On Causality (and covariance)

Here I was planning to write all these Very Important Posts I had postponed for too long, when I stumbled upon this fabulous cartoon on causality (via Espen Andersen) – very apt for Saturday Morning.

Now, to stay with the deeper questions for a second or two: you may of course ask what the detour from those blog posts I planned to write signifies about my personal epistemology, or in plain talk: am I a scatterbrain, or is the web turning me into one? That's one of the serious things I was planning to blog about, but first this cartoon (translation below):

For those who can't read the text: During a convivial gathering there is talk of the unhygienic aspect of using galoshes. One of those present chips in: "Yes, I've also noticed this. Every time I've woken up with my galoshes on, I've had a headache."

Now, Espen's more serious point is that there's a lot of bogus stuff which passes for causality these days, to which he adds (and I'm paraphrasing him here): 'I mean, we all know that the supposed global temperature increase we're currently experiencing is due to the baby-boom-generation hitting the climacteric age with corresponding hot flushes, right?'

Since I've studied both philosophy and statistics, the latter as part of my politics degree, I found Espen's post particularly funny: always great to see issues academics argue about, with a huge arsenal of technical terms, brought down to a level where it's clear how it impacts on daily life. When I studied game theory I kept promising myself to write one short story for each or the different games to make them more comprehensible (and fun: it was very dull reading), e.g. a love story called "Priscilla's dillema", but alas, never got around to that.

Comments

[Candadai Tirumalai]

Two or so years ago, someone wrote on a blog that "Correlation is not Causality." I thought it a succinct way of making a point. Since then I have come across it quite often, and have wondered at what point a phrase or idea loses its potency from frequent use. Variation, I suppose, is one way of extending its life.

[nobody you know]

Why wonder? You can look it up!
http://www.google.com/search?q=%22Correlation+is+not+Causality%22

[Espen]

Boingboinged! http://www.boingboing.net/2008/03/29/cartoon-explains-the.html

[a_spod]

:lol: I think you got a better statistics education from a degree in journalism than I did from one in physics. (And damn, they taught you Game Theory?! That's not fair!)

Anyway, you sent me on an odyssey looking up Post Hoc and Cum Hoc fallacies on the web. http://www.logicalfallacies.info/cumhoc.html lines up declines in piracy with rising temperatures. (You'd think higher sea levels would mean *more* piracy. ;-) http://www.fallacyfiles.org/cumhocfa.html has an oft cited example by David Hackett Fischer. But Epsen's and the toon are both good 'uns.

However in any debate on statistics shouldn't somebody provide the data which shows there is ‘a lot of bogus stuff which passes for causality’? (Otherwise I'm gonna shout ‘Confirmation bias’.) To be honest, humanity is soo hard wired to see causal connections (and then prove, or disprove them) that I suspect these fallacies retain a constant rate. But perhaps our more secular society is looking for less supernatural explanations, so we're more aware of these fallacies? Or perhaps, because of the internet, we can see things on a global scale that were always there locally?

Anyway, if in doubt, I just blame the Big Bang: it's the cause of everything... :D