14. Hayes Process Macro Model 4 with Multiple Mediators and Covariates

The tutorial guides on the use of Model 4 of the Hayes Process Macro for Mediation Analysis using Two Continuous Mediators, Continuous Independent, and Dependent Variables and Two Covariates.
For lecture notes and much more, visit
researchwithfa...
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
#ProcessMacro #HayesMacro #MediaitonAnalysis #covariates

[@ASMM1981EGY]

I can't stop praying to Allah for you to bless you Dr. Fawad. All respects and love from EGYPT. I'm so proud of you our Muslim scientist brother.

[@researchwithfawad]

Alhamdullilah. Thank you.

[@SanayaKhanna-t9l]

So well you have explained.. Really..your videos are very helpful

[@researchwithfawad]

Thanks. I am glad you liked it.

[@lailsss]

I don't think I got your answer right: wich model summary you'll use to talk about the R2 of the X and M on the Y. Do you use 'total effect model' or do you use the part before that (with path b and c') or the one before that with (path a), or do you report them all. you answer was 'the final mode' but which one is that?

[@researchwithfawad]

Thanks for watching. For R-Sq you can use the model summary with the DV (Impact of IV, MEDs on DV).

[@merveyuksel3048]

Hi from Turkey, Fawad. First of all, thank you for your videos helping me very in my phd thesis. I am wondering if the independent variable is categorical, how i understand which category (control or experiment) is significant or not?

[@researchwithfawad]

Thanks for watching. You can take any category as reference.

[@utkubeyazt5568]

The narration was great, very instructive, thank you. There's one point I don't understand. I cannot find the t values in the output that you gave in the report at the end of the video. How did you get the t values 2.0046 and 3.3942.

[@researchwithfawad]

Thanks for watching. You can divide the effect size by BootSE to calculate the t-statistic.

[@raz2936]

so clear!

[@researchwithfawad]

Thanks for watching.

[@chantellewong1952]

Hello Dr Fawad. Words cannot express my gratitude to you, your videos are really helpful. And I am currently working on running mediation analysis with outcome variable of three categories. May I seek your guidance and advice on how can it be done using SPSS? Thank you very much.

[@researchwithfawad]

Thanks. I am glad you liked the videos. Are the categories ordinal?

[@chantellewong1952]

@@researchwithfawad thanks for your reply. The cartrgories are nominal

[@researchwithfawad]

Please use Multinomial Regression

[@chantellewong1952]

@ResearchWithFawad Thank you for your suggestion. I am really a rookie in statistics. Do you mean using multinomial regression instead of the PROCESS macro, or is there an option of multinomial regression in the PROCESS macro? How can the indirect effects be estimated? Thanks a bunch for your genuine help

[@raifalves.0]

Hello Fawad, Thanks for sharing this valuable insight with us. I have a question, I doing a mediation analysis for emotions (hope and fear) in a 2 (Framing: positive vs negative) x 2 (Focus: Promotion vs prevention) IVs. I’m interested in the interaction effects of these two IVs, so I created an interaction variable (Framing*Focus). Once I’m using Process macro, is it correct if I just add the new interaction variable as Y and add the two original IVs as covariates?

[@researchwithfawad]

Thanks for watching. Why would you create an interaction if that facility is already in the macro and this is what the macro is for. Covariate have a different use.

[@raifalves.0]

@@researchwithfawad Thanks for the prompt reply, i wondered about creating the interaction because macro only allows me to add one IV per time, but since my model is testing the interaction of the two IVs, I was concerned about how to do this analysis on macro.

[@utkubeyazt5568]

I have one more question, my c' value is non sigificant, that is, the total effect of x on y is significant and the direct effect becomes non significant at the end. But when I check indirect effects, there is 0 between the LLCI and ULCI of one of my mediators, the other mediator does not have 0. What does this mean? Is the mediator with 0 a partial mediator and the one without 0 a full mediator?

[@researchwithfawad]

The one that has a 0, the mediator, is not mediating the relationship. Whereas the other is mediating the relationship.

[@jumaniqbal8206]

Amazing as always sir. JazakAllha khairan. Can you please tell me what codes did you assign to age, gender which were your covariates?

[@researchwithfawad]

Thanks. I am glad you liked it. I used 0 and 1 for Gender and for Age it was 1, 2, 3, and so on for each category.

[@duyguserbetci8962]

Hi Fawad, thanks so much for this video, super helpful! Do you know how to test assumptions for this model?

