I would heavily recommend "doing the science". It's a lot of fun, and it has one of two outcomes: You confirm what you think is true, in which case your argument has a LOT more weight (as so few people will "do science"), or you disprove your idea, which means you get to find a way to make your design even better. Win-win!

It really is a lot of fun, too. We had a university assignment to investigate a particular public-facing system, and while most of my peers turned to Google I hit the streets with a clipboard. I had a really great afternoon, learnt a huge amount about the system, and discovered data that no one else did.