It is often argued that employees satisfied with their jobs perform better, which in turn will lead customers to be more satisfied.  The evidence on this “satisfaction mirror” hypothesis is decidedly mixed in the private sector.  At the same time, the “satisfaction mirror” hypothesis appears to remain widely accepted in public administration.  We put it to an empirical test for one of the core public services by assessing whether job satisfaction of police officers is mirrored in citizen satisfaction with these officers’ work.  We use Davis et al. (2009)’s data (released by the Inter-university Consortium for Political and Social Research as study number 29742), which consist of community surveys, officer surveys, voluntary contact surveys and involuntary contact surveys conducted in Knoxville, Tennessee, Kettering, Ohio, and the Broward County, Florida Sheriff's Office from 2008 to 2009.  Our model specification with citizen satisfaction as the dependent variable is informed by the expectation-disconfirmation model, which has become the prevalent theoretical approach to explaining citizen satisfaction (Van Ryzin 2004; James 2007; Poister and Thomas 2011; Morgeson 2013).  As the data do not include a direct measure of expectations, we use personal characteristics of the citizens in contact with the police as proxies for their expectations of police effectiveness and professionalism.  Using a structural equation model, we find no evidence in accordance with the “satisfaction mirror.”  Instead, we find that citizen race and gender predict their satisfaction with the police.  We also test for heterogeneity between the three study settings.