Moderation Analysis using SmartPLS: Analyze and Interpret the Moderating Effect

[@britoroy442]

He is so much helpful and responsive.🤩

[@researchwithfawad]

Thanks.

[@EngineeringandInfo]

Great Video. You are really doing awsome job. Sir, is there any way of getting crack version of smart PLS with all features.

[@researchwithfawad]

Thanks. I am glad you liked it.

[@abebenegeri1823]

Very helpful video. Thank you so much, Dr. Fawad!!!!

[@researchwithfawad]

Pleasure.

[@usmanshehzad6099]

Dear sir, I love the way you explain different research related concepts. I Really appreciate your efforts..!!! Stay blessed and keep sharing Knowledge.!!!! ❤ Thanks

[@researchwithfawad]

Thanks. I am glad you liked it.

[@chefberrypassionateresearcher]

Prof. Should we do measurement invariance testing before moderation anlaysis?

[@researchwithfawad]

Please read more on MICOM and when it is required.

[@anitalow3628]

Dear Sir,....according to your webinar that you are developing a system in doing Lit view Matrix software. When will you share the software and also make a tutorial video on how to use it? Looking forward

[@researchwithfawad]

Hi. Thanks for the interest. It is in testing phase and hopefully will be available soon.

[@anitalow3628]

@@researchwithfawad looking forward the release of software (used for organizing LR matrix)...thanks

[@hanahoah7325]

Sir if i have 2 hypothesis: 1. Commitment has a positive and significant effect on supply chain performance >>> (one tailed) 2. Trust moderates the relationship between commitment and Supply Chain Performance >>>(two tailed). The questions are: 1. what choice option that i should choose when i want to run bootstarpping in smart pls, two tailed or one tailed? 2. Or i should run bootstrapping two time with different choice, the 1st one is one tailed option for hypothesis number 1, and the 2nd is two tailed option for hypothesis number 2? And please include the references from ur response sir. Thx a lot

[@researchwithfawad]

Why is your moderation two tailed. I guess literature will support a one tailed hypothesis. However, run two tailed but for h1 you can divide the p vale by 2 to get one tailed results

[@hanahoah7325]

@@researchwithfawad Oh yeah, bcs i was tried it before but the result just differences in p values sir. Sir, i want to ask for more input regarding this because my hypothesis was using one tailed for moderation, but because of the background I made, my lecturer suggested using two tailed. Where the background/reason I use trust as a moderator between the relationship of trust with supply chain performance is because of the journal literature that says "commitment without trust is vulnerable" (but in this reference journal the dependent variable here does not discuss supply chain performance, but other variables). Maybe the reason my lecturer suggests me using two tailed is because the moderation reference that I use does not use the context of supply chain performance. So do you think, with that background, can my hypothesis 2 sounds "trust positively and significantly moderates the relationship between commitment and supply chain performance" >>> but there is no previous journal reference that empirically proves it can moderately positively and significantly in supply chain performance context? Your respons it would help me a lot sir, thx u.

[@researchwithfawad]

If literature supports a positive moderating effect, go for it.

[@gpdang]

Great explanation

[@researchwithfawad]

Thanks. I am glad you liked it.

[@sabrafarrukh9857]

Thank you for sharing very informative session specially graphically explanations of moderator. Kindly guide me if I have IV formative 3 Higher order construct.DV and moderators are reflective construct, while adding Moderator on DV what we will select Either Product indicator or Two-stage options. I am checking moderating effect on direct and Indirect path both. I have a mediator also in the model and that is also reflective construct.

[@researchwithfawad]

Thanks. I am glad you liked it. Use product indicator when IV and Moderator are reflective. Use two stage approach instead with formative indicators

[@sabrafarrukh9857]

Thank you so much for very quick reply. And what if IV formative and moderator is reflective?

[@shontenegeri4495]

Dear brother Dr. Fawad thank you so much and be blessed for the excellent videos. Next, I have some doubts that is when we undertake testing moderation/mediation should we keep only those constructs that are involved in moderation/mediation and run only that part of the model? Or should we run the entire model?

[@researchwithfawad]

Use Moderator with all other variables when testing the measurement model, add moderating effect for testing the structural model.

[@shontenegeri4495]

@@researchwithfawad Thank you very much and May God bless you more and more dear brother Dr. Fawad🙏🙏🙏🙏🙏

[@fatihah_pg]

Sir, I have 3 iv with one moderator (ordinal). how do I need to transfer the data to the excel sheet? do I need to report it 3 separate excel sheet or can I get it with one excel sheet to analyse low medium or high? Thank you for your video :)

[@researchwithfawad]

You will need to have three separate graphs as the single moderator is moderating three relationships.

[@ramatuissifu1877]

Thanx for the video.but Pls check ur sound I can't hear u well

[@researchwithfawad]

Pleasure. I am glad you liked it. I agree, sound was poor. But now i have improved the recording facilities. New video uploaded in past 10 months are better. I hope you enjoy the videos.

