Once you’ve written a clear and specific question, you cast a wide net to find studies that might be related to your topic.
While you’re searching, be aware that the studies that make it to publication don’t reflect the entire body of research on a topic. Studies might go unpublished for a lot of reasons: perhaps their results aren’t statistically significant or their conclusions aren’t novel enough to get attention. This is called publication bias.
You can help combat publication bias by seeking out unpublished research, checking out work presented at academic conferences, and considering other “grey literature” such as reports by government agencies and non-profits.
Another common problem in the published literature is reporting bias, when a study is published without all of its results, perhaps because some results are considered secondary or insignificant. In some cases, results are excluded for more nefarious reasons, especially if researchers leave out findings of harm. Failure to include these missing results in a systematic review can bias your findings.
You can combat reporting bias by following up with the authors of your studies to see if any results were left out. You can also compare the published results of a study with its protocol, which is written ahead of time. It’s easy to look up protocols for funded trials at Clinicaltrials.gov.
Once you’ve gathered a diverse collection of studies, you take a closer look, removing any studies that don’t address your specific question or don’t meet your inclusion criteria. You may need to look at the studies you find pretty carefully to do this right.
For more information, check out our Search and Screen Your Studies course.