Publication bias

Studies with statistically significant results are more likely to get published than those with non-signifcant results. Meta-analyses that are exclusively based on published literature may therefore produce biased results. This type of bias can be assessed by a funnel plot.