How are you controlling your experiments? (also I wonder if you and I mean the same thing by 'experiments'). If you're not controlling them you don't know what (which) variable(s) are dependent. So you don't actually know what you're changing. For instance, Valium, being a benzodiazepine, also produces muscle relaxation, anterograde amnesia (difficulty creating new memories) and sedation. So, was it really that your psychological state was altered? Or that speech production was easier because muscles were less tense? Or did you actually not remember stuttering? Or were you not alert enough to notice your stuttering? (And I do hope you were not driving while on valium). Not to rain on your parade, but single-subject drug trials are incredibly complex. I wouldn't be convinced that you found anything. Meditation has some of the same limitations (albeit, it is probably safer and more legal to experiment with a regiment of meditation than regiments of controlled substances). So, keep a journal of how much you meditate, the quality of the meditation, how you feel before / after. At the end of the day record how much stuttering affected you this day in a 1-10 scale. Record yourself speaking in controlled environments (e.g., reading out loud, conversation with a familiar person). After the fact count your stutters (in a 300 word segment), and compare how your mood changed over time. That's something like what I would do.