Introduction

The LEADS model links dissent to a number of variables: the MP's electoral security, career advancement, and the like. These observable implications are developed within the context of an individual-level model of behaviour, but many can be tested equally well at the aggregate level. If, for example, advancement suppresses an MP's propensity to dissent, then parties in which opportunities for advancement are relatively abundant should experience less dissent than parties in which such opportunities are scarce. There is also scope at the aggregate level to test arguments about dissent that are not directly addressed by the model. The confidence convention stands out in this regard: it does not play a role in the model (by assumption, the MP's vote is not critical to the survival of the government), but it is a central feature of parliamentary government, and one that might be expected to influence the level of dissent that parties experience. Indeed, the impact of these sorts of institutional variables is often better studied at the aggregate level, where there is typically more room for institutional variation than at the individual level.