Looking at the data on slide 12 of your presentation, how did you control for potential confounding variables in this experimental setup?
