How to lie with statistics

Sales figures in music industry press releases are a lot like those chain e-mails you get about Microsoft tracking and rewarding how many people you forward it to, or how you can get a new car for free, in that there's an easy way to tell whether the claims are true. They're never, ever true. And if you think, well, maybe it might be true, then it's not true, either.

So, it's with great skepticism that I read an article in today's McPaper saying that hard rock is back, screaming with a vengeance. According to the article, hard rock album sales in the first half of 2003 are up more than double from the first six months of last year.

Now, on the surface, you might ask "what's the problem with the numbers?" Sales increased from 10.9 million in 2002 to 36.9 million in 2003. That's more than double, clearly. There are three main problems here. First, and this is endemic to every single statistic you'll ever see on the music business, there is no mention of how many hard rock records were released during each six-month span. If there were more than double the number of albums, the sales figures become a more conflicted message. Most records released on major labels are going to sell some nominal amount of records, From Zero notwithstanding, so it could be a scarcity issue. Fans would have bought more albums last year had more albums been released. Granted, that still speaks to the popularity of the genre, but in a different way. More pessimistically, if you flood the market with hard rock records, they're going to sell, because sales are tied more to relentless marketing than to actual tastes in the marketplace. It's the nature of the beast.

Second, which albums? An interesting cut of the data would be to look at the number of gold or platinum albums released in each six-month period, rather than talking about how many weeks rock albums spent at the top of the charts. Then, and you would there's enough detail in the sales data of the big retailers to see this, look at how many hard rock albums were purchased WITH those big-sellers. It's entirely possible that this "explosive" growth is just an artifact of a front-loaded release schedule that included Metallica, Led Zeppelin, and the inexplicably popular Linkin Park.

Which brings us to the third point, and that's whether or not a six-month period is at all appropriate for considering these numbers. There's absolutely no mention of the hard rock sales in the SECOND half of 2002, which makes me think that they were large enough to invalidate the whole argument. With the second Disturbed record hitting in that time period, I would imagine that to be the case.

None of this is meant to say that hard rock isn't somehow "back," in the sense that Rob Halford is "back" in Judas Priest. It may be that it never really left.