A new survey of US workers indicates that about 15% of them have drink near or during working hours
. The headline, of course, says that "15% work under the influence," which is not what the survey showed. The survey included being hungover and having a drink within 2 hours of being at work (I hope they specified *before* work, but it's also possible that this includes after) as 'drinking on the job,' a loose definition at best.
An important aside, since the release mentions drinking at lunch, is that having a drink at lunch is entirely the norm in much (most?) of the world, and does not appear to be detrimental to productivity there. I'd like to see if the alcohol consumption/hangover data actually relate to performance data; I'm sure they do, after a certain threshold, but not immediately.