#SmartPLS4

Multigroup Analysis (PLS-MGA) using SmartPLS4. The video focuses on the concept of PLS-MGA, running, interpreting, and reporting multigroup analysis in SmartPLS4.
www.smartpls.c...
Khawaja Fawad Latif, Rabail Tariq, Dilnaz Muneeb, Umar Farooq Sahibzada
& Shakil Ahmad (2022): University Social Responsibility and performance: the role of service
quality, reputation, student satisfaction and trust., Journal of Marketing for Higher Education, DOI:
10.1080/08841241.2022.2139791
Link: www.tandfonlin...
Do you have any questions, tips, or ideas? Let me know in the comments section below!
About:
Research With Fawad is a Research initiative to help young and early career research learn the research tools and techniques. It is an initiative to share research knowledge and help researchers learn to research the easy way.
Learn more about Research With Fawad at www.researchwi...
Join The LinkedIn Group: / 8948539
Like on FaceBook: / researchwithfawadfb
#Multigroup #PLSMGA #SmartPLS4 #PLSSEM

[@fawadkulachi8199]

Ap k liye dil se dua niklati hai Sir, ALLAH ap se RAZI HO..Ameen

[@researchwithfawad]

Alhamdullilah. Jazak Allah Khair

[@JayaKumar-cb4kl]

Thanks a lot for uploading this video, very well-explained regarding PLS SEM Multigroup Analysis Prof.

[@researchwithfawad]

Thanks. I am glad you liked it.

[@啾啾-h3y]

I am a Chinese master student struggling for my thesis, appreciating for your sharing and reply , It helps me a lot!！

[@researchwithfawad]

Thanks for watching.

[@rashid3187]

Sir please make a video on fsqca analysis with primary data... A much needed session

[@cheep-ih2yg]

Barakallah prof!! MasyaAllahhh 😍😍😍

[@researchwithfawad]

Alhamdullilah. I am glad you liked it.

[@shakilahmad2341]

Thanks a lot Sir for your informative session

[@researchwithfawad]

Thanks for watching.

[@RashaElnaggar-no5ie]

many thanks for this video

[@researchwithfawad]

You are welcome

[@ankeetasharma9357]

Thank you so much. You really explained sooo good and also guided well about reporting results.

[@researchwithfawad]

Glad it was helpful!

[@vaishnavibalaji6811]

Thanks a lot Sir 🙏

[@researchwithfawad]

Most welcome

[@manpreetkailay5382]

Very informative Fawad sir thank you so much

[@researchwithfawad]

Thanks for watching.

[@sybertronlionhawk]

VEry well explained, thank you!

[@researchwithfawad]

Thanks. I am glad you liked it.

[@kanishkapathak6626]

Great job sir

[@researchwithfawad]

Thank you. I am glad you liked it.

[@khalilhussain4288]

You are such a wonderful teacher. Stay blessed. Is it possible to do MGA on a model with multiple endogenous formative constructs and reflective constructs?

[@researchwithfawad]

Thanks for watching. Yes, why not. You can create groups and run the model.

[@TaniaAfroze-c2e]

Can a continuous variable like age be grouped to do MGA?

[@啾啾-h3y]

dear professor,in your paper ,in the part of mutiple group analysis, you said " The findings revealed that all the differences were found significant (the difference in the impact of USR on ST was partially significant). " I wonder if "partially significant " means the p value(0.052 in your result) is more than 0.05 and below 0.1? Looking forward to your reply!

[@researchwithfawad]

Yes it is less than 0.10

[@AyaKasber]

Thank you for the video! Why didn’t you report the MICOM? is it ok to use the bootstrap MGA without ensuring the invariance? And what is the difference between the permutation and the bootstrap MGA? Thank you in advance

[@researchwithfawad]

Thanks for watching. Yes, MICOM may also be reported.

[@thesisandstudyhelpbyharoon1014]

I enjoyed your videos. Please can you tell me what is difference between MGA and Moderation analysis. For example you might have tested country wide difference in a moderating model.

[@researchwithfawad]

Thanks for watching.

[@TaniaAfroze-c2e]

Assalamualaikum. In my research, I have two groups, 1) agro, 2) manufacturing. CR and AVE of the Agro group are.638 and 0.382 respectively. But the complete model and manufacturing group show those above the threshold level. Should I conduct MGA, and if yes, what would be the justification/reference?

