WTF does “correlation does not equal causation” mean anyway?
You may have heard scientists throw this around: “correlation does not equal causation.”
It’s a favourite line.
Other than trying to appear smarter than you (which is NOT always the case… I promise!), this saying does have a pretty solid meaning.
The purpose of it is basically to prevent from jumping to conclusions.
…And conclusions are jumped to All. The. Time. 🙁
In a nutshell, “correlation does not equal causation” means that just because we notice two things happening at the same time, even though logically they look related, it doesn’t necessarily mean that one caused the other.
So, let’s chat about WTF those terms mean, which studies show correlation and which show causation. Then I’ll give you a few fascinating examples.
WTF do “correlation” and “causation” mean anyway?
A mutual relationship or connection between two or more things. (ref)
Essentially, coincidence. When two things appear together or at the same time. They’re “associations” that may or may not have anything to do with each other. (Even if they seem to be related.)
Have you seen a bunch of those charts that show ridiculous graphed correlations? Things like how the number of pool drownings increases in the same years that Nicolas Cage is in more films? Here’s a link to see a bunch of these “spurious correlations.”
Yes, these may seem like ridiculous links (as the amt of mozarella cheese increases, so do the # of civil engineering doctorates);
These are everywhere in health science. Especially nutritional science.
Most of the reason for this is due to study design (HINT: Observational studies).
Last definition is “cause” which is,
A person or thing that gives rise to an action, phenomenon, or condition. (ref)
Unlike with correlation, causation is confidence that, without one thing, the other would not happen. A caused B, so without A, B would unlikely occur. It’s fairly strong proof that one thing is actually responsible for the other. That A had a direct or indirect role in bringing about the result, B.
In health science, this is determined with very strong data (HINT: Experimental studies).
Here’s a story, told by Sherry Nouraini, PhD that outlines the differences between “correlation” and “causation:”
— Sherry Nouraini (@snouraini) May 24, 2017
WTF does study design have to do with “correlation does not equal causation?”
Science can be super-creative!
Yes, there are some (lots!) of basic types of studies used to conduct health research. For a super-simple summary, check out my most favourite infographic on the internet. This infographic shows some of the most common types of studies.
Note the arrow on the left pointing down. You can see how the strength of the evidence increases as you go down this list.
This is why having at least two different randomized controlled trials (RCTs) are necessary before we can reasonably and confidently consider making health recommendations. Better yet, there would be many of these RCTs, even reviews of them, with supporting studies of different types looking at the same thing that come to a similar conclusion.
In other words, you need STRONG evidence to be sure of causation.You need STRONG evidence to be sure of causation. #causation #health #science Click To Tweet
In the most basic sense, there are two main types of research that can be done:
- “Observational” where you look at what’s going on (and do some math); and,
- “Experimental” where you make a change and see what happens (and do math).
A key thing to note on that infographic is that word in the small parentheses under the different study types.
Do you see it?
It says “observational” or “experimental.” These two major ways of doing research are key for knowing what study results are “correlation,” and which can show “causation.”Observational study = correlation; Experimental study = causation. #correlation #causation #studies Click To Tweet
Even within each study type, there really are a tonne of different ways to conduct research. And these are why many studies should be done before we can be sure of something. The higher quality, the better.
- Who/what to do the study on (e.g. People? Animals? Cells?);
- How to reduce bias (e.g. Randomization? Blinding?);
- What to measure (e.g. Diseases? Blood/urine tests? Symptoms?);
- How to measure them (e.g. How subjective/objective? How sensitive/specific should the tests be?);
- How to statistically interpret the results of a study; and,
- How to be realistic (read: cautious) when it comes to applying the results to real life.
Yup! Scientists can be creative beings.
Credible Health Research
Download this list of credible health resources and tips on where to research health topics.
How TF does “observational” = “correlation?”
An observational study observes.
There is no intervention. No experiment. No one who participates needs to change one iota of their lifestyle for the study. They just go ahead and live their lives, while answering a questionnaire, and/or providing blood/urine samples, and/or getting checkups and allowing their medical histories to be seen.
Observational studies ask questions and take measurements.
There are no “interventions.” Participants don’t have to take a pill, or stop drinking coffee/tea for 2 weeks to then promise to drink 5 cups/day for the following 8 weeks, etc.
Nothing like that.
Observational studies just let life go on, ask questions, and take measurements.
THIS is why observational studies = correlational results.Observational studies ask questions and take measurement = correlation. #healthscience #observational #correlation Click To Tweet
And this is the predominant type of study when it comes to nutritional science (which we’ll talk about below).
