If we were to claim that evening illuminance plays no role in sleep, we would certainly have gained some attention. In fact, our recently published meta-analysis shows that there is no direct correlation between illuminance and sleep.
It should be noted that illuminance is measured in lux, which in turn is based on the SI unit candela. The origins of this unit can be traced back to the light produced by a pure spermaceti candle, which weighs 76 grams and burns at 7.8 grams per hour. (Spermaceti was made from a substance derived from heads of the sperm whale). That this photometric unit cannot directly be related to human sleep quickly becomes understandable. But the situation is different if we choose a unit that takes into account the non-visual effects of light. Then we do find evidence of a link between light exposure and sleep. Based on the melanopic EDI, our meta-analysis shows that evening light affects sleep latency, sleep efficiency, and deep sleep. The estimated melanopic EDI in the range of 100-1000 lx showed a clear dose-response relationship for sleep latency and sleep efficiency. The melanopic EDI may therefore be a robust predictor of non-visual responses to sleep. After all, we never would have claimed that light per se has no effect.
Link to the publication: https://journals.sagepub.com/doi/full/10.1177/14771535221078765