Abstract. In the field of disaster risk reduction (DRR), there exists a proliferation of research into different ways to measure, represent, and ultimately quantify a population's differential social vulnerability to natural hazards. Empirical decisions such as the choice of source data, variable selection, and weighting methodology can lead to large differences in the classification and understanding of the "at risk" population. This study demonstrates how three different quantitative methodologies (based on Cutter et al., 2003; Rygel et al., 2006; Willis et al., 2010) applied to the same England and Wales 2011 census data variables in the geographical setting of the 2013/2014 floods of the River Parrett catchment, UK, lead to notable differences in vulnerability classification. Both the quantification of multivariate census data and resultant spatial patterns of vulnerability are shown to be highly sensitive to the weighting techniques employed in each method. The findings of such research highlight the complexity of quantifying social vulnerability to natural hazards as well as the large uncertainty around communicating such findings to stakeholders in flood risk management and DRR practitioners.