Anomalous transport in tokamaks is usually explained in terms of turbulent fluctuations. Predictions for the confinement properties of tokamaks would be placed on a more sound basis if the level of fluctuations and the resulting transport could be calculated. At present there is no clear agreement on whether the responsible fluctuations are electrostatic or magnetic. This paper discusses the existing evidence and arguments, both direct and circumstantial, that might help to resolve this question.