Research in Brief: Correlation vs Causation

NARIC’s Research In Focus series highlights new and interesting findings from NIDILRR-funded studies, presented in lay language summaries. The series covers a wide array of topics, and aims to present peer-reviewed research in readable formats, so our readers can learn about issues that affect them every day. The Research in Brief companion series breaks down some of the concepts readers might come across when exploring research in disability and rehabilitation. This issue introduces the concepts of correlation and causation.

One of the central rules of research is that correlation does not equal causation. But what does this mean? Let’s look at a simple example. 

Let’s say you’re a scientist who is looking at the rates of sunburn in a group of people. You discover that people who reported having sunburn had also had ice cream recently. From this you conclude that eating ice cream increases your chance of getting sunburn. 

Is this conclusion correct? Well, the numbers may point to a relationship between eating ice cream and getting sunburn that doesn’t automatically translate to one thing leading to another. A correlation is when there is a relationship between two variables, in this case, ice cream and sunburn. But causation is defined as when one thing causes another; it requires substantial proof and in scientific studies is rarely found. 

Most authors of research papers will state their results using correlations, such as most likely to or is associated with to show that there is a relationship between two variables. The truth is that there could be many other variables contributing to the relationship that we just don’t see. Therefore, to make a judgment that A causes B would be incorrect without proof that no other variables could be in play. 

Going back to our earlier example. Can you think of something else that may contribute to both increased ice cream consumption and sunburn? 

How about sunny weather? Unless specifically checking for the variable sunny weather, scientists may not be aware of the effect sunny weather has on both sunburns and ice cream consumption. Therefore, a claim that eating ice cream causes sunburn would be incorrect and missing a crucial piece of information that would truly explain the relationship. 

Let’s look at an example where scientists have unequivocally said: A causes B. In this case, you may have even seen it yourself at the convenience store. A package of cigarettes may come with the message, smoking causes cancer, printed clearly on the box. 

In order to come to this conclusion, scientists in the 1950’s recruited hundreds of thousands of people and took note of their smoking habits and their cause of death over a span of many years. These studies were repeated continuously until over a million people were recruited. The results consistently showed that people who smoked were more likely to die of cancer. Finally, in 1964, the surgeon general wrote a report with the conclusion that smoking was a serious health issue that the United States had to address and counteract. The history of research into lung cancer and smoking is a good example of rigorous, repeated research leading to public health policies that can save lives.

A robust scientific article will be careful to note that their findings are correlations between variables and not necessarily a direct causal relationship. Even studies that test a treatment or an intervention will describe the results carefully until they have run many trials with successful results. You can look to the discussion section of the paper to find what researchers think may be the cause of certain relationships or potential factors that may have led to their results. They may also identify areas for future research, which may help find causal evidence. Any study that states that one variable causes another without an abundance of evidence should be looked at warily. 

NARIC’s Research In Focus series explores many studies that examine links between disability and full participation in the community. Here are a few examples:

For People with Traumatic Spinal Cord Injury, Relationships and Social Supports May be Linked to Life Satisfaction

A Novel Music-Based Exercise Intervention May Help Improve Physical Functioning in Adults Post-Stroke

For People with Serious Mental Illness, Getting Out and About May be Good for the Brain

Hafsa Abdirahman MPH is a public health scientist and freelance medical writer and editor. She believes that access to evidence-based, quality health information can save lives and she’s worked throughout her career to put this belief into practice.

This entry was posted in Research In Brief and tagged , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.