Another day, another paper wrongly using SEM instead of SD. I‘m going to start editing the PDF and drawing longer error bars going forward.