[@ramatuissifu1877]

@@researchwithfawad thank you very much. Am grateful

[@AzamKhan-pq3vq]

Dear sir just want to know the interpretration of third value that came after bootstrapping RA----OP

[@researchwithfawad]

That is the standard deviation of bootstrapping estimates. The lower the better. The standard error is the approximate standard deviation of a statistical sample population. The standard error is a statistical term that measures the accuracy with which a sample represents a population by using standard deviation. In statistics, a sample mean deviates from the actual mean of a population; this deviation is the standard error of the mean.

[@SitaraCh]

Assalam-o-Alaikum Sir, Your videos are really helpful for me as a research student. Bundle of thanks for your contributions. May ALLAH give you the best rank in this world and the world hereafter for your selfless services Ameen. I am an MS student and have to submit my thesis in this month. I humbly request you to please make a video on moderated mediation (Model 7 of Process macro) explaining its concept, analysis, interpretation and result reporting. Thank you. Take Care

[@researchwithfawad]

Thanks l. I am glad you liked it. Ameen and thanks for all the prayers. I will try to make videos using process macro soon.

[@Dr.MuhammadAsgharAli]

Great work sir, Thanks for uploading your video on moderation analysis. I can understand the basic mechanism, sir can we add multiple moderations between single IV and DV relationships? for instance if we take the big five personality traits (Introvert, openness to experience, etc) as moderators between the relationship of X (IV) and Y(DV) variables. We need to generate interaction terms and perform moderation analysis for all five dimensions at the same time or we can perform moderation of each dimension one by one. Which is the best strategy? Please advise, waiting for your detailed response to address this confusion. Thanks

[@researchwithfawad]

You can add multiple moderators between a single relationship. Run the whole model together

[@Dr.MuhammadAsgharAli]

@@researchwithfawad Sir thanks so much for your positive response. I will be grateful if you please can provide any article or book refrence to prove this stance. JazzakALLAH, Need your help, Thanks,

[@dishachoudhary9261]

@@researchwithfawad but sir how can we convert ordinal scale into categorical for moderating effect

[@shayistamajeed9889]

Are there any specific preconditions for moderation analysis as is the case with mediation?

[@researchwithfawad]

No. but on should understand the concept in detail plus how to run moderation for categorical and continuous variables.

[@talhamansoor7108]

That's great dear Sir

[@researchwithfawad]

Thanks. I am glad you liked it

[@MB-ui7er]

Respected sir, I need an excel sheet shown at 12:15. Please if you can share the link here. Further, I am learning a lot from each of your videos. Hope this journey goes on. May Allah bless you.

[@researchwithfawad]

Thanks for Watching. You can get it online. Please type in Google James Gaskin Stats Tool Package.

[@villasinivelautham3239]

Hi Dr. Fawad, if my moderation p value is not significant and my beta value is -0.088. How shall I report this? No moderation effect and MV dampens the relationship between IV and DV?

[@researchwithfawad]

Thanks for your message. You could say that although the moderating variable dampes the relationship but it was found insignificant

[@bhattisaleemraza1592]

Nice video.

[@researchwithfawad]

Thanks. I am glad you liked it

[@karlmiller466]

Hi Fawad, how do you stratify by group for different variables like age, gender and employment status? Sorry I’m deviating from the current topic.

[@researchwithfawad]

Please check your email

[@norizanashobri2754]

@@researchwithfawad i am also have the same question with Karl. pls help me. thank you

[@joelkovos7355]

Hi, I also have the same question😌 I’m using age as a moderator. I was planning to divide age into two groups like: 35> and

[@a.yusyaa]

@@researchwithfawad Hello sir, I have the same question with Karl (Agw and Gender), can you help me too?, Thank you

[@vivekdsingh9008]

Dear Fawad sir. I am struck during analysis. Should we initiate PLS algorithm for the whole model with moderator - by creating moderating effects in DV or without creating moderating effects.

[@researchwithfawad]

Please perform without the moderating effect. Once you add the moderating effect, take the difference of R-Sq from the original model without moderating effect.

[@vivekdsingh9008]

Thanks. Sir how can we get this excel file

[@AliAhmad-cp2gw]

Thanks sir👍

[@researchwithfawad]

Pleasure l. I am glad you liked it.

[@amberpervaiz1609]

Sir I have a question, we can performe the moderation analysis in two ways. 1 way is to use the process option , 2nd way is to use pls-sem option . I have tried both options, but results are different in terms of coefficient sign and significant. Is it possible that I may use the direct and specific indirect coefficient of pls-sem moderation, while the conditional direct effect of process model..??? I am waiting for your kind reply. It's urgent

[@researchwithfawad]

Thanks for watching l. Specific Indirect effect is for mediation, not for moderation.

[@amberpervaiz1609]

@@researchwithfawad and conditional direct effects for ??

