Based on extensive clinical questionnaire data, this paper explores the relation between research subjects' self-assessed understanding and actual knowledge of a largescale predictive screening study, and its implications for the proper handling of information and consent routines in longitudinal studies. The intitial data show that low self-assessed understanding among participants was correlated with limited knowledge, concern over participation and collected samples, less satisfaction with information, and feeling passive or negative towards the study. Among those reporting high understanding, a non-negligible number displayed a lack of knowledge regarding central aspects of the study. Regarding high assessed understanding, the multivariate analysis identified the main predictor variables to be knowledge, having a positive attitude towards participation and the study itself, being satisfied with information, having a stable psychosocial background and feeling calm regarding the handling of samples. These findings indicate that to evaluate participants' understanding through self-assessment may not be reliable method. Self-assessed understanding may rather be a good indicator of general attitudes than a tool for analysing content. The data also show that actual understanding varies considerably among participants, suggesting that more effort needs to be put into adjusting the information to the needs of different subgroups. It is argued that when doing this, researchers in longitudinal studies must be careful not to exhaust participants with excessive information that they do not want and that may cause them to drop out. The ethical relevance of obtaining repeated consents in longitudinal research is discussed.