My test presented the non-BeeReader passage before the BeeReader version, so there's definitely some randomization going on. Perhaps the aggregate A/B numbers

My test presented the non-BeeReader passage before the BeeReader version, so there's definitely some randomization going on. Perhaps the aggregate A/B numbers make a more compelling case for using BeeReader? I wonder if the color combo choice has any affect on the speed/comprehension of the text.