R - Mediation with Bootstrapping Lecture + Example

[@miao9732]

thank you very much for this video, I got three questions:1 should all variables folllow normal distribution when perform mediation analysis? 2. should I check the multicollinearity for mediator and dependent variable? 3. whether the sobel test is equal to bootstrap confidence interval, if so, can I choose one of them to report? looking forward to your reply, thanks!!!

[@StatisticsofDOOM]

Yes that's an assumption of linear regression - normality of the residuals. No, you can't have multicollinearity with the DV, having a good relationship with the DV is the entire point of regression. No, they are not equal, you should probably use the bootstrap results as a better estimate.

[@miao9732]

@@StatisticsofDOOM Dear Dr. Erin, thank you so much for the reply!!! all the points are clear now.

[@Arthur-fz5dw]

Thank you so much for this course! It is interesting and could be useful, however, the first link in the description doesn't work and I can't find the script.

[@StatisticsofDOOM]

Working on updating all the links now on the new website. I think this is the file you are looking for osf.io/nsr72/ - also the power point is the c7 one in that same folder.

[@Arthur-fz5dw]

@@StatisticsofDOOM Fantastic, thank you very much!

[@BlackZeroHawk]

Idk if you realize this or not but we cant see the coding you do since you are showing half of your screen. So in min 30 when you are trying to explain R its hard to understand cause we cant see anything.

[@StatisticsofDOOM]

Yes unfortunately the screen cut off ... maybe because of powerpoint? Not sure of the culprit. You can download the scripts to follow along at the link in the description.