Causality

Standard

Most people don’t grasp causality except in the simplest possible terms. If I throw a rock at a window and the window breaks, most will agree that the second event was directly caused by the first. But what if sometimes the window doesn’t break? Then we would say that throwing rocks at windows will probably cause the window to break. Now step back a bit. Suppose we come across a bunch of windows — some are broken and some are not but in every case we find a rock nearby. What can we say about that?

We can still say that probably those rocks caused the windows to break. But what if there is only a rock present 80% of the time? Can we now attribute broken windows to thrown rocks? Yes, we can — some of the time. Now we might say the rock was a possible cause – among others — for the broken windows. We would then postulate other causes —- such as baseball bats or booted feet or flying birds and would look for evidence to back up those hypotheses. We shouldn’t say, however, the windows broke spontaneously or by chance or for no reason at all. We would expect there to be a cause.

A leads to B is the general rule we apply in the physical world. Unless you are working in quantum physics where all the rules of causality seem to have been, well, thrown out the window.

A quite different thing happens when you talk about causality in the social world. There is seldom a direct line you can draw from Event A to event B and, like in quantum physics you are always dealing with probabilities rather than linear events. So this cause may be 50% responsible for that event and the rest was caused by something as yet unknown. Or,  A causes B but C is a countervailing force so we don’t see as much B as we would expect from the quantity of A.

Because social causation — and even biological causation — is often so muddy we find it easy to reject causal arguments where we don’t happen to like the conclusion. That doesn’t mean the causality isn’t present it just means we can dismiss the evidence even more readily.

This was at play for years with respect to cigarette smoking. You couldn’t show that everyone who smoked would get cancer only that it increased the probability. So, some people denied the causal link (mostly cigarette companies). No-one thinks that now. Too much evidence from too many studies.

But when I suggested that access to guns increased deaths by gunshots — the first thing someone said was correlation is not causality. True. Correlation may exist when two events occur for which no logical causal chain can be constructed. For example — ‘when the leaves fall off the trees, unemployment rates go up’ is a correlation, two unrelated events that occur at the same time (in Canada at least). One doesn’t cause the other. However if we postulate that increasing the number of guns available increases the chances that people will use guns in spontaneous acts of violence, we can see the causal link. It may not be the only cause but it is a cause.

That’s basic social science.

And that’s ten minutes.

Advertisements

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s