A funnel plot is performed to compare studies included in a meta-analysis. This can help to determine if the conclusions of the meta-analysis are valid.

The funnel plot depends on 2 axes at right angles to each other:

(1) vertical axis: sample size, standard error, precision

(2) horizontal axis: effect measure (odds ratio, risk ratio, etc)


If all of the studies are all plotted on the same pair of axes, then the points should form a pattern relative to the effect measure of the studies with the largest sample size.


If the points are symmetrical and roughly form the shape of a funnel, then the studies are homogeneous and the probability of significant bias is low.


If the plot is asymmetrical or skewed, then the cause needs to be determined.


Causes of asymmetry in a funnel plot:

(1) selection bias for inclusion in the meta-analysis (publication bias, language bias, citation bias, multiple publication bias)

(2) true heterogeneity between studies (intensity of intervention, differences in underlying risk)

(3) data irregularities (inadequate analysis, fraud, poor study design)

(4) artefact due to the choice of effect measures

(5) chance


If the studies included in the meta-analysis are few in number or have small sample sizes, then the funnel plot can be unreliable.

To read more or access our algorithms and calculators, please log in or register.