Designing a personal experiment that could change your mind — one year on — setting out what I have, and where I think it stops being reliable.
Two things I would like separated before anyone answers on Designing a personal experiment, because they get bundled and then argued about as one thing.
The first is descriptive: what has actually been observed, by whom, and how. The second is causal: why. I am asking about the first only.
Post #3 is right about the mechanism and I think understates the practical bit.
Practical answer on Designing a personal experiment, since the theoretical one is upthread: do the simplest check first, write down the result, and only then decide whether the complicated explanation is needed. It usually is not.
The interesting personal experiments are the ones where the answer was no. They are also the ones least likely to be written up.
I would not lead a decision with this, but I would not ignore it either.
Picking up post #3: that is the part I would want checked first.
Record everything that changed, not just the intervention. Personal experiments are confounded by life, and the confounders are only recoverable if they were written down.
Coming back to post #3, because the follow-up matters more than the original answer.
Two people in this thread mean different things by Designing a personal experiment and are disagreeing about the definition while believing they are disagreeing about the facts. Worth pausing to define it.
Post #7 and I disagree about the size of the effect, not about the direction.
Generalisability: a robust n-of-1 result applies to you. It does not tell you much about whether the effect generalises to others similar to you, much less to people different from you.
The most useful reply I ever got about Designing a personal experiment was a request to state my units. It sounds like pedantry and it has saved me twice.
I had written a reply contradicting post #10 and deleted it. Here is what survived.
Regression to the mean explains a great many personal experiments. People start when things are at their worst, and things at their worst tend to improve.
I have deliberately not rounded that, because the rounding is where the argument starts.
The arithmetic in post #12 is right; the assumption feeding it is the part to check.
Designing a personal experiment that could actually change your mind: that is the standard for an n-of-1 design. An experiment designed so that any result confirms what you already believed has not changed anything.
The uncertainty is in the assumption, not in the calculation.
Having read the whole Designing a personal experiment thread before replying: the question in the first post has not actually been answered yet, and three of us have answered a nearby one instead.
A washout period between conditions is what stops one bleeding into the next, and its length should be set by the half-life rather than by convenience.
Everything in post #17 holds. The case it does not cover is the one I have.
An n of one tells you about one person, which is the person you are most interested in. That is the whole value and it is not nothing.
Reporting the observation and leaving the explanation open deliberately.
Reporting the whole series rather than the interesting segment is the discipline that makes personal data worth reading. Selective reporting is the default without effort.
The reasoning is more useful than the number, which is why I have shown it.
Duration should be set by how long an effect would take to appear rather than by patience. Stopping early is the commonest reason a personal experiment answers nothing.
That is what the documentation says. What happens in practice is usually close.
Narrowing post #23, because the general version has more than one answer.
Designing a personal experiment is a question about a distribution, not about a value, and treating it as a value is what produces the confident wrong answers.
Adding a reference point for Designing a personal experiment. Mine is a single case, collected without controls, and I am posting the method alongside it so it can be discounted appropriately.
That reframing is the whole thing. The facts I already had.
Post #26 answers the question as asked. The question underneath it is different.
Nothing in a personal experiment is medical advice to anyone including yourself, and several people here have found that a prescriber will happily discuss a well-documented one.
Post #28 is the version of this I will quote in future. One addition.
Multiple outcomes measured without a primary one means something will move. Nominate the primary in advance and report the rest as secondary.