We're all familiar with the argument that police forces and the like profile because it's effective. People of a kind tend to commit more crimes than others, so they get focused on. This could be based on anything a 'lifetime of experience' has suggested - profiling black people for theft, gays for child molestation or women for shoplifting, middle-aged white men for fraud and corporate crimes.
The results of a study by the British intelligence service (MI5) suggest this cannot be applied to terrorism. As a state with a considerable history of anti-terror work, they're good people to listen to. I've put a link to an article summing up the general thrust of the report below.
If you haven't picked it up yet, I'm inclined to agree with the Brits. Supporters of profiling appear to think that terrorist groups are stupid, that they lack the sophistication to figure out what a security service is looking for and supply something different. Thus profiling for terrorism actually weakens national security, rather than strengthening it, because it adds the implication 'not terrorist' to anyone who doesn't fit the profile. That's bad news when you're trying to instil eternal vigilance.
Or perhaps the Brits just weren't looking closely enough; there was an implication in the newspaper article that most terrorists have low-paying jobs. Is it time we started profiling society's poor instead?