Maybe they're exposed to light at night because they're awake at night more often, possibly shift workers, which we already know is unhealthy. I doubt just having light on is causing the effect.
Yes they tracked hours of light exposure (above some threshold? I don't see that they say.) and found this result in the 90-100th percentile. So almost certainly night shift workers.
They grouped the population into percentiles with 0-50% having quite low exposure (0.62 lux median, range of 0-1.21) vs the 90%-100% percentile being 105 lux (range of 48.3-6400). You can compare the daylight light exposure to the nighttime and basically see what a shift worker would be.
But they already deconfounded for shift workers, so that's irrelevant. And they also showed the amount of light exposure for both night and day.
