Data was collected from five evacuations from the same school building, conducted in Spain between 2011 and 2014. Children from 6 to 16 years old were observed during the evacuation exercises. Four of the evacuations were unannounced, while one was semi-announced: the staff being aware that the drill would be conducted on a particular day, while the students were unaware. Information was gathered on the key factors deemed to influence evacuation performance: a description of the geometry, the population involved, the procedures employed and the organization of the drill itself. This information should allow interested parties to gain a reasonably detailed understanding of the initial conditions of each of the five trials. Evacuation data was also collected, focusing on the pre-evacuation times, the routes employed, the speeds adopted and the arrival times. Where data is ambiguous, flawed or omitted, this is documented in an attempt at transparency. To demonstrate an application of this data, we performed a series of small test cases using the Pathfinder, STEPS and EXODUS evacuation tools. The purpose of this work is to (1) provide insight into the configuration of these models for equivalent scenarios; (2) examine any variation in the simulated conditions given equivalent initial conditions; and (3) make public the configuration files (e.g. architectural files, raw experimental data, etc.) and analysis to contribute to the understanding the emergency movement of school pupils and the subsequent use of this data as part of modelling validation exercise. It is suggested that the very transparency of this process is relatively novel in the area of egress modelling.