Research on public demands for assistance and distribution of police resources indicates that the system for deployment of uniform officers designed to respond to differences in workloads resulting from time of day, location, and type of patrol has been hampered by resistance to change and insufficient data. Critics have argued that the use of uniformed patrols is unjustified, as research has tried to establish whether patrols can satisfy citizen demands for assistance. Other studies have shown that special patrols for reducing adolescent truancy may have reduced absenteeism but probably have had little effect on the number of crimes committed during school hours. Moreover, according to preliminary evaluations, an expensive campaign to reduce car thefts by informing the public about security measures appears to have had little effect on modifying the habits of automobile owners. Similarly, campaigns to reduce burglaries by informing the public about possible security measures have done little good, although such campaigns may be effective for commercial establishments. Studies on the effectiveness of crime prevention efforts encounter problems related to the experimental method, such as lack of control over the experimental conditions of the study and unforeseen interference with the carefully planned circumstances of the study. It appears from the research that the usefulness of police crime prevention rests less with the precise nature of the measures taken than with symbolic confidence in their efficiency. However, by showing that police preventive work is based on a false conception the effectiveness of such police work will be diminished and the fear of crime will increase. Notes, an appendix on police research conditions in Great Britain, and 37 references are supplied. --in French.