[Tom Salvitti]

Thanks a lot Joseph Coveney.

After doing this, how to produce the forest plot in which, in addition to the effect size of the individual studies with the new IC, the over all one is seen? tHANKS A LOT!!!!

[Joseph Coveney]

I don't do forest plots, sorry. It doesn't look too overwhelming, though: you might take a look through the graphing code sections of some of the user-written meta-analysis packages in order to get some pointers on how to do it.

Keep in mind that those packages are typically for meta-analysis of randomized controlled trials in which, say, odds ratio is the effect measure for each study. In your case, you don't have that, and so I'm guessing that you'll want to plot the odds for each study (and confidence interval) rather than prevalence (and its back-transformed interval).

[Tom Salvitti]

yes..Thanks the same

[Leonardo Guizzetti]

As suggested in #12, you can also proceed with -metan- if you prefer a meta-analysis package, which will also give you a forest plot.

metan pos tot , proportion transf(logi)

Partial output:

. metan pos tot , proportion transf(logit)

Studies included: 17
Participants included: 33182

Meta-analysis pooling of Logit-transformed Proportions
using the common-effect inverse-variance model

--------------------------------------------------------------------
Study                | Proportion  [95% Conf. Interval]   % Weight
---------------------+----------------------------------------------
1                    |     0.032      0.013     0.079       2.45
2                    |     0.087      0.034     0.203       2.31
3                    |     0.015      0.007     0.035       3.12
4                    |     0.012      0.002     0.067       0.63
5                    |     0.029      0.005     0.149       0.61
6                    |     0.006      0.001     0.031       0.63
7                    |     0.005      0.001     0.030       0.63
8                    |     0.063      0.021     0.168       1.78
9                    |     0.002      0.001     0.005       3.16
10                   |     0.018      0.007     0.045       2.49
11                   |     0.013      0.002     0.068       0.63
12                   |     0.003      0.002     0.005      19.56
13                   |     0.014      0.011     0.017      44.97
14                   |     0.002      0.001     0.003      13.90
15                   |     0.022      0.006     0.076       1.24
16                   |     0.001      0.000     0.003       1.27
17                   |     0.005      0.001     0.029       0.63
---------------------+----------------------------------------------
Overall, IV          |     0.008      0.007     0.009     100.00
--------------------------------------------------------------------

[Tom Salvitti]

thank you

[Tom Salvitti]

Thanks Leonardo Guizzetti I run the instruction but it gives me errorr and says to me "option proportion not allowed". I use STATA 16..


What can I do??

[Joseph Coveney]

I run the instruction but it gives me errorr and says to me "option proportion not allowed". I use STATA 16..

Works for me.

.ÿ
.ÿversionÿ17.0

.ÿ
.ÿclearÿ*

.ÿ
.ÿlocalÿdirÿhttp://fmwww.bc.edu/repec/bocode/

.ÿ
.ÿ#delimitÿ;
delimiterÿnowÿ;
.ÿlocalÿfile_listÿ
>ÿÿÿÿÿÿÿmetan.ado
>ÿÿÿÿÿÿÿmetan.mata
>ÿÿÿÿÿÿÿmetan.sthlp
>ÿÿÿÿÿÿÿmetan_binary.sthlp
>ÿÿÿÿÿÿÿmetan_continuous.sthlp
>ÿÿÿÿÿÿÿmetan_model.sthlp
>ÿÿÿÿÿÿÿmetan_pooling.ado
>ÿÿÿÿÿÿÿmetan_proportion.sthlp
>ÿÿÿÿÿÿÿmetan_hlp_run.ado
>ÿÿÿÿÿÿÿmetani.ado
>ÿÿÿÿÿÿÿmetani.sthlp
>ÿÿÿÿÿÿÿmetannt.ado
>ÿÿÿÿÿÿÿmetannt.sthlp
>ÿÿÿÿÿÿÿforestplot.ado
>ÿÿÿÿÿÿÿforestplot.sthlp
>ÿÿÿÿÿÿÿlabbe.ado
>ÿÿÿÿÿÿÿlabbe.sthlp
>ÿÿÿÿÿÿÿlmetan.mlib;