Here are a few ways to do an observational study:
- We look at people who eat a certain way (based on their answers to a questionnaire), and see what types of diseases/test results/symptoms they have. It’s like a “snapshot in time.” (These are cross-sectional observational studies.)
- Or we do it the other way around. We take a group of people with a certain disease/test results/symptoms, and then ask them how much of different foods/drinks they’ve consumed. So, we’re looking backwards (retrospectively) to see if we can figure out what they may have done that may be related in some correlational/associational way. (These are case-control observational studies.)
- Or, even better, we can take a group of people, ask them what they eat and have them give blood/urine samples, or share their medical records from time to time. We keep collecting this information regularly for many years (prospectively). Then, we see who gets which disease/test result/symptom over time. (These are cohort observational studies.)
You’ll notice that none of these have interventions. No one has to change what they’re doing. No one is asked to eat a certain way for a certain time. No one starts taking a pill (active ingredient or placebo) or doing a certain amount of a certain exercise.
NOTE: The “food frequency questionnaires” often used in these studies are not very accurate anyway.
The information we gather from these types of studies are observations.
Sometimes, observations make sense. For example, people who drink more alcohol may be more likely to get several cancers, including liver. It makes sense that, since alcohol is metabolized by the liver, that over-consumption over time can eventually lead to liver cancer. BUT, this is still a correlation.
Here’s why: It’s possible that people who drink more alcohol, may possibly also eat more junk food, or smoke. Maybe they don’t sleep as long, or always have the best mental health. Maybe they do! But these studies don’t account for these factors, and this can be a part of why increased alcohol intake is associated with these cancers.
These factors, that can affect the results, are called “confounding” factors. There can be dozens, or hundreds, of other factors that can affect the correlation.
Scientists can account for many of these confounding factors, but it’s impossible to account for all of them
Why do we do “observational” studies if the information isn’t going to show causation?
According to Darya Pino Rose, PhD:
“Why are processed foods so bad?”
The answer in some ways is just as simple: we don’t know. Unfortunately that isn’t very satisfying intellectually, and our attempts to create theories around the observational fact that obesity and most chronic diseases only appeared after the introduction of processed foods has opened the door to a world of speculation and confusion. (ref)
According to Julia Belluz at Vox:
This study design can be very valuable — it’s how scientists learned about the dangers of smoking and the benefits of exercise. But because these studies aren’t controlled like experiments, they’re a lot less precise and noisy.(ref)
According to Sarah Ballentyne, PhD:
The human body is fantastically complex, and the only way to understand it is to design studies to answer one very specific and detailed question at a time. The insight gleaned from tens of thousand of studies can then be combined to form a bigger picture, a complete understanding. Each little detail will eventually fit into place (or a flaw in a study will be identified).(ref)
How TF does “experimental” = “causation?”
Experimental (or interventional) studies do just that: An experiment. They intervene in someone’s life. They don’t just ask questions and take measurements – they require a change (preferably just one single change).
Whether that change is to take a pill (or placebo), or use a certain product, or even be admitted to a “metabolic ward” and given exact amounts of foods to eat (more on that below), it’s an intervention.
And when you intervene in someone’s life, to make that change, you can be more certain that it’s that change that caused the effect, and not just an association (or coincidence).Experimental/interventional studies require a change to try to find causation. #experiment #causation #healthscience Click To Tweet
Even better, have several experimental studies looking at the same thing in a different way, and having similar results to be clear on a cause-effect relationship.
This where review studies come in. There are several different types of reviews, including the highest quality systematic reviews. What they have in common is that they take several studies, critically evaluate them, and come up with an overview on the topic.
PRO TIP: If you search PubMed to research a topic, i.e. you’re not looking for one study in particular, start by filtering out the non-review studies. It is easier (and smarter, IMHO) to start your research with a smaller number of higher quality studies that are an overview of the topic you’re looking for.
Here’s what it looks like:
Why TF should we do different kinds of studies?
Scientific knowledge builds on itself. There is a foundation of health knowledge that is based on dozens, if not hundreds, of studies. In fact, there may be a bunch of different studies (of all different types) that point in the same direction. And that is pretty solid.
So, before we would ever realistically change a health recommendation, many high-quality studies need to be done. Then, look at the results of those studies to see if they point to the same conclusion. All of that evidence should be weighed to come to a solid conclusion.
And, as a practitioner, looking at the “balance of the evidence” gives context, and that can increase credibility.
NOTE: When there is ample evidence, even without proper clinical trials, some conclusions *should* be made. And this is especially true when doing an experimental study would be unethical, like forcing certain people to eat one way. So, let’s talk about nutrition studies.
Credible Health Research
Download this list of credible health resources and tips on where to research health topics.