[@ayodejinajeemiziaq9166]

Thanks you for this effort !! May Allah reward you abundantly. Meanwhile, can one use Smartpls for Secondary data like Interest rate, inflation and Gdp analysis

[@researchwithfawad]

Thanks. I am glad you liked it. Yes you can. But it is better to use the software primarily used for secondary data

[@hoangyennhi8157]

Thank you so much, Professor, for all your effort in explaining this to the audience. I would like to ask for your suggestions regarding my case below (I use R for all the analyses): My moderator variable is the countries (developed/developing). First, is it necessary to test for measurement invariance between the two groups? I am struggling to find code to run measurement invariance tests in R; could you suggest any available code? When I test between the two groups, one group does not fit the measurement test (Cronbach alpha

[@researchwithfawad]

Thanks for watching. MICOM is not absolutely necessary but you may be asked to do it. As for low reliability, you can make your case that since it is acceptable for the complete sample and in the other group, it can be referred to as acceptable.

[@hoangyennhi8157]

Thank you so much Professor !!@@researchwithfawad

[@kimtan2429]

Do you have a reference for not using MICOM?@@researchwithfawad

[@hoangyennhi8157]

@@researchwithfawad Thank you so much Professor, I would really appreciate if you could provide code to run MICOM in R. Thank you so much !

[@anugrewal9790]

Greetings sir In case of my pls sem model I have 1 independent variable 1 dependent variable 4 mediator variables 1 moderator variable 3 demographic control variables (dummy created for them) So do I need to use mga as specifically it’s not part of my hypothesis to check whether age,gender,education differs as I have taken these as control variables ,as my hypothesis are to check moderation and meditation of metric variables only.

[@researchwithfawad]

There is no need for MGA. Please add control variables as they are if they are ordinal. Otherwise, create dummies and add them to the model.

[@anugrewal9790]

@@researchwithfawad thank you sir

[@mwabasifanu9843]

Nice video kindly assist me with comparing 3 different groups I have enjoyed all your 👏

[@researchwithfawad]

You can do it in similar manner, generate groups and run the model as in the video .

[@michaelfumar863]

Great video, but if you dont mind me asking, how can i compare more than 2 group?

[@researchwithfawad]

You can any multiple comparisons. Just create more groups.

[@matsersilvaworld5224]

Dear Prof, what is the difference between using control variables and MGA?

[@researchwithfawad]

I would recommend you to read the paper on MGA. Please search Google Scholar for MGA using PLS SEM.

[@goharali6966]

Good video. Thank you.

[@researchwithfawad]

Thanks. I am glad you liked it.

[@ShahzafIqbal]

Great.

[@researchwithfawad]

Thank you.

[@Alhamzah_F_Abbas]

Well explained. prof I just wondering why you didn't use MICOM or permutation for MGA, as I know in SamrtPLS3 we need to use MICOM, but here in PLS4 is slightly different. can you clarify?

[@researchwithfawad]

Yes MiCOM is also recommended

[@ranggakrooschannel6199]

Hello, Professor Fawad. My name is Aziz from Northeast Normal University. How are you doing? I would like to ask you about Goodness of Fit (GoF) in your research paper (University Social Responsibility and performance: the role of service quality, reputation, student satisfaction and trust). In your paper you said "In this study, the Goodness of Fit (GoF) for the overall hypothesized model was 0.641, while 0.572 and 0.457 for Pakistan and China respectively, indicating that the model satisfied the global criterion of 0.3 (Henseler & Sarstedt, 2013)". My question is how did you get the GoF value for the overall hypothesized model? Thank you very much and I wish you always healthy. May Allah bless you.

[@researchwithfawad]

Thanks for your interest. Please watch: kzbin.info/www/bejne/lWfNXpmuaK1-jcU

[@jonpark3857]

Thank you very much for the video, professor! Can I use the same procedures from your video to perform a multigroup analysis for four groups? I have four groups with a 2 (short vs. long) x 2 (yes vs. no) experimental design, and I'd like to compare them: group 1 (short, yes), group 2 (short, no), group 3 (long, yes), and group 4 (long, no). I've read a smartPLS book that mentions the "Omnibus test of group differences (OTG) approach" for comparing more than two groups. However, your other video covered a three-group comparison using multigroup analysis, which has left me slightly confused. Is it possible to perform a four-group comparison (PLS-MGA) using smartPLS 4? Your video has been incredibly helpful, and I would greatly appreciate your response to this question.

[@researchwithfawad]

Yes, you can perform for 4 groups using SmartPLS4.

[@fenglin7864]

Professor, I followed your guidance, and as a result, one of my papers has been accepted by a high-impact factor SSCI journal. Thank you so much. I have a question. I ran "Consistent Bootstrap MGA", and its result was better than the Bootstrap MGA you presented in the video. May I know if it is fine to use "consistent bootstrap MGA instead of Bootstrap MGA in this case, please?

[@researchwithfawad]

Yes, you may use them. I am happy to know that the sessions helped. Congratulations

[@indahdewi7152]

thanks for uploading this video, very well-explained. i want to ask when we do MGA test to see any differences from 2 groups, bootstrapping in advanced setting, should i choose one-tailed or two-tailed test type?

