-- Selection bias results from deficiencies in study design

-- Depends on study design - in general - disease or exposure status influences selection of subjects to a different extent in groups being compared

-- Since subjects are selected for a prospective cohort study before disease occurrence - selection bias not a serious problem (but can get selective loss to follow-up)

-- Self-selection bias - differential participation of subjects after selected for case-control study; selected exposed persons agree to participate but many selected unexposed persons refuse to participate; if disease is associated with participation in unexposed group- bias will result (may elect not to participate because they are sick