Nutrition studies are usually observational
Because we simply can’t, and shouldn’t, force people to eat a certain way for a long time, this is why nutritional science is very “correlational.” I mean, how are you going to ensure that someone sticks to a very prescribed amount of certain foods day in and day out?
You might be able to do these on a short-term basis. And you can do this when it comes to supplements/placebos. But, when it comes to dietary research, it’s really, really hard.
It’s possible to do a “feeding” or “metabolic ward” diet where you actually provide specific amounts of food for a specific amount of time. But that’s very expensive, and hard to find lots of willing participants.
Here’s an example of one of those rare experimental nutritional studies where people were fed a specific diet for 8 weeks. Researchers compared a low fat diet for 4 weeks to a low carb diet for the other 4 weeks to see how they affected weight and body composition.
All subjects were confined to the metabolic ward throughout the study with no access to outside food. Subjects knew that it was imperative that they eat all of the food provided.
They also had to spend 2 consecutive days each week in a “metabolic chamber” to measure how much energy they were burning. They had to do exactly 90 minutes of stationary cycling every day.
They recruited 17 overweight and obese men.
This kind of experimental nutritional study is rare because it’s very hard to get people to agree to having everything measured. It’s also very expensive.
Here’s a quick interview with the author.
If you don’t require a change (or intervention); how can you prove something actually caused something else (i.e. “causation”)?
Answer: You pretty much can’t.
Hence, “correlation does not equal causation.”
Here’s one example:
— Leesa Klich Health (@LeesaKlich) May 18, 2017
Fascinating example about protein and your bones
NOTE: This is “CHAPTER 4: High Protein and Urinary Calcium” from my ebook “The Real Deal About Calcium and Your Bones.”
You can download a free copy of the whole ebook (it even has an ‘amount of calcium in different foods’ chart!):
The real deal about calcium and your bones
A science-based, holistic approach to optimal bone nutrition.
Click here to download the free ebook, complete with 40 scientific references.
Yes, it’s true that eating higher protein meals causes our body to excrete more calcium in the urine (hypercalciuria). It’s been shown in many studies for the past hundred years.
After this observation, the question then was, “Where the heck did this higher-than-normal urinary calcium come from?”. And that was a great question!
Of course, one of the main hypotheses at the time, was that the calcium came from the body’s main storage of calcium, namely bone. Since our bones store 99% of the body’s calcium on a protein matrix, that’s a logical place where this calciuria may very well be coming from.
This was the hypothesis, and it went on to consider whether there might be a link with the acidic ash that results from the metabolism of proteins versus other foods.
This hypothesis got many, many people on board with the idea that excess meat consumption contributes to loss of bone density. (By the way, there are still a lot of people why try to sell an “alkaline” diet to prevent or reverse osteoporosis…I don’t think they’ve looked at much research done in the past 10 years).
How can we test if this hypothesis is true? How can we figure out what the actual cause of the hypercalciuria was? Where exactly did that urinary calcium come from?
We can test this by using radio-labeled calcium!
So, back in 2004 a study was published that solved this mystery using safe levels of non-harmful radiation.
Scientists took a group of women and for 10 days they were on either a moderate or a high protein diet, ingesting 600 mg of calcium per day. Then they had them eat a meal where the calcium was radioactive so they could measure (via the radiation) where that calcium from the meal actually went.
The urine was radioactive!
In fact, most of the increased calcium in the urine was radioactive. Yup, it came from the meal, NOT the bones!
What does radioactive urinary calcium mean?
Well, this is when we learned that eating calcium with protein actually increases the body’s absorption of calcium, from the normal level of about 19%, up to 26%.
Here’s exactly what they say:
“The high-protein diet caused a significant reduction in the fraction of urinary calcium of bone origin…”
You heard that right, reduction in the amount of urinary calcium from the bones!
So, by eating higher amounts of protein with a calcium-containing meal, you can help your intestines absorb more of that calcium from the meal!
How does the calcium get into the urine?
Of course, your kidneys are always doing their job to regulate your blood so it’s always just about perfect – its pH, electrolytes, mineral concentrations, etc. So, when you have enough calcium in your blood at the moment (because you just absorbed a whole whack of it from your high protein meal), your kidneys filter some of that excess calcium out into the urine so as to maintain perfect blood levels of calcium.
So, what we actually learned was that protein helps us to absorb more calcium from our meals!
Yes! Similarly to vitamin D (which we all know is critical for increasing absorption of calcium), protein also increases our absorption of calcium.
In case you’re wondering, other studies have also shown that high protein diets do not negatively affect bone health.
