Problems of scientific cosmology only rarely occur in the works of Karl Popper. Nevertheless, it was a subject that interested him and which he occasionally commented on. What is more important, his general claim of falsifiability as a criterion that demarcates science from non-science has played a significant role in periods of the development of modern physical cosmology. The paper examines the historical contexts of the interaction between cosmology and Popperian philosophy of science. Apart from covering Popper's inspiration from Einstein and his views on questions of cosmology, it focuses on the impact of his thoughts in two periods of controversy of modern cosmology, the one related to the steady state theory and the other to the recent multiverse proposal. It turns out that the impact has been considerable, and continues to be so, but also that the versions of Popperian methodology discussed by cosmologists are sometimes far from what Popper actually thought and wrote.

This article examines in detail a remarkable but short-lived cosmological theory of 1959. The theory depended crucially on a hypothesis that could be, and was, tested in the laboratory. I use the case to discuss the nature of testing in cosmology and to argue against ideas about astronomy suggested by Ian Hacking. The case of the electrical universe exemplifies how disagreements can be settled by good experiments and also how experiments of wide-ranging theoretical significance need not be biased by either theory or the experimentalists’ predisposed ideas. I argue that the episode supports Allan Franklins “evidence model” and disputes accounts based on the sociology of scientific knowledge.