Subjects were 34 male and 36 female volunteers from introductory sociology classes. Arrest reports obtained from the local police department related the incidents leading to the arrest of three white male suspects for trespassing and simple assault, trespassing and possession of a controlled substance, and driving while intoxicated. Fictitious names and photographs of the offenders were provided the subjects. Six facial expressions for each of the offenders displayed happiness, sadness, anger, disgust, fear, and surprise. Ten subjects were randomly assigned to each of six groups which viewed one of the facial expressions of each of the three offenders. A seventh group of ten subjects was given only the objective arrest information. Each group was asked to determine the guilt or innocence of the three suspects based on material supplied and then predict the likelihood of each suspect's committing another offense. All subjects judged the three suspects guilty. The predictions of recidivism indicated that the objective information contained in the arrest reports was used more than the subjective information derived from facial expressions. The results suggest that when secondary sources of information are available for predicting future behavior, they may not be as important to the observer as the objective information. In the debriefings, however, many subjects referred to the look on the suspect's face to explain why they decided as they did about guilt and recidivism. This information suggests a direction for subsequent research; i.e., whether facial expressions may provide a rationale for interpersonal judgments, even though not actually determining the judgment. Tabular data and eight references are provided.