Search

In 2001 Republican Latino candidate for mayor Orlando Sanchez won 73% of the heavily Democratic Latino vote in Houston while in Miami, Cuban Democrat Manny Diaz won 70% of the staunch Republican Latino vote. In 2003, Green Party candidate for mayor Matt Gonzalez captured 66% of the Latino vote in San Francisco while in Colorado Springs a majority of Latinos voted for Republican Lionel Rivera. These elections renew the debate over Latino vote preference, and call into question whether Latinos follow partisan cues or ethnic cues when casting a ballot. I argue that while party is still an important predictor of vote choice, for Latinos with a high degree of shared ethnicity, party ties are less significant than ethnic attachment. Using a unique survey fielded by the Tomás Rivera Policy Institute in 2002, I model crossover vote preference for Latino candidates in California and New York using probit and ordered probit regression. While strong partisans were significantly less likely to crossover and vote for the ethnic candidate of the opposing party, Latino voters who scored high on ethnic attachment were significantly more likely to ditch their party for the Latino candidate. By incorporating more flexible measures of ethnicity (as opposed to a dichotomous measure), this paper finds that, in addition to SES and party, ethnicity can be an important predictor of candidate preference. [ABSTRACT FROM AUTHOR]

The effect of ethnicity upon political participation presents an interesting paradox to political scientists. No consensus exists on the extent to which the interaction between contextual factors and individual level characteristics allow individuals to overcome barriers to participation, or how this relates to patterns of mobilization. While it is well documented that contacts and mobilization are effective in turning out the vote, few scholars have asked the question of who gets mobilized, and why? In this paper, we use a modified model of the effects of group size and group empowerment for Latinos in the United States to predict which Latinos reported being mobilized prior to the 2000 Presidential election. The modifications include a departure from an emphasis on the presence of Black or Latino mayors as a measurement of empowerment. Instead, we propose using a multiple-level Latino representation model of empowerment. Using a five state survey of registered Latino voters collected after the 2000 elections, we test this alternative model of empowerment to determine if Latino registered voters, living in districts with Latino representatives, are more likely to be contacted and asked to vote. Further modifications include making a distinction between type of contact and the addition of other characteristics that have been found to significantly impact Latino participation. The strategic nature of party contacting leads organizations to contact those individuals who are already the individuals most predisposed to participate. Probit regression analysis finds that Latinos represented by co-ethnics are more likely to be mobilized than Latinos represented by White or Black elected officials, providing additional support for the notion that descriptive representation empowers minority communities. However, the relationship is not linear. Latinos with only one Latino representative are the most likely to receive mobilization while those with two or three representatives are less likely. [ABSTRACT FROM AUTHOR]