Control for all variables
Consider who gets their name on a lot of papers and participates in a lot of committees ("co-operates with others"): middle-age and older (takes a while to understand the system and make the contacts). I could already have told you that this group gets out for an evening beer less often than young newcomers. This article doesn't have a direct link to the paper it discusses, but did it break down by age?
Further contributing factors: active conference and standards-body participants travel a lot, and probably don't manage to stay in touch with their local pub regulars (or even local coworkers) so much as their less-active colleagues, so they wouldn't have anyone to talk with if they did pop in for a pint or two. The only people they see regularly live elsewhere.
Then, there are possibly social differences between the bar scene in Prague and that in Oxford, which may mean the results would not transfer. It may not be the actual beer that's the problem.