A statistical model has been derived which can be used either to test the interpretation of a given magnetic stratigraphy or to make time estimates from the paleomagnetic data. The model is based on the presently accepted statistical properties of the earth's magnetic field and the stochastic nature of the paleomagnetic sampling process. Given knowledge of the number of paleomagnetic sites, the distribution of the sites, an estimate of the geologic time span sampled, and the mean length of the magnetic polarity zones, the expected success of the sampling program can be calculated and compared with its actual performance. In a number of test cases the theoretical results have agreed with realized results. The mean time span for polarity intervals in the late Neogene has been established from existing data t 120,000 years. Once the mean time span for polarity intervals has been established, estimates for stratigraphic time span can be made by means of the statistical model. Of three models studied, a time uniform sampling program is the most efficient for establishing a local magnetic polarity stratigraphy; the least efficient strategy is an exponential distribution of samples.