[@researchwithfawad]

You can use one tailed if based on literature you can assume that one group may have a stronger relationship than other.

[@rabiarasheed6073]

Assalam sir, very well explained. Just 1 issue I am facing when I create group and I select all items of variable it gives error.

[@researchwithfawad]

WAS. What is the error.

[@rabiarasheed6073]

@@researchwithfawad Thank you. Issue is that I want to measure the group difference in one measurement variable i.e. difference in mean caring burden (measured using 10 items) based on gender or type of stress training given. Can I perform this analysis in PLS-SEM using MGA or any other test?

[@dkeke88]

Thank you for the the very informative video. I tried to run MGA test on 2 groups. I found direct effect of a construct (e.g. PE) is not significant to BI in Group A nor in Group B. However, direct effect of PE to BI is significant in the complete group. how can I interpret this? Thank you..

[@researchwithfawad]

Thanks for watching. This is because the sample size is low in groups where as the overall sample is high.

[@dkeke88]

@@researchwithfawad Thank you very much for your reply

[@bilalpanji4205]

it is acceptable when AVE score was below 0,5 ?, can you gimme some recomendation article journal about that

[@researchwithfawad]

If it is over 0.40 and CR over 0.70, then you are good.

[@chefberrypassionateresearcher]

Sir, what to do with those outer loadings which are marked in red and which are in negative values?

[@researchwithfawad]

Please watch kzbin.info/www/bejne/nIXdaICKg5JkfMU

[@chefberrypassionateresearcher]

@@researchwithfawad Thanks Professor

[@odontuyabayarsaikhan9403]

Hello, I have conceptual model with several variables and designed within-subject experiment to test my hypothesis. Participants will be shown visual stimuli of human influencer and virtual influencer, have to evaluate the advertisement value. I would like to compare the results. Can I use MGA for data analysis? And report similarly like this article? My data is not longitudinal. I have seen research articles suggests ANOVA for repeated measures, but I find Smart pls very user friendly for new researchers like me.

[@researchwithfawad]

Hi. You may use MGA if there are relationships to be assessed that may be different between groups. If you have only one variable/construct you may create dummy variables.

[@odontuyabayarsaikhan9403]

@@researchwithfawad thank you so much for your reply.

[@matsersilvaworld5224]

Do we need to have minimum sample for each group to perform MGA?

[@researchwithfawad]

Yes, should be at least 30.

[@dikshachoudhary3898]

when i m doing its show singular matrix problem

[@researchwithfawad]

If you are using dummy variables. Keep a variable out of the model as a reference.

[@thesisandstudyhelpbyharoon1014]

Assalam o ALaikum Sir, I have just emailed you a question, would you please guide me on MGA and moderation analysis. I would be grateful.

[@researchwithfawad]

Thanks for watching.

[@bhadrakv730]

Hi Sir, why do we use bias corrected bootstrap instead of percentile in MGA?. in rest of the mediation analysis videos we had used percentile

[@researchwithfawad]

Both can be used, but Bias Corrected gives more stable solution and shall be used when Bootstrapping.

[@manishaverma7456]

Video is not clearly visible

[@researchwithfawad]

Please check your Internet

[@ninsssssss]

Multigroup analysis is it same with moderator ?

[@researchwithfawad]

Use MGA for moderation if all the relationships are moderated hy a categorical variable.

[@soniamaheen9266]

Can MGA be used for longitudinal data?

[@researchwithfawad]

For that use repeated measures in SPSS for any other tool.

[@anacatarinamorais5021]

If the reliability and convergent validity are ok, but MICOM test fails to some of the constructs, would it be ok to the MGA? What is the best alternative?

[@researchwithfawad]

You can use MGA.

[@cata.maican]

Hi, Considering that we have three or more groups (A, B, C, ...), what is the recommended approach for comparing data among these groups? Thank you!

[@researchwithfawad]

Thanks for watching. If you want an overall score to assess if there are differences across the three groups, use One Way ANOVA and later use Multiple Comparisons. if you want comparison between groups and assess if overall there are significant differences, you can use SmartPLS MGA

[@nomanrasheed9644]

i am assessing relationship between variables for two groups i.e. rural and urban but i smart pls does not produce p value, sd and t values of rural group just produces path coefficients whereas in other columns, i got just n/a. kindly help me.

[@researchwithfawad]

Have you checked the sample size in each group, plus I hope you haven't used them as dummies as well.

[@nomanrasheed9644]

@@researchwithfawad each group has sample size more than 300 and that variable hasn't used as dummies.

