It is now theoretically possible to obtain absolute accurate values of number-average molecular weight of complex polymers (e.g., branched polymers or copolymers) using size exclusion chromatography (SEC) with only a detector that measures the difference between the eluting polymer solution viscosity and the viscosity of the pure mobile phase (a differential viscometer [DV] detector). However, both precision and accuracy of these "DV M(n)BAR" values are of concern. In this work, the precision of NBS 706 polystyrene was found to be two to three times worse for the DV M(n)BAR than for the conventionally calculated M(n)BAR. Also, regarding accuracy, the DV M(n)BAR values were affected by the location of the universal calibration curve along the retention volume axis (a problem intimately associated with the problem of specifying the correct interdetector volume), the sensitivity of the DV detector to low molecular weights present in the sample, and axial dispersion. Each of these sources of error are examined in turn and two methods of calculating M(n)BAR values are proposed.