Your last statement is contradictory to the rest. Observation without controlled experiment is not science, its speculation. Controlled experiment always comes into it, you aren't completing the scientific process until you systematically test that hypothesis against observation. Anyone can pull a hypothesis out of the air and make predictions, until its applied to observation the science isn't finished. That isn't to say we can't make predictions, and look for their evidence as you were saying. How did we know what evidence an asteroid impact leaves? We studied it, we may not have been in control of the experiment but we did observe it (or its results).
The author does make a good point, a lot of theories, studies, or things "proven" that hit media are many times not scientifically sound. They contain too many conclusions or predictions on too little (or poor) evidence.