[R-meta] help

[Michael Dewey]

I do not think there is anything much you can do here which is different 
from any other situation where you have missing data. Some form of 
imputation would be possible if you have other variables which would 
enable you to do it.

I wonder though whether it is going to be worth spending too much time 
on what is an ecological analysis. You are not looking at the effect of 
age but at the effect of being enrolled in a study which had people of a 
particular average age.

If you post again please post in plain text, not HTML as your data 
arrived here in a very scrambled format. In fact outputting it using 
dput() may be even better.

Michael

On 02/11/2017 00:39, Yalemzewod Gelaw wrote:
> Dear team,
> 
> 
> 
> As usual, I come today looking your help. Thank you for the metafor
> package.
> 
> I am doing metareg for dealing with the source of heterogeneity. The
> moderator (mean age) has missing values: 22 studies out of 68 didn’t report
> the mean age of the study group. Whn I ran the model the error message
> displayed as below;
> 
> 3.6          35.4        28.4        32           37.6        39
> 1.05        31.75      NA          NA          NA          NA
> NA
> 32           37.8                34.9        NA          30
> 30.5
> 29.8        35           30           24.8        32.7        35
> NA          28.7        34.2        NA                34.9        28
> 28           31.3        NA          35           Na           35
> NA          28.5        38.1        NA          NA          33.3
> 34.2        6.2          NA          25.5        33.4        32.6
> 29.5        NA          42           37.2        33.7        NA
>    8.5
> 39.5                NA          NA          NA          NA          NA
> 26.6        35.2        NA          33.3        8.5          35
> 
> ress <- rma(yi,vi,mods = ~meanage_year, data = datz)
> 
> Warning message:
> 
> In rma(yi, vi, mods = ~meanage_year, data = datz) :
> 
>    Studies with NAs omitted from model fitting.
> 
> So, my question is what functions of metafor do help to impute this missing
> values. I wonder if you can also provide examples done before.
> 
> *Sincerely, *
> 
> 
> *Yalemzewod Assefa Gelaw  (Yalem)*
> 
> *Ph.D. Candidate, the University of Queensland, Australia *
> *Email: yalassefa at gmail.com <http://yalassefa@gmail.com/y.gelaw@uq.edu.au>*
> 
> 
> *"Change is good ... but it hurts "*
> 
> 	[[alternative HTML version deleted]]
> 
> _______________________________________________
> R-sig-meta-analysis mailing list
> R-sig-meta-analysis at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-meta-analysis
> 

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