Mathematical modelling has been used extensively to understand the spread of SARS-CoV-2, and to project the potential impacts of different control measures. For instance, what effect do school closures have on R0? Social contact data play a key role in the parameterisation of these models, informing patterns of age mixing, and social mixing by location. These data are also widely used in models of other infections, such as tuberculosis and influenza.

Despite this, little research has been conducted into what effect choices made in how data are collected and analysed have on model projections and estimated intervention effects.

The student will use a combination of data analysis of existing social contact datasets and mathematical modelling to explore a number of issues around the design and analysis of social contact surveys. This includes the choice of recall period, if and how the duration of contacts should be incorporated into analysis and modelling, and how the definition of contact used effects estimates.

This work comes at a critical time, with the COVID-19 pandemic driving a surge in social contact data collection and analysis, and with technology availability leading to new methods of data collection (for instance the 2018 BBC Pandemic Experiment).