The level paradox points to the mismatch between the multilevel nature of e-collaboration phenomena and the fact that most research on it is single-level. Before exposing this contradiction, it is important to define the terms that are going to guide the discussion. When we build theories we should specify which entities need to be considered and are involved in the explanation of the phenomenon of interest (Whetten, 1989). Such entities to which research wishes to generalize are the focal units or level of theory (Hitt, Beamish, Jackson, & Mathieu, 2007; Rousseau, 1985). According to Rousseau (1985), two types of levels exist for research on a focal unit: the level of measurement and the level of analysis. The level of measurement represents the unit to which the data are directly attached (Hitt et al., 2007; Rousseau, 1985). In contrast, the level of analysis “is the unit to which the data are assigned for hypothesis testing and statistical analysis” (Rousseau, 1985, p. 4).