Association is not Causation

The death of Disco music. The rise of Rap. The release of glyphosate/RoundUp herbicide. Drops in numbers of butterflies and pheasants in Pennsylvania. These all occurred on the same timeline. Are they related?

Statistically, yes. They are related but the only thing they have in common is parallel time courses. At the same time there was also a drop in eating fondue and the popularity of bell bottom pants

Where is this going? The point is that just because two things happen at the same time it doesn’t mean they are even connected, let alone causal. There may be a statistically significant link between pretzel eating and alcoholism just because bars have free pretzels but it doesn’t mean cause and effect. All statistical significance tells you is that the connection is probably not due to chance alone. It doesn’t tell you what the connection is.

Pretzel eating causing alcoholism, or vice versa for that matter, seems ridiculous but there are many claims being made about horses that spring from the same faulty reasoning and they often are based on the most dangerously misleading or misunderstood information out there – actual scientific studies. Lay magazines reporting  on scientific studies can be worse yet.

Good science follows the laws of logic and is very careful about how statements are worded but even so it is taken the wrong way. In a scientific study, “associated with” means there is a relationship between two things but it does not tell you what the relationship is. For example, statistically, peanut butter is associated with jelly and both are associated with bread.  None of these factors causes the other. They are simply found together. Association is not causation.

Not understanding that can result in a lot of trouble. An example was the observation that in  some parts of the world when horses develop laminitis the fructan levels may be high in fall pastures. The press was all over this. Even though subsequent studies showed fructan related laminitis is a gut overload phenomenon and about 90% of laminitis cases are not, the fructan “cause” of laminitis refuses to die.

Another example is that lactate causes fatigue and muscle pain. The truth is that high levels of lactate from anaerobic metabolism occur at the same time as peak exertion but the lactate doesn’t cause the fatigue any more than ashes cause fire. It is a marker of the problem, not the cause.

More recently, a discussion arose over a lay article linking metabolic syndrome to environmental exposure. This was reporting on an actual study that found dioxin level in horses’ blood was linked to triglyceride (fat) levels.  However, there was no connection with glucose or insulin and dioxin happens to be a fat soluble toxin that is present  literally everywhere (including at the north and south poles). The actual article stated there “may” be an influence; not that there is one.

The message here is to learn to think critically. Don’t believe the slant every lay article puts on things. Go to the original published study. While you are there, read each sentence carefully and be alert for words like “may”, “might”, “could”, “tends” or “trends” because these qualifiers mean any connections fall short of actually proving cause and effect.

Association is not Causation.

Eleanor Kellon, VMD

 

About Dr. Kellon

Graduate of University of Pennsylvania Veterinary School. Owner of Equine Nutritional Solutions, www.drkellon.com, industry and private nutritional consultations, online nutritional courses. Staff Veterinary Expert at Uckele Health and Nutrition.
This entry was posted in Equine Nutrition. Bookmark the permalink.

2 Responses to Association is not Causation

  1. Lucinda says:

    Thank you for illuminating this issue!

  2. Carla Caudill-Waechter says:

    Bravo!! Thank you for your brilliant metaphors!

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 )

Google photo

You are commenting using your Google 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 )

Connecting to %s

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