The examination value of the railway traffic action effect is calculated from long-term monitoring data without resorting to load models and structural analysis. Monitoring data are obtained from measurements on a short span RC slab bridge. The action effect is represented by the distribution of its maxima per train passage, whose left tail is modeled by a shifted exponential distribution. The examination value is calculated by extreme value theory, using additionally a confidence upper bound in order to take into account the monitoring duration (described by the number of recorded maxima exceeding a certain threshold). (C) 2016 Elsevier Ltd. All rights reserved.