This suggests that if someone who normally goes to bed at 1 a.m. goes to bed at midnight instead and sleeps the same duration, they could cut their risk by 23%; if they go to bed at 11 p.m., they could cut it by about 40%.
Someone didn't get the memo about correlation/Causation. People raising early are likely to be less depress, does not equate with "raise early and you cut the depression rate by 23%"
This sound more as an explanation