Chad Orzel has some very penetrating thoughts on the Sherman interview, and on Blogcritics in general. He should join us, shouldn't he?

On Sherman:

It sounds impressive, but it's a classic attempt to befuddle the public with statistics. 19% plus 41% is only 60%, and having covered "purchasing more" and "purchasing less," the only options left for the missing 40% are "purchasing about the same number of CD's" or "don't know/ no answer."

40% would be an awful lot of "no answer" responses, so at least some of those are probably "about the same." If it's half "about the same" and half "no answer," well, then, we have what statisticians would call "a wash" — 41% purchasing less, 39% purchasing about the same amount of music, if not more. If you put all those people into the "about the same" category, an arguably more accurate spin on these numbers would be that 59% of people surveyed bought at least as much music after file-sharing as before.

Now, 41% is still much larger than 19%, so it's reasonable to believe that there's been a net reduction in sales, but that's not necessarily true. If those 41% halved their music buying, and the 19% doubled it, it would basically be a wash. If the 19% doubled their purchasing, and the 41% reduced their spending by a third, it'd be a net gain. Without knowing what they actually asked, and the relative sizes of the increases and decreases, it's impossible to say whether file-sharing is really decreasing sales. Sherman's initial citation of this study is the sort of dodgy use of statistics that should raise a red flag for anyone who can do math.

The right thing to do here would be to use this study as a starting point, and use it to craft a sensible approach to file-sharing technology....

Which brings us around to the problem with Blogcritics (which, I hasten to add, I think is a brilliant idea)-- what we've got here is an assemblage of a hundred random people writing about music, each from their own perspective. It's a great idea, but it's not really a useful tool, yet. It'll take a good while before it becomes truly useful, as we, the readers, will need to see a bunch of reviews from the various critics before we can really judge whether Ken Layne liking an album will mean that I will like the same album, or my wife will like it, or whoever.

A humble suggestion, then, or at least as humble a suggestion as you're likely to get from anyone arrogant enough to run a web log: The process could maybe be sped up a bit by getting the various blogcritics to provide some sort of reference point from which to judge their tastes. Something like a Ten Favorite Albums list (at the risk of sliding into High Fidelity territory), with a few sentences (say, 75 words or less) saying what they like about each album. Then provide a link to that list from the blogcritics site, either at the end of the reviews, or off the front page, so that when, say, Brian Linse says an album reminds him of London Calling, we can get an idea of what, exactly, he means by that.