To discover the validity of crime reconstruction on the basis of witness testimony, 28 simulated interrogations in 2 West German police departments were conducted. The test persons (volunteers of varied social background) were shown color sound films of several simulated offenses and then asked to act the part of victim or witness. The interrogations, which were recorded on tape, were conducted by actual police officers. The validity of witness depositions was then tested against detailed synopses of the films established for the purpose of the experiment. The results of the study were arranged according to the three basic components of an interrogation: (1) the preliminary questioning, (2) the writing phase, and (3) the finished product. On the whole, the crime reconstructions were far more complete and reliable than was generally assumed. The net informative value of each interrogation (i.e. the completeness and correctness of the information) was found to depend on a number of factors including the witness' powers of observation, and the officer's recording technique, memory retention capacity, method of summarizing results, and, above all, ability to relate to and cooperate with the witness. Interrogating techniques could be improved by the officers (1) placing themselves in the position of witness and offender, (2) informing witnesses clearly of what is expected of them and (3) encouraging witnesses to narrate the crime freely and without interruption. The study contains numerous statistical charts, detailed transcriptions of witness testimonies, an extensive bibliography, and an index. The instructions given to witnesses before the interrogation and the film synopses are appended.