I'd like to be able to quantify this phenomenon with a record of how many individual facts it takes to make someone see that another person's anecdote is not reliable. This would have to include the relative strength of each fact and how well the person knows the story-teller. I suspect the relationship between the two has some scale to it or measurable effect upon the efficacy of the anecdote's being believed.

My sister, who has known Uri Geller for some time and visited his house numerous times ( he buys her art...she's not there as a 'believer' ) , told me of one ocassion when she was absolutely certain Geller had bent a spoon without touching it. The spoon, so I was told, fell on the floor......and continued to bend significantly while lying there.

Well...being of scientific mind, I threw every question I could at this. Was it just a peculiar angle the spoon was seen from ? Was she distracted ? Etc etc. The trouble is that my questioning led to almost a sort of ' so you're calling me a liar ?' response.

The problem with anecdotal evidence is not just that one was not there.....but that one cannot be certain that one would not oneself have been deceived even if one had been there.

I have an anecdote from Professor Arthur Ellison, in which he describes an incident in which he and Arthur Koestler and one other, were all witness when Geller bent a Yale key that was placed flat on a table with Koestler's thumb on the big end. Ellison swears blue that he saw this happen.

I hate having to rely on someone else's one-off report.....no matter how reliable that person may be and no matter how much they may swear blue that they know what they saw. There's a particular annoyance that comes from reading or hearing of such things. I think James Randi expresses it quite well in some of his articles.