Social & Preventive Medicine

Bias is unlikely to invalidate cohort studies used to assess risk of exposure because -