The "Post hoc" problem
"Post Hoc" is short for "Post hoc ergo propter hoc" which is a classic logical fallacy - it is latin for "after this, therefore because of this."
The problem is that the "Post hoc" principle often works in practice, and I believe that it works often enough that in the process of human development we have become genetically programmed to think in those terms. Unfortunately although
"After this, therefore it MAY have been caused by this"
is a perfectly reasonable inference, we often act as though
"After this, therefore it MUST have been caused by this"
were also valid. And it isn't. Sometimes a third factor caused both events and sometimes the appearance of correlation is a pure coincidence.
A classic recent example has been caused by the fact that the age at which the NHS has chosen to give children the MMR vaccine happens to be close to the age at which children who are going to develop autism often begin to show symptoms.
Let me declare my own prejudice on the subject of MMR: my late mother had been, presumably wrongly, diagnosed with rubella when she was a child, and therefore she was not inoculated against it. Unfortunately she then developed rubella when pregnant with my sister, who as a direct consequence was born severely handicapped and died at the age of one week. If MMR had been available and in use when my mother was an infant, the overwhelming probability is that my sister would be alive today.
But I appreciate that some of those parents whose children developed autism shortly after being given the MMR vaccine are as passionately convinced that the vaccine caused it as I am that the net effect of the vaccine is overwhelmingly positive.
Yesterday I saw this graphic on the subject:
.png)
Part of the problem is that the human brain evolved as a pattern-spotting mechanism designed to alert us to potential dangers and opportunities. And there are some patterns which appear to indicate danger or opportunity sufficiently often that there was an evolutionary advantage in selecting for people or proto-humans who employed them, possibly because in a pre-scientific era they were the least worst options our ancestors had.
Suppose you are an early hominid: your tribe has chosen or been forced to migrate to a new area with new fauna and flora. You are short of food and need to find out whether new items you might add to your diet are safe to eat. Taste and smell will take you so far, but in some cases you have to rely on trial and error.
You eat something and feel OK and reasonably conclude it must be all right. You eat something and are violently sick, and you avoid that food in future. A member of your tribe eats something and falls down dead and the rest of the tribe avoid it like the plague. And in the majority of cases will have been right to do so, though it is next to certain that some ancient tribes will have had the tradition that a particular item was unsafe to eat because some poor devil had an unfortunately timed heart attack or stroke which had nothing whatsoever to do with the particular food concerned.
It is my opinion that some patterns of thought which are not strictly correct will have been close enough to the truth, perhaps the best available approach, in more primitive times, to create natural selection in favour of those who use them. I believe that in consequence, a preference for those patterns of thought is hardwired into our brains, even though those patterns are not always right.
I further believe that one of those patterns is a propensity to give a bit too much weight to "Post hoc Ergo Propter Hoc" style thinking as though it were a golden rule which is always right, rather than a useful indication which can often be right.
The problem is that the "Post hoc" principle often works in practice, and I believe that it works often enough that in the process of human development we have become genetically programmed to think in those terms. Unfortunately although
"After this, therefore it MAY have been caused by this"
is a perfectly reasonable inference, we often act as though
"After this, therefore it MUST have been caused by this"
were also valid. And it isn't. Sometimes a third factor caused both events and sometimes the appearance of correlation is a pure coincidence.
A classic recent example has been caused by the fact that the age at which the NHS has chosen to give children the MMR vaccine happens to be close to the age at which children who are going to develop autism often begin to show symptoms.
Let me declare my own prejudice on the subject of MMR: my late mother had been, presumably wrongly, diagnosed with rubella when she was a child, and therefore she was not inoculated against it. Unfortunately she then developed rubella when pregnant with my sister, who as a direct consequence was born severely handicapped and died at the age of one week. If MMR had been available and in use when my mother was an infant, the overwhelming probability is that my sister would be alive today.
But I appreciate that some of those parents whose children developed autism shortly after being given the MMR vaccine are as passionately convinced that the vaccine caused it as I am that the net effect of the vaccine is overwhelmingly positive.
Yesterday I saw this graphic on the subject:
Part of the problem is that the human brain evolved as a pattern-spotting mechanism designed to alert us to potential dangers and opportunities. And there are some patterns which appear to indicate danger or opportunity sufficiently often that there was an evolutionary advantage in selecting for people or proto-humans who employed them, possibly because in a pre-scientific era they were the least worst options our ancestors had.
Suppose you are an early hominid: your tribe has chosen or been forced to migrate to a new area with new fauna and flora. You are short of food and need to find out whether new items you might add to your diet are safe to eat. Taste and smell will take you so far, but in some cases you have to rely on trial and error.
You eat something and feel OK and reasonably conclude it must be all right. You eat something and are violently sick, and you avoid that food in future. A member of your tribe eats something and falls down dead and the rest of the tribe avoid it like the plague. And in the majority of cases will have been right to do so, though it is next to certain that some ancient tribes will have had the tradition that a particular item was unsafe to eat because some poor devil had an unfortunately timed heart attack or stroke which had nothing whatsoever to do with the particular food concerned.
It is my opinion that some patterns of thought which are not strictly correct will have been close enough to the truth, perhaps the best available approach, in more primitive times, to create natural selection in favour of those who use them. I believe that in consequence, a preference for those patterns of thought is hardwired into our brains, even though those patterns are not always right.
I further believe that one of those patterns is a propensity to give a bit too much weight to "Post hoc Ergo Propter Hoc" style thinking as though it were a golden rule which is always right, rather than a useful indication which can often be right.
Comments
It's fantastic, usually you will see it used to appeal to a certain group. For example "google avoided enough tax to send every student to college with no fees" Never is it mentioned that even if Google did not and paid it, it most likely would not have been used to cancel tuition fees, or to employ x number of nurses either.
But cost the economy is my favourite, the government never took it so the economy lost it. no. The government never took it but it stayed in the economy just working it's way around.
The lunar missions lost the US economy billions of dollars. No. The government spent billions into the economy, they dud not blast the dollars to the moon, they placed them into the economy by buying rockets from rocket firms and things.