Previous studies have investigated the close association that exists between CAG repeat number and the age at onset in SCA2 = spinocerebellar ataxia type 2. These studies have focused on affected individuals. To further characterize this association and estimate the risk of a carrier developing SCA2 at a particular age as a function of a specific CAG repeat size, we have analyzed a large group of 924 individuals, including 394 presymptomatic and 530 affected individuals with a CAG repeat length of 32-79 units. Using a Kaplan-Meier survival analysis, we obtained cumulative probability curves for disease manifestation at a particular age for each CAG repeat length in the 34-45 range. These curves were significantly different (p < 0.001) and showed small overlap. All these information may be very valuable in predictive-testing programs, in the planning of studies for the identification of other genetic and environmental factors as modifiers of age at onset, and in the design of clinical trials for people at enlarged risk for SCA2.