Week 5 Lab: Take It Away, Put It Back
This week
The simplest honest experiment one person can run alone, demonstrated on breakfast.
Before you run this
- When to stop
- Stop and return to eating normally if you feel faint, if you cannot concentrate by mid-morning, or if the removal phase stops feeling like an experiment and starts feeling like a rule. Four days is the cap, and shorter is fine.
- What this is not
- This is not a fasting exercise and there is no target. You are not trying to eat less, lose anything, or discover that breakfast is unnecessary. Anyone with a history of disordered eating, diabetes, or any condition affected by meal timing should take the alternative instead, and so should anyone who simply does not want to.
- If this one isn't for you
- Run the same design on something inert: your first coffee's timing, the radio on your commute, or a lamp. The method is the lesson, not the meal.
What you bring to the lab
The lesson below is how you get there. The lab is where you show it.
- your design has three phases and you can say what each one is for
- the removal phase has an end date fixed before you begin
- you report the return phase even when it is inconvenient
Three phases. Measure the thing as it is. Take something away. Put it back.
That is a withdrawal design, and it is the most useful experiment available to a person working alone. Its logic is simple: if the effect leaves when the thing leaves and returns when the thing returns, the thing is a plausible cause. Coincidence has to work much harder to produce that pattern twice.
Why the third phase is the one that matters#
Most people run two phases and stop. They remove breakfast, feel worse, conclude breakfast matters, and go back to eating it, without measuring the going back.
But the second phase is when you were also getting used to a new routine, and possibly sleeping badly, and possibly annoyed about the whole exercise. Putting the thing back is what separates breakfast from change. If the effect does not reverse, you learned something much more interesting than you were expecting.
So the return phase is assessable, and it is assessable specifically because it is the phase everyone skips.
About the self-experimenting tradition#
You are joining an old and slightly alarming lineage. In 1984 Barry Marshall could not persuade anyone that a bacterium caused stomach ulcers, so he drank a culture of it, became ill within days, and eventually shared a Nobel Prize.
The lesson usually drawn is about courage. The lesson worth drawing is about control: Marshall had one subject, no comparison, and no way of not knowing what he had done. His experiment was decisive because the effect was enormous and fast. Yours will not be, so you need the design to do work that drama cannot.
The limits of this particular week#
Read the boundary at the top of this page before you plan anything. Breakfast is used here because it is a convenient, reversible, everyday variable that almost everyone has an opinion about, not because this course has a view on whether you should eat it.
If the removal phase starts to feel like an achievement rather than a measurement, that is the signal to stop. Run the lamp instead. The method transfers perfectly and nobody needs to be hungry to learn it.
Evidence on file
- 309 people not blinded
Over sixteen weeks, being told to eat breakfast and being told to skip it produced no significant difference in weight loss, and it made no difference whether someone already ate breakfast.
Dhurandhar EJ et al., The effectiveness of breakfast recommendations on weight loss, American Journal of Clinical Nutrition 2014;100(2):507-513- History, not measured
Barry Marshall drank a culture of Helicobacter pylori in 1984 to test whether it caused gastritis, developed the illness within days, and shared the 2005 Nobel Prize in Physiology or Medicine for the work.
Marshall's own account; Nobel Prize in Physiology or Medicine 2005 (Marshall and Warren)
Teaching team
- Dr Wren Halliday, convenor
- Priya Ramanathan, tutor