.ÿ#delimitÿcr
delimiterÿnowÿcr
.ÿ
.ÿforeachÿfileÿofÿlocalÿfile_listÿ{
ÿÿ2.ÿÿÿÿÿÿÿÿÿlocalÿsubdirÿ=ÿsubstr("`file'",ÿ1,ÿ1)
ÿÿ3.ÿÿÿÿÿÿÿÿÿcopyÿ"`dir'`subdir'/`file'"ÿ`c(pwd)'/`file',ÿreplace
ÿÿ4.ÿ}

.ÿ
.ÿquietlyÿinputÿbyteÿsidÿintÿtotÿbyteÿpos

.ÿ
.ÿmelogitÿposÿ||ÿsid:ÿ,ÿbinomial(tot)ÿnolog

Mixed-effectsÿlogisticÿregressionÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿNumberÿofÿobsÿÿÿÿÿ=ÿÿÿÿÿÿÿÿÿ17
Binomialÿvariable:ÿtot
Groupÿvariable:ÿsidÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿNumberÿofÿgroupsÿÿ=ÿÿÿÿÿÿÿÿÿ17

ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿObsÿperÿgroup:
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿminÿ=ÿÿÿÿÿÿÿÿÿÿ1
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿavgÿ=ÿÿÿÿÿÿÿÿ1.0
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿmaxÿ=ÿÿÿÿÿÿÿÿÿÿ1

Integrationÿmethod:ÿmvaghermiteÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿIntegrationÿpts.ÿÿ=ÿÿÿÿÿÿÿÿÿÿ7

ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿWaldÿchi2(0)ÿÿÿÿÿÿ=ÿÿÿÿÿÿÿÿÿÿ.
Logÿlikelihoodÿ=ÿ-49.337124ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿProbÿ>ÿchi2ÿÿÿÿÿÿÿ=ÿÿÿÿÿÿÿÿÿÿ.
------------------------------------------------------------------------------
ÿÿÿÿÿÿÿÿÿposÿ|ÿCoefficientÿÿStd.ÿerr.ÿÿÿÿÿÿzÿÿÿÿP>|z|ÿÿÿÿÿ[95%ÿconf.ÿinterval]
-------------+----------------------------------------------------------------
ÿÿÿÿÿÿÿ_consÿ|ÿÿ-4.770905ÿÿÿ.3262176ÿÿÿ-14.62ÿÿÿ0.000ÿÿÿÿ-5.410279ÿÿÿÿ-4.13153
-------------+----------------------------------------------------------------
sidÿÿÿÿÿÿÿÿÿÿ|
ÿÿÿvar(_cons)|ÿÿÿ1.350316ÿÿÿ.5867997ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ.5761466ÿÿÿÿ3.164736
------------------------------------------------------------------------------
LRÿtestÿvs.ÿlogisticÿmodel:ÿchibar2(01)ÿ=ÿ114.15ÿÿÿÿÿÿProbÿ>=ÿchibar2ÿ=ÿ0.0000

.ÿnlcomÿoverall:ÿinvlogit(_b[_cons]),ÿnoheader

------------------------------------------------------------------------------
ÿÿÿÿÿÿÿÿÿposÿ|ÿCoefficientÿÿStd.ÿerr.ÿÿÿÿÿÿzÿÿÿÿP>|z|ÿÿÿÿÿ[95%ÿconf.ÿinterval]
-------------+----------------------------------------------------------------
ÿÿÿÿÿoverallÿ|ÿÿÿ.0084015ÿÿÿ.0027177ÿÿÿÿÿ3.09ÿÿÿ0.002ÿÿÿÿÿ.0030749ÿÿÿÿ.0137281
------------------------------------------------------------------------------

.ÿ
.ÿmetanÿposÿtotÿ,ÿproportionÿtransf(logit)

Studiesÿincluded:ÿ17
Participantsÿincluded:ÿ33182

Meta-analysisÿpoolingÿofÿLogit-transformedÿProportions
usingÿtheÿcommon-effectÿinverse-varianceÿmodel

