A study was undertaken to develop the use of survivorship methods in characterizing the magnitude of culling and in testing for differences in culling among dairy herds. A prospective observational study was conducted on nine herds representing 19482 cows. The cull rate derived from survivorship data was estimated as the weighted slope of the cumulative proportion of cows remaining in a herd after first parturition, where cumulative proportion was computed using a cohort life-table with intervals of 1 month. Cull rates ranged from 9.0-13.8% per 12 months of age, compared with culling density rates of 22.2-39.7 culled per 100 cow-years. Comparison of ranks of density rates, weighted-slope rates and median ages at culling among the herds illustrated that the measures were not interchangeable. An advantage of a survivorship approach to measuring culling was illustrated by the use of the Cox proportional hazards model that tested for differences in cull rates among herds. Results suggested that variation in culling among herds during the first lactation, and particularly during the first part of the first lactation, may be an important consideration in future studies of optimal culling practices. In addition to being able to compare culling among herds, a survivorship approach to measuring culling provides an estimate of the rate of removal that is not biased by age, in contrast to currently used methods.