How to do a mediation analysis in R...with visuals!

Link to the added variable plot video:
• Controlling/Conditioni...
Learning Objectives:
Conditioning versus mediation
Three (two?) step process of mediation
How to do traditional mediation analysis in R
How flexplot visualizes mediation
How to visualize with flexplot
This video relies on flexplot, which can be downloaded from R using the following commands:
install.packages("devtools")
devtools::install_github('dustinfife/flexplot', ref="development")
And here's a paper I wrote about my eight step approach to data analysis: psyarxiv.com/r...
Undergraduate curriculum playlist (GLM-based approach): www.youtube.co....
Graduate curriculum playlist (also GLM-based approach): www.youtube.co....
Exonerating EDA paper: psyarxiv.com/5...
Download JASP (and visual modeling module): www.jasp-stat.org

[@nickvarney8148]

This video is incredible. You described this better in 11 minutes than my professor did in two hours. Thank you!

[@allenshaw9996]

The music (and your energy) makes it so much better! 🤣

[@ericomarcon]

This video is pure gold. I would just recommend some assumptions' checking at the beginning of the analysis. Otherwise, 10/10

[@DrAnaWeidenauer]

you are not only a good teacher, you are hilarious

[@mikhaeldito]

Bootstrap video, please! And then proceed to parametric and semi-parametric bootstrapping for mixed-effects models :D

[@stephendye2924]

Agreed, that would be useful

[@capricious8929]

EXTREMELY HELPFUL! THANK YOU!

[@HarmonicaTool]

What about covariates / control variables? are they just added in both models? Do we need to normalize predictors? U love how fast and on the point this lesson is but a hint to limitations / precautions would improve even more. That being said: Thanks so much!

[@KN-tx7sd]

as usual excellent, Bootstrap video, please! Also are the variables are all continuous or is any of them categorical. Can mediation be done when few variables are categorical in nature.

[@janja7471]

Great vid! And yes, pls do Bootstrap!

[@symeonmandrinos9528]

Thank you for your effort

[@jesusgarza-noriega1394]

I fucking love this guy

[@kriegsmandot1]

Thanks so much, this is great! Why did you include exercise in the plot? Seems like that's a question unrelated to our variables of interest?

[@nl7247]

Thank you very much ❤

[@noemimicheli6803]

What if the mediator is a categorical binary variable? Thank your for your videos!!

[@capricious8929]

What if I have two independent sample groups and two mediators, and 1 dependent variable I want to test out?

[@IsobelFrench]

Where can I get the dataset used in this video please? Thanks!

[@erickcampos50]

Great video! It's possible to do mediation using Jamovi and GLM? And yes, I would like to see a video about bootstrap

[@QuantPsych]

Not at the moment.

[@davidwang8971]

Hi, thank you for this amazing video!! I have one question, is it possible to do medaition analysis with "multiple mediators" (assuming they are uncorrelated with each other) in this package "mediation"? Many thanks.

[@caviper1]

Gracias man.

[@ito945]

Thank you very much for your videos. I would like to do the mediation analysis using multiple imputation for missing value using the "amelidiate" function in "mediation" package. If you know how to use this function , please give me the information.

[@perevilaltahuguet5383]

What do you think of the lavaan package?

[@QuantPsych]

Love it.

[@LaraFontana-dh5oh]

Using mediation function, the ACME is significant, but on the graph the two slopes are almost identical... why?

[@QuantPsych]

Do you have a large sample size? What's the effect size? It might be significant, but the effect size is tiny. This is probably what's happening.

[@SanazDabiri-f8n]

Can you also visualize longitudinal mediation with flexplot?

[@carolinaochoa3309]

Hi, thanks for the great work! I work with time to event data for the outcome. My exposure and mediatior as continuous variables. I used survreg function, which is suppoted by the mediation package. I'd like to represent my mediation model in a plot, like the mediate_plot you just showed. Do you have already a code for plotting mediation on this kind of outcome? Thanks!

[@nadjaburgel9931]

How do I do a mediation when my mediator is a latent variable consisting of four measured variables?

[@nicujo]

Hello from Denmark! Thank you very much for your videos, I found them VERY instructive. I am studying the univariate and multivariate playlists but I cannot install the flexplot package. Is it still available? Thanks for your effort :)

[@QuantPsych]

It is: devtools::install_github("dustinfife/flexplot")

[@edafade]

Is it possible for flexplot to handle moderated mediation analysis with multiple predictor variables? I don't think it can visualize this, as we don't want to control for variables necessarily after the pipe.

[@AmanzholBekmagambetov]

On 5:30, Aren't we supposed to establish first the relationship between Depression and Quarantine as a base model and only after that test full model?

[@QuantPsych]

Yes, that is customary.

[@caviper1]

Q: Can "mediate_plot" function be used with a binary response variable? If so, how?

[@qizhengli7952]

what could I do when r reports the error "mediator model is not yet implemented", THANK YOU!!!

[@rogerakanji5154]

Hi I am currently having a go at mediation path analysis model but I am stuck in the "error sem" My specify model is shown below: specmod

[@jodi9434]

Are mediation models technically additive?