Week 8 Lab: Everything Else That Happened
This week
You cannot randomise your own life, so this week we go looking for what else moved while we were not watching.
What you bring to the lab
The lesson below is how you get there. The lab is where you show it.
- you list at least ten things that changed during your week 5 experiment
- you rank them by how plausibly each could explain your result
- you identify one you could have measured and did not
A clinical trial assigns people to conditions at random, which is how it makes everything other than the treatment average out. You cannot do this. You get one life, in order, and it does whatever it was going to do.
Confounding is the consequence: something you were not tracking moved at the same time as the thing you were, and you have no way to tell which one produced the result.
The list#
Open your week 5 records. Write down everything that also changed during those days. Not the plausible things: everything.
Weather. A deadline. Who you ate with. Whether you were annoyed with someone. Daylight. A cold coming on. Money. A new season of something. Whether you were avoiding a piece of work by doing this course’s homework instead.
Ten minimum. Most people reach thirty and stop out of embarrassment.
Then rank them: for each, could it plausibly have produced the effect you measured? You will find two or three that could have done the entire job on their own, and the honest version of your week 5 conclusion has to name them.
Why we chose the evening#
The component this week is the social evening, because other people are the confounder nobody controls for and everybody has. Your day is not a closed system. It contains a housemate who came home in a mood, a friend who cancelled, a sibling who called. None of that appears in a sleep tracker, and all of it shows up in your data.
The move that rescues this#
You cannot randomise, but you can sometimes find a comparison that nature arranged for you.
In 1854 John Snow was trying to show that cholera travelled in water rather than air, against the obvious objection that poor neighbourhoods had both bad water and everything else bad. So he found streets in south London where two different water companies supplied neighbouring houses, one drawing from a polluted stretch of the Thames, one from upstream. Same streets, same poverty, same air, different water.
He had not assigned anybody to anything. He had found a place where the assignment had already been made, for commercial reasons, by someone with no interest in cholera.
That is the move to look for in your own life: two stretches of time, or two parts of a week, where the thing you care about differs and most of the rest does not. It is rarer than you want and better than nothing, and finding one is this week’s real exercise.
Evidence on file
- 16,608 people blinded
Observational studies had found that women taking hormone therapy had less coronary disease. The randomised trial found it caused more, along with more breast cancer and more clots, and was stopped early.
Rossouw JE et al., Risks and Benefits of Estrogen Plus Progestin in Healthy Postmenopausal Women, JAMA 2002;288(3):321-333- History, not measured
In 1854 John Snow compared cholera deaths between households supplied by two different water companies in the same south London streets, one drawing from a sewage-polluted stretch of the Thames and one from further upstream, which let him separate the water from the neighbourhood.
John Snow, On the Mode of Communication of Cholera, second edition, London 1855
Teaching team
- Tobias Veck, tutor