I very much doubt you'll experience data loss, and if you do, you should have your backups (duh).

Every xml parsing modules has about 0% chance of losing data on its own (it's only reading after all, and that can't be damaging :D).

Now, look at the reports of cpantesters, and check your target platform(s) for success/failure (lack of test reports don't mean much, inspect reports closely as not all cpan-testers know what they're doing -- true even more today due to CPANPLUS)
http://testers.cpan.org/search?request=dist&dist=XML-Parser
http://testers.cpan.org/search?request=dist&dist=XML-LibXML
http://testers.cpan.org/search?request=dist&dist=XML-Twig

Do a codereview like monks have already suggested, but also be sure to review the tests (and the more tests there are, the better).

Look for bug reports (http://rt.cpan.org).

Now, also look at who wrote the module. I frankly wouldn't do too much of a codereview on a module like Text::xSV which is written by tilly. Some of SCHWERNs stuff I might inspect a little closer though ;D

MJD says you can't just make shit up and expect the computer to know what you mean, retardo! I run a Win32 PPM repository for perl 5.6x+5.8x. I take requests. ** The Third rule of perl club is a statement of fact: pod is sexy.