This study shows the results of an analysis of infection data of 21 Dutch dairy farms, where during 10 years each culled animal was tested for the presence of Mycobacterium avium subsp. paratuberculosis. Twenty of the farms were part of a vaccination study. By allocating animals of the herd according to the S-L-I concept (Susceptible, Latently infected and Infectious), the transmission parameter ß of the infection was estimated by generalised linear modelling. This parameter ß is the average number of new infections caused by one initial infection per unit of time. Through ß, the reproduction ratio R₀ can be derived. The effect of management on the farm, time since start of vaccination, and several assumptions about the infectious period, were studied. The present analysis indeed showed that management had a significant effect on ß and R₀, and the advised hygienic measures decreased the within-herd transmission on these 21 farms. Furthermore, the transmission decreased significantly with time since the start of vaccination, probably due to a combination of calf vaccination and improved management on the farms. When assuming that infected calves are infectious during a certain period immediately after infection, the model fits much better to the data. According to this model, the level of infectivity of (infected) calves is much higher than that of heifers or cows. This does not mean that calves shed more bacteria, because the effect can be explained by more contacts with the other calves on the farm. If the role of calves in transmitting the infection to other calves is indeed as important as this statistical study suggests, is now being studied in transmission experiments, in which calves are housed in-between faecal-culture positive cows during 3 months. After that period, these calves will be transferred and housed together with new calves during another 3 months.