Prevalence of hypovitaminosis D from previous NHANES reports did not account for assay changes and drifts over time. Thus, published NHANES reports on vitamin D status for the U.S. population were likely either over- or underestimated. We investigated changes in vitamin D status in the U.S. using assay-adjusted serum 25-hydroxyvitamin D [25(OH)D] data from NHANES, 1988-1994 (n = 18,641) and three cycles of NHANES, 2001-2006 (n = 23,424). Changes in geometric mean serum 25(OH)D and prevalence estimates for various serum 25(OH)D cut points (<25, <30, <40, <50, and <75 nmol/L) were determined. From 1988-1994 to 2001-2006, geometric mean serum 25(OH)D significantly decreased by 9% in all participants, 12% in men, 14% in blacks, 16% in 12- to 15-y-old adolescents, 16% in 20- to 30-y-old adults, 13% in nonsupplement users, and 12% in persons with BMI >80th percentile (P < 0.001). From 1988-1994 to 2001-2006, prevalence of serum 25(OH)D <30 nmol/L increased from 5 to 10% in all participants, from 3 to 8% in men, from 22 to 38% in blacks, from 3 to 8% in 12- to 15-y-old adolescents, from 5 to 12% in 20- to 30-y-old adults, from 6 to 14% in nonsupplement users, and from 8 to 17% in persons with BMI >80th percentile (P < 0.001). Previous NHANES reports overestimated the increase in prevalence of hypovitaminosis D. The recent decline in vitamin D status in the U.S. is more likely due to increased prevalence of obesity and other lifestyle changes but not to changes in milk consumption.