No more pop generation gap, claims this survey by Pew Internet. “There’s now broad agreement across the generations about one realm of American culture that had been an intense battlefield in the 1960s: the music.”

Unfortunately for Pew, its headline chart for this study – “The Generation Gap, 2009” – completely contradicts this spin. Second on the list of “things young and older people are different in” is music – ahead of moral values, religion, and indeed everything but those pesky computomators. Scroll down to the charts of which ‘iconic musical performers’ each agegroup likes and, yes, it would seem that there is broad disagreement and bafflement over many if not most of them.

So how come Pew’s analysis on this one is so odd? For the same reason so, so many surveys of music taste are failures: they insist on trying to run analyses by genre.

I’ll hold up my hand and say that I’ve fallen into this trap too – when working on a study of teenagers last year I ended up putting a bunch of genre questions in. In my case I made the typical music head’s mistake of trying to include too long a list, especially when it came to dance sub-genres. Pew go in the other direction, and boil music down to seven flavours, which is where the problems start.

65% of people listen to rock – a generational unifier, hurrah! But without caveats this could mean almost anything: The Eagles, Slipknot, Andrew WK, Elvis, Motorhead, The Pains Of Being Pure At Heart (there’s no “indie” or “pop” on the list, of course). Even more problematic is the category “R&B / Rhythm And Blues”: there is a very strong possibility that a 55-year old and a 20-year old ticking this box are simply not talking about the same kind of music, at all.

Hence the mix up, I think: music is still divisive, but often it’s the different styles cropping up within the hollowed-out macro-genres which cause that division. The lesson for Pew (and all of us): researching music tastes is a strange and difficult row to hoe.