The impact of life style characteristics and chemical exposure conditions of about 80 nursing women on the contents of Organochlorine compounds (OCC) in their breast milk is analyzed and described by a robust statistical procedure. Information on independent variables such as age, diet, smoking, occupation and chemical exposures is obtained by a questionnaire. Concentrations of HCB alpha-, beta-, gamma-HCH, Heptachlorepoxide, Dieldrin, DDE and PCB in milk samples are measured and serve as dependent variables. By logistic regression we quantify the influence of combinations of several independent variables on the probability of being extremely contaminated with at least one OCC. We also study the pattern in which OCC are associated with each other. There are distinct pairwise correlations of the four main contaminants: HCB, beta-HCH, DDE and PCB. These correlations entail that a woman with a high PCB concentration, for example, is more likely of being also contaminated above average with the remaining contaminants. We define a simple load score and compare its different theoretical distributions under the assumption of presence or absence of the correlations. By this we can show that the proportion of heavily contaminated women is underestimated, if the OCC associations are ignored. We therefore recommend that the pattern of OCC-correlations in human milk is further studied in more representative samples. These investigations should also comprise occupationally exposed women.