Bias resulting from a loss of baseline subjects at follow-up (drop-out), and newcomer bias resulting from subjects entering the study at the follow-up stage, were investigated in a three-community coronary risk factor follow-up study. The study consisted of a cross-sectional baseline study on 7188 participants aged 15 to 64 years, a four-year intervention period and a follow-up cross-sectional study in the same communities on 6283 participants aged 19 to 68 years. The overall non-response rate of 45% in men and 42% in women varied from 30 to 79% in the various age and sex groups, with the biggest drop-out rate occurring in the youngest age group of 15 to 24 years. At baseline drop-outs were more likely to have lower educational qualifications than those who participated in both the baseline and follow-up studies (stayers) and included significantly more smokers than non-smokers. Coronary risk factors of newcomers were not different from that of the stayers at follow-up except for slightly, but not significantly, higher smoking rates in newcomers. These findings suggest that drop-out and newcomer bias need to be assessed and its effect studied before final evaluation of data in community follow-up studies.