This is not my field of expertise at all. However, as someone pointed out, all of us get "gut feelings" about someone lying. Without actual proof of lying, how are these gut feelings any different than any other "feelings" people get about any number of things we consider woo?
To me the most compelling evidence of someone lying is when the story is either self-inconsistent or inconsistent with what we know about the subject being discussed. The more you can get someone to talk, the better the chance of revealing this.
If I know someone, then it's easier to tell when they are lying. The more I know someone, the easier it is. Usually there's a change in how they tell a truthful story versus an untruthful one. This assumes, of course, that the person is not an experienced liar who can do it very smoothly and never be found out.
Beyond that I'm not convinced that investigators aren't assuming they are seeing indicators of lying because they have other more reliable information to indicate the person is lying (consistency, for example). Like I said, my experience is limited, but I cannot tell you how many times I've seen an investigative TV show where the investigator points out the manner in which a person says something as an indicator of untruthfulness, and I disagree. They say, "Would an innocent person say or do this?" A lot of times my answer is, "Yes, I would say or do that. People I know would say or do that."
So, as a skeptic, I have to ask where the evidence is proving these theories? I'm willing to be convinced.
Also, someone pointed out that writing Internet posts doesn't count. Where is the evidence for that? Some posts I write are well considered - I pick and choose my words. Other times, like in my joking response to LongTabberPE, it was competely off the cuff - just like normal conversation. It surprised me that my joke was completely missed by someone stating how reliable statement analysis is.