Not to mention that study after study shows that high protein diets are associated with higher bone mineral density and lower fracture risk! Especially when those people are eating adequate amounts of calcium.
In fact, in one large study, the people who ingested the most protein and at least 800 mg/day of calcium had an 85% reduced risk of fracture compared with those who ate the lowest amount of protein!
There are several thoughts as to why higher protein diets are actually better for your bones, such as whether it’s only due to increased calcium absorption, or perhaps also due to increased muscle mass, or suppressing parathyroid hormone, or other reasons. It’s not quite clear exactly all the reasons why.
Some studies have suggested that higher protein diets are good for bone health, but mostly when adequate calcium is also ingested.
Once again, calcium is critical for your bones.
The bottom line is that your bones need the calcium, and the protein (and a bunch of other vitamins and minerals too). Don’t skimp out on any of these the vital nutrients!
- High protein intake helps your body to absorb more calcium from your foods.
- Research within the last 10 years shows that higher protein diets seem to be beneficial to bone health, particularly when enough calcium is consumed.
Credible Health Research
Download this list of credible health resources and tips on where to research health topics.
Fascinating example from a study on coffee and hormones
This study by Kotsopoulos, et al. (2009) looked at caffeine/coffee/decaf/tea intake and blood levels of certain sex hormones.
They tested blood samples provided during the Nurses’ Health Study (NHS) for levels of estradiol, estrone, testosterone, etc. This included 32,826 nurses.
Kotsopoulos et al. also looked at the amt of caffeine/coffee/decaf/tea the nurses reported having when they completed questionnaires during the study.
The researchers used information collected in a snapshot of time. One blood measurement, and one reported caffeine/coffee/decaf/tea intake per nurse.
Then, once they had the numbers, they looked for links.
Any links found would be “correlations,” which means they coincide with each other. They seem to go together. We can’t be sure how they’re linked.
There was no experiment conducted here. It was purely an observation. They were just looking at these factors, doing lab tests & statistics to see if there were any links.
The nurses were free to live their lives without any intervention.
***This. Is. Key.***
The study was observational, not experimental/interventional.
This is why the results from observational studies like these are “correlations” (Not causation.)
On the other hand, an experiment (clinical trial if in people) would require an intervention. The nurses would have to agree to change one particular thing and this would “intervene” in their regular life.
For example, this could look like: having them agree to drink a 3 cups of coffee every day for 3 months, and measuring blood levels of hormones before the study, then every month for those 3 months.
If an actual clinical trial is done, where you are deliberately changing or controlling for something (like coffee intake), and you can measure before and after (and dose-dependancy, etc.) AND you find a link, you can be more certain of the cause.
It goes beyond observing what’s happening (correlation), to actually effecting a change to see what will happen (causation).
The study found:
– No links between caffeine/coffee/decaf/tea intake and blood levels of certain sex hormones in postmenopausal women;
– Premenopausal women with highest caffeine/coffee = lower luteal estrogen, but no assoc’n with follicular estrogen;
– Premenopausal women with highest tea intake = higher follicular estrogen.
Correlation does not equal causation.
(just kidding). 🙂
“Correlations” are associations that happen when two things occur at or near the same time. These results come from observational studies. These are made when we observe and measure things without purposefully effecting a change to see what happens. We’re watching, learning, and doing some math to see which things are associated with each other.
“Causations” are when the research is strong enough to actually show that one thing made something else happen. These results are much stronger to show that something actually caused something else. These results come from experimental or interventional studies. When we make a change to see what happens.
Signing off and toasting: To rocking the “correlation does not equal causation” card.
Over to you
What do you think? Did I explain this uber-fascinating concept (or can I improve it)?
Are the differences between observational (that yield “correlational” results) and experimental/interventional studies (that yield “causational” results) clear?
Do you feel confident that you can rock the “correlation does not equal causation” card?
I’d love to know (in the comments below)!
- Association is not the same as causation
- I asked 8 researchers why the science of nutrition is so messy. Here’s what they said.
- Why nutrition science is so confusing. [Infographic] 9 reasons eating well isn’t as straightforward as we’d like it to be.
- Everything Causes Cancer: How much can science really tell us?
- Processed Food vs Real Food: Why Nutrition Science is So Confusing (and what to do about it)
- The one chart you need to understand any health study
- The Evidence-Based Medicine Pyramid!
I'm Leesa Klich, MSc., R.H.N.
Health writer – Blogging expert – Research nerd.
I help health and wellness professionals build their authority with scientific health content. They want to stand out in the crowded, often unqualified, market of entrepreneurs. I help them establish trust with their audiences, add credibility to their services, and save them a ton of time so they don’t have to do the research or writing themselves. To work with me, click here.