For an experiment related to the Eindhoven Marathon 2019, we added preliminary support for sleep logging.
As theoretical background, we looked at shortPSQI, which identifies some key components:
- Component 1: sleep latency
- Component 2: sleep duration
- Component 3: sleep efficiency
- Component 4: sleep disturbances
Our idea was to focus further by:
- omitting everything related to sleep inefficiencies (the idea being that for a sporter we could focus on the actual sleep, together with the other data we collect already). Specifically, we assumed that latency and efficiency could be ignored for our study.
- ignoring sleep disturbances (as they should have been reflected largely in sleep duration, assuming that episodes of waking up during night are taken away from the duration).
Essentially, that left us only with shortPSQI component 2 (sleep duration). The shortPSQI instrument, considers the following ranges in relation to it’s question 4 (“During the past month, how many actual hours of sleep did you get at night?”): (>7 (0), 6-7 (1), 5-6 (2), <5 (3)).
Then, we did go beyond the scope of shortPSQI by adding a component related to sleep regularity (which has been shown to be an important factor independent from sleep duration).
Finally, we reframed everything to the week-level.
The proposed questions and corresponding answering scale then became:
- During the last week, how many actual hours of sleep did you get on average per night?
- less than 5 hours: -2
- between 5 and 6 hours: -1
- between 6 and 7 hours: +1
- more than 7 hours: +2
- During the last week, how many half hours did you deviate from that average?
- more than three half hours: -2
- between two and three half hours: -1
- between one and two half hours: +1
- less than one half hour: +2