Skip to content

Correlation and Causation – easily explained!

Correlation refers to the relationship between two statistical variables. The two variables are then dependent on each other and change together.  A positive correlation of two variables, therefore, means that an increase in A also leads to an increase in B. The association is undirected. It is therefore also true in the reverse case and an increase in variable B also changes the slope of A to the same extent.

Causation, on the other hand, describes a cause-effect relationship between two variables. Causation between A and B, therefore, means that the increase in A is also the cause of the increase in B. The difference quickly becomes clear with a simple example:

A study could likely find a positive correlation between a person’s risk of skin cancer and the number of times they visit the outdoor pool. So if a person visits the outdoor pool frequently, their risk of developing skin cancer also increases. A clear positive association. But is there also a causal effect between outdoor swimming pool visits and skin cancer? Probably not, because that would mean that only outdoor swimming pool visits are the cause of the increased risk of skin cancer.

It is much more likely that people who spend more time in outdoor swimming pools are also exposed to significantly more sunlight. If they do not take sufficient precautions with sunscreen or similar, more sunburns can occur, which increases the risk of skin cancer. The correlation between outdoor swimming pool visits and skin cancer risk is not causal. 

Das Bild zeigt ein Kartoon mit einem Pool und einem Mädchen mit Hautkrebs.
Example of an Association between Outdoor Swimming Pool Visits and Skin Cancer | Source: Author

A variety of curious correlations that very likely do not show causation can be found at tylervigen.com.

Das Liniendiagramm zeigt zwei Linien, die eine Korrelation darstellt. Der Margarinekonsum und die Scheidungsrate nehmen in dem Zeitraum beide proportional ab.
Correlation between divorce rate and margarine consumption in Maine (USA) | Photo: tylervigen.com

For example, there is a very high association between the divorce rate in the American state of Maine and the per capita consumption of margarine. Whether this is also causation can be doubted.

What are the Types of Correlation?

In general, two types of contexts can be distinguished:

  1. Linear or Non-Linear: The dependencies are linear if the changes in variable A always trigger a change with a constant factor in variable B. If this is not the case, the dependency is said to be non-linear. For example, there is a linear correlation between height and body weight. With every new centimeter of height gained, one is very likely to also gain a fixed amount of body weight, as long as one’s stature does not change. A non-linear relationship exists, for example, between the development of sales and the development of a company’s share price. With a 30% increase in sales, the stock price may increase by 10%, but with the subsequent 30% increase in sales, the stock price may only increase by 5%.
  2. Positive or Negative: If the increase in variable A leads to an increase in variable B, then there is a positive correlation. If, on the other hand, the increase in A leads to a decrease in B, then the dependency is negative.
The picture shows the different kinds of Correlation.
Different Types of Correlation | Source: Author

How is the Pearson correlation calculated?

The Pearson correlation coefficient is most commonly used to measure the strength of the correlation between two variables. It can be easily calculated using the following values:

  1. Calculating the mean value for both variables
  2. Calculation of the standard deviations
  3. Deviation from the mean: The respective deviation from the mean must be calculated for each element of the two variables.
  4. Multiplying the deviations: Element by element, the deviations are then multiplied together and this is added up for all elements of the data sets.
  5. Divide by standard deviation: This calculation is then divided by the product of the two standard deviations and the number of data records, which is reduced by one.

\(\) \[r = \frac{ \sum_{i \in D}(x_{i} – \text{mean}(x)) \cdot (y_{i} – \text{mean}(y))}{(n-1) \cdot SD(x) \cdot SD(y)}\]

where:

  • Σ represents across all observations in the datasets.
  • xi and yi are the individual observations for variables x and y.
  • mean(x) and mean(y) are the mean values for variables x and y.
  • SD(x) and SD(y) are the individual standard deviations.
  • n is the number of observations and n-1 correspondingly the number reduced by one.

What is the (Pearson) Correlation Coefficient and how is it interpreted?

The Correlation Coefficient indicates how strong the association between the two variables is. In the example of tylervigen.com, this correlation is very strong at 99.26% and means that the two variables move almost 1 to 1, i.e. an increase in the consumption of Margarine by 10% also leads to an increase in the divorce rate by 10%. This is illustrated in the screenshot above, as margarine consumption and the divorce rate decrease almost in parallel. This shows that a decrease in margarine consumption also leads to a decrease in the divorce rate.

The coefficient can also assume negative values. A coefficient smaller than 0 describes the Anti-Correlation and states that the two variables behave in opposite ways. For example, a negative association exists between current age and remaining life expectancy. The older a person gets, the shorter his or her remaining life expectancy. 

What problems are there when investigating correlation and causality?

When researching the relationships between two variables, it is important to bear in mind the problems that frequently occur to avoid misinterpretations or false results.

A classic mistake here is to infer causality from a correlation. A correlation simply describes a relationship between two variables which means that a change in one variable leads to a change in the other variable. This may or may not mean causality. To prove perfect causality, additional evidence is required, which can be obtained through randomized experiments and is usually very time-consuming.

Another problem can be reverse causality, where the direction of causality is misinterpreted. In such a case, the assumed effect of the causality may be the cause of the causality. The example described of the supposed causality between ice cream consumption and skin cancer is a reverse causality, as the supposed cause, namely the consumption of ice cream, is also an effect.

