Two experiments were conducted to test the hypotheses that political candidates who are more (a) credible, (b) likeable, (c) demographic-ally similar, and (d) similar in issue stands to the electorate than their opponents will gain more votes in an election. In Exp. 1 an initial survey of 55 students in general psychology was employed to gain information about students' demographic characteristics, their attitudes about issues, and traits which students felt likeable and credible people possessed. From this information, written descriptions of four pairs of fictional student candidates were produced. The candidate pairs differed in either (a) credibility, (b) likeability, (c) demographic similarity with the electorate, or (d) similarity in issue stands with the electorate. These four descriptions were then distributed in approximately equal numbers among 726 students in general psychology under the guise that an election of a student representative was taking place. The student was asked to vote for one of the two candidates described. Exp. 2 was similar to Exp. 1, except that the initial survey to gain information was conducted among 93 registered voters residing in Indiana County, Pennsylvania, and candidates descriptions requiring a choice of candidate were collected from another 130 Indiana County voters. Statistical analysis of the election results showed that for both college students and registered voters, the more credible, likeable, and attitudinally similar candidates won significantly more votes. Further statistical analysis indicated that for college students, attitudinal similarity was the most powerful determinant of voting, followed by credibility and then likeability. By contrast, no differences were found among the four factors in their ability to determine the votes of registered voters. Finally, candidates' likeability had a significantly greater effect on the votes of registered voters than those of college students.