[R-meta] Capturing the variability at the lowest level of longitudinal studies

[Michael Dewey]

Dear Simon

I am not the world's greatest expert on these designs but I would have 
included ~1 | esid not ~time | esid

Incidentally you posted in HTML not plain text which leads to messages 
getting scrmbled.

Michael

On 01/05/2021 00:41, Simon Harmel wrote:
> Hi All,
> 
> I'm running a meta-analysis on a sample of longitudinal studies. However,
> there is so much variability both in the number and in the value of the
> time points used in each study.
> 
> My general data structure looks like:
> 
> study  yi  time  esid x1
>    1      .1    1     1
>    1      .2    4     2
>    2      .3    2     3
>    2      .4    2     4
>    2      .5    2     5
>    3      .6    3     6
> 
> I was wondering what --rma.mv()-- syntax can capture the variability at the
> lowest level in my data?
> 
> Currently, I'm using the following, which I know can't capture the
> variability at the lowest level in my data:
> 
> m1 <- rma.mv(yi ~ time + x1, V = V, struct = "HAR",
>                     random = ~time | study,
>                     data = data)
> 
> I also tried the following:
> 
> m2 <- rma.mv(yi ~ time + x1, V = V, struct = c("HAR","HAR"),
>                             random = list(~time | study, ~time | esid),
>                             data = data)
> 
> But I get a warning that says:
> "Each level of the outer factor contains only a single level of the inner
> factor, so fixed value of phi to 0. "
> 
> 
> I appreciate your expertise,
> Simon
> 
> 	[[alternative HTML version deleted]]
> 
> _______________________________________________
> R-sig-meta-analysis mailing list
> R-sig-meta-analysis using r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-meta-analysis
> 

-- 
Michael
http://www.dewey.myzen.co.uk/home.html