Replication data for 'The competitive dynamics of invalid voting in French legislative elections, 2002-2022'
Description
French Ministry of Interior election data collated across legislative elections from 2002 to 2022. Abstract for paper: "Despite a recent increase in studies, spoiled and blank ballots remain an under-theorised aspect of voting behaviour. The French case provides an opportunity, conceptually and empirically, to analyse a number of competitive dynamics which should influence rates of vote invalidation. The stable institutional context of two-round legislative elections between 2002 and 2022 provides electoral district-level data to examine how outcomes from the first round encourage vote invalidation at the subsequent run-off. Starting from a multi-level time-series cross-section specification, this paper finds that the more that candidate choices and performance deprive voters of second-round choice, the higher the rates of invalid voting. In particular, electoral thresholds ousting otherwise competitive candidates increase invalid voting. Despite the radical shift in party system array across this 20-year period, however, it finds that the candidate alignments which best characterize those voter pools deprived of choice remain relatively stable over time."
Files
Steps to reproduce
French Ministry of Interior data for French legislative elections were downloaded from the data.gouv.fr website, and were reshaped to list results by political nuance and election district. Voting blocs were derived using the coding detailed in Appendix 1 of the paper. The Stata 'final_do' do file produces all related tables and figures in the paper. The 'raw data coding' do file (included in the 'datasets' zip file) provides the full reshaping process for legislative elections (as well as presidential elections) as well as merging a number of external datasets with socio-economic measures, constituency boundary changes [manually derived from online sources], GIS data for district population density, etc.. These datasets were also obtained from data-gouv.fr . All data files used in the 'raw data coding' do file are included in the 'datasets' zip file.