The risk of fatality among paired motorcycle drivers and passengers involved in a crash was modeled using an estimating equation. The data were obtained from the Fatal Accident Reporting System, 1976-1989. Overall, helmet use was associated with a 35% lower risk of fatality. The effectiveness of helmet use appeared to increase over the study period, from 14% in 1976 to 51% in 1989. The measured effect of helmet use was greater in noncollision injury events than in collisions, and tended to be greater in collisions with objects that were soft or not fixed than in collisions with motor vehicles or hard objects; in side or rear collisions than in frontal collisions; and in crashes with minor or moderate damage to the motorcycle than in crashes with severe damage.