Using science is about iteratively getting closer and closer to the truth, it is not about absolutes - since conclusions can become outdated by actually being wrong, or get supplanted by better conclusions. It's about delivering conclusions with high confidence, not absolute facts. In a vacuum, science itself is a perfect, unfailing tool. But since human beings use it, it's used imperfectly. And, when used correctly, the process is driven by data, not ideas. Science itself (meaning the tool) shouldn't be questioned, but the conclusions people reach and the way it is used should be/are questioned. What conclusions we trust should be dependent on factors like how often that conclusion is reproduced, how thorough the methodology is, and how many limitations were taken into account. The average person should have some understanding of this, so that they don't blindly believe in things and so that they aren't fooled into thinking there's a scientific consensus on a matter that does not have a consensus.