Correlation

At HumanaNatura, our goal is to both inspire and inform, so that each of us makes progressively healthier and higher quality choices over time. This strategy is in keeping with the scientific foundations of our natural health system, and with research suggesting that a mix of good facts and feelings best steers us toward improved life and health.

A new study by Harvard University affords a nice informational or teachable moment, one that can help us better understand and make use of the health research we encounter. In the new study, public health researchers found that reported soft drink consumption in teens was closely associated with reported violent behavior. For our discussion, the key words from the study are “associated” and “reported.”

When researchers state an association or correlation like this, they are indicating that two or more things have been observed moving together in a pattern. Correlations can be positive (with things moving in the same direction) or negative (moving in opposite directions) but they cannot be neutral (since no movement means no correlated or concurrent change). Importantly, association or correlation never means that causation or cause and effect has been established (that A causes B, or the reverse). In reporting on the new study, the researchers took pains to highlight that they have not shown causation between soft drinks and teen violence, in either a forward or backward direction.

When studies like this talk about a reported behavior or condition, they mean just that. Participants were asked one or more questions and gave a reply or report. As you might suspect, what we say we do and what we actually do can be substantially at odds with one another, either because we are intentionally withholding or exaggerating information, or because we have a distorted recollection or sense of the information. A much more reliable source of information is observed or measured behavior or data, and even better than this are observations and measurements that are double-blind (where neither the observed person nor observer is privy to key details of the measurement process).

If correlational and reported behavior studies are each less valuable than available alternatives, why have them at all? First, because they are often easier and much less expensive to perform. Second, they can suggest areas for more intensive follow-up research. And in the case of correlational research in particular, while it does not provide causal information, it can lead to insights that are quite useful. In this case, researchers have discovered that quite innocuous information about soda consumption may be a signal for teenagers that are at risk of acting violently, potentially leading to better directed social service interventions.

If you would like to learn more about research techniques to investigate correlation and causation, check out Correlation Does Not Imply Causation.

Please share this post…and the science of creating healthier modern life!

  1. Types & typos « NaturaLife
  2. Sugar science « NaturaLife
  3. Fit Brains Go Naturally With Fit Bodies « NaturaLife
  4. Natural Limits of Red Meat « NaturaLife
  5. Re-Focusing Your Creativity! « NaturaLife

Leave a Reply

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

WordPress.com Logo

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

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s

  • Healthy Natural Eating Made Easy Via OurPlate!

  • HN-100: Your Own Waiting Personal Health Revolution

  • Natural Truth – Our Health Idea Rating Campaign

  • Help Your Career Via HumanaNatura Certification!

  • Health Leadership Opportunities In Your Community!

  • A Comprehensive Approach For Health Centers

  • HumanaNatura’s Mark Lundegren

  • Shop & Support HumanaNatura @ The HN Marketplace

%d bloggers like this: