Stephen M. Swindle, "The Supply and Demand of the Personal Vote: Theoretical Considerations and Empirical Implications of Collective Electoral Incentives," Party Politics, 8 (May 2002), 279-300.

First Paragraph:
By the very nature of democratic elections, representation occurs through individual candidates who, in turn, almost always belong to a political party. When entering a voting booth, a voter is faced with two primary considerations: who the specific candidates are and, simultaneously, the party to which those candidates belong. On the one hand, the voter may have an affinity for an individual candidate, and therefore will have the incentive to cast a candidate vote to express that preference. On the other hand, the voter may be concerned that the party he or she supports gains as many seats in the legislature as possible, and therefore will have the incentive to cast a party vote to express that preference.

Figures and Tables:
Table 1. An ordinal scoring system of electoral systems
Table 2. Nominating errors of large parties in Ireland and Japan
Table 3. Cost of nominating errors
Figure 1. Average standard deviation of co-partisan votes, by election
Table 4. Logistic regression of vote dispersion on electoral performance (dependent variable: SEATLOSS)

Last Paragraph:
The importance of this distinction between candidate-driven and party-directed personalism should not be minimized. For although ultimately voters may be receiving the same types of goods and services, the specific motivations for the provision of those goods and services and the consequences of their provision can be dramatically different. When personalism is candidate-driven the primary underlying motivation for its provision is its perceived ability to provide individual candidates with electoral advantage. That is, personalism of this type is seen as a tool to mobilize a personal electoral reputation separate from what the party label can provide and that will ultimately allow individual candidates to effectively differentiate themselves from their competitors (including co-partisans). When personalism is party-directed, however, considerations of individual electoral reward are secondary to the collective electoral objectives of the party. This does not necessarily imply that there are no electoral advantages to providing personalistic goods and services, but only that these electoral rewards are partyspecific rather than candidate-specific. To think of these types of personalism synonymously, therefore, obscures important theoretical and substantive differences that should be recognized.