Help me please, I'm just learning to do meta-analysis using the R program. When I use summary measure with mean differences (MD) mode "metacont(Ne, Me, Se, Nc, Mc, Sc, data = data1, studlab = paste(Author), sm = "MD", common = T, random = T, hakn = T, prediction = T, subgroup = Sbg)" everything works normally.
However, when I change the summary measure to standardized mean differences (SMD) mode "metacont(Ne, Me, Se, Nc, Mc, Sc, data = data1, studlab = paste(Author), sm = "SMD", common = T , random = T, hakn = T, prediction = T, subgroup = Sbg)" an error message appears as follows "Error in (function (yi, vi, sei, weights, ai, bi, ci, di, n1i, n2i, x1i , : Fisher scoring algorithm did not converge. See 'help(rma)' for possible remedies."
Thank you
As referenced in this link,
https://www.metafor-project.org/doku.php/tips:convergence_problems_rma,
you need to increase the number of iterations to achieve convergence.
Hello Puguh Oktavian. I'm having the same problem you had regarding the Fisher conversion for the data with the "meta" package. How were you able to resolve it?
Thanks for your attention
Hello Gustavo Waclawovsky I solved the problem by changing method.tau to "DL" which refers to the DerSimonian-Laird estimator.
Glad you managed to resolve it! I will try to follow this path. Thank you for the immediate response.
You are welcome; The following is my code for running meta-analysis, hopefully it can help.
metacont( n.e = Ne, mean.e = Me, sd.e = Se,
n.c = Nc, mean.c = Mc, sd.c = Sc,
median.e = mede, q1.e = e1, q3.e = e3,
median.c = medc, q1.c = c1, q3.c = c3,
data = datatir1, studlab = paste(Author), prediction = T,
subgroup = NULL, subgroup.name = NULL,
method.tau = "DL", method.random.ci = "classic",
sm = "MD")
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Writing