Week 7 Lab: You Are the Placebo

19 April 20272 min readhigh risk

This week

Every experiment in this course has the same unfixable flaw, and this week we look straight at it.

Before you run this

When to stop
Stop and go back to your usual intake if you get a headache that will not shift, or if you cannot work. Caffeine withdrawal headaches are common, ordinary and unpleasant, and they are a reason to stop rather than a stage to push through.
What this is not
There is no reduction target and no abstinence phase. Nobody is quitting anything. If you do not drink caffeine, do not start for this week, and if you are pregnant, taking medication affected by caffeine, or prone to migraine, take the alternative.
If this one isn't for you
Blind a condition that carries no physiological load at all: two unlabelled playlists, two desk lamps at different temperatures, two teas that taste alike. The interesting part is the concealment, not the chemical.

What you bring to the lab

The lesson below is how you get there. The lab is where you show it.

  • you can describe how you concealed the condition from yourself
  • you record your guess about which condition you were in, before you check
  • you report how often your guess was right

Here is the flaw, stated plainly: you know what you did.

Blinding means not knowing which condition you are in, so that what you expect cannot colour what you record. It is the single most important safeguard in experimental work, and as a person experimenting on yourself you essentially cannot have it. You poured the coffee. You set the timer. You know.

That makes you the placebo, not a victim of one. The effect is real, it is yours, and it is not caused by the thing you think caused it.

An attempt anyway

This week you will try to conceal a condition from yourself, and mostly fail interestingly.

The setup: someone else prepares your drink from two possibilities, in identical cups, on a schedule you do not see. You record your alertness as usual. And critically, you also record your guess about which one you got, before you are told.

That guess is the real measurement. If you can reliably tell, you were not blinded and your results carry your expectations. If you cannot, you have bought yourself something rare and worth having.

The oldest version of this

In 1784 Louis XVI appointed a commission (Franklin was on it) to investigate mesmerism, then fashionable in Paris and said to work by manipulating an invisible fluid. The commission’s method was to blindfold subjects, so that they could not tell when the treatment was being applied.

The effects followed the subjects’ beliefs, not the treatment. People convulsed when told they were being mesmerised and nothing was happening; nothing happened when they were mesmerised and not told.

It is, as far as anyone can tell, the first blinded experiment. It was designed by people who had no statistics, no controls in the modern sense, and no mathematics of chance, and they got the essential move right on the first attempt, because the essential move is not mathematical. It is the decision to stop trusting yourself.

Living with it

You will not solve this. Every result you report for the rest of the semester carries some unknown amount of your own expectation baked into it, and the honest response is to say so in the report rather than to pretend the design was cleaner than it was.

Evidence on file

114 trials not blinded

Across trials that randomised patients to a placebo or to nothing at all, placebo had no significant effect on binary outcomes. The exception was self-reported pain. Note that placebo against nothing is the one comparison nobody can blind.

Hrobjartsson A, Gotzsche PC, Is the Placebo Powerless?, New England Journal of Medicine 2001;344(21):1594-1602
History, not measured

The commission appointed by Louis XVI in 1784, with Benjamin Franklin among its members, tested mesmerism by blindfolding subjects so they could not tell when they were being treated. The reported effects followed what subjects believed rather than where the treatment was applied.

Report of the Royal Commission on Animal Magnetism, Paris 1784 (Franklin, Lavoisier, Bailly and others)

Teaching team