How things change in science
A friend draws my attention to an authoritative 1986 paper on how things change in science, addressing the theories of scientific change of the 1960s and 1970s (Kuhn, Feyerabend, Lakatos, Laudan, and others):
... they all agree, contrary to Popper, that scientific theories tend to persevere despite mere empirical problems, and that theories are underdetermined by data, and that metaphysical, theological, or other nonscientific factors play a role in theory assessment, and that theory evaluation is at least partly relative to the competition available, and that paradigms/research programs/... are never abandoned without a replacement on hand, and lots of other interesting stuff. In fact 17 items of agreement are listed, though there is some near-overlap.
My friend thinks that the agreement of theorists should be taken as a reliable claim about how science works (whether it works well or not). He is particularly concerned that lay people not go away with the idea that Karl Popper's "falsifiability" counts for much in practice. He notes that both Lakatos and Feyerabend say that any grand idea in science can be made to sound right if enough clever scientists work on it (p. 26).