Hi @jeffb , I wondered if you might show up. 
We are well aware of the difference, and there are numerous textbooks on the subject of âexperimental designâ. The difference is usually described in terms of observational and prospective data/experiments. Prospective data is the Gold Standard when you can get it, and thatâs why we do clinical trials for new medical treatments (and new treatment donât have any history to observe). You CAN do experiment using historical/observational data too, but it looks different.
The discovery of Tiktaalik, is one such example. Iâve think Iâve shown you this before so I we be brief: Neil Shubin and his team noted a gap is the fossil record between fish and land walking tetrapods (four-legged critters). They hypothesized that such a creature should have existed, and that fossils showing transitional features might be found (prediction). They consulted geological record (independent data) for locations where such fossils could exist (ancient tidal marches), and organized an expedition to go fossil hunting. He found the fossil Tiktaalik we now know as Tiktaalik, confirming the prediction.
This is an example of abductive inference in Paleontology. The experiment is repeatable, looking for similar fossils in other locations. So no, no one is denying a difference, the objection is to the claim that you cannot do real science with historical data.
