We follow the approach of expressive voting and consider that voters with more extreme ideology can enjoy higher utility after voting. However, along with effect of ideology, voters also take the difference of his ideology and candidate¡¦s into account. Given the above assumptions, two candidates choose their ideology before voters decide whether and for whom to vote. Two candidates¡¦ ideology converges to the middle point if voters¡¦ utility generated by expressive voting approach is less important than by instrumental voting approach. In the opposite, two candidates¡¦ ideology diverge and the voters with mild ideology refuse to vote in point if voters¡¦ utility generated by expressive voting approach is more important. We examine the ANES data of ten U.S. presidential elections from 1972 to 2008. The nested logit model is used to estimate the corresponding coefficients of voters¡¦ utility generated by expressive voting approach and instrumental voting approach. The data supports our conjecture that voters with more extreme ideology are more likely to cast their votes.

In this thesis I explore a novel context sensitive conceptualisation of civic duty according to which the conduct (or misconduct) of elected officials affects whether eligible voters feel either a duty to vote (DTV) or a duty to abstain (DTA). Specifically, I argue that under conditions of corruption the norm of electoral accountability may override peoples’ sense of DTV in which case they instead feel a DTA. This context sensitive account is contrasted with a Kantian account of civic duty according to which eligible voters feel a duty to always vote, regardless of contextual factors. The empirical results provides tentative support for the claim that corruption not only decreases eligible voters’ sense of DTV but also increases their sense of DTA. This thesis thus contributes not only to the advancement of the conceptualisation of civic duty in relation to voter turnout, but its results also has important implications for how the rational choice perspective approaches the cost/benefit analysis commonly associated with the voting decision. In the latter case these results indicate that abstainers too may act out of duty and can therefore be assumed to gain positive utility from their abstention. However, the possibility that abstention (just as voting) yields unique costs and benefits has to my knowledge never been acknowledged in the rational choice literature on voter turnout. I therefore conclude by presenting a novel suggestion of how the potential costs and benefits of abstention can be incorporated into the calculus of voting.