Week 4 Nap Lab: It is not the sandwich

A Nap Lab that tests the lunch explanation directly, and checks your week 1 prediction against two weeks of your own recorded data

15 March 2027

This is the first lab where you test a claim most of the room arrived believing. It is also the first time your own recorded data is old enough to argue with you, which is the more useful half of the session.

Before the lab#

• You will be assigned a meal condition, ordinary lunch or deferred lunch, and you will not get to pick. If deferring a meal is not sensible for you — and for some of you it is not — tell us in advance and you will be assigned to the ordinary condition. There is no penalty and no explanation required.
• Bring the sealed trough estimate from week 1. If you have edited it, say so. An honest “I changed it in week 2” is a usable data point; a quietly improved prediction is not.
• Keep caffeine ordinary. We moved that variable last week and we are not moving two at once.

In the lab#

Repeated measures across the afternoon. Karolinska and reaction-time median at intervals through the session, both conditions running side by side in the same room. Nobody naps: this lab is about locating the trough, not treating it.

Read the shape, not the endpoints. The interesting thing is where each person’s curve bottoms out, not how low it goes. Two people can dip equally hard three hours apart, and a policy built for the average of them serves neither — which is week 12’s problem, seen early.

What would falsify it. Before we look at the pooled numbers, the room commits out loud to what the deferred-lunch group’s curve should look like if lunch is the cause. Writing that down first is the whole method: a prediction made after seeing the data is not a prediction, and this is the cheapest place in the semester to feel the difference.

Your prediction against your data. Finally, you compare your week 1 estimate with what the last fortnight of your Nap Log actually shows. Expect this to be uncomfortable for roughly half the room. Being wrong here costs nothing and is worth a paragraph in your analysis.

Afterwards#

Your afternoon series goes into the Nap Log, which now has enough in it to analyse rather than merely collect. Start looking at your own data as data — the analysis is the graded half, and a fortnight of careful entries with nothing said about them is worth very little.

If your trough is not where the lecture said it would be, do not smooth it out. A clean minority result, properly recorded, is the most valuable thing an individual student contributes to this course.