<p>Identifying the sources of traffic congestion is essential for its effective mitigation. In this study, we rigorously evaluated the method for identifying congestion sources introduced in our earlier work via traffic simulation. Because ground-truth locations of congestion sources cannot be observed in real-world data, previous empirical assessments of the method have been incomplete. Moreover, although the method has used average speed as input due to its accessibility, average speed is a less comprehensive traffic indicator than traffic density. This raises concerns about the method’s practical reliability. In contrast, traffic simulation permits both the deliberate introduction of congestion at specified links and the easy collection of traffic density and average speed across all road links, thereby offering a suitable test bed for thorough validation. The experimental results demonstrate that the method can accurately identify the true congestion sources regardless of whether traffic density or average speed is used as input.</p>