Results found that mock jurors reacted more favorably towards the defendant when the facial composite resembled him than when it did not, but reacted similarly when the composite did not resemble the defendant versus no composite presented. In addition, assessments of the strength of the composite evidence varied depending on the other case evidence presented. Mock jurors selectively incorporated evidence into their judgments about the credibility of the eyewitness and the culpability of the defendant. They responded to strong versus weak facial composite evidence as anticipated: a good match composite was viewed as more favorable to the prosecution than a poor match composite, and weak facial composite evidence was largely ignored. It is noted that previous research on the reliability of eyewitnesses' facial composites suggests that some skepticism is warranted when evaluating their accuracy. The current research sought to examine the degree to which mock jurors evaluated facial composite images as a source of incriminating or exonerating evidence against a defendant, particularly when they were inconsistent with other evidence presented. Data were derived from a test group of 120 undergraduate students attending a large southwestern university in the United States. These mock jurors read a fictitious criminal trial transcript of a person charged with armed robbery. Experiment 1 manipulated the facial composite evidence such that participants viewed a facial composite which was either a good or a poor match to the defendant. Experiment 2 also manipulated the quality of the facial composite evidence, as well as the other case evidence against the defendant. After reading the transcript, participants made several testimony-relevant judgments about the case. Figure, tables, and references