The electric field strength and shear rate dependence of the apparent shear viscosity of electrorheological (ER) suspensions can often be represented by a function of only the Mason number. A Mason number defined for magnetorheological (MR) suspensions by direct substitution of magnetostatic variables for electrostatic variables does not produce a similar collapse of shear viscosity data for MR suspensions. We show that a Mason number defined in terms of the suspension magnetization can be employed to produce a collapse of experimental data at various magnetic field strengths and shear rates. As for ER suspensions, this Mason number can be calculated from experimentally measured quantities.