Despite advances in the scientific methodology of witness testimony research, no sound measure currently exists to evaluate perceptions of testimony skills. Drawing on self-efficacy and witness preparation research, the present study describes development of the Observed Witness Efficacy Scale (OWES). Factor analyses of a mock jury sample yielded a two-factor structure (poise and communication style) consistent with previous research on witness self-ratings of testimony delivery skills. OWES subscales showed differential patterns of association with witness credibility, witness believability, agreement with the witness, and verdict decision. Juror gender moderated the impact of communication style, but not poise, on belief of and agreement with the witness. Results are discussed with attention to application of the OWES to witness research and preparation training.