[p1necone]

This is a very important thing to consider when interpreting this statistic.

"Heart attack rates go up 24% after daylight savings changes" is not the same thing as "There are 24% more heart attacks due to daylight savings". You can't really know the weight of magnitude vs distribution without actually stopping daylight savings.

[clircle]

or by conducting some causal inference with a load of identification assumptions