Of course there are uncontrolled variables, but there are those in every experiment. With large enough samples, those uncontrolled variables become less significant. That's how statistical sampling works--if you expected everyone to be exactly the same, then you would only have to ask one person.

I understand that this is a personal outlook, but I feel like it is trying to make a more general point than "This is my story".

BTW: Would you consider economics a "soft science"? I study economics with some background with econometrics (statistics, essentially, where the entire focus is on controlling for unknowns)