Bojan Todosijevic, "Issues and Party Preferences in Hungary: A Comparison of Directional and Proximity Models," Party Politics, 11 (January, 2005), 109-126.

First Paragraph:
The assumption that citizens elect their representatives according to the political goals they wish to be realized has often been regarded as essential for representative democracy. When scholars started to analyse empirically the determinants of citizens' voting choices, it became obvious that the process included more than simply rationally selecting from available policies.

Figures and Tables:
Table 1. Party membership of the sampled MPs
Table 2. Eight policy preference items
Table 3. Average scores on eight policy issues among MPs and voters
Table 4. Position of parties on the Left-Right continuum assessed by the MPs and voters
Table 5. Regression coefficients (beta) for proximity and directional effects of eight specific issues (separate equations)
Table 6. Directional and proximity effects of the issues (first block) and of the issues and L-R self-placement. Entries are standardized beta coefficients and R2
Table 7. Regression coefficients (beta) for proximity and directional effects (separate equations) of issues and L-R self-identification, for respondents with higher and lower educational levels

Last two paragraphs in the conclusion:
Examination of the interaction of political sophistication with the two models of issue voting partly confirmed the hypothesis that issue voting is less common among the politically less sophisticated. However, Maddens's (1996) hypothesis that the proximity model, by requiring greater political sophistication, should be more pronounced among the better educated was not confirmed. While both effects are stronger among the better educated, the difference in favour of the directional model remains at all levels of political sophistication. These findings apply to the effects of issues and to L-R self-placement.

If the present findings are compared with findings in other countries, then Hungarian voters do not appear exceptional. The directional model in general seems better equipped to capture the effects of issue voting.