Jessup Playbooks Logo

How do I evaluate evidence from studies?

Not all evidence is created equal. Learning the different levels of evidence will help you evaluate scientific studies and other materials.


Profile Photo
L Skinner

Need Help?

Ask a Librarian @
Betty Sue Jessup Library

501 College Drive
Charlottesville, VA 22902

What's the best evidence?

You're reading a magazine and find an article entitled "Study Shows Drinking Coffee Is Good for Your Health." Later, mug of coffee in hand, you are reading a news website and happen upon this article: "Study Shows Drinking Coffee Is Dangerous to Your Health."  You pour your coffee into the sink and wonder which article is true.

Is there a way to figure out which of these studies is right?

The best method to decide on the value of a study is to check it against the Hierarchy of Evidence, a tool that ranks the strength of a study's evidence based on the way the study was designed.

The Hierarchy of Evidence

The types of evidence listed below are ranked from weakest to strongest. Keep in mind that sometimes a study may not be able to provide certain types of evidence for various reasons, including legal and/or ethical ones.

  • Anecdotal & Expert Opinions
    • Anecdotal evidence is someone's personal experience or viewpoint, and it doesn't necessarily represent typical experience. An expert's opinion (by itself or as part of a news article) is considered a weak form of evidence when it doesn't have studies to back it up.
  • Animal & Cell Studies (experimental)
    • Animal research cannot always predict effects in humans, which is why human trials are often required. Experiments on isolated cells can yield effects that are different from those in the body.
  • Case Reports & Case Series (observational)
    • Case reports (written record tracking a single subject) and case series (written record tracking multiple subjects) can provide evidence that may lead to the discovery of new conditions or of side effects of treatments.  These two study types only provide evidence of correlation (a relationship between two variables), not causation (cause and effect).
  • Case-Control Studies (observational)
    • Case-control studies are retrospective (meaning variables are tracked back to search for origins or causes) and involve two groups of subjects: one has the condition,characteristics, or symptoms and the other doesn't. These studies are quick, relatively inexpensive to conduct, and don't need too many subjects. These studies prove correlation but not causation.
  • Cohort Studies (observational)
    • Similar to case-control studies. Two groups are compared over time to notice any difference in outcomes. One group has the condition, characteristic, symptom, treatment, etc., and the other one doesn't. This type of study requires large groups of subjects and can take a long time to implement.
  • Randomized Controlled Trials (experimental)
    • For this type of study subjects are assigned to a test group (the group that receives the treatment) or to a control group (which typically receives a placebo). RCTs can be blind (participants don't know whether they're in the test or the placebo group) or double blind (neither participants nor experimenters know who is in what group). Blinding is done to remove bias.
  • Systematic Review
    • Systematic reviews follow an orderly procedure to collect and critically evaluate design, methods, data, conclusions, and more from multiple RCTs. Systematic reviews reduce bias and provide a more complete and accurate picture, and as such are considered the most reliable type of evidence.

Need Help?