The discovery of large, millennial-scale oscillations (1--10 kyr period) in climate records from ice cores and marine sediments in the North Atlantic has motivated the search to understand their cause and evaluate the geographic extent of this variability. One uncertainty in comparing marine geologic records of millennial-scale variability is the extent to which mixing of sediment by organisms at the seafloor (bioturbation) may attenuate millennial-scale events in the stratigraphic record. Evidence of attenuation of millennial variability can be found in cores with sedimentation rates in the range of 5--15 cm/kyr. Modeling the attenuation of a hypothetical input time series similar to GISP2 using a moderate level of mixing provides estimates of attenuation as a function of sedimentation rate. The amplitude of a 4 kyr duration event is reduced by 50% for sedimentation rates ~10 cm/kyr and reduced by 20% for sedimentation rates of 20 cm/kyr. Minimal attenuation (<5%) is achieved only when the sedimentation rate exceeds 50 cm/kyr for a 4 kyr duration event and 70 cm/kyr for a 2 kyr duration event. The intensity of mixing plays a role in determining how much attenuation occurs. The amplitude of a 4 kyr duration event in a 10 cm/kyr core is reduced by 75% under strong mixing or reduced by 25% under weak mixing. This study indicates that the original amplitude of Dansgaard-Oeschger (DO) millennial-scale events observed during the last glacial may be twice the observed amplitude for cores with sedimentation rates 10--20 cm/kyr that have experienced moderate bioturbation. Regional comparisons of the amplitude of millennial-scale variability will require cores with weak mixing or sedimentation rates exceeding 50 cm/kyr in order to avoid biased estimates of the amplitude of millennial-scale (4 kyr duration) variability. ¿ 2001 American Geophysical Union