Comparative study of the impact of party leaders and their personality traits on election outcomes is diverse and non-cumulative-it has followed a number of directions, and the literature is conflicted in almost all of these areas. One of the main branches within the leaders literature focuses on the evaluation of personality traits, and scholars are unable to agree about which traits voters take note of; whether traits actually affect election outcomes; if they do which are more important; and why or how traits are evaluated in the first place. When one looks more closely at the literature, it becomes clear that what is needed is a larger comparative analysis that incorporates election studies across time and space, in order to be able to make more definitive conclusions about the role of leader evaluations. However, one of the challenges to this endeavor stems from the nature of the data available. Different traits are included in different election studies over time, and often even where trait continuity does exist, the question format changes, making comparability over time and across space very difficult. This paper examines the impact of question format, using pooled data from election studies conducted in two countries, Canada and the United States. The analysis suggests that question format does indeed have an effect on respondents' evaluations of party leaders, but more research is needed before we can make conclusions about the implications of these effects for the comparative study of the role of leaders. ..PAT.-Unpublished Manuscript [ABSTRACT FROM AUTHOR]