ABSTRACT
Party identification has long been central to controversies about voting behaviour in Britain, but there is no agreement about how the concept is best measured. Building on previous research this article presents evidence from a new study, which demonstrates that the traditional BES question is seriously flawed. In addition, it examines evidence from the BES itself, which supports this claim. The article then argues that, when read in conjunction with evidence from aggregate level studies and focus groups, these findings cast yet more doubt on the validity of the traditional BES measure. The article ends by arguing that we should invest more time and more effort into measuring this vitally important concept.
