Ryan Snead posted on Monday, July 24, 2017 - 7:04 am
Hello! We want to investigate multiple mediation. The design links individuals with providers and there's a 2nd-level var in the model. ICC = .27. I was hoping for guidance on syntax and interpreting the findings. Mainly, how to specify %between%/%within%/TYPE. Is there a method to adjust clustering w/o using multilevel?
We have a sample of 100. 53 providers, 28 have gt 1 nested. I would like to run a posthoc power analysis given the sample size and parameters. Are there examples on how to perform a Monte Carlo simulation meeting my needs? Is this possible? I saw example syntax on the website, but couldn’t tell if they closely aligned with our model.
Below is a sample of syntax for unclustered path model. emoexaus is 2nd-level var. Thank you for all of your help.
MODEL: waigoal ON emoexaus emoavail; satswprg ON waigoal; dysenrld ON momage hispanic satswprg waigoal emoexaus emoavail;
MODEL INDIRECT: dysenrld IND emoavail; dysenrld IND emoexaus; dysenrld IND waigoal;
You can use Type=Complex to handle clustering without twolevel.
Models are shown in the articles by Preacher and Zyphur on our website under Papers, Multilevel SEM - such as
Preacher, K., Zyphur, M. & Zhang, Z. (2010). A general multilevel SEM framework for assessing multilevel mediation. Psychological Methods, 15, 209-233. download paper contact first author show abstract
Regarding Monte Carlo - all UG examples have montecarlo counterparts on our website.