Back to: Assess for Risk of Bias

This is also where you would report the results of sensitivity analyses that separately analyze those studies with low risk of bias and determine whether conclusions differ from analyses including studies with high or unclear risk of bias

Publication bias occurs when not all studies that are conducted on a certain research question are published and the results of the studies that are published are different from those not published. As a consequence, if a team of systematic reviewers only include published studies, the systematic reviewers will not get the full picture of the research that’s been conducted. Instead, they will get a warped picture, or a biased picture. This bias is called publication bias.

Unfortunately, not all research is published and there’s clear evidence that a study is more likely to get published if it has more positive results (or results that indicate that the treatment is effective). The most likely consequence to your systematic review is that you may conclude that a treatment is more effective than it really is because the studies that weren’t positive are missing.

What to do about publication bias

To reduce the impact of publication bias, be sure to include unpublished research and other “grey literature” in your systematic review search. We talk more about how to do this in our “search and screen” course.