The number of new sexual partners per year (partner frequency) is a key factor in the spread of sexually transmitted diseases. Data from two Norwegian population-based surveys conducted in 1987 and 1992 were used to estimate recent (in the previous 3 years) and earlier partner frequency and to examine covariates affecting the distribution of partner frequency. Seventy-two percent of respondents reported having no new partners per year, and 2% reported having more than three new partners per year. Results from a Poisson regression model indicated that a low partner frequency was associated with being married or cohabiting, being female, greater age, and late sexual debut. Partner frequency was lower in 1992 than in 1987 (rate ratio = 0.8, 95% confidence interval 0.7-0.9). In comparison with earlier life, there was a large reduction in partner frequency for married/cohabiting individuals. In contrast, there was either no change or some increase in partner frequency for single persons.