The U.S. government currently publishes workplace fatality rates, using employment as the denominator. However, employment may not be a good proxy for worker exposure to risk if groups of workers differ in their hours worked. Using micro data from the Census of Fatal Occupational Injuries and Current Population Survey, this paper presents alternative fatality rates calculated using employment and hours worked. Rates are compared for worker groups defined by gender, age, occupation, and industry. In general, both measures identify the same groups of workers as facing especially high and low fatality risks. The rank correlation of the employment- and hours-based fatality rates for 222 detailed occupations is very high, with a value of 0.99. However, for a few groups, the hours- and employment-based rates may differ more than 10%. Most notably, workers below age 20 and above age 64 have rates that are 60% and 37% higher, respectively, when hours is used in the denominator. This suggests that hours--the conceptually preferable denominator--should be used when possible to calculate workplace fatality rates.