Existing methods for inferring the ventilation age of water masses in the ocean using radiocarbon data neglect the effects of diffusive mixing. in the presence of varying atmospheric Delta(14)C, this neglect produces spurious time dependence in the estimated ventilation ages. To correct this deficiency we propose a new method for estimating the ventilation age from sediment core radiocarbon data. The new method is formulated in terms of parameterized age distributions that account for the effects of advective and diffusive transport in the ocean. When applied to simulated radiocarbon data from an OGCM, the method is able to closely reproduce the modeled ventilation age, whereas other methods are not. We also applied the method to sediment-core radiocarbon data from the deep subarctic northeast Pacific. We estimated the model parameters using a Bayesian approach that allows for a careful quantification of the uncertainty in the inferred ventilation age. Results from the sediment-core analysis show larger ventilation ages during the last glacial maximum compared to the Holocene, although uncertainty on individual age estimates is high. Large excursions in ventilation ages for one of the cores may also indicate a change in the reservoir age of North Pacific surface waters during the last deglaciation.