Using a natural quasi-experimental pretest/posttest design, residents in randomly selected homes in a suburb of Los Angeles were surveyed about their perceptions of their neighborhoods with respect to cohesion and sense of community. Responses from the pretest surveys--administered before the construction of a freeway that would bisect the city--were compared to the responses from the posttest survey six years later, administered two years after completion of the freeway. Respondents living adjacent to the new freeway--residents who experienced a fourfold increase in the average decibel levels in their neighborhoods since the freeway opened--reported both a lower sense of community and smaller neighborhood areas as compared to residents not living adjacent to the freeway and as compared to the results from the pretest. The analysis of the data incorporated geographic information system (GIS) software to allow for the analysis of phenomenological neighborhoods--neighborhoods as defined by respondents. This Resident-defined Neighborhood Mapping methodology permitted us to analyze neighborhoods as the respondents outlined them, not as they were preconceived by someone outside the neighborhood. It is suggested that this new methodology may prove useful in advancing the field of neighborhood research by detecting neighborhood-level change that traditional methods may miss.