This paper explores the methodological approaches most commonly adopted in the scholarly literature on Massive Open Online Courses (MOOCs), published during the period January 2008-May 2014. In order to identify trends, gaps and criticalities related to the methodological approaches of this emerging field of research, we analyzed 60 papers selected across four relevant scientific databases plus one journal in the sector of eLearning that published a special issue on this topic. The emerging picture is that of a research field in its infancy, heavily relying on theoretical research and case studies, which is just beginning to identify suitable methods to deal with large cohorts of learners, very large amounts of data and new ways of learning. The state-of-the-art is also quite fragmentary, due to the different epistemological and ontological conceptions of the authors of the papers about the nature of the issues faced and the way they should be studied. In this paper, we compare the problems related to the definition of the methodological approaches in this research field with the Greek myth of Proteus, the elusive, multiform divinity whose wisdom would only be revealed to those capable of catching him. We therefore advocate the need for catching Proteus, that is, better identifying and choosing the methodological approaches for MOOC research as a condition for its development.