Tuberculosis skin test conversions were examined among New York State prison employees. In late 1991 and early 1992, a total of 24,487 employees of the New York Department of Corrections were subjected to baseline and 1 year follow up tuberculin skin tests. The exposure variables used in the calculation of conversion rates were prison, or worksite, and job category. Prisons were classified as low or high risks for tuberculosis infection based on the number of inmates diagnosed with active tuberculosis in 1992. Most of the prison employees were white and male, with a mean age of 40 years. Of the 24,487 employees tested for tuberculosis infection, 466 conversions were recorded, resulting in an overall incidence rate of 1.9%. Conversion was most common among older, nonwhite prison workers who lived in New York City. The conversion incidence rate of employees increased with an increasing number of inmate tuberculosis cases at the worksite. The incidence rate increased from 1.4% at prisons with no known inmate cases of active tuberculosis to 2.6% at prisons with a high number of inmate tuberculosis cases. Compared to prisons with no known inmate cases, the risk of employee conversion was significantly elevated in prisons with low or high numbers of inmate tuberculosis cases, with adjusted odds ratios (ORs) of 1.66 and 2.20, respectively. The rate of employee conversion did not differ significantly according to job category in prisons with no known inmate tuberculosis cases. However, in prisons with known inmate cases of tuberculosis, the risk of tuberculosis conversion was significantly elevated among corrections officers and medical personnel, with adjusted ORs of 1.63 and 2.37, respectively. The authors conclude that occupational exposure can cause tuberculosis infection in prison employees. Prison employees should be tested routinely for tuberculosis infection.