Metanalysis (random effects model error "Convergence not achieved during tau2 estimation"*)

[Isabella Vanorio]

Hello all,

I am running a metanalysis using a precomputed effect size and standard error.
I have no issues declaring the data for the metanalysis but when I select the random effect models for these analysis I get the error "Convergence not achieved during tau2 estimation". Summary command runs when I use a fixed-effect model but this type of model is not appropriate for the analysis of the data.

**************code*************
meta set Meanatbaseline SDbaseline if ( _Study_design=="A" ), studylabel(_author_year) studysize( _pop) eslabel( baseline )
(75 missing values generated)

Meta-analysis setting information

Study information
No. of studies: 15
Study label: _author_year
Study size: _pop

Effect size
Type: Generic
Label: baseline
Variable: Meanatbaseline

Precision
Std. Err.: SDbaseline
CI: [_meta_cil, _meta_ciu]
CI level: 95%

Model and method
Model: Random-effects
Method: REML


. meta summarize
Convergence not achieved during tau2 estimation


thank you for your help
IV

[Claire Rich]

Hi all,

I'm wondering if anyone has insights on this issue? I'm running into something similar- running random effects model w/ reml.

Claire