Informing Non-Response Bias Model Creation in Social Surveys with Visualisation

Through an ongoing process of co-design and co-discovery we are developing and using visualization to explore large amounts of auxiliary data from unfamiliar sources to understand non-response bias in social surveys. We present auxiliary data in their geographical contexts and show how this can complement traditional data analysis and provide a more comprehensive understanding of the data. This is helping select variables for non-response modelling. These processes are not just limited to non-response analysis, but have potential to be used in wider quantitative analysis in social science.