In transportation modeling, intrazonal trips are frequently omitted during trip assignment. These trips are not assigned to the network because their origin and destination are in the same zone. However, in reality, intrazonal trips use the network and take up some of its capacity. This omission is due to the spatial aggregation problem. Omitting these short trips from assignment models affects the level of service of the network and biases the estimation of main assignment outcomes. The issue of intrazonal trips omission has received limited attention in transportation research. In this paper, we address the problem of ignoring intrazonal trips in traffic assignment models by applying a stochastic approach in order to characterize the statistical impact of their omission. Our results show that the omission of intrazonal trips has a significant impact on main assignment results. Network speeds, volumes and congestion levels vary significantly with the omission of intrazonal trips. The extent of this impact depends on the road’s category in the network hierarchy. As regards level of service, local streets are more sensitive to the omission of intrazonal trips than the primary network. These findings reveal the existence of a bias due to the omission of intrazonal trips in assignment models and raise doubts about the accuracy and reliability of assignment results from standard four step transport models especially when the spatial zoning is coarse.