Models for the spread of sexually transmitted diseases, including AIDS, rely on sexual partner mixing patterns in the population. From data acquired through partner notification for infection with Chlamydia trachomatis among young people in Gothenburg, Sweden, it was possible to construct contact matrices for the variables ‘rate of partner change’ and ‘social class’ of the partners in pair formations. These matrices show a restricted mixing between these two variables for young heterosexual women. Therefore random mixing models will probably overestimate rate of spread and possibly final size of an epidemic of a sexually transmitted disease.
