The results of major studies in four specific areas are reported. First, a study of the Kansas City Police Department (Missouri) on reaction time indicates that excessive attention should not be devoted to the rapid response time of police and that deployment strategies should consider such factors as type of crime, delays in reporting by victims, and urgency of requests for assistance. Another Kansas City study on preventive police patrols questions the usefulness of police patrols in marked cars for preventing crime, and a Rand Corporation study of the criminal investigation process indicates that special detectives frequently contribute little to the solution of cases and that many of the main leads in serious cases result from ordinary police work. Finally, a Forensic Sciences Foundation report shows that the results of tests in many crime laboratories are unacceptable for one reason or another. To gain insight into the allocation of police resources, the National Institute has set as priorities investigations regarding the collection of information on police operations and response time, on the objectives of police activities and police operational policy, and on management strategies designed to realize better practical results. Factors which affect the quality of research results are the varying quality of the researchers, negative reactions of practicing police officers to negative research results, and unrealistic expectations of immediate results from research findings. In conclusion, the author emphasizes the importance of comparing the research findings of different countries and of encouraging collaboration among various countries on research projects. No references are cited. --in French.