[@qaisermohiuddin413]

Sir as I am going to put two studies one in individual level and second is on team level both have different variables, but the DV is same. Do I have to use Multigroup analysis? Is multilevel and multigroup is the same thing ?

[@researchwithfawad]

You will do multilevel analysis. Multilevel and multigroup are different.

[@qaisermohiuddin413]

Sir, is there any paper or video of you on Multilevel? As i have not find yet@@researchwithfawad

[@janabkhan3592]

Assalam o alaikum sir I have a question, hum literature say questionnaire adopt karty hain laikin sirf 1 question main 2 statements ko split kar k 2 statements bana dain tou kia is k liye koi additional steps ki zaroorat hoge jabky change tou bilkul minor hai main question share kar daita hoon. I do not transfer personel knowledge and experience into organizational knowledge. Sir is question main experience our knowledge ko split kia gaya hai sirf

[@researchwithfawad]

No additional steps required.

[@vaishnavibalaji6811]

Sir when I run Age Group in Multi group analysis it's showing as singular matrix why it's showing like that sir, is there any solution to this?

[@researchwithfawad]

Please watch kzbin.info/www/bejne/nHi3YYiqaqeIjtk

[@vaishnavibalaji6811]

@@researchwithfawad Thank you Sir

[@MehvishRiyaz-k8z]

Asalamualaikum. Is the micom procedure necessary.. can we directly use mga in pls4.. i m doing mga for gen x and y. I directly did mga in pls4.. was it recommended?

[@researchwithfawad]

It is recommended but not necessary.

[@manishaagarwal7434]

Hello sir Please help me in understanding How can I do comparative analysis of three organization

[@researchwithfawad]

Create groups as in the video.

[@rymarshorts]

How does this work in a study that involves a variable that is both self-reported and externally assessed?

[@researchwithfawad]

If you want to compare self-reported and externally assessed for a particular relationship, you can create a grouping variable for the two categories and perform MGA.

[@rymarshorts]

@@researchwithfawad Thank you very much. Would it be possible to have the pairing of the self-report and external assessment? How would it be done?

[@researchwithfawad]

Pairing?

[@rymarshorts]

@@researchwithfawad self reported leadership practice of principal and the assessment of teachers. There are let's say 20 principals and under each principal there are 9 teachers.

[@FowokemiAlaba]

please, how do i calculate Qsquare using smartPLS4? iv tried, but, no blindfolding here or, i couldnt find how to calculate

[@researchwithfawad]

Please watch kzbin.info/www/bejne/hYCxgmmkndynZtU

[@FowokemiAlaba]

@@researchwithfawad thanks very much

[@poornisakrabani1563]

Dr. Fawad, if I want to test the impact of both gender and age, must I do the analysis separately or can I run the analysis simultaneously? Kindly advise. Thank you

[@researchwithfawad]

You can run the comparison in a single analysis.

[@poornisakrabani1563]

@@researchwithfawad Thank you very much, Dr. I really appreciate your prompt response.

[@mridultrivedi5411]

If the reliability of one group does not establish like in your study (China). Is it ok to proceed?

[@researchwithfawad]

Yes it is, better to report reliability for the overall sample as well.

[@annusingh704]

Sir, could u please share this paper access link

[@researchwithfawad]

Thanks for your interest. Here is the link www.tandfonline.com/doi/abs/10.1080/08841241.2022.2139791

[@dishachoudhary9261]

sir what is hensler mga and permutation mga

[@researchwithfawad]

Hensler's MGA (Multiple Group Analysis) and Permutation MGA are statistical techniques used in multivariate data analysis to test for differences between groups or populations. Hensler's MGA is a non-parametric technique that tests for differences between multiple groups based on ranks of observations. Hensler's MGA is particularly useful when the assumptions of parametric methods are not met. Permutation MGA is a resampling technique that tests for differences between multiple groups by randomly permuting the group labels and calculating a test statistic for each permutation. The main difference between Hensler's MGA and Permutation MGA lies in their approach to testing for differences between groups.

[@dishachoudhary9261]

@@researchwithfawad Thanks for application...Sir does pls sem also perform hensler mga test..which step should follow..

[@elviraoktanurisa7048]

Hi sir, i want to ask. If from the measurement model we found there are unreliable or not valid item but each groups has different unvalid item. What should we do? Can we delete it or just let it be? Because if we're deleting it we can't conduct face to face multigroup analysis. Thank you

[@researchwithfawad]

Hi. Thanks for watching. I would keep them for further analysis and will only delete items that have low loading in all the different groups under investigation.

[@iFanuca]

Hello @@researchwithfawad . I had the same question, but can you provide any reference that justify that procedure? Again, thank you so much

[@goharali6966]

Urdu please.

[@researchwithfawad]

I am afraid, i dont make in Urdu.