The core of the authors' research program consists in the assumption that complex statistical interactions of sociological and psychological factors with pathophysiological and standard risk factors are more important for the early detection, prediction and prevention of cancer than one or just a few medical risk factors, e.g. the relationship between smoking and lung cancer. As the analysis of these relationships can only be done in interdisciplinary research, interpretation and discussion of social scientific, molecular biological and epidemiological aspects was attempted. In the first part we explicate a model of relevant propositions from the psychosomatic literature to explain the relationship between psychosocial factors of carcinogenesis. A discussion of the authors' research program contains additional variables and propositions. It appears to the authors that among the psychosocial and sociological factors of carcinogenesis, the molecular biological variables are especially relevant, as they form the intervening link between the social science variables and cancer. Discussion of the methodology and the sampling of the authors' empirical studies follows. Two prospective studies, carried out in Yugoslavia and West Germany, are described in detail. The data of these studies are used in the empirical analysis. In the next part the empirical results of different multivariate analyses are presented: The effects of the sociological and psychological variables on carcinogenesis are described; standard risk factors and biochemical-molecular biological factors are brought together, special attention being given to the interaction between these two groups of variables.