Thanks for bringing up N-of-1. My team and I have been working on operationalizing N-of-1 for lay-users to DIY lifestyle medicine.
I'm loath to criticize any writer, particularly when they put in as much effort as you have for your article. So, please do not take my comment as negative criticism.
There is a lot more to N-of-1 than what you demonstrate with your experimental data. In fact the test statistic you use is questionable as a stand-alone in N-of-1.
The purpose of N-of-1 is to have the subject become their own intervention and control group. That is, over consecutive periods of applying the intervention (drinking alcohol) and withholding it (staying sober) we look at the daily data of each phase. In your case, the simplest scenario would be the AB design. "A" being the baseline phase (e.g. 14 days of drinking alcohol) and "B" being the intervention phase (not drinking alcohol). We would then compare the data , not just with a test of significant differences between means, but also exploring phase trends relative to each other. Only such a configuration qualifies as an N-of-1 study design, and it is the simplest of them. Each of those designs uses specific statistics. Comparison of means often doesn't catch intervention effects, particularly if the signal-to-noise ratio is small.
We have made several short explanatory videos about the subject of DIY lifestyle medicine on our site: www.adiphea.com
particularly one that you and your readers might find interesting addresses the N-of-1 method in a lay-friendly way:
I like very much that you have recognized N-of-1 as a key toindividualizing lifestyle medicine. And I encourage you to pursue his path further and in-depth.
