Conceptual difficulties in the use of statistical inference in citation analysis

In this comment, I discuss the use of statistical inference in citation analysis. In a recent paper, Williams and Bornmann argue in favor of the use of statistical inference in citation analysis. I present a critical analysis of their arguments and of similar arguments provided elsewhere in the literature. My conclusion is that the use of statistical inference in citation analysis involves major conceptual difficulties and, consequently, that the usefulness of statistical inference in citation analysis is highly questionable.