[@researchwithfawad]

Thanks for watching. As for assumptions, you can check for VIF and other regression assumptions like Durbin Watson, using normal regression.

[@lailsss]

Is the 'total effect model' including the mediator or is it just the path of X-> Y

[@researchwithfawad]

Total effect is the sum of direct and indirect effect.

[@lailsss]

i want to test a confounder in process macro. I want to see if SES has a signficant impact on X and if SES has a significant impact on Y. How do I report this?

[@researchwithfawad]

If you want to test control variables, you can use covariates if the variable is continuous.

[@LA-cx2od]

Can I also use a dochotome variable?

[@researchwithfawad]

Yes you can use dichotomous variable as covariate.

[@lailsss]

Hi, I was wondering which model summary you'll use to talk about the R2 of the X and M on the Y. Do you use 'total effect model' or do you use the part where you see b and c and the part of path a', or do you report them all?

[@researchwithfawad]

You can use the final model with the IV, MED, and DV

[@LA-cx2od]

Is the final model the 'TOTAL EFFECT MODEL' or something else? And do I also look at this part if the direct path is not significant?

[@researchwithfawad]

If you are testing mediation, you will need to first look at the Specific Indirect effect then the Direct Effect. Total effect may be reported but it is not required.

[@deekshasingh2257]

Sir how to analyse simple mediation in process macro with IV (dichotomous- Yes, No)?

[@researchwithfawad]

Use the same process, add categorical IV in X.

[@khanv5083]

Sir can i use process macro if i have more than one independent variable? If yes then how?

[@researchwithfawad]

This is the limitation of Process Macro. You can use SEM tools like SmartPLS or AMOS.

[@abuhanifahayob2048]

Dr Fawad, what is I have 2 iv; is it possible to use model 4?

[@researchwithfawad]

No. You will need to use SmartPLS or AMOS

[@greatmind3842]

Thanks so much! Your illustrations and presentations are unbeatable. I'm preparing a poster for a conference presentation and writing manuscripts for my papers. Please can you send me a link to the paper containing the Mediation Analysis table and reporting, or any other peer-reviewed paper?

[@researchwithfawad]

Thanks for watching. You can visit my profile for the papers scholar.google.com/citations?user=7rdmg9kAAAAJ&hl=en&oi=sra Or visit researchwithfawad.com and search mediation for sample.

[@gisellereidgr]

Hi, I wanted to know if you had a video on model 17, which includes 1 mediator but 2 moderators?

[@researchwithfawad]

Thanks for your interest but i dont have a video on model 17 yet.

[@maikonagaoka330]

Hi, I learn mediation anaysis through your video, they are so helpful. I apprecoate it. I have a question, i want to put dichotomise variable in mediation analysis, what model is fit to accept dechotomise variable? i tried this in model 4 in mediaton variable, but it didn't work. Could you give me advise for me? Thnak you.

[@researchwithfawad]

Thanks. I am glad you liked it. What was the issue?

[@maikonagaoka330]

@@researchwithfawad I put continuous variable as x ,dchotomous variable as M and other continuous variable in y, and process macro said, error:process does not allow dichotomius mediators. So I think i cannot use dichotomous variable in mediaor in model 4. can I put dichotomous variable in mediation anaysis? if so what model?

[@Alhamzah_F_Abbas]

Great content. prof. regarding control variables when we have to use them, are they depend on our aim or purpose in the thesis or paper?

[@researchwithfawad]

Pleasure. I am glad you liked it. There is no specific criteria, but you may use different demographic variables as control and see how they impact the results.

[@ASMM1981EGY]

@@researchwithfawad Dr. Fawad I'd like to ask a question regarding interpreting and reporting Covariables/Confounding Variables/Controlled Variables. In Mediation when we calculate the R2 value multiplying it by 100 to get a percentage of the effect/relationship between IV, Mediator & DV. How to report/write about the deduction/controlling of the Covariables from our model. Reporting the Mediation effect while deducting the effects of Covariables.

[@emmav1092]

Hi, I was wondering if you have a video on Hayes Process Macro with two independent variables (not multicategorical) or if this is even possible. Thanks!

[@researchwithfawad]

Hi. Actually it is not possible. For multiple IVs or DVs you will have to use SEM.

[@benn5454]

@@researchwithfawad Hi, Part 2 of Hayes (2022) outline how run PROCESS macro with multiple Xs and Ys. Am I misunderstanding Hayes (2022) or do you disagree with his methodology?