Well, the basic problem is that real science is technical, detailed, and tends to use a lot of math. Most people don't have the time or energy to learn the math, let alone all the details and technicalities. In fact, going through that learning process is a big part of the training to be a real scientist. And, unfortunately, that training only works for one field. If you go through the training to be a physicist, you won't have the time to go through the training to be a biologist (even though there is some overlap).
So, we have a specialized population of people that have gone through this training. But policy decisions have to be made by people who have *not* gone through that training. And the general public, while 'interested', wants to avoid all that training because they have other things to think about.
So, we have 'science writers' who have often (not always) gone through some of the basic training, but don't have the details and the technicalities (or, often even the math). They are, in essence, journalists that attempt to communicate ideas to the public. The problem is that the journalists, themselves, don't *really* understand these ideas. They are communicating a poorly understood set of ideas to people who want to know even less: they want the end result, not the details.
And, of course, to make the story interesting, these journalists have to 'play up the controversy' and make things seem more unusual than they actually are. They are selling the *story*, not the science.
And, of course, there are also many 'science writers' who really don't understand much about what they are writing. These give poor descriptions and cater to those who don't want to believe the science. SO, politics enters in and we get those who want things one way when the science says they are another. And, for the general public, who has little training, figuring our who the 'experts' are is, in and of itself, a job that requires more thought than they want to do. So they 'go with their gut' and the result is bad policy all around.