I have selected 9 variables and three of them have publication bias in my meta-amalysis (egger's test). so is there any effective methods could reduce these potential influences? thanks very much.
Dear Weibin,
Actually by Egger's test you can not determine weather publication bias exists or not.
Egger's test and funnel plots show that publication bias (if exists) can be clinically important.
There is no way to decrease publication bias other than better search of the literature and inclusion of grey literature.
Best wishes,
Ramin
The Duval and Tweedie's Trim & Fill correction can help adjust the overall effect size for the extent of publication bias.
Thanks a lot for your answer, Ramin.
You are right. Egger's test and funnel plot can not totally determine whether publication bias exists or not, but the asymmetry funnel plot do revealed a potential publication bias or other bias. So I am woundering if there any methods could avoid or reduce these bias and improve the credibility of the results.
Dear Weibin,
I think you misinterpret Ramin's answer. I think he is trying to say that any type of publication bias analysis (like the funnel plot or Egger's test) is in fact incapable of actually detecting publication bias. However, the thing they can do is showing the potential influence of a publication bias on your meta-analytic result.
Best,
Henk
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis