Power analysis for a mediational analysis using bootstrapping

[Neuropsych2be]

---

Members don't see this ad.

Does anyone know of a good program or resource for doing a power analysis for an analysis of mediation using the bootstrapping technique. I need to estimate the sample size I would need for a mediation analysis. Any good software or resources for this??

 

[neuronic]

Maybe G Power, but it has been a long while since I used it.

 

[ADDICTED2STATS]

Does anyone know of a good program or resource for doing a power analysis for an analysis of mediation using the bootstrapping technique. I need to estimate the sample size I would need for a mediation analysis. Any good software or resources for this??

You can run a monte-carlo simulation in MPlus. I use this for all my power analyses. There is a paper on the MPlus website describing how this is done. In addition, Lesa Hoffman has some excellent materials on this exact topic. You can find it by Googling something like Hoffman UNL Mplus mediation power analysis... you get the idea. If you use her stuff, be sure to give credit where it's due.

 

[CheetahGirl]

Does anyone know of a good program or resource for doing a power analysis for an analysis of mediation using the bootstrapping technique. I need to estimate the sample size I would need for a mediation analysis. Any good software or resources for this??

Here's a great article:

Hayes, A. F. (2009). Beyond Baron and Kenny: Statistical mediation analysis in the new millennium. Communication Monographs, 76, 408-420.

I believe Hayes wrote a macro for SPSS. If I use the bootstrapping method, then I'll probably go this route. I'm not there yet or I'd share more info. Good luck with it though!

 

[Ollie123]

Don't think the Hayes paper covers power analysis (unless I'm confusing it with another).

Really have to learn MPlus. University just got a copy for the grad lab so I may play around with it a bit. GPower can't do it. PASS may be able to (not sure but there's a free trial so you can find out).

If you don't want to splurge on software, its certainly possible in R. Downside being I suspect you'd need to write the code yourself. Shouldn't be too hard if you don't mind a brute force approach (i.e. create tons of simulated datasets to analyze, run the analysis and track yes/no significance decisions, set it iterate and alter the parameters until you achieve your desired outcome). Depending on the complexity of your results though, could take a few hours to run, so make sure to include intermittent save procedures in case of a crash.

 

[cara susanna]

MPlus is my new favorite program 🙂 Although I still have a lot to learn about how to use it.

 

[Pragma]

MPlus rocks once you get the hang of it 🙂

 

[Neuropsych2be]

Thanks to everyone for the suggestions! I've run across Mplus but have not used it.