Free Trials - Open registration Vs. Credit Card required

I think your general principal is provide a "minimum barrier to entry".

The harder you make it, the less people will adopt it. There were several a BayCHI podcasts and several automate my small business podcasts on this topic (over the last 3 years) looking at barriers on your site and the rate of "drop off" (or "improvement you could make") by reducing the number of steps before the user achieved the result and choosing to adopt your site.

I think this principal applies in your senario, if the first thing you place in front of a user is a form with lots of required fields and "give us your credit card" then the majoirty of people will turn away.

Generally I would suggest letting them on the site and use a tracking cookie, then when they do something minor you collect a name, then collect a bit more, over time you build up the profile. You collect their credit card when they go to make their first transaction, because its the first time they will be committed to your site ... prior to that your an unknown, and a large percentage are likely to leave.

The stackexchange sites are a good example of the softly softly approach (in fact, their old podcast may have Joel and Jeff talking about this). You land on the site and it doesn't know you, but you can do stuff, then you want to upvote or comment so there is a little bit of a signup using OAuth, then you get more involved and there are badges and points and you fill in your profile ... etc.

As the other answers have indicated, you absolutely must test this as it vary widely by your industry, offer, reputation and product. In general the harder it is to "get" a product or the more options in the field, the less barriers people will accept to sign up while the more value a tool is known to have, the easier it is to get people to stick more out there to try it.

That said, don't overlook the value of a perceived non-paying user.

Whether it's in evangelizing the benefits of your service to other potential customers or a conversion down the road [your sales cycle could take days, even months], there's lots you can do with a sign up and little you can do with an abandoned form report.

Microprofiling as Robin suggested can be a good way to begin to prioritize and segment your sign ups based on what they've done while not excluding people from "dabbling". As you learn why customers buy, you can then start to move the right people along

Finally, if you are seeing really strong results in unpaid sign-ups you may consider this as opportunity to evolve your offerings and introduce micro-features. Social & Mobile Gaming does this extremely well -- paid apps can sell well, but free apps with various "upgrades" appeal to a much wider audience, hooking them in and giving them a reason to upsell themselves but without the absolute of a "paid" account.

Dharmesh's answer is spot-on as always -- however, I would add a few things:

We currently offer a no obligation free trial. The theory being that we will have higher total numbers >by allowing customers to experience our product and service risk free. So far conversion rate is low, >but about on par with the industry standard.

While I don't know your business, higher conversion rates are the goal -- but revenue is. By that I mean, ceteris paribus, you would probably take a lower conversion rate if it meant that the people you are converting are spending more.

However, over the last few weeks we've gotten more popular and better search exposure thanks to >ongoing blogging and some media spots. Consequently, we're also getting more people starting free >trials that really have no intention of ever converting (high school students, college kids doing >research, or people we personally email after signing up to help with any questions and never >respond).

You are getting people outside of your target market wasting your time & theirs. This may be a good reason to "qualify" people by asking for cc info. See my first comment.

What makes this situation unique is that our site has a large community component, and I don't want to >disturb what we're building in that area by people who aren't serious about using our service (i.e., >welcoming new members, new users posting silly questions, those sorts of things).

What is the potential future value of those people? Will they ever fall into a market segment you want/can address?

Lately, I've been thinking of moving to a credit card required free trial system. However, I don't >want to lose out on potential customers who don't want to take the risk, despite our moneyback >guarantee.

What about losing out on customers who want to give you their credit card now? There are lots of cases where customers want to see price & be asked for credit card info as it increases perceived value of the product & signals stability.