When examining the correlation, confounding variables should be considered in order to obtain correct figures for the correlation. Confounding variables are third variables that have an influence on both the cause and the effect. If these are not taken into account, this can lead to distorted correlation coefficients. A multivariate analysis that takes more than two variables into account can provide a remedy.

Not quite as well known as the problems mentioned so far is the post-hoc fallacy, in which causality is wrongly assumed because there is a temporal sequence between events. Just because one event follows another does not necessarily mean that there is an effect. There may be other reasons that lead to this observed relationship.

Simpson’s paradox occurs when the data is divided into several groups and a correlation is observed between the groups. The paradox describes the fact that this correlation reverses or even disappears when these groups are combined. Therefore, the effects of group assignments should be taken into account in the analysis, as these can influence the correlation between variables.

The ecological fallacy is the error of concluding individuals based on correlations found at the group level. Such predictions should generally be cautious, as statistical conclusions about individuals often lead to false assumptions.

Another common pitfall is the omission of variables, also known as omitted variable bias. Incorrect calculations of the correlation coefficient can occur if important variables that show a relationship are omitted. Therefore, all factors that are measurable and related to the research should always be considered before the analysis. If these are omitted, the results may simply be incorrect.

These problems should be known before setting up a study or experiment in order to avoid these errors and obtain meaningful data.

How do you prove Causation?

To reliably prove causation, scientific experiments are conducted. In these experiments, people or test objects are divided into groups (you can read more about how this happens in our article about Sampling), so that in the best case all characteristics of the participants are similar or identical except for the characteristic that is assumed to be the cause.

For the “skin cancer outdoor swimming pool case”, this means that we try to form two groups in which both groups of participants have similar or preferably even the same characteristics, such as age, gender, physical health, and exposure to sunlight per week. Now it is examined whether the outdoor swimming pool visits of one group (note: the exposed sun exposure must remain constant), changes the skin cancer risk compared to the group that did not go to the outdoor swimming pool. If this change exceeds a certain level, one can speak of causation.

Why are experiments important to prove causation?

Real causality can only be found and proven with the help of so-called randomized controlled trials (RCTs for short). Here are some important reasons why these experiments are essential for proving causation:

  • Control: Only in an experiment possible confounding factors that influence the outcome variables can be controlled for. In a study, participants are randomly assigned to a so-called treatment and control group. Only the treatment group is then exposed to the influencing variable to determine causality. This ensures that the effects were only caused by the influencing variable and are not based on differences between the groups.
  • Replication: Due to the precise description of the experiment, RCTs can be easily replicated by other researchers. This makes it possible to investigate whether the same or similar results are obtained when the experiment is repeated, which in turn increases the robustness of the results and underlines their generalizability.
  • Accuracy: Only in experiments are all possible confounding variables measured, as far as this is possible, which maximizes the accuracy of the results.
  • Ethical considerations: To avoid unethical decisions and false conclusions, causal relationships should not be based solely on observational studies. This can lead to false prejudices.
  • Political implications: In many cases, policy decisions are based on causal relationships. To avoid serious legislative changes or bans being based solely on correlations, these should instead be confirmed in independent and meaningful experiments.
  • Scientific progress: RCTs can lead to new scientific findings that can be understood and interpreted by other researchers. Experiments are used to test hypotheses and make new suggestions that can change and improve our entire lives.

In conclusion, it can be summarized that experiments and especially RCTs are essential to prove causation and calculate its strength. Particularly in areas such as medicine, psychology, and economics, such studies lead to significantly better results and reliable figures. Observational studies, on the other hand, are important for making assumptions or formulating initial hypotheses, but their significance is significantly lower.

This is what you should take with you

  • Only in very few cases does a correlation also imply causation.
  • Correlation means that two variables always change together. Causation, on the other hand, means that the change in one variable is the cause of the change in the other.
  • The correlation coefficient indicates the strength of the association. It can be either positive or negative. If the coefficient is negative, it is called anticorrelation.
  • To prove a causal effect one needs complex experiments. 
Median

What is the Median?

Learn about the median and its significance in data analysis. Explore its computation, applications, and limitations.

Arima

What is the ARIMA Model?

Master time series forecasting with ARIMA models: Learn to analyze and predict trends in data. Step-by-step guide with Python examples.

Game Theory / Spieltheorie

What is Game Theory?

Discover the power of game theory and its real-world applications in policy making, negotiation, and decision-making. Learn more in this article.

Multivariate Analysis / Multivariate Analyse

What is Multivariate Analysis?

Unlock the power of multivariate analysis: Explore techniques to analyze and uncover relationships in your data in our comprehensive guide.

Bayesian Statistics / Bayessche Statistik

What are Bayesian Statistics?

Unlocking insights with Bayesian statistics: Optimize decision-making and quantify uncertainty for robust data analysis.

Confidence Intervals / Konfidenzintervalle

What are Confidence Intervals?

Quantify uncertainty and make informed decisions with Confidence Intervals: Measure the reliability of estimates and enhance statistical analysis.

Other Articles on the Topic of Correlation and Causation

  • Detailed definitions of the terms can be found here.
Das Logo zeigt einen weißen Hintergrund den Namen "Data Basecamp" mit blauer Schrift. Im rechten unteren Eck wird eine Bergsilhouette in Blau gezeigt.

Don't miss new articles!

We do not send spam! Read everything in our Privacy Policy.

Cookie Consent with Real Cookie Banner