I have a different theory as to the disconnect between media and actual science. Basically, it stems from two ideas. One, in general people do not understand that correlation is not causation. Two, that what is true in one circumstance may not be true in another.
As an example of the first, we'll take cancer research. Scientist X does research and finds that expression of Gene Y is elevated in certain types of cancer, and publishes a paper about it. Note that there's a lot of qualification in just that simple sentence. As to my first point, correlation is not causation: Gene Y
may be involved in a certain type of cancer, but by no means
causes it. It could be that both cancer and Gene Y elevation are a response to the actual cause. It could also be that elevation of Gene Y is a response to cancer. Finally, it could be a coincidence, though this is fairly unlikely given a rigorous methodology including multiple confirmatory experiments (not always the case).
My favorite example of the second instance is from roughly a decade or two ago. Scientists doing research into obesity discovered in mice that there was a defect in a hormone that triggered the hunger response. Giving them the correct version of the hormone made them lose weight. A lot of people thought that this might be a cause for obesity. It was picked up by most media outlets, and a lot of people got rightfully excited. Except that in humans, the problem isn't with the hormone, but with a faulty receptor, which is much more difficult to treat. The thing is, mice and humans are similar, but not necessarily the same.
You see this a lot as well with potential treatments for illnesses. Just because something works in carefully controlled lab experiments doesn't mean it works in humans. That's why we have multiple stages of testing, and why a significant number of all potential treatments don't make it past the first stage. As an aside, that's a big part of why biochemical research is the way it is.
As an aside, OB-positive mice are
so cute.
The difference is that the scientific community has standards that they enforce. When a scientist publishes a paper and other scientists read it, they don't just go, "That sounds good, I like it! I'll go with it." They try to repeat the experiment or they research their own data and see if it fits with the theory. Or they even create their own experiment to prove the same hypothesis.
Not just that, but there are very real consequences for things like plagiarism because of the same system. Scientists who plagiarize typically don't stay scientists, because they face societal repercussions from the scientific community. The most severe is perhaps that your data, and by extension you, can no longer be trusted, and basically get blacklisted. In that aspect, it's self-policing.
Evolution and geology and several other topics are what we often call "historical sciences" because it is difficult to develop predictable tests for hypotheses, and test them today.
I can not, as a scientist, agree with that statement at all.
The process of evolution can be (and is) tested in many ways. Furthermore, directed evolution is a scientific method in which evolution is harnessed to create creatures (usually bacteria, quickest lifecycle) more suited for an environment.
Geology (and by this I presume you mean geological phenomena) can be tested in numerous ways - indirectly through modeling, directly through many different techniques.
You can take any science and apply it to historical data. That doesn't make those sciences outside the realm of modern testing.