Political Inclusion of Latino Immigrants in the United States: The Limited Effects of Naturalization on Political Participation

Continuing our new series of collaborations with political science journals, we are pleased to present the following guest post from University of Georgia political scientist Ines Levin discussing her article “Political Inclusion of Latino Immigrants: Becoming a Citizen and Political Participation” that appears in the current issue of American Politics Research.  In conjunction with this post, SAGE will make the article freely available to all for the next 30 days; you can download it here.


The ongoing debate on the Senate immigration bill raises important questions for political science. In particular, what are the consequences of providing a path to citizenship for the economic and political advancement of immigrants? Just as it is often taken for granted that legal status and citizenship contribute to economic mobility, the notion that acquisition of citizenship contributes to the political incorporation of immigrants is often taken as a self-evident truth. Acquiring citizenship is certainly necessary for the political inclusion of immigrants, since it is the only way for individuals to gain access to basic forms of political participation in a democracy, such voting and contributing to electoral campaigns. But is citizenship sufficient for ensuring broad political incorporation?  If this were the case, one would expect naturalized immigrants to be more deeply engaged than their non-citizen counterparts when it comes to non-electoral community and political activities open to all immigrants. These questions are addressed in my current article in American Politics Research.

Levin - figure 1

A naïve comparison of the behavior of naturalized and non-naturalized immigrants based on data from the 2006 Latino National Survey indicates that acquisition of citizenship might indeed stimulate involvement beyond the ballot box. While both naturalized and non-naturalized Latino immigrants are very likely to say that they would work with others through groups or organizations to deal with issues that need to be addressed, the naturalized are more likely to report that they participated in the activities of a group, tried to contact a government official, or volunteered at their child’s school. In particular, the naturalized are more likely to say that they participated in the activities of groups including non-Latino members or that they contacted non-Latino officials relative to their non-naturalized counterparts.

Levin-Figure 2

But the unequal involvement of naturalized and non-naturalized immigrants is not necessarily caused by differences in citizenship status. Immigrants who have and have-not acquired citizenship come from different countries (and also from different regions within the home country, in the case of Mexican immigrants), settle in different areas of the U.S., and have different levels of pre-immigration political involvement. Both types of immigrants also have unequal access to politically relevant resources long thought to affect political participation such as educational attainment, income, and English language skills. Moreover, the naturalized have usually spent more years in U.S. (and in their current homes) and are typically older. Differences like these might explain both observed inequalities in non-electoral participation and differences in citizenship status. Indeed, the probability of acquiring citizenship conditional on the above-mentioned attributes is considerably lower among the non-naturalized than among the naturalized, suggesting that the process of assignment to citizenship status is anything but random.

Levin - Figure 3

One way to address the question of whether differences in political engagement are driven by citizenship itself, or by differences between the kind of people who have naturalized and those who have not, is to “match” the research subjects on measures such as length of residence. After using propensity score matching to control for observable differences between naturalized and non-naturalized immigrants, inequalities in political participation are greatly reduced. While the naturalized are still more likely to participate in most non-electoral activities than similar non-citizens, differences are small and no longer significant. The exception is contacting government officials, although a sensitivity analysis indicates that differences in propensity to contact are highly sensitive to bias that could have been caused by differences in unobserved factors between both immigrant types.

Levin - Figure 4

In sum, when one compares similar naturalized and non-naturalized immigrants, the resulting evidence lends little support to the hypothesis that naturalization leads to greater political involvement. Although it seems intuitive that going through the naturalization process should lead to greater attachment to the American identity and principles – including a sense of the importance of fulfilling one’s civic duty – a number of factors might reduce the effectiveness of naturalization for making more involved citizens, including: absence of civic infrastructure in settlement areas, experiences of discrimination, and inadequate access to resources. As long as immigrants are exposed to some of these limiting factors, the provision of a path to citizenship might not be enough to ensure that the naturalization process produces civically engaged citizens who are capable of bearing participation costs.

Note: I thank Sean Ingham and Alex Street for their very useful comments on this blog post.

4 Responses to Political Inclusion of Latino Immigrants in the United States: The Limited Effects of Naturalization on Political Participation

  1. RobC August 16, 2013 at 4:02 am #

    Although the free download link goes to the wrong article, on the basis of what Professor Levin has written here, the use of propensity score matching appears to be a valuable technique for analyzing these issues. In fact, even propensity score matching “to control for observable differences” probably doesn’t go far enough, because it is unable to match for characteristics that are not easily quantifiable. Specifically, the process of becoming a naturalized citizen requires a desire to engage in a civic activity and a considerable amount of initiative and follow-through to achieve that goal. Those qualities may not be equally present in the sample of non-naturalized immigrants, and it’s reasonable to imagine they may be generalizable to how likely someone is to engage civic activities other than attaining citizenship. Note that my point is not that going through the naturalization process leads to a greater inclination to fulfill one’s civic duty, but rather that people who choose to go through the naturalization process very likely already have that inclination.

    • John Sides August 16, 2013 at 10:12 am #

      RobC: I fixed the link. Thanks for letting us know. You’ve hit on the central challenge of using “observational” (i.e., not experimental) data to do causal inferences.

  2. Andrew Gelman August 16, 2013 at 3:44 pm #

    Nice graphs.

  3. John H. August 19, 2013 at 3:46 pm #

    It’s cool to see a case where p-score matching *works*, that is balances the sample characteristics after matching. Very nice!

    But to me this seems like a case where the available outcomes don’t directly speak to the research question, or perhaps demand a different interpretation. I’m not sure dealing with other people’s issues, volunteering at school, or participating in groups gets at non-electoral forms of political participation, as opposed to volunteerism or social activity. And contacting govt officials seems to be significantly different after matching, at least given the selection assumptions.

    So why don’t we interpret this as a confirmation of the theory? Irrelevant outcomes (i.e., interacting with people) are unaffected, while relevant ones (i.e., contacting political officials) are positively nudged?