Witnesses to crimes are often asked to help law enforcement officers construct a facial composite, or visual representation, of a perpetrator’s face. There are several techniques that allow composites to be created by those lacking artistic talent. The current study evaluated five facial composite techniques, E-FIT, PROfit, FACES, sketch, and Evofit for their composite naming accuracy. The study utilized a methodology reflecting composite construction under realistic conditions by allowing an interval of 2 days to pass between seeing an assailant and constructing a facial composite. Participants were 50 individuals who had not previously constructed a composite and were paid for their participation. Participants were shown a photograph of a celebrity for 1 minute and were asked to construct a facial composite of that celebrity 2 days later. Outcome data were gathered by asking independent observers to name the composites through the use of either sorting or line-up techniques. Results showed that composites were only named in 3 percent of the cases, overall. Sketches had the highest composite naming rate at 8 percent. Sorting provided a better match than line-up. The findings suggest that the manually-generated sketches used in past police work are better than the computerized systems in terms of composite identification and, further, that line-up techniques are poor ways of conducting facial composite identification. Future research should focus on improving composite quality. Tables, references