Thank you for answering. Yes , I tried the commands above. Attached is the funnel plot with the suggested commands. I guess is better to present the log odds, since there is no option to choose the FTT and back transformation in the funnel plot. I guess I was trying to use STATA provided command to build the funnel plot with sample size on Y axis.
I was not aware that random effects model was not appropriate for a funnel plot, as I understood, the choice between fixed-effect and random-effects models for the summary estimate in funnel plots should be determined by the underlying assumptions about study heterogeneity and the objective of the meta-analysis. Would you be so kind to share literature about the random effects being not appropiate? I would love to learn more.
Thank you again,
Virginia
I was not aware that random effects model was not appropriate for a funnel plot, as I understood, the choice between fixed-effect and random-effects models for the summary estimate in funnel plots should be determined by the underlying assumptions about study heterogeneity and the objective of the meta-analysis. Would you be so kind to share literature about the random effects being not appropiate? I would love to learn more.
Thank you again,
Virginia
Attached Files
Comment