{"id":312221,"date":"2021-07-12T13:05:10","date_gmt":"2021-07-12T11:05:10","guid":{"rendered":"https:\/\/www.scribbr.nl\/?p=312221"},"modified":"2022-10-10T12:38:54","modified_gmt":"2022-10-10T10:38:54","slug":"correlation-vs-causation","status":"publish","type":"post","link":"https:\/\/www.scribbr.com\/methodology\/correlation-vs-causation\/","title":{"rendered":"Correlation vs. Causation | Difference, Designs & Examples"},"content":{"rendered":"

Correlation<\/strong> means there is a statistical association between variables. Causation<\/strong> means that a change in one variable causes a change in another variable.<\/p>\n

In research, you might have come across the phrase \u201ccorrelation doesn\u2019t imply causation.\u201d Correlation and causation are two related ideas, but understanding their differences will help you critically evaluate and interpret scientific research.<\/p>\n

<\/p>\n

What\u2019s the difference?<\/h2>\n

Correlation<\/strong> describes an association between variables<\/a>: when one variable changes, so does the other. A correlation is a statistical indicator<\/a> of the relationship between variables. These variables change together: they covary. But this covariation isn\u2019t necessarily due to a direct or indirect causal link.<\/p>\n

Causation<\/strong> means that changes in one variable brings about changes in the other; there is a cause-and-effect relationship between variables. The two variables are correlated with each other and there is also a causal link between them.<\/p>\n

A correlation doesn\u2019t imply causation, but causation always implies correlation.<\/p>\n

Why doesn\u2019t correlation mean causation?<\/h2>\n

There are two main reasons why correlation isn\u2019t causation. These problems are important to identify for drawing sound scientific conclusions from research.<\/p>\n

The third variable problem<\/strong> means that a confounding variable<\/a> affects both variables to make them seem causally related when they are not. For example, ice cream sales and violent crime rates are closely correlated, but they are not causally linked with each other. Instead, hot temperatures, a third variable, affects both variables separately.<\/p>\n

The directionality problem<\/strong> is when two variables correlate and might actually have a causal relationship, but it\u2019s impossible to conclude which variable causes changes in the other. For example, vitamin D levels are correlated with depression, but it\u2019s not clear whether low vitamin D causes depression, or whether depression causes reduced vitamin D intake.<\/p>\n

You\u2019ll need to use an appropriate research design<\/a> to distinguish between correlational and causal relationships.<\/p>\n

Correlational research designs<\/a> can only demonstrate correlational links between variables, while experimental designs<\/a> can test causation.<\/p>\n

Correlational research<\/h2>\n

In a correlational research design, you collect data on your variables without manipulating them.<\/p>\n

Example: Correlational research<\/figcaption>You collect survey data<\/a> to investigate whether there is a relationship between physical activity levels and self esteem. You ask participants about their current levels of exercise and measure their self-esteem using an inventory.<\/p>\n

You find that physical activity level is positively correlated with self esteem: lower levels of physical activity are associated with lower self esteem, while higher levels of physical activity are associated with higher self esteem.<\/figure>\n

Correlational research is usually high in external validity<\/a>, so you can generalize<\/a> your findings to real life settings. But these studies are low in internal validity<\/a>, which makes it difficult to causally connect changes in one variable to changes in the other.<\/p>\n

These research designs are commonly used when it\u2019s unethical, too costly, or too difficult to perform controlled experiments. They are also used to study relationships that aren\u2019t expected to be causal.<\/p>\n

Example: Correlational research<\/figcaption>To study whether consuming violent media is related to aggression, you collect data on children\u2019s video game use and their behavioral tendencies. You ask parents to report the number of weekly hours their child spent playing violent video games, and you survey parents and teachers on the children’s behaviors.<\/p>\n

You find a positive correlation between the variables: children who spend more time playing violent video games have higher rates of aggressive behavior.<\/figure>\n

Third variable problem<\/h2>\n

Without controlled experiments, it’s hard to say whether it was the variable you\u2019re interested in that caused changes in another variable. Extraneous variables<\/a> are any third variable<\/strong> other than your variables of interest that could affect your results.<\/p>\n

Limited control<\/a> in correlational research means that extraneous or confounding variables serve as alternative explanations for the results. Confounding variables can make it seem as though a correlational relationship is causal when it isn\u2019t.<\/p>\n

Example: Extraneous and confounding variables<\/figcaption>In your study on violent video games and aggression, parental attention is a confounding variable that could influence how much children use violent video games and their behavioral tendencies. Low quality parental attention can increase both violent video game use and aggressive behaviors in children.<\/p>\n

But it\u2019s not something you control for, so you can only draw a conclusion of correlation between your main variables.<\/figure>\n

When two variables are correlated, all you can say is that changes in one variable occur alongside changes in the other.<\/p>\n

Spurious correlations<\/h2>\n

A spurious correlation<\/strong> is when two variables appear to be related through hidden third variables or simply by coincidence.<\/p>\n

Example: Spurious correlation<\/figcaption>In Germany and Denmark, statistical evidence shows a clear positive correlation between the population of storks and the birth rate spanning decades. As the stork population fluctuates, so does the number of newborns. How do you account for this pattern?<\/p>\n

The Theory of the Stork<\/a> draws a simple causal link between the variables to argue that storks physically deliver babies. This satirical study shows why you can\u2019t conclude causation from correlational research alone.<\/p>\n

In reality, the correlation may be explained by third variables (such as weather patterns, environmental developments, etc.) that caused an increase in both the stork and human populations, or the link may be purely coincidental.<\/figure>\n

When you analyze correlations in a large dataset with many variables, the chances of finding at least one statistically significant<\/a> result are high. In this case, you\u2019re more likely to make a type I error<\/a>. This means erroneously concluding there is a true correlation between variables in the population<\/a> based on skewed sample data.<\/p>\n

Directionality problem<\/h2>\n

To demonstrate causation, you need to show a directional relationship<\/strong> with no alternative explanations. This relationship can be unidirectional, with one variable impacting the other, or bidirectional, where both variables impact each other.<\/p>\n

A correlational design won\u2019t be able to distinguish between any of these possibilities, but an experimental design can test each possible direction, one at a time.<\/p>\n

Example: Directionality problem<\/figcaption>The variables of physical activity and self esteem can be causally related in three ways:<\/p>\n