Then as second step, you can evaluate the data yourself, or even question it. See where you can find the source, or see other data. This step is more checking if the statistical evaluations are correct, or if there is p-hacking or any other problem. It must also follow basic common science. And if not, this should be explained.
What I usually find (>90% of the articles that I find interesting) is that the reports are conflicting and even the data is conflicting. The evidence/science is often incomplete and filled in with unfounded theories. And knowing this, it reduces the amount of information that is important in the articles.
So technically logical thinking is a way to compress all the data, without loss of information.