John Bartle, "Partisanship, Performance and Personality: Competing and Complementary Characterizations of the 2001 British General Election," Party Politics, 9 (May 2003), 317-345.

First Paragraph:
In the wake of every general election, political commentators and party politicians alike rush to offer 'instant' interpretations of the outcome on the basis of the impressions that they formed during the long or short campaigns. The 2001 general election was no exception. Some regarded the election as reflecting a long-term shift in partisan advantage from a once hegemonic Conservative Party to an ascendant Labour Party. Butler and Kavanagh (2001: 259) for example suggested that the election reflected 'the creation of a new electoral landscape' and others that Tony Blair had made a reality of Harold Wilson's claim that Labour is the "natural party of government"'. Labour ministers claimed that the outcome was a vindication of Labour's record in office and a mandate for investment in and reform of the public services (Butler and Kavanagh, 2001: 254). Distraught Tory backbenchers on the other hand attributed the outcome to the Tory party's failure to reposition itself in the centre of British politics and project an attractive 'new' image. Still others blamed the unattractiveness of the Tory leader; a view that was partly supported by William Hague himself, who attributed the Tories' defeat -- at least in part -- to a personal failure to persuade people that he was the 'alternative Prime Minister' (Battle, 2002: 191).

Figures and Tables:
Figure 1: Assumed causal order among explanatory variables
Table 1: The incidence of party identification in 2001
Table 2: Attitudes toward enduring conflicts and policies in 2001
Table 3: Attitudes toward the single currency in 2001
Table 4: Non-partisan economic evaluations in 2001
Table 5: Retrospective evaluations of the government's performance in 2001
Table 6: Party and leadership images in 2001
Table 7: The imporance of explanatory themes
Table 8: The imporance of explanatory themes [sic]
Table 9: Alternative assumpions about the causal order

Last Paragraph: [last two paragraphs]
Taken as a whole, these findings underline the importance of producing improved measures of partisanship. In the absence of multi-wave panel data that speak to the validity of the two measures of party identification, it is unclear whether the new or the traditional BES question is the more valid measure of party identification. Moreover, it is arguable that party loyalties are inherently unstable and that they are often updated in accordance with recent political experiences and opinions (Fiorina, 1981). It may therefore be the case that whatever indicator of partisanship is preferred, including such a variable as an explanation of vote, conceals the prior effect of short-term factors on partisanship. Analysts who support this view would undoubtedly demand that the party identification theorists provide us with more compelling models of partisanship.

While the characterizations of the 2001 general election contained in Tables 8 and 9 accord with some basic intuitions about the contingent nature of Labour's 2001 election victory (Bartle, 2002), they are necessarily subject to great debate. These findings also underline the need to determine the most appropriate causal order between party and leadership images, or to determine those cases where leadership images influence party image and vice versa.