I am starting a blinded n-of-1 melatonin trial today. I have two bottles, A and B, with identical-looking pills that are either 3mg melatonin or placebo (provided by iTrial, a new product for blinded self-experimentation which I am testing out). I will take either A or B for 30 two-day periods that are randomly assigned (60 days total trial).

Outcomes are measured via a daily survey:

(Primary outcome) How restful was your sleep last night (1-10)?

What was your total sleep time (hours)?

How many times did you wake up in the night (0, 1, ...)?

How easy was it to wake up this morning (1-10)?

Throughout the trial, I will NOT KNOW what I had actually taken.

I currently use melatonin occasionally, maybe like 1x per week when I want to sleep well, and I feel like it helps a lot. But that could just be a placebo effect.

YES if: Any of the outcomes show an improvement with p <= 0.07

NO if: None of the outcomes show an improvement with p <= 0.07