Most people are wrong about most things. This is especially true of the people who are brought to your attention by newspapers and television. It doesn’t matter how smart they are, or how well-read, or how thoroughly educated. There aren’t very many fields of endeavour where you can get ahead on the sheer strength of being right. Our expert classes succeed instead by cultivating the correct allies, publishing the right papers in the right journals, working on the right problems, winning the right grant funding, and making the right friends. People who enjoy these trivialities are precisely the people for whom being right is not a priority.
Above all, experts prefer to work within and propagate safe, consensus positions. This is because they have primarily careerist goals, which are best pursued secure from the criticism of colleagues. Being wrong is not nearly so important as seeming wrong, which can cost you promotion. Once you realise that experts are little more than consensus-establishing and -propagating professionals, statements about what the science says or what the literature shows acquire a totally new meaning.
Forget, then, about expert opinion. There is no substitute for doing your own research. In everything that matters to you, you must consider the actual theories that are presented to you for yourself. And, particularly in areas of limited evidence, you’ll be less interested in which theories are wrong (though that matters too), than in the subtler problem, of which theories are more or less probable than the alternatives.
Most of the theories that are put about, are not really theories at all. They are, instead, arguments, designed to justify or advocate for specific policies. Arguments are not genuine attempts to understand anything; they are attempts to convince other people to think in a certain way.
People assemble arguments like they would a house. They develop a program (the plan), collect evidence in favour of this program (the materials), and finally they present their program with all the evidence adduced in neat footnotes (the construction). This approach is reasonable enough, if all you want to do is persuade, but if you want to understand how a given model of reality fares against others, it is the wrong way.