David Harrell  is the Editor of Digital Audio Insider and has blogged about the economics of digital music since 2006. With his indie rock band the Layaways , he has self-released three albums. The latest, "The Space Between," is now available for free streaming and download from Last.fm.

The best I can remember, the class was called something like "the scientific method." A graduate biology seminar, it consisted of meeting once a week to watch an old science fiction movie, followed by a discussion about the treatment of science in the film. For the students, it was a low-stress way to add a couple semester hours to our schedules, and for the tenured professor who "taught" the class, it seemed like an incredibly easy way to pad his course load.

But unlike most of my college classes, there's something specific that's stayed with me. It was the professor's contention that there is no such thing as a truly repeatable experiment, at least not for biological sciences. The classic scientific method depends on the notion of repeatable results -- running the same experiment again should give the same results as those obtained in previous experiments. His point was that no matter how careful a researcher was, there'd always be some alteration in a small detail, such as a different batch of food for the lab animals or the health of the laboratory personnel. He wasn't implying that most scientific research wasn't valid, just that there was no such thing as a perfect do-over for most experiments.

Why am I yammering on about science experiments in a music business blog? Well, as anyone reading this post knows, over the past few years, the music industry -- from the largest record companies and the biggest selling artists down to the level of self-released artists -- has been experimenting like crazy. Free music, pay-what-you-want music, "360" deals, exclusive deals with Wal-Mart sans digital distribution, iTunes exclusives, etc. And after every large-scale experiment (Radiohead's "In Rainbows," Kid Rock's no-iTunes strategy, AC-DC's current Wal-Mart exclusive), music industry analysts, the news media, and bloggers attempt to assess the relative success of the approach.

Yet in all of these cases, "experiment" is probably a misnomer --there's no "control group" receiving the placebo treatment. Absent a trip to a parallel universe where you could buy Kid Rock's last album (or the single) from the iTunes store, we really don't know for certain if Kid Rock helped or hindered his total sales.  (It's possible, of course, that some record companies have been using control groups of some sort. You could make an album available in iTunes for a specific country or region, and compare sales to those in non-iTunes regions. Though the demographics and fan bases probably aren't identical across regions, so you're still guessing somewhat…)

And even if we knew for certain if an individual experiment was a relative success, it's not necessarily transferable. Just because something worked for Radiohead doesn't mean it would work for R.E.M. Further, in addition to the non-interchangeable nature of audiences and albums, the music retail environment itself is changing on a daily basis.

Yet music is art, not science, and even if these different business tests aren't controlled, repeatable experiments, there does seem to be a few obvious takeaways. One thing that seems certain is that for acts of a certain stature, deviations from the standard sales approach will result in increased attention, perhaps enough to generate additional sales. Offering free music certainly seems to have helped Trent Reznor SELL a lot of music, music that is readily available for no charge. And maybe AC-DC's new album wouldn't be selling as well without the promotional push behind the Wal-Mart exclusive.

One problem, however, is diminishing returns. Radiohead got the attention it did for its "In Rainbows" experiment not because it was the first act to offer its music on a "pay what you want" basis. Rather, because they were the first act with that level of name recognition and artistic credibility to do so. Post-Radiohead, a similar experiment by a well-known act might not get the same attention.

What's less obvious, however, is if free or pay-what-you-want music has the same impact on the other end of the scale. That is, for self-released acts like my own, who are all trying to figure out how to compete in an environment of seemingly endless listening choices.

If you spend some time on the CD Baby message boards, you'll see that some self-released musicians that are quite indignant over the idea of "free music." They'll point to the both hard work that went into writing and recording their music and the hard-earned dollars that funded the recording, mixing, mastering, and manufacturing of their music. They simply want a chance to earn some of that money back.

But offering free music doesn't mean you don't expect to get paid for it. For musicians at ANY level, the fundamental challenge is twofold. First, you need to get people to hear your music. Then you have to convince them to buy it. For relatively unknown artists, without access to commercial radio and the mainstream music press, offering free music along with paid versions of it seems the easiest way to increase your listening audience and, eventually, your paying audience.

Anecdotally, there are plenty of examples of how free and paid music can co-exist. A few years ago, when the Strokes released their second album, they offered a free 192k mp3 of the lead single on their website. Yet that song remained their top-selling track on iTunes, despite the existence of a free equivalent. And when I look at our cumulative iTunes sales, the two tracks we've sold the most copies of are songs we've made available as free downloads.

These examples don't necessarily prove anything -- there's no way to know if the Strokes would've sold more (or less) iTunes downloads if they didn't offer a free version of the track. Nor do they reveal why some consumers are paying for music they could have legally obtained without purchasing it. Maybe some purchasers of the iTunes tracks were simply unaware of the free versions, or maybe they are deliberately choosing to support the artist by purchasing the tracks.

For our new album, we're taking the free music plunge, inspired in part by a post by David Rose on this blog. For their 2004 release "Conductor," the Comas had a level of critical and promotional success that most small bands would kill for -- things like an 8.0 Pitchfork review and strong support and airplay from KEXP (one of the biggest/best known CMJ-reporting stations).

Yet they sold less than 6,000 total units, including physical CDs and album downloads. Would the Comas have done better with a free strategy? Again, without access to a parallel universe, there's no way to know to know for certain.

But in our case, given our very modest sales, the upside seems to outweigh the minimal downside of potential lost sales. And while it won't be a controlled or repeatable experiment, there are plenty of things we can quantify over the next few months: our number of Last.fm listeners, web traffic, friend statistics for MySpace and Facebook, and -- we hope -- increased sales over our previous release.