Earlier this fall, a couple of beer writer friends and I joined the throng attending the Great Canadian Beer Festival in Victoria, B.C. We'd scoped out the Canadian beers we wanted to taste, and wandered from booth to booth, spending tokens on two-ounce samples.

When it turned cooler and the crowd turned friskier, we decided it was time to head into town. I asked one friend what he wanted to do next. "I feel like having a beer," he said.

I laughed. We'd been at a beer festival for five hours, I pointed out.

"Ah, we've been tasting beer," he said. "Now I want to have a beer."

So the three of us found a quiet bar with an average selection of beers, and each ordered a pint of local brew. We sat and talked, drank, and ordered a second. The experience was a complete contrast to the festival we'd just attended: a full pint was different from a small sample, and the context was relaxed conversation, not focused tasting.

I've attended many festivals, and conducted scores of formal beer tastings, where the purpose is to try small amounts of a wide range of beers. I find the events to be a great way to broaden people's understanding of beer, and expose them to new flavors. Still, my friend's comment on having a beer, rather than tasting one, set me thinking.

I recently got around to reading Malcolm Gladwell's "Blink," a clever book that examines the split-second insights we experience, sometimes based on very little information. These first impressions can sometimes be more accurate than painstaking evaluations based on much more systematic and thorough data-collection. But there are other times when these first impressions let us down.

As an example of the latter, Gladwell recalls the "Pepsi challenge," in which consumers tasted small samples of Pepsi and Coca-Cola in blind tests, and were shown to prefer Pepsi in significant numbers. The Coca-Cola Company was so alarmed by its competitor's campaign that it eventually reformulated its recipe. After careful study, the company launched "New Coke," which more closely resembled the flavor customers seemed to prefer. It flopped.

Coke got it wrong on several fronts. Pepsi was sweeter, and tasters preferred the sweeter flavor -- at first taste. But the point Coke missed was that a flavor that was initially appealing didn't necessarily wear well: a sip was good, but a whole can was a different matter.

Coke also missed the importance of context. When customers knew which brand they were drinking, that affected their experience of taste. Coke misread the results of an isolated taste test as a rejection of the old Coke brand. Customers actually felt a powerful loyalty to the Coke they knew: its flavor, its appearance, its branding -- the whole experience. Context was critical.

So, what do we do when we taste an ounce or two of a beer, when we concentrate on the first flavor impression, and strip away the normal social context of beer drinking? It seems likely to me that we get excited by extreme flavors we might not want to stick with for an entire pint. Highly hoppy beers, very sweet beers, high alcohol beers can be very impressive. But it may be that these extreme tastes won't wear well.

It also seems possible that a beer tasting, like a wine tasting or any other academic exercise that takes food and drink away from their social context, may give us distorted information. Am I giving up on tastings? No, not at all. But I'm going to give myself a second challenge: When I sample a beer, I'm also going to ask myself whether this is a beverage I'd enjoy having by the pint with a couple of friends. A first impression may not tell the whole story.