Funnel plot

A funnel plot is a graph that allows you to check a suspected publication bias in the context of a meta-analysis.

The common representation is a scatter diagram in a Cartesian coordinate system wherein the treatment effect is plotted on the x -axis and the size of the study on the y- axis. Already by simply viewing a funnel plot can be evaluated: a symmetrical shape arises from a balanced study publication, which reflects the results in the context of the natural statistical variation. Therefore, larger studies should achieve more precise results that are closer to the mean of all study results. If the image by many small studies on one side of the mean asymmetric, this argues strongly for non-published or carried out studies with opposite results.

Criticisms

A major problem of the funnel plot is that it does not capture the quality of the included trials as a very simple tool. If studies with higher precision actually produce different results than those with lower so would result in a false impression of publication bias. These criteria must, therefore, be reviewed by the analyst himself. It may also happen that can not be publications on a topic reduce to a single number, as for example the selection of subjects can (gender, age, ...) or other circumstances vary.