Swiss Elections with Randomized Ballot Ordering, 1999-2023
Description
This dataset was assembled to test the hypothesis that a party list's vote share and its campaign spending is influenced by its ballot position. It shows that party lists receive a significant gain in vote share but don't spend significantly more or less on campaigning the higher their ballot position. The effect is not influenced by party affiliation of the list, turnout and number of competing party lists or district magnitude (i.e. number of candidates on a party list). It contains data from all (available) Swiss OLPR elections (national, cantonal, communal; legislative and executive) with randomized ballot ordering conducted from 1999 to 2023. There are 4457 party lists (observations) competing in 1011 elections. A party list's vote share is operationalized as the number of times the list was cast. A party's campaign spending is operationalized as the sum of the self reported budgets of its candidates. Ballot position is operationalized through a fraction ranging from -1 to 0 (the party list's absolute ballot position divided by the number of party lists competing in the election multiplied by negative one) and by a variety of dummy variables (first, second and last position; first and second half of the ballot).
Files
Steps to reproduce
All elections with randomized ballot ordering were determined by examining cantons' and communes' electoral laws on how ballot order is decided. Election results (as well as turnout and district magnitude) were obtained through cantons' and communes' election offices (downloaded on their websites or received upon request). Campaign spending data was taken from the Swiss Election Study (Selects)'s Candidate Survey, which reports individual candidates' budgets, and aggregated to party lists.