[@susanna5254]

Hello Sir. I estimated my model again. The results for my moderation effect are: b=0.014, p=0.416. So an insignficant but positiv effect. When I go to 'conditional direct effects' all 3 of them are significant p

[@researchwithfawad]

This is due to addition of new variables.

[@susanna5254]

@@researchwithfawad but I havent added new variables? b=0.014, p=0.416 are the values for the moderator for graphical output and path coefficient and b=-0.330, -0.358, -0.344 (for the 3 different levels of the moderator) are the values when i look at "conditional indirect effects". But I havent added new variables

[@susanna5254]

Hello again. Another question, i hope you can help. I also tested my model for moderation effect.When i look at "path coefficents" there is no significant moderation effect (original sample = -0.017 and p=0.389=. But if I go to "conditional direct effects" where it reports my effects under condition of the moderator with +1, mean and -1, every 3 of them are significant?? (p=0.000) and the values for orignal sample are different. What does that mean? I dont get it... do i have a moderating effect or not..?

[@researchwithfawad]

Overall there is no moderating effect. But at different levels of the moderator the effect changes and is significant.

[@susanna5254]

@@researchwithfawad Thank you so much Sir. I couldnt get trough my master thesis without your videos and your help.

[@cheep-ih2yg]

hi prof, may i ask where to add moderating effect on smartpls 4? because when i click right, it doestn show the option

[@researchwithfawad]

Please watch kzbin.info/www/bejne/n6S5cquYdpp3hac

[@Dr.MuhammadAsgharAli]

Sir, can we break down moderation analysis from two to three levels in smart PLS? Normally moderation graph shows low and High values in Smart PLS, Can we identify Low, medium and high values? Please advise

[@researchwithfawad]

Yes, if you add moderating effect and then run PLS Algorithm, select Slope Analysis from output, it will present the slope at mean level as well.

[@dishachoudhary9261]

Sir,ordinal scale can be run as moderator without converting it in pls sem .is your all constructs is in ordinal?????

[@researchwithfawad]

Yes, you can consider it as a metric variable and use it as moderator

[@dishachoudhary9261]

@@researchwithfawad thanku so much for responding sir....you mean we can put it as it and can see the moderating effect without converting it into categorical form

[@researchwithfawad]

There are sometimes different opinions. Use it as categorical moderatorin smartpls. This will out you in a safe position and will help avoid criticism.

[@shumailakashif5650]

AOA, If we have to perform moderation analysis of reflective-formative HOC, can we use the same aproach (IV & DV are refelctive-formative HOC)?

[@researchwithfawad]

Yes but take LVS when you haveHOC

[@shumailakashif5650]

@@researchwithfawad thanks

[@researchwithfawad]

Pleasure

[@BusinessCoachp]

Dear Sir, Kindly guide me about how to get cracked version of PLS Smart, as public sector universities in Southern Punjab do not have this software for their scholars and researchers.

[@researchwithfawad]

I am afraid there is no cracked version. One has to purchase the software.

[@jumaniqbal8206]

Asalamualaikum sir Here you didnt change the bootstapping aub samples to 5000 . Can we do on 500 sub samples as well

[@researchwithfawad]

I did that for the video, as 5000 sub samples takes more time to run. For final analysis one should have 5000 sub samples

[@glowingstars8163]

The value of independent variable, moderator or interaction u add in excel file is not same as pls table value. Why?

[@researchwithfawad]

Thanks for watching. Therein, i am just trying to explain another method for plotting, it is just an example. You can use the same values.

[@AsmaKhan-yz1mp]

sir you can share complete reference of James Gaskin???

[@researchwithfawad]

Please refer to James Gaskn Website. STATWIKI

[@AsmaKhan-yz1mp]

thank you

[@rsantos4103]

Hello. Could you please share your excel with that legend?

[@researchwithfawad]

I used James Gaskin Stats Tool Package, You can visit his website to download.

[@naveedfaraz9688]

Sir would you please share that Excel file containing formula for ploting the moderation. Thanks

[@researchwithfawad]

Please email me on [e-mail] Or you can download it from www.kolobkreations.com/Stats%20Tools%20Package.xlsm

[@ajayanchan2085]

Please post same done in spss

[@researchwithfawad]

Hi. Yet to do categorical moderation in SPSS. Soon hopefully

[@ajayanchan2085]

@@researchwithfawad please do at the earliest...look forward to it ...thank you

[@sheikhfamilylahore]

Can you please share the data file for our practice?

[@researchwithfawad]

Kindly email me on [e-mail]. I will share a sample data file

[@lovingyourself20_20]

Can u pls share the data file?

[@researchwithfawad]

Please drop me an email on [e-mail]

[@britoroy442]

@Research With Fawad please check your email, I am requesting you to sent me the James Gaskin graph making excel file that you showed in the video.

[@researchwithfawad]

Please check your email.