Abstract
Why do legislators in some political parties choose to vote coherently over time and over a varied set of policy proposals, while less so in others? Are there structural differences between these parties that increase or decrease the vote cohesiveness? And are there specific situations where the cohesiveness breaks down?
Previous research has proposed four sources for observed party unity in legislatures: preferences, party discipline, agenda-setting and division-of-labour. I link these sources to specific hypotheses regarding the effect of situational and structural variables, which I put to test using multi-level logit regression models on roll-call data from the 2009-10 European Parliament.