1

There is a saying that "Correlation does not imply causation." I am trying to find the best way, preferrably in a word or short phrase, to explain when one thing really does have a causal effect on another, but the effect may not be strong.

This might be more clear with examples. People tend to get better with experience, so years of experience really might cause someone to be more skilled. But there are so many other factors (intensity of that experience, amount of study and practice, etc.) that the effect of experience might easily get swamped out and someone with fewer years of experience might be more skilled.

Similarly, exercise can help cause weight loss. But diet and genetics often play a bigger role than amount of exercise in determining how much someone will weigh.

One phrase that I think comes close to my meaning is "loosely coupled", but I don't think that quite captures it and for people with a software development background it will carry a very specific meaning. So what is the best way to express this?

4
  • tangentially, minimally, moderately, greatly affects.
    – SrJoven
    Aug 13, 2014 at 19:12
  • It is of course not true that "correlation does not imply causation". What's true is that correlation does not equal causation. That observed correlations lead us to test for causation tells us that indeed the one implies the other, and we test for the truth of that implication.
    – Jim Mack
    Aug 13, 2014 at 19:54
  • I think you may be interpreting the word "imply" differently from the way I mean it. In formal logic and in the scientific community saying that "X implies Y" means that if X is true than Y must absolutely be true. Wikipedia has a good discussion of the meaning and origins of that phrase here: en.wikipedia.org/wiki/Correlation_does_not_imply_causation Aug 13, 2014 at 20:45
  • ... I use that definition when I'm teaching maths. But it is improper to use it on a linguistics website without warning. It's not the principal sense in everyday English. Aug 13, 2014 at 20:50

4 Answers 4

3

I think you're looking for "influence". As in:

Certainly diet and exercise influence weight regulation, but genetics and ingrained eating habits predispose many of us to gain weight.

Alternatives for predispose include bias, sway, and skew, all of which mean to incline (ie make more liable) something towards a specific condition, without totally determining the outcome.

3
  • Inform is used to mean 'have a strong influence (on)': 3. To be a pervasive presence in [AHD] //3. to pervade or permeate with manifest effect: A love of nature informed his writing. [RHK Webster's] Aug 13, 2014 at 20:56
  • Oh, good one. Do you think inform can be used for physical, real-world processes like weight gain or skill acquisition?
    – Dan Bron
    Aug 13, 2014 at 20:57
  • No, I think that the DO has to be a sentient's opinions / decisions / manifested skills /etc. Hence not an 'answer'. Aug 13, 2014 at 21:13
1

Related (“Standing in relation or connection”), linked (“connected, either with links, or as if with links”), and connected all are possibilities. (Note that related, like correlated, does not imply causation; and to some extent the same criticism applies to linked and connected. For specificity, perhaps say causally related, linked, or connected.)

1

My go-to way to categorize such impacts is by stating that the additional source as having a secondary or tertiary impact.

-1

Since you are trying to discount something from being the sole cause in a causality, you could use the term "not entirely." As in "The reason for [effect] is not entirely down to [cause].

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.