"how do you deal with the problem that it is hard to know what good science is?"
In what way is it hard to know what good science is? Simply apply Sagan's baloney detection toolkit, or ask the following simple questions:

Is the hypothesis falsifiable?

Was there a placebo/control group used?

Are the experimental results published and available for open scrutiny and experimental reproduction?

The biases of the scientist in terms of his personal beliefs and his funding are irrelevant, because if the scientific method is followed, the results are valid science in any case. Unfortunately, many people ignore key components of the scientific method and then attempt to pass off their results as science, which weakens the credibility of real scientists in the popular view.