Labs
Twelve weeks, twelve methods, one day taken apart and rebuilt
You arrive with a record of your own day and leave with a result about it, or, more often, with a reason the result cannot be trusted.
These are twelve methods, not twelve topics, each introduced once and each depending on the ones before it. They run in three moves:realisation in weeks 1 to 4, where you learn to see your own day accurately; change in 5 to 7, where you alter one thing and test it properly; and rethinking in 8 to 12, where you work out how much of the result was the change and how much was everything else.
The Day You Didn't Design
We record one ordinary day and change nothing about it, which is harder than it sounds
Week 1 · 22 February 2027 · naturalistic baseline
What Is "Tired"?
Before you can measure how tired you are, you have to decide what you mean, and that decision quietly settles what you are able to find
Week 2 · 1 March 2027 · operationalisation
Two Instruments, One Night
Your phone, a paper diary and your own memory will disagree about last night, and none of them is the referee
Week 3 · 8 March 2027 · measurement error
The Observer in the Room
You cannot watch your own day without changing it, so this week we measure the size of the change instead of pretending it away
Week 4 · 15 March 2027 · reactivity
Take It Away, Put It Back
The simplest honest experiment one person can run alone, demonstrated on breakfast
Week 5 · 22 March 2027 · ABA withdrawal design
Twenty-Five Minutes
The most repeated productivity instruction in the world was one student's answer to one question, so this week we test it properly
Week 6 · 29 March 2027 · counterbalancing
You Are the Placebo
Every experiment in this course has the same unfixable flaw, and this week we look straight at it
Week 7 · 19 April 2027 · blinding
Everything Else That Happened
You cannot randomise your own life, so this week we go looking for what else moved while we were not watching
Week 8 · 26 April 2027 · confounding
Why You Felt Better Anyway
Almost everything appears to work, because almost everybody starts on their worst day
Week 9 · 3 May 2027 · regression to the mean
The Subject Has Changed
We run week 5's experiment again, on the same person, and get a different answer
Week 10 · 10 May 2027 · non-stationarity
Forty Things at Once
Track enough about yourself and something will always correlate with something, which is how a dashboard lies to you with true numbers
Week 11 · 17 May 2027 · multiple comparisons
The Day You Don't Measure
When to stop experimenting, why every answer you found has an expiry date, and what you are left holding
Week 12 · 24 May 2027 · stopping rules and exit