[edanm]

This has been really questioned, and I think largely debunked.

The "effect" shown in the study that demonstrated this was probably caused by batching of certain kinds of defendants/cases by time of day, e.g. put the longest (most difficult) cases first, then have a bunch of easy no decisions right before the break. (That's not what really happens, it's just an example, I don't remember what really happens).