In 1999, almost one in three couples were composed of two persons with the same social status; this is twice what would be observed if couples were formed at random. This proportion was slightly lower for couples married during the 1990s than for those formed during the 1930s and 1940s. At the same time, the socio-professional structure of the population has evolved, which largely explains the evolution of socio-professional homogamy over the study period. However, when these changes in socio-professional structure are taken into account, the evolution of homogamy remains unclear: it is difficult to say whether men and women today form relationships with a partner of the same socio-professional category more or less often than in the past. In the majority of couples, the partners also have identical levels of education. However, the proportion of couples composed of two people with the same level of education is lower for recent unions than for those formed during the 1950s. This evolution results in part from significant changes in education, and in particular the rise in the level of education. Yet independently from this structural effect, there has been a downward trend in educational homogamy. There are various factors that favour educational homogamy, but starting a relationship whilst still being a student is a determining factor.