Path analysis in R using Lavaan (video 1; Sept, 2019)

[@arunkumardubey120]

Dear Mike, Your effort is truly phenomenal. Thanks for sharing. Love from India. Best wishes

[@nguyentho9467]

Thank you so much for this extremely helpful video. You explained all the basics so clearly - really appreciate your time and effort! Saved me hours or days!

[@GabrielaGuzman-m4c]

This is the greatest video! Thank you so much!

[@paty_lnunes]

This video helped me a lot. Thank you!

[@MrDevid1975]

Thank you for your video Dr. Mike. what will be the codes if there is only one IV and two mediators (serial) and one outcome variable?

[@Bravemaster3]

Thank you for the video. Really well done.

[@kasunpathirana9410]

Very clear explanation. Thank you sir

[@Artyom109Zinchenko]

Hi Mike, thanks a lot for the video! I just wanted to ask why don't you try working with Rstudio? This will remove the necessity to copy-and-paste code from a text file and go back and forth between pages. In any case, thank you!

[@mikecrowson2462]

Hi Artyom, you are welcome! And thanks for suggestion about RStudio! I may use it in some of my future videos. Best wishes!

[@AndresGonzalez-rw7oe]

Hi Mike, thank you for sharing your knowledge. This video is very useful!

[@laurasantiago798]

Thank you so much--life saver

[@mikecrowson2462]

You are very welcome Laura!

[@georgeellis9626]

Thank you for this example.

[@mikecrowson2462]

You're very welcome! Thanks for visiting!

[@alex000cozar]

Nice explanation

[@lizhenye1715]

Hi Mike, this is very helpful! Thank you very much! I got a question: I created a CLPM model by lavaan. Now I want to compare the strength of the lagged effects (i.e X1 on Y0, and Y1 on X0). However, I am lost. Could not find the instruction in your video..

[@annanoer5627]

Thanks so much! Is this the same as cross lagged model?

[@김안나-k7b]

Thank you! I love this!

[@mateusragusequadros4530]

Thanks so much for this video, helped me a lot. I have a question aboute se operationalization of the formula of model writing: in lavaan is possibol use two columns of my data (e.g. axis1 and axis2 of a PCoA, that both discribes my interest variable) as a single variable (let's say "descriptor 'x'")? In anothers softwares I know it is possible, but I cant find that possibility in lavaan. Thanks, in advance, for your atention!

[@brenna3126]

Super helpful, thank you so much

[@bradleypan-weisz2105]

Hi Mike. Thank you for the video. I'm curious though -- is it completely necessary to estimate the variances of the exogenous variables? If it is not necessary, is there a reason why you estimate them?

[@mikecrowson2462]

Hi Bradley, the variances and covariances of the exogenous variables are generally estimated in SEM. If you use the 'sem' function instead of the 'lavaan' function in lavaan, you don't have to be as explicit about it. But they are still being estimated in the total number of parameters. In a path model such as this, the parameters that are estimated are: (a) paths, (b) variances of exogenous variables, (c) covariances of exogenous variables, (d) variances of disturbance terms (i.e., prediction errors associated with endogenous variables, and (e) any covariances among disturbances (if any are included in the model). Hope this helps!

[@bradleypan-weisz2105]

@@mikecrowson2462 Thank you for your quick response!

[@ssrrapper]

Question...at 10:54, to estimate the covariances between ses, mastery, and perfgoal, you created the line "mastery~~perfgoal+ses". Does it matter which variable is on the left side? Could you have typed "perfgoal~~mastery+ses" or "ses~~mastery+perfgoal"?

[@brittjane4486]

Thanks, this is great. This would be easier to see the output in R Studio - is there a reason you don't use it?

[@mikecrowson2462]

You're very welcome. Regarding your question: No, there's no reason I don't use R-studio beyond the fact that I'm just used to doing things the way I learned with R. I'm probably more like folks who still use LISREL matrix command language instead of using SIMPLIS language (which is way easier). I guess I'm a bit "old-fashioned"(why drive when I can take a horse and carriage, right? haha). However, I'll certainly consider incorporating it in future in R presentations - especially if it might help folks see things on the screen. Thanks for the suggestion. Best wishes!

[@Ashenange1]

Hi Mike, these videos are really helpful. I was wondering if you were planning on making a video that shows how to incorporate interaction terms or if you have one and I missed it?

[@xuyang2776]

Hello,Mike. Could you tell me how to get the residual vairances of a MSE by lavaan()? Thanks

[@saharsalehi8070]

Hi Mike. Thank you for the informative clip you made. At exactly 4:10-4:15, you mention a program you used to make the path model. May you please help me to understand what software you are mentioning?!

[@mikecrowson2462]

Hi Sahar, I used the AMOS Sem program (it is integrated with SPSS) to draw out the model (it saved me from doing this in Powerpoint, which is a pain). I actually did not run it or anything in AMOS. AMOS just has a good drawing program that is easy to use - and I tend to like to include visuals when I am teaching. Unfortunately, I have yet to see any options in R that will allow such nice drawings. There are options, but they are not the best in my opinion. Nevertheless, lavaan is way more powerful than AMOS in terms of options. Cheers!

[@saharsalehi8070]

@@mikecrowson2462 Many many thanks for your reply.

[@raimo1799]

Hi Mike! I really enjoy each of your tutorials they have always been helpful to me. By the way, I am trying to follow this video for my own analysis but unfortunately, it is not showing up. Actually, when I run the model with my variables, it appears the same in the console box, and when I try the following step (fit, for the analysis results), still there is no results appearing but an error notification. Could you please assist me? Hoping forward to read from you soon.

[@juliethasagun1634]

I think it would be helpful if your post your error notification.

[@annanoer5627]

Hello ! How do I overcome this: "Error in Sigma.inv * sample.cov : non-conformable arrays"

[@HarryJayCavite]

thank you so much

[@yergali]

Thank you! Just in case, install.packages("lavaanPlot") does not work anymore, do you have any ideas to solve it?

[@mikecrowson2462]

You're very welcome. Well, it looks like I will have to do some research on that possibility to see if there is anything new out there. I put this video together awhile back, so maybe things have changed. I'll check and update if I find something newer! Cheers!

[@Cjstistajh]

Hi, I wonder what if it is a non recursive model? Is it the same?

[@udimakov3702]

A very nice presentation Could you please supply a link for the data and text files? Cheers!

[@mikecrowson2462]

Hello. Thanks for visiting. The links to the data and text files are underneath the video description. cheers!

[@assefawasie9612]

dear mike it is good lecture thank you please help note link

[@Jaji1948]

What is SES?

[@mikecrowson2462]

SES is socioeconomic status. Cheers!

[@meenakshichakraborty2352]

Thank you so much for this extremely helpful video. You explained all the basics so clearly - really appreciate your time and effort! Saved me hours or days!