Week 10 Nap Lab: The device and the night

A Nap Lab comparing what your tracker said against what you experienced, and against the crude instruments this course has used all semester

10 May 2027

You have ten weeks of Karolinska readings and reaction-time medians taken with instruments that look primitive next to a wrist tracker. This lab is where the two meet, and the outcome is usually less flattering to the tracker than people expect.

Before the lab#

• Bring device output for one night if you have a device, and your honest written impression of that same night. Write the impression down before looking at the app, which is the one instruction people find hardest.
• No device? You are needed. You will work with the supplied anonymised datasets, and you will be the group asking the questions that device owners have stopped asking.

In the lab#

Two accounts of one night, twenty-five minutes. Device against recollection. Where do they agree, where do they diverge, and by how much? Total sleep time usually agrees reasonably. Stage breakdown and “time awake” frequently do not, and someone in the room is always told they slept well on a night they remember as terrible.

Device against this course’s instruments, twenty-five minutes. Pull the 20-minute nap entries from your Nap Log and see what the device made of them, if anything. Many devices will not have registered them at all. A nap that the instrument cannot see is the cleanest demonstration of the lecture’s point, and you have ten weeks of them.

Following one claim, twenty minutes. In groups, take a device’s own marketing or support material and find a claim its sensors cannot support. Be precise about the gap — “it cannot see brain activity, so a REM figure is inferred from movement and heart rate” is an argument; “it’s not accurate” is not.

Which instrument for which question. We close by sorting questions rather than devices. For did I sleep at all, the tracker is fine. For how do I feel and perform now, the nine-point scale you have been complaining about since week 1 is better, and it is better because it measures the thing you actually care about instead of inferring a proxy for it.

Afterwards#

Add the comparison to your Nap Log analysis if you are still revising it, and carry the habit into the Nap Policy Proposal: a proposal that promises to measure its own effect has to say with what, and “we’ll give everyone a tracker” is now an answer you can see through.