40.7 percent. Voter registration also increases slightly. The patterns are similar for the
Salvadoran/Guatemalans in terms of community civic and political life and voter registration.
U.S.-born Salvadorans/Guatemalans are somewhat less likely to be members of community
organizations than are the non-naturalized or the naturalized Salvadorans/Guatemalans.
[Tables 2 and 3 approximately here]
The analysis presented here does not include the most common form of political activity
among citizens in the United States – voting. We are in the process of validating self-reported
voting and will add a test of voting propensity to this analysis once we have completed that
validation.
Modeling 1.5 and 2
nd
Generation Civic and Political Behavior
We test the impact of parental unauthorized migration and naturalization in multivariate
logistic regression models that include the standard predictors of political behavior in Latino
communities (Wolfinger and Rosenstone 1980; Calvo and Rosenstone 1989; DeSipio 1996a;
Shaw, de la Garza, and Lee 2000; Pantoja, Ramírez, and Segura 2000; Barreto, Segura, and
Woods 2004). The models include several respondent demographic characteristics, a measure of
political recruitment, a measure of relative political connectedness to the United States and to the
country of origin/ancestry, and two measures of immigrant acculturation.
The demographic measures are age, years of education, household income, number of
months at the current address, and gender. Across all racial and ethnic populations in the United
States, increasing age and increasing levels of formal education generally predict increasing
political engagement (Wolfinger and Rosenstone 1980; Verba, Schlozman, and Brady 1990).
also leads to higher levels of political engagement for the population as a
4
As with many surveys, a significant share of IIMMLA respondents refused to provide information on their
personal or household incomes. In order to maintain these respondents in the analysis, we include income as a
categorical measure in model and include “refused” as a category. While we would assume that higher income
respondents were somewhat more likely than lower income respondents to refuse to disclose their household
incomes, we do not believe that the refused category is a reliable category for analysis.
9
Convention