Cause and effect

Here’s a city. On the West side there are a lot of police, and there’s also a lot of crime there. On the East side, there are much fewer police officers. And there is also a lot less crime being committed. Hmmmm.

It seems as if… Can it really..? The more police, the more crime! It’s the police that causes the crime? Or… no. That seems implausible.

Maybe the other way around? The more crime there is in an area, the more police are sent there? That sounds more reasonable. But we can’t be really sure, can we? Just maybe the police in this area are corrupt, and it really is the police that cause the high crime rate?

It’s not entirely easy to determine, how a causal relationship works… Is is A that causes B, or the other way around? Here’s another example of a possible causal relationship: Most children who are born are well and healthy, but not all. Women who have a cappuccino maker at home in their kitchen, are more likely to give birth to a healthy child, compared to women who don’t. Hmmm. This sounds too weird to be true.

Could a cappuccino maker cause healthy babies? No, not likely. The other way around then? Can healthy newborn babies cause cappuccino makers? No, not that either.

What if there is a third factor, that lies behind both things? If you want to pause the film, and make a guess, do it now. It probably has something to do with how much money you’ve got. Families with a good income are more likely to own an expensive coffee maker. And richer families are also on average healthier, and give birth to healthier babies.

So, a more likely causal relationship looks like this: Economic prosperity causes both more gadgets at home, and better health. The connection between coffee makers and children’s health is thus probably caused neither by coffee makers nor by babies, but by a third factor. Let’s look at one more relationship, to make this crystal clear. People who get good grades in school, get - on average - higher salary when they grow up and start work. Grades are related to income.

And we can be certain that it’s not the high salary that causes the high grades, because the grades come before the salary. And while it certainly sounds plausible that high grades cause you to get a better paid job, we still can’t be sure, can we? Could there perhaps be another, unknown third factor? Pause the film and think about that! How about this one: Determination… ...ambition, persistence.

People who are prepared to work hard to obtain a goal, might be more likely to get both higher grades and a higher paying job. All these examples show that it can be hard to know whether a causal relationship really exists, when it comes to issues in social sciences. In the natural sciences, you can often design experiments, and test your hypotheses under controlled circumstances. Imagine designing experiments to test if these are real causal relationships. That would be both expensive and impractical.

But, it is possible to test if a causal relationship is operating, in the social sciences too. If you are to test the hypothesis that: ‘A causes B’, and you want to be really sure, it takes some advanced statistics. But you can come a long way just by testing if the relationship meets these five criteria: One: That which causes the change - A - happens before the effect - B. Two: If the amount of A increases, then B also increases. Three: If the amount of A is reduced or disappears, then B is also reduced or disappears.

Four: There is a reasonable explanation why A would cause B. Five: There are no other, just as reasonable ‘third-factors’, which could have caused both A and B. Causal relationships. A bit harder than they seem. Especially in social sciences.