I performed a reasonably rigorous experiment for my AP statistics class in high school - I tested a ton of different sodas and their diet equivalents across my

I performed a reasonably rigorous experiment for my AP statistics class in high school - I tested a ton of different sodas and their diet equivalents across my classmates, and of the 8 sodas tested almost all had a statistically significant bias in favor of the non-diet version. For coca cola, I also tested the non-caffeinated versions - the non caffeinated regular Coke was the top performer. The only exception to this was diet SunKist - it handily beat regular SunKist, by a large margin.