Intervention strategies aimed at fear reduction derive from either a "crime-centered" approach or a "quality of life" approach. Crime-centered models emphasize crime-related neighborhood problems as constituting a major determinant of fear of crime. In the quality of life approach the hypothesis is that non-crime problems in a neighborhood are the main determinants. In this article the validity of both points of view is tested. It turned out that neither a pure quality of life model nor a purely crime-centered model could provide an adequate explanation for the data observed. Both models prove to be invalid. These empirical results demand an integrated model predicated on a "crime-centered plus" kind of approach to fear of crime. Toward the end implications of this model are discussed relative to police performance. 2 tables, 1 figure, 13 references. (Publisher abstract)