people will tell me things like "but, our market surveys suggested that people smoke, on average, three grams a day".

we just went over this with the bradley effect.

i think it's a pretty sad reflection of society, but i think everybody knows it's true - if you call up people on a saturday night and get them to answer your survey, they're probably going to exaggerate the amount of pot they've smoked. they don't want you to think they're a prude, or a loser, or something. as they answer your survey on a saturday night...

so, those surveys are of limited value due to the presence of an obvious bias: nobody wants to admit they don't smoke pot at all, and pretty much everybody is going to exaggerate the amount they've smoked in the past. it's sad, but it's true.

the actual sales data is far more instructive, and nobody should be surprised about the disconnect.