Different approaches may be considered in the evaluation of past occupational exposures: questionnaires only, questionnaires and experts (such as industrial hygienists or occupational physicians), questionnaires and job exposure matrices (JEM). These approaches are described in this paper and their use in the framework of population-based case-control studies is discussed. The use of experts to assess exposure from questionnaires is generally considered to be the reference method. Its major drawbacks are its cost and the shortage of experienced experts. On the other hand, JEM reduce the cost of the study and provide an objective way of evaluating exposures. The main drawback of JEM is the fact that they can produce misclassification. The development of a JEM includes the choice of the job classification system which influences the degree to which the matrix can be used in other studies, the precise definition of exposures, and the choice of entries which provide more information than the oversimplified exposed/non-exposed dichotomy. Recent results on JEM methodology (statistical methods of analysis and evaluation of JEM quality) should encourage both their development and their use in epidemiological studies.