IYH Dear Giuseppe Dario Testa

My solution suggestion assumes you are using Stata v14 or later (h/t Bruce Weaver for his pertinent scoping).

The error occurred using the reffect option wthe metatrim command in STATA. The reffect option is used to specify that the effect size estimates are in the form of raw effect sizes (e.g., Cohen's d, Hedges' g) rather than the standardized mean difference (SMD). metatrim in STATA does not support the reffect option.

Here is the syntax that you used:

db metatrim . metatrim _ES _seES, reffect eform Note: default data input format (theta, se_theta) assumed. subcommand meta __000005 is unrecognized r(199);

To correct the error, you can use the metan command instead of metatrim to estimate the trimmed mean and its standard error for a set of raw effect size estimates.

metan _ES _seES, trim(0.1) eform

In this syntax, _ES and _seES are the variables containing the effect size estimates and their standard errors, respectively. The trim(0.1) option is used to specify that the trimmed mean should be estimated using a trimming proportion of 0.1 (i.e., 10% of the effect size estimates will be excluded from the analysis). The eform option is used to display the results in the effect size metric rather than the standardized mean difference.

See pp.6- Article Metan: Fixed- and Random-Effects Meta-Analysis

Hello Giuseppe Dario Testa. What version of Stata do you have? I ask, because if it is recent enough (v16 or later, IIRC), it will have a suite of new (official) meta-analysis commands, including this one:

• https://www.stata.com/manuals/metametatrimfill.pdf

PS- I recommend posting questions about how to do X using Stata to Statalist. If you do decide to post there in future, see the very helpful FAQ first.

• https://www.statalist.org/forums/forum/general-stata-discussion/general
• https://www.statalist.org/forums/help