The molecular kinetic theory (MKT) of dynamic wetting, first proposed nearly 50 years ago, has since been refined to account explicitly for the effects of viscosity and solid-liquid interactions. The MKT asserts that the systematic deviation of the dynamic contact angle from its equilibrium value quantitatively reflects local energy dissipation (friction) at the moving contact line as it traverses sites of solid-liquid interaction. Specifically, it predicts that the coefficient of contact-line friction zeta will be proportional to the viscosity of the liquid n(L) and exponentially dependent upon the strength of solid-liquid interactions as measured by the equilibrium work of adhesion Wa(0). Here, we analyze a very large set of dynamic wetting data drawn from more than 20 publications and representative of a very wide range of systems, from molecular-dynamics-simulated Lenard-Jones liquids and substrates, through conventional liquids and solids, to molten glasses and liquid metals on refractory solids. The combined set spans 9 decades of viscosity and 11 decades of contact-line friction. Our analysis confirms the predicted dependence of zeta upon n(L) and Wa(0), although the data are scattered. In particular, a plot of ln(zeta/n(L)) versus Wa(0)/n (i.e., the work of adhesion per solid-liquid interaction site) is broadly linear, with 85% of the data falling within a triangular envelope defined by Wa(0) and 0.25 Wa(0). Various reasons for this divergence are explored, and a semi-empirical approach is proposed to predict zeta. We suggest that the broad agreement between the MKT and such a wide range of data is strong evidence that the local microscopic contact angle is directly dependent upon the velocity of the contact line.