). Applied structural equation modeling using AMOS: Basic to advanced techniques. Routledge.
Thank you so much for such a wonderful explanation. I have one query. My HTMT ratios are perfect but in Fornell-Larcker Criterian some values are negative which probably results in negative betas which are not supported by literature. Can you please explain how to change negative betas into positive?
Pleasure. I am glad you liked it. Negative values show inverse relationship.
Dear Sir, I would like ask you, if we need to eliminate any particular item from a construct to solve issues related to Convergent Validity, do we use the same eliminated model or rather use the original model for Discriminant Validity test?
Please use the one with item eliminated.
Thank you so much for your helpful videos professor.
Is it possible to teach us about best worst modeling (or max diff) with SPSS. Thank you in advance
Thanks. I am glad you liked it. Hopefully soon i can make some on the mentioned topic.
in smartpls 4, do we need to report the confidence interval bias corrected?
We dont need to.
Thank you so much sir
Thanks for watching
Dear sir, if we have to delete a few items, should we run again for other test such as convergent validity?
Yes please run again.
Well explained
Thank you.
The best one
Thank you.
Thank you very much ... very worth tutorial
Glad it helped
Yes but my question is in smart pls 4 fornel larker criterion table dependant variable loadings are showing on upper row and independant is on lower row does it have issues
No. This is no issue. If you want to arrange it. Just add a serial number when you rename them. For instance
1. Job Satisfaction
2. Job Commitment
@@researchwithfawad thanku sir
I still have a question about you removing LC1 and LC2, why do you consider it problematic
because i have the same problem, where the value is all over 0,10 but i still have 1 htmt over 0.90 so im confused what other variables considered problematic
Reason being that LC2 is loading well onto another factor, its loading is over 0.70 on another factor.
I followed this but the second I deleted the responses with relationship of two problematic variables, on PLS I got an issue of latent variables using different scales. Is there a solution?
What is the issue?
Hi sir, is it okay to delete the item in construct to meet the HTMT ratio?
Yes, you can delete a items to improve HTMT.
@@researchwithfawadsir may I know how to explain in write up about reason for delete item
@@researchwithfawad sir the loading between the problematic item should less than .01 right sir..if less .01, then I can delete the item right?
Delete if difference of loading is less than 0.10. The item is thus loading onto two constructs. Hence, can be deleted.
What you have explained for cross loadings, i don't get it. Can you repeat it pls?
You can also watch
ruclips.net/video/O2Jxm-ibQO8/видео.html
can you please tell from where can i extract excel data ? because my working file doesn't show the relation of variables
Are you referring to the results in excel format?
Sir, after doing that, in which part I should report it?
In measurement model when reporting Discriminant validity.
@@researchwithfawad
Thank you
May Allah bless you
@@researchwithfawad Sir do I need to provide a citation or justification for that ?
Halo Sir. I had tried to solve the discriminant validity, but until kept two indicators, HTMT criterion still not reach. Can I ignore HTMT validity ? and any reference like my case?
You can report Fornell and Larcker
@@researchwithfawad thank you for the response sir
Thank you Dr. I have problem on this.
I hope your problem was resolved.
@@researchwithfawad Sir, still not yet...Anyway, Im using Amos...😪
Check for SD in Excel
Check for Cross Loadings using EFA
Check for very high Modification Indices
@@researchwithfawad Thanks Dr. How to check EFA in amos? Do you have email address? Thank you
If the model is fit in Amos?Does it mean that here is no validity issue?
Is there any paper that related to this explanation?
Good Question and suggestion.
Well this bunch of information is not in one single paper but i accumulation of knowledge over the past 5-6 years of working on SEM tools like SmartPLS/SEMinR/AMOS.
Primarily you can use the following references
Collier, J. E. (2020). Applied structural equation modeling using AMOS: Basic to advanced techniques. Routledge.
Hair JF, Hult GTM, Ringle CM, Sarstedt M. Thousand Oaks, USA: Sage Publications; 2014. A Primer on Partial Least Squares Structural Equation Modeling (PLS-SEM)
Hair Jr, J. F., Hult, G. T. M., Ringle, C. M., Sarstedt, M., Danks, N. P., & Ray, S. (2021). Partial least squares structural equation modeling (PLS-SEM) using R: A workbook.
I cleaned my dataset until it the sample size came to 80, is that ok sir?
The sample size is too low for SEM.
@@researchwithfawad dear sir, sorry for follow up question, but I'm using 87 samples for my reasearch. is it still too low? because i have deleted the items and left 3 indicators for each construct but still got 1 htmt higher than 0.9 although fornell lacker criterion test has passed the ideal criteria
SD 0.25 threshold citation pls sir
Please refer to the following text
Collier, J. E. (2020). Applied structural equation modeling using AMOS: Basic to advanced techniques. Routledge.
Asalaaam:
Hello sir i have still problem in HTMT
You may need to collect more data.
I am not able to solve.
Anyone can help?
AR CB EHM IS IW PEOU PU SUIT TRU
AR
CB 0.607
EHM 0.607 1.083
IS 0.551 0.659 0.701
IW 0.555 0.846 1.01 0.751
PEOU 0.566 0.57 0.582 0.741 0.621
PU 0.576 0.43 0.422 0.681 0.498 0.918
SUIT 0.577 1.278 0.891 0.6 0.721 0.541 0.424
TRU 0.543 0.499 0.495 0.678 0.582 0.705 0.797 0.47
You may need to collect more data.