A principal problem of studies on public opinion is that many reinforce the myth of consensus with pseudoscientific methods. In many cases, such studies are commercial ventures. Still, studies of ideologies are essential to understanding of penal matters and events within prisons. Both the institution and the effectiveness of norms must be examined carefully. Most studies, however, fail to develop a conceptual basis for their surveys: they assume naively that they need only ask interesting questions of a selected population sample. As a result, they do not consider whether questions are pertinent to the views of respondents. Methodological shortcomings of survey instruments, to quantitiative matters such as sampling and survey instruments, to qualitiative matters such as interview factors, and to reorganization of research steps. A review of studies indicates that the public images of social control of crime depend on the aims assigned to repression of crime rather than on knowledge of the law. Views of social control are strongly influenced by social attitudes and values on subjects such as conformism and optimism or pessimism about human beings. Finally, attitudes toward social control are associated in a complex manner with the group structure of society. It is recommended that commercial surveys be discouraged and that the scientific community carefully assess studies of penal justice and public opinion. Very precise methods should be designed to deal with particular problems and research should be repeated at regular intervals and compared on an international level. The author proposes the Council of Europe as the framework for such a research system. Tables and notes are supplied. A bibliography of approximately 200 entries is appended.--in French.