In this paper we estimate relative geomagnetic field paleointensity between 71 and 12 kyrs BP as recorded in three deep sea sediment cores from the Blake Outer Ridge, western North Atlantic Ocean. Paleointensities were estimated by normalizing sediment natural remanent magnetization separately to (1) magnetic susceptibility, (2) anhysteretic remanent magnetization and (3) saturation isothermal remanent magnetization (SIRM). In one core, paleointensities were estimated within short time windows which display uniform sediment magnetic characteristics, and offsets between windows were removed to minimize environmental biases. We find that most features of our records are preserved regardless of normalizer choice, but we chose SIRM as the best normalizer for the final paleointensity estimates. Our three records preserve and agree upon a number of short-duration (~103 years) paleointensity features even though the cores are separated by almost 250 km. We conclude that these are real geomagnetic field signals of at least local extent. We also identify a number of differences between our records which must be artifacts of sediment remanence acquisition or paleointensity normalization. Such artifacts occur as either (1) baseline shifts between time intervals with slightly different sediment magnetic characteristics or (2) differences in amplitude of short duration events. In spite of these environmental biases, the number and ages of relative paleointensity highs and lows are preserved. Thus sediment paleointensity estimates may be used locally for high-resolution chronostratigraphic correlation. Correlation of our paleointensity records with other records from the same extended region (North Atlantic Ocean-western Europe) indicates that major paleointensity features appear in all regional records. However, differences in the ages and disagreement in magnitude and number of individual features call into doubt the use of these relative paleointensity records for high-resolution chronostratigraphic correlation on a broader regional scale or for quantitative estimation of past geomagnetic field variability. ¿ 1998 American Geophysical Union