Look-alike modelling in violence-related research: a missing data approach

Abstract

Violence as a phenomena has been analysed in silo due to difficulties in accessing data and concerns for the safety of those exposed. While there is some literature on violence and its associations using individual datasets, analyses using combined sources of data are very limited. Ideally data from the same individuals would enable linkage and a longitudinal understanding of experiences of violence and their (health) impacts and consequences. However, in the absence of directly linked data, look-alike modelling may provide an innovative and cost-effective approach to exploring patterns and associations in violence-related research in a multi-sectorial setting. We approached the problem of data integration as a missing data problem to create a synthetic combined dataset. We combined data from the Crime Survey of England and Wales with administrative data from Rape Crisis, focussing on victim-survivors of sexual violence in adulthood. Multiple imputation with chained equations were employed to collate/impute data from different sources. To test whether this procedure was effective, we compared regressions analyses for the individual and combined synthetic datasets on a binary, continuous and categorical variables. Our results show that the effect sizes for the combined dataset reflect those from the dataset used for imputation. The variance is higher, resulting in fewer statistically significant estimates. We extended our testing to an outcome measures and finally applied the technique to a variable fully missing in one data source. Our approach reinforces the possibility to combine administrative with survey datasets using look-alike methods to overcome existing barriers to data linkage.