As a researcher/investigator I often dwell on this subject. This article from ParaScience.com is worth browsing:
http://www.parascience.org.uk/articles/evidence.htm
So what exactly constitutes "evidence" of paranormal activity? And how much evidence is required before one can say a specific location is "haunted"?
Well, to me credible evidence is anything that you cannot logically explain - things that move on their own would be the most credible to me, EVP's are good too, anything else is kinda iffy. "Creepy" feelings can be psychosomatic; "wierd" creaking and bangs are most often attributed to plumbing, or wood expansion, the wind, etc. There are many logical explainations for most "hauntings" - tho when that chair moved by it's at the lighthouse on Ghost Hunters, well, I was pretty impressed. If it was faked it was EXTREMELY well done (unlike that episode they had at the Queen Mary with the faked bedcover incident). I don't think they are fakers tho, so I cautiously trust their evidence. I guess it all comes down to - if you can't explain it, or think of a logical probability for it, then you have to consider that it might, just MIGHT, be paranormal.