Health Sciences and Human Physiology and Nutrition Research

This page covers some common questions about how to evaluate and read research articles. 

If you're looking for information on how to determine if an article is peer reviewed, check out the Peer-Reviewed Articles page. 

Who is represented in the sample and who is missing? 

Think critically about the representation of race, gender, disability, and economic status in human studies. 

How was this sample collected? 

Often samples can be self-selecting or inherently skewed. For instance, if you conduct interviews on diet in a pricey health foods store, you're going to get a very different sample than you might in a different location. Your results are going to be skewed by the population included. 

Does the methodology make sense for the question being asked? 

You get the data you ask for. Do the tools used in this study provide data that would actually answer the research question?

Do the conclusions seem reasonable based on the results? 

If 55% of students said they liked this webpage, would it be okay if I said a vast majority of students liked it? No. Check that the conclusions don't overstate the results. 

Do the limitations of the study invalidate the results?

Every study has limitations, but consider them critically and make sure you address them when you talk about conclusions from an article. 

Is this a primary or secondary source?

Most studies should be based on original research, which would make them primary sources. Secondary sources, like book reviews, are based on other studies and may not be acceptable sources for your project

The use of evidence in order to provide the best care for patients is often referred to as Evidence-Based Practice (EBP). Articles used in EBP should have a population, an intervention (or variable of interest), and gather and interpret the results to offer a recommendation. 

When discussing "evidence based practice" and articles that are "evidence based," there are several common models that put types of studies in a pyramid according to how much evidence they provide. For example a case study with one subject may not provide as much evidence as a randomized controlled trial with 400 participants. The amount of evidence is determined largely by the number of participants, but also the presence of a control group. Studies at the top of the pyramid are also supposed to have the least room for bias.

Here's the one pyramid model of evidence (CC-0): 

A few things to keep in mind:

• check the methods section of the article to see what type of article it is/where it fits on the pyramid
• depending on your question, a case study or small qualitative study may be better than a large quantitative study
• if a large sample lacks diversity or doesn't cover the population you're interested in, its results could still have limited applicability for your research
• the pyramid here only represents one way of evaluating evidence. It is neither foolproof, nor absolute. Be sure to consider your assignment or research question requirements.