What something looks like before you interfere with it. Without one, you have nothing to compare a change against, so the first week of this course changes nothing on purpose.
Introduced in week 1 · The Day You Didn't Design
Glossary
Every technical term this course uses, in plain language
Each term is linked from the first week that uses it. If you meet one the course never explained where you first met it, that is a fault in the course, not in you.
- baseline
- blinding
Not knowing which condition you are in, so that what you expect cannot colour what you record. Doing this to yourself is close to impossible, which is a real limit on every experiment in this course.
Introduced in week 7 · You Are the Placebo
- confounding
When something you were not tracking moved at the same time as the thing you were, so you cannot tell which one mattered. Weather, deadlines and other people confound almost everything.
Introduced in week 8 · Everything Else That Happened
- counterbalancing
Alternating the order you try two things in, so that “I did this one second” cannot be mistaken for “this one worked better”.
Introduced in week 6 · Twenty-Five Minutes
- mesmerism
A treatment fashionable in Paris in the 1780s, whose practitioners held that an invisible fluid runs through every living body, that illness is that fluid flowing wrongly, and that passing their hands over a patient could put it right. No such fluid was ever found.
Introduced in week 7 · You Are the Placebo
- multiple comparisons
Testing many things at once and reporting whichever one looks interesting. Track forty measures and something will correlate with something by chance alone. This is how dashboards mislead.
Introduced in week 11 · Forty Things at Once
- non-stationary
Changing over time. A person is non-stationary, so an answer that fitted you in March may not fit you in May, and that is not a mistake in your method.
Introduced in week 10 · The Subject Has Changed
- operationalise
To turn a vague word into something you can actually write down. “Tired” is not a measurement; “minutes to fall asleep” is. Choosing badly here quietly decides what the rest of your experiment can find.
Introduced in week 2 · What Is "Tired"?
- orthosomnia
Sleeping worse because you are tracking your sleep. Sleep clinicians named it after a run of patients whose main problem turned out to be anxiety about their own sleep data.
Introduced in week 3 · Two Instruments, One Night
- placebo
A real effect produced by expecting an effect. It is not imaginary. It simply is not caused by the thing you believe caused it.
Introduced in week 7 · You Are the Placebo
- pre-registration
Writing down what you expect, how you will measure it, and when you will stop, before you begin. It is what stops you quietly changing the question once you have seen the answer.
Introduced in week 6 · Twenty-Five Minutes
- reactivity
The way measuring something changes it. People behave differently when watched, and this still holds when the person watching is themselves.
Introduced in week 4 · The Observer in the Room
- regression to the mean
Unusual measurements tend to be followed by ordinary ones. Because people start new routines on their worst days, almost any routine appears to help.
Introduced in week 9 · Why You Felt Better Anyway
- sample size
How many people, or how many days, a finding rests on. A result from four people and a result from four thousand are not the same kind of fact, and most everyday advice does not tell you which one it is.
- withdrawal design
Measure, add the thing, then take it away again. If the effect arrives with the thing and leaves with it, the thing is the likely cause. It is the simplest honest experiment one person can run alone.
Introduced in week 5 · Take It Away, Put It Back