Skip to main content
. 2005 Jan 25;92(3):434–444. doi: 10.1038/sj.bjc.6602358

Figure 5.

Figure 5

log RR – log relative risk; s.e. – standard error (A) Funnel plot of data on survival in studies using IHC. As discussed in the text, there is evidence of publication bias – asymmetry due to lack of negative studies with high standard error and low statistical power. (B) Funnel plot of the same group of studies after trimming and filling. The dummy studies are indicated by circles within squares, and the genuine studies, as in (A), by circles.