It is quite easy, and quite common, to get different words and terminologies mixed up. Though the English language does indeed contain many words that can be used interchangeably, people often erroneously think of two different concepts as the same. One of these common misconceptions includes mixing up correlation with causation.
Even if someone does know that the two are different, they might have trouble explaining how they’re different. You might look at say, a graph of some data, and then you notice that one value always changes when another one changes. Now you might get confused as to whether this is because the two values correlate with each other, or once causes the other.
Correlation
Correlation is when two events can be logically connected to each other without actually directly influencing one another. So if statistical data shows that whenever event A happens, event B does too, we can confidently state that both events correlate with each other.
This of course, does not mean that event A cause event B or vice versa. We cannot claim that event B took place because event A took place. To claim that we would need to do more research. And that is where causation comes into play.
Causation
Causation is basically what people mistake correlation for. Where two events taking place at the same time could be said to be correlated, causation actually focuses on events that cause other events. If adequate proof is provided, we claim that event A causes event B to take place. This is causation.
So in this case, the two events did not take place at the same time because of some unseen factor. Instead, an event actually physically caused another to occur. Causation produces reactions, while correlation seems to produce reactions but can be logically separated.
The Summertime Example
Let’s take a common example used to explain the differences between the two. Let’s say that you have some statistics in front of you, and the stats show that the number of people drowning goes up when the number of ice creams being sold rises. Can we claim that the sale of ice creams causes people to drown? That would be claiming causation between both events on no solid grounds.
For now, the data only correlates. We know that when ice cream sales go up, that particular period of time also sees an increase in people drowning. Now let’s consider a factor that we have ignored up till now; both events seem to occur when it is a hot sunny day outside. With this we find out two other factors in play; on hot sunny days more people go out for an ice cream and more people go to the beach to cool off. Again, we do not have solid proof that one causes the other.
So what we do now is, on the next sunny day we close off the beach. When we do this we notice no drownings take place but ice creams still get sold. With this bit of research in our hand we can safely claim that there is no causation between ice cream sales and drownings, but merely correlation.
Bald Men And Long Marriages
Let’s take a fun little example to discuss both these concepts further. You have some data on marriages at hand, and you notice something peculiar. You see that men who are bald seem to have been married longer than men who still have hair. This shows the difference between ‘data’ and ‘information’. This is just raw data that you have with you. If you were to take this data at face value, you would come up with the weird conclusion that bald men seem to enjoy successful marriages for some reason.
However, if you were to do just a tiny bit of rudimentary digging for some more data, you would find another factor that plays into the marriages: age. Men who are bald are much older than men who still retain their hair. Because bald men have been alive longer, they have had more time to remain married. Again, correlation does not equal causation. Result: Being bald does not improve your chances of being married longer. Sorry boys.
Chicago And Houston Crime Rates
Let’s take another example, just to clarify how wrong it is to interchange correlation and causation. The city of Chicago has a higher crime rate than Houston. Chicago is also much colder than Houston, with Houston being in Texas whereas Chicago is in Illinois. Can we then safely assume that colder temperature leads to a higher crime rate?
Of course not, that would be really stupid. And that perfectly encapsulates just how different correlation and causation are. Causation always has a solid link between one event causing another; like rain causing the ground to get wet or sunny days raising the temperature of an area. Correlation can literally be plotted between anything; for example, the graph of budget spending on NASA and the graph of suicides by strangulation is almost exactly the same. But NASA being given money isn’t driving people to suicide is it? That’s basically correlation, just two events happening at the same time with the same intensity due to different factors.
Conclusion
Those examples should have hopefully cleared up any remaining confusion about the difference between correlation and causation. If you want to try experimenting yourself, go right ahead. You’d be astounded how many coincidences take place each day that people might mistake as causation rather than just correlation.