3 ms·
The results are interesting. It seems like there is a lot of clustering going on in the regression data. It could be worth investigating if instead of a linear
by diatone 5y ago
The results are interesting. It seems like there is a lot of clustering going on in the regression data. It could be worth investigating if instead of a linear relationship between sleep quality and consistency with academic performance, there is a logarithmic (or some other superlinear) relationship instead. This would help confirm the intuitive notion that there are diminishing returns to optimising one's sleep.
- contravariant 5y agoYou'll want to have a stronger link before you can worry about what kind of relation it is exactly. And if you ignore the people with scores <60% you lose most of the correlation entirely, which to me suggest there's something causing both bad sleep and bad academic performance rather than there being some continuous relation. In fact the distribution looks bimodal which suggests there's something splitting the population in 2.
- alisonkisk 5y agoBinodality appears in feedback loops: bad sleep makes good sleep harder to get.