--------------------------------------------------------------------
Studyÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿProportionÿÿ[95%ÿConf.ÿInterval]ÿÿÿ%ÿWeight
---------------------+----------------------------------------------
1ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ0.032ÿÿÿÿÿÿ0.013ÿÿÿÿÿ0.079ÿÿÿÿÿÿÿ2.45
2ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ0.087ÿÿÿÿÿÿ0.034ÿÿÿÿÿ0.203ÿÿÿÿÿÿÿ2.31
3ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ0.015ÿÿÿÿÿÿ0.007ÿÿÿÿÿ0.035ÿÿÿÿÿÿÿ3.12
4ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ0.012ÿÿÿÿÿÿ0.002ÿÿÿÿÿ0.067ÿÿÿÿÿÿÿ0.63
5ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ0.029ÿÿÿÿÿÿ0.005ÿÿÿÿÿ0.149ÿÿÿÿÿÿÿ0.61
6ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ0.006ÿÿÿÿÿÿ0.001ÿÿÿÿÿ0.031ÿÿÿÿÿÿÿ0.63
7ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ0.005ÿÿÿÿÿÿ0.001ÿÿÿÿÿ0.030ÿÿÿÿÿÿÿ0.63
8ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ0.063ÿÿÿÿÿÿ0.021ÿÿÿÿÿ0.168ÿÿÿÿÿÿÿ1.78
9ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ0.002ÿÿÿÿÿÿ0.001ÿÿÿÿÿ0.005ÿÿÿÿÿÿÿ3.16
10ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ0.018ÿÿÿÿÿÿ0.007ÿÿÿÿÿ0.045ÿÿÿÿÿÿÿ2.49
11ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ0.013ÿÿÿÿÿÿ0.002ÿÿÿÿÿ0.068ÿÿÿÿÿÿÿ0.63
12ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ0.003ÿÿÿÿÿÿ0.002ÿÿÿÿÿ0.005ÿÿÿÿÿÿ19.56
13ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ0.014ÿÿÿÿÿÿ0.011ÿÿÿÿÿ0.017ÿÿÿÿÿÿ44.97
14ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ0.002ÿÿÿÿÿÿ0.001ÿÿÿÿÿ0.003ÿÿÿÿÿÿ13.90
15ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ0.022ÿÿÿÿÿÿ0.006ÿÿÿÿÿ0.076ÿÿÿÿÿÿÿ1.24
16ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ0.001ÿÿÿÿÿÿ0.000ÿÿÿÿÿ0.003ÿÿÿÿÿÿÿ1.27
17ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ0.005ÿÿÿÿÿÿ0.001ÿÿÿÿÿ0.029ÿÿÿÿÿÿÿ0.63
---------------------+----------------------------------------------
Overall,ÿIVÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ0.008ÿÿÿÿÿÿ0.007ÿÿÿÿÿ0.009ÿÿÿÿÿ100.00
--------------------------------------------------------------------

Testÿofÿoverallÿeffectÿ=ÿ0:ÿÿzÿ=ÿ-60.810ÿÿpÿ=ÿ0.000


Heterogeneityÿmeasures,ÿcalculatedÿfromÿtheÿdata
withÿConf.ÿIntervalsÿbasedÿonÿnon-centralÿchi²ÿ(common-effect)ÿdistributionÿforÿQ
---------------------------------------------------------
Measureÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿValueÿÿÿÿÿÿdfÿÿÿÿÿÿp-value
---------------------+-----------------------------------
Cochran'sÿQÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿ164.96ÿÿÿÿÿÿÿ16ÿÿÿÿÿÿ0.000
ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿÿÿÿÿÿÿÿ-[95%ÿConf.ÿInterval]-
Hÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ3.211ÿÿÿÿÿ2.724ÿÿÿÿÿ3.681
I²ÿ(%)ÿÿÿÿÿÿÿÿÿÿÿÿÿÿÿ|ÿÿÿÿÿ90.3%ÿÿÿÿÿ86.5%ÿÿÿÿÿ92.6%
---------------------------------------------------------
Hÿ=ÿrelativeÿexcessÿinÿCochran'sÿQÿoverÿitsÿdegrees-of-freedom
I²ÿ=ÿproportionÿofÿtotalÿvariationÿinÿeffectÿestimateÿdueÿtoÿbetween-studyÿheterogeneityÿ(basedÿonÿQ)

.ÿquietlyÿgraphÿexportÿforest.png,ÿreplace

.ÿ
.ÿexit

endÿofÿdo-file


.

What can I do??

Maybe you didn't install the correct meta-analysis user-written command?

Again, if you're going to stick with this kind of graph, then I recommend trying to tweak it to plot odds instead.

[Leonardo Guizzetti]

I posted a working example (as it commonplace here by also showing the output) so I agree with Joseph that you likely need to install (or update) -metan-.

This is the version I was using, and I seem to recall David adding the -proportion- in a more recent version.

. which metan
c:\ado17\plus\m\metan.ado
*! version 4.05  29nov2021
*! Current version by David Fisher
*! Previous versions by Ross Harris and Michael Bradburn