Activity interference is widely considered to be a central mechanism linking exposure to noise and emergence of annoyance. Salient sound events in particular may divert attention from the task at hand, thereby reducing task performance. Sound events caused by traffic noise intruding the dwelling are therefore often found to be a main cause of community noise annoyance. In this work, experimental and simulation results on activity interference caused by traffic noise are compared. On the one hand, an experiment on activity interference by traffic noise was conducted in a realistic setting resembling an at-home situation. Subjects were instructed to read, while being exposed to a combination of road and railway traffic noise. The number of train pass-by events, the distance to the railway track and the emergence of train events above the background noise was varied among subjects. After completion of the reading task, the subjects had to evaluate their perceived disturbance due to passing trains and to report how many trains they noticed in retrospective. On the other hand, a computational model of auditory attention was used to determine the number of train pass-by events that subjects would notice, solely based on the acoustic stimuli used in the perception experiment. Using an optimized stochastic function that simulates the attention spend on the reading activity of the subjects, the model was able to replicate trends found in the empirical results, and estimated the number of noticed train events quite well.