Social contact studies are used in infectious disease epidemiology to infer a contact matrix , having the mean number of contacts between individuals of different age groups as elements. However, M does not capture the (often large) variation in the number of contacts within each age group, information is also available in social contact studies. Here, we include such variation by separating each age group into two halves: the socially active (having many contacts) and the socially less active (having fewer contacts). The extended contact matrix and its associated epidemic model show that acknowledging variation in social activity within age groups has a substantial impact on the basic reproduction number, R₀, and the final fraction getting infected if the epidemic takes off, τ. In fact, variation in social activity is more important for data fitting than allowing for different age groups. A difficulty with variation in social activity, however, is that social contact studies typically lack information on whether mixing with respect to social activity is assortative (when socially active mainly have contact with other socially active individuals) or not. Our analysis shows that accounting for variation in social activity improves model predictability, yielding more accurate expressions for R₀ and τ irrespective of whether such mixing is assortative, but different assumptions on assortativity give rather different outputs. Future social contact studies should, therefore, also try to infer the degree of assortativity (with respect to social activity) between peers and their contacts.