Causality vs Correlation

This is a superbly fun topic!

Causality – the notion of cause and effect. The first thing causes the other and the other is caused by the first.

Correlation – Two items occurring together or seen together.

Correlation is easy.  You do a ‘rain-dance’ on Tuesday and it rains.  This is a correlation.  The world gets this wrong all the time.  Some people see causality.  The world easily offers you correlations between things.  It is much harder to find causality.

Comparing two cities you notice one has more crime.  You want to find out why.  You observe that the city that has more crime also has more police.  You decide that more police equals more crime.  You think you have found the cause.

It can get even messier.  Let’s say that data was collected over the next year in that city and it was shown that both crime fell and the amount of police fell.  That could lend credibility to those that think and know that police numbers cause crime numbers.  They would tote that statistic around all day.  Proof, I tell you, proof!  This fascinating position is aided by confirmation bias and cognitive dissonance. (see my post on all that here

This may be the false thinking that works in understanding some pain states, for example looking at an MRI image of a spine and noticing a disc bulge and thinking the link is causative or feeling better after a certain treatment; chiropractic for example. Chiropractic outcomes rely on correlation (and placebo), and have yet, in the past 150 years since it was made up, to be shown to be causative of outcomes.

If you were to walk into the clinic with rounded shoulders, a kyphotic thorax and a forward head posture and complain of neck pain I could easily state that your poor posture was causing all this pain and we would fix that.  Here its the police-posture and the crime-pain.

Perhaps in truth that presenting posture could be an effect from the neck pain, the cause of which had nothing to do with poor posture.  This truth, in our example: more crime-pain causing more police-posture.  Confounding our correlation vs causality inquiry; the patient could have even had poor posture before the problem.  How does this fit in to thinking about the patient and what to ‘fix?’  Both scenarios above are plausible and possible.

You need experience with this sort of presentation.  You need to also know the evidence, both about diagnosis and treatment.

This certainly makes the job of treating pain hard and complex.  Kyle Ridgeway PT,DPT has a great post at ptthinktank.com about how experience and evidence play into figuring all this out.  Let me quote him:

Evidence is more important than experience.
Evidence can not replace experience.
You can’t have evidence based practice without experience.
Experience is meaningless without evidence. 

Experience is bias.
Evidence is rigid.
Experience lacks rigor and control.
Evidence lacks experience applying to the individual.

Really interesting stuff.  Perception is all we have.  Informed perception is perhaps the balance that we seek.

**I am about to embark on another clinical internship, as part of the DPT program, to further gain this experience and apply what we have learned in academics.  I’m going into ‘sponge-mode’ and going to soak all I can.  I will attempt, however, to swish the experience around a bit and get a feel for the nuances as best I can.  It is a daunting task as a student, who neither has been exposed to all the pertinent evidence nor the hours of experience.  My goal will be as follows  “… it is experience bounced off the wall of thought and thrown into the sea of reflection that ultimately returns the reward of expertise.”  

– Matt D

 

Advertisements

4 thoughts on “Causality vs Correlation

  1. Pingback: Let me fix that for you! | ptbraintrust

  2. Great post, and you have identified some important issues, grey areas, and that ever present principle of uncertainty. I appreciate you linking to my recent post on clinical “experience.” Many times these conversations about complex, hard to define constructs such as expertise, experience, pain, etc are over simplified or made inaccurately complex via logical leaps.

    I have also written about the importance of thinking about your thinking and science based practice here: http://ptthinktank.com/2014/01/06/metacognition-critical-thinking-and-science-based-practice-dptstudent/

    as well as the importance of data quality and appropriate data analysis in clinic and research trials:
    http://ptthinktank.com/2014/05/15/data-quality-garbage-in-garbage-out/

    The importance of language, professionally (http://ptthinktank.com/2014/05/07/precision-in-language/) as well as when interacting with patients (http://mattlowpt.wordpress.com/2014/04/14/the-importance-of-language/#more-57) is of the utmost importance. Words matter.

    Keep analyzing and thinking. Great stuff here.

    Like

    • Thanks Kyle, I enjoy reading your stuff indeed.
      Being quite ‘green’ in the PT field I am trying to make connections between subjective and objective findings and am hypothesizing that there are causative factors there. Got to find the connection somewhere, things need to match up. Research can certainly help (such as in SIJD presentation of Fortin’s and Tuber’s area: Van Der Wurff, 2006) and sometimes you are on your own and must rely on clinical experience.
      Knowing that experience can mislead your thought process and that research can be flawed is very important to know. Although it can be a lot to try to digest when justifying decisions in real time.
      I believe the reflective process (as I think you do as well) is a valid way to sift through the information.
      Thanks for the comment!

      Like

  3. Pingback: A Year in Review | PTbraintrust

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