Actually anecdote is exactly what is most trustworthy in the context of another's mind. The primary reason is, in order to disprove my hypothesis that meditation is good for me, you must be me. Given you can't be me, disproving my assertion requires disproving what I'm saying about my own thoughts and feelings. While we, as humans, are exceptionally good at speaking for others, it's a horrifically inefficient process to render someone else's feelings. Chances are, you'll be wrong about how I feel, given you aren't me and simply don't have enough time left in this world to accomplish being me. And trust me, you don't want to be me.

Note that my "reality" here is lack of circumstantial evidence for the harmful effects of meditation, which would just help identify there was a problem to begin with. In theory, this could be put to the scientific method to prove or disprove, given you could measure someone's mental state using something other than "he seems unstable today", which you can't, without speaking for someone else.

Anecdote is also often the first sign that a phenomenon exists. It shouldn't be dismissed out of hand, but if all the evidence you have is anecdotal, that's your cue to gather something more robust.