Several potential sources of variation in steric exclusion chromatography (SEC) are discussed: flow rate stability of the pump, column temperature, data processing, and the chromatography itself. Many of these variation sources may be minimized by careful attention to analytical techniques. Flow-rate stability, however, is a function of the chromatographic hardware used and is not as readily subject to control. Experimental techniques are described whereby relative standard deviations below 0.3% in the routine determinations of molecular-weight averages may be achieved on an optimized chromatographic system.