The intellectual roots of this assumption are examined, and an emerging perspective developed from direct observation of patrol operations is presented. The research questions are addressed with data that were collected from observations of patrol activities in 60 residential neighborhoods in 24 agencies. The analysis, which is referenced to officers' shifts of duty, shows that the effect of supervision is in the direction posited by the authors of traditional police administration texts, but the magnitude of the effect is not overly strong. The findings are used to inform the evaluation of a common policy option directed at improving officer performance through supervision. (Author abstract)