Last November, a trio of scientists published a Nature piece in which they detailed the top twenty "tips for interpreting scientific claims." The article might come off as a bit infantilizing, but there's also some good stuff on the list: "regression to the mean can mislead," "seek replication, not pseudoreplication," &c, &c.
But I missed the response! In December, Chris Tyler—the director of the Parliamentary Office of Science and Technology—offered a rebuttal of sorts. The resulting list ("Top 20 things scientists need to know about policymaking") is worth a glance for any researcher who has wrung his hands over the sorry state of "science policy."
When policy makers talk about "science policy", they are usually talking about policies for things like research funding, universities and innovation policy. Researchers additionally use "science policy" to talk about the use of research evidence to help deliver better policies in a wide range of areas. I find it helps to distinguish between "policy for science" on the one hand, and "science for policy" on the other.