As originally designed the RECAP experiment created and examined the calls-for-service records of the 500 most active addresses (in terms of calls for police service) in Minneapolis, Minn. The data source was a series of archived "dump tapes" created by the Minneapolis Emergency Communications Center. This study corrects the main flaws of the original randomization, address fragmentation, and the presence of duplicate call lines. T-test procedures comparing reranked pairs of addresses (236) in reconstructed databases confirm the original findings of no statistically significant differences between treatment and control groups. Further analysis finds no differences in subgroups by type of address and none by type of crime. Statistically significant reductions are found for residential addresses with low levels of treatment (77); diminished statistical power restrains generalization. Qualitative sections examine RECAP in terms of officers' problem identification, resistance encountered, relationships to external agencies and within the police department, and problemsolving technologies. The experiment's definition of a problem as "an address with a high call load" did not always produce targets amenable to intervention. Problems at RECAP addresses were conglomerations of human interaction; the effectiveness of traditional crime-prevention activities was minimal. RECAP methods of problemsolving were primarily disseminating information, persuasion, coercion, and limited law enforcement actions. Resistance throughout the police department limited the effectiveness of actions that required support from patrol units. Interdependencies, process factors, and ethical issues raised by the problem- oriented tactics are discussed. A comparison of incident- driven policing to the problem-oriented policing is outlined. Appended tabular data and 135 references