Here's an article from an Australian source claiming that, essentially, rock stars die young due to drug abuse and the pressures of fame (article via Click Opera). It seems rather suspect (and not only because the author doesn't know how to spell Jimi Hendrix's name).
First, rock'n'roll simply hasn't been around long enough for its earliest exponents to live long enough to die at a fine old age. Someone who was 25 in 1955 would be celebrating only his 78th birthday this year, only a few years older than the average life expectancy in the West. Any average calculated now will inevitably skew young: no rocker has lived to be 100, because no one in their middle forties was making rock'n'roll in the mid-1950s.
Another factor that would skew the age young: the survey includes rock stars who'd died up until 2000...but obviously, it could not include those who were still alive in 2000. Those folks might live to be 50, or 60, or 100: we just don't know yet. If I say, what's the average age of death for folks under the age of 30, you'd be right to question what the answer proves...obviously, people who die under age 30 tend to die of misadventure, since very few people under 30 die of natural causes. And equally obviously, that age is going to be...under 30. Effectively, this is what the study's methodology does. (Similarly but unrelatedly, scare articles that point out that suicide is the most common or second-most common cause of death among teens shouldn't surprise: few teens are going to drop dead of a heart attack or lung cancer. If a teen's going to die, it's probably going to be from an external cause: murder, car crash, suicide.)
Sorry: it's simply too soon to make any statement about whether rock stars, on average, live shorter or longer lives than average.