Skip to Main Content

Research

Evaluating your search results

You will always need to evaluate resources to establish if they are suitable for your purpose: 

RELEVANT?  This is likely to be your first consideration. You don't have to read the whole article to evaluate relevance - look at the title, then the abstract and if those pass your relevance test access the full-text.  Then maybe read the introduction and/or methodology and conclusion to get a good idea of the contents before committing time to read the whole thing.  It can also be helpful to display your search question next to your computer as you review search results.

RECENT?  Is the information too old to be useful?  Sometimes older articles are "seminal", influential and still relevant. Not every piece of research has to be repeated every 5 years - but be very aware of what might have changed in terms of medical practice or social situations since an older article was published.

RELIABLE? Just because an article was published in a scholarly journal does not mean you can accept its findings without proper consideration.  Is the article a piece of original research? If so is the methodology sound, does it include a large enough sample of the population being considered etc. If it is a review article does it have something useful to say and are the findings backed up by any original research cited?

 

Evidence hierarchy

When evaluating information sources you can consider the evidence pyramid.  Information at the top of the pyramid is considered the most evidence-based because it has drawn from the widest pool of evidence in a systematic way.  More information ...

When reviewing accounts of research you can consider:

  • Effect size: What was the effect size of the primary outcome? Was it statistically significant?
  • Sample size: Was the sample size large enough to show a difference? Power analysis should be used to determine the number of subjects required to identify a pre-determined difference between groups, usually for the primary outcome
  • Intention to treat (ITT) analysis: is primarily an analysis of all participants in the group to which they were originally assigned regardless of whether they dropped out of the study or changed groups. This term may also be used for the imputation of data for participants who were lost to follow-up.
  • Other sources of bias: e.g.conflict of Interest- meta-analyses have shown that industry-sponsored studies tend to show increased effect sizes