What I learned from gatecrashing the Australasian Statistical Conference, day 1: Who needs experiments? Who needs truth?

(Those readers who consider themselves "normal people" might like to take a few days away from reading this blog. Don't worry, some serious movie-watching is coming soon.)

• How can you deal with observational data? By adjusting mean and variance for the fact that it's observational! I'm see-sawing between "this is idiocy" and "this is genius idiocy".
• The Australian Bureau of Statistics (and I guess most official statistics agencies) fudges its numbers! It does this for noble reasons (avoiding the release of personally identifiable information), but it's still fudging, perturbing table cell counts from their measured values. That they can do this consistently without biasing or affecting grand totals is quite ravishing though.
• If you're giving a talk on your research, you should know not to make it so basic as to be freshman level, and also not to make it so detailed that it's of no plausible use to anyone besides yourself. If, however, you veer discontinuously between these two no-nos, I for one may find it oddly thrilling.
• At the Sky City hotel, at which I couldn't afford to stay even if my parents weren't putting me up, a condom from the minibar is five bucks. Crikey, for that price in Thailand you could get two hookers and a vasectomy.*

*EBV does not condone paying workers of any description less than a living wage.