MiSoC’s methodology work covers the following areas: the use of machine learning techniques from computer science for the analysis of social science data; the use of genomic, epigenetic and proteomic data to help understand the relationship between education and mental health; and the use of genomic databases to understand the causes of historical demographic trends.
Research area leader: Paul Clarke
MiSoC’s new programme of methodological work brings data science into the MiSoC programme across all research areas, and deals with crucial measurement issues key to our research programme. Our machine learning projects involve applying machine learning techniques a) to reduce the sorts of bias one can find when using conventional causal techniques (e.g. propensity scores, instrumental variables regression), and b) to identify groups of people who are affected in very different ways by treatments or social exposures.
Professor of Economics, European University Institute
Professor of Demography, University of Bologna, Italy (Co-I)
Professor of Social Statistics, University of Essex (Co-I)