While there is politics in everything to some degree, peer reviewed scientific journals try to remove as much as possible. With a good editor the analyses include complete review of the statistical process, including mechanisms to detect bias.
I have read many articles in many journals, even written a few, and to try to imply that these type of articles are not superior to the crap we hear and see on television, or in lay media, is patently ridiculous.
With correct training, you can read the articles and see these processes. In addition, many of these articles are presented at professional meetings where they come under the scrutiny of the best minds in the areas they are writing about, including those doing similar research and who know intimately how to dig out the true facts and bias. Some of these other professionals are actually in some degree of competition to arrive at solutions, so they are even more critical of the works.
