Deportation threat and political engagement among latinos in the Rio Grande Valley

ABSTRACT

How does the threat of deportation affect Latinos’ political engagement and participation? Little scholarship systematically analyzes deportation effects upon Latino political engagement. This article explores how the threat of deportation raised under the Trump administration affects cognitive and electoral behaviour in politics among Latinos in the Rio Grande Valley (RGV). By utilizing original data from an RGV public opinion survey conducted in 2018, we examine how deportation threat, measured as knowing a deportee or detainee as well as worrying about deportability affects democratic engagement and political activism of Latinos in the RGV. We find the fear of being deported depresses the level of attention to politics, but increases the frequency of discussion. Also, it discourages voting. Our findings identify a potential hurdle of political behaviour among Latinos in the U.S.-Mexico border region which has been directly affected by immigration policy threat under the Trump administration.

KEYWORDS:

• Threat
• deportation
• detention
• Latino/a political engagement and participation
• Rio Grande Valley
• Trump administration

Acknowledgments

The authors would like to thank Dr. Laryssa Mykyta, the previous co-director of the Center for Survey Research and Policy Analysis at the University of Texas Rio Grande Valley, for her advice and expertise in designing and fielding our survey.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 see, https://www.census.gov/

2 The RGV Sector has nine stations and also includes monitoring by air and sea (https://www.cbp.gov/border-security/along-us-borders/border-patrol-sectors).

3 In 2019, in the Hidalgo port of entry, 8,267,555 people entered through personal vehicles, and 276,103 through buses while another 2,347,699 entered on foot (see, https://www.bts.gov/content/border-crossingentry-data).

4 In January 2019, former President Trump declared a security and humanitarian crisis on the border in South Texas and visited McAllen, a city in the RGV (Jervis 2019).

5 Conceptually, our expectation could be viewed as Latino linked fate. As discussed by Vargas, Sanchez, and Valdez (2017), Latinos living in mixed status families or who may have undocumented friends and acquaintances recognize that they are affected by punitive immigration policies. Although we did not include measures of linked fate in our original survey, we will include them in future iterations of our survey.

6 Based on random-digit-dialing (RDD), we obtained a random sample that approximates demographics of the local communities. According to U.S. Census, Hidalgo County has about 64 per cent of the entire population of the Rio Grande Valley, followed by Cameron (31 per cent), and Starr (5 per cent). In our sample, Hidalgo has about 58 per cent of all responses, followed by Cameron (28 per cent), and Starr (15 per cent). We see that Starr County has been slightly over-represented. In terms of Hispanic ratio, our sample tightly approximate the racial profile of the community. Our sample has about 91 per cent Hispanics while the Census data says these three counties have about 94 per cent Hispanics. From the US Census data, we created a population weight variable and re-estimated the models. The results, however, do not change the conclusion of this analysis. The results were presented in the Appendix (Table A.6).

7 255 observations for the household income variable were missing. This could impose a serious challenge to our analysis as the reduced sample could introduce the potential self-selection bias. As related variables, such as education or ideology, also contain a rather high volume of missing observations (146 for education and 125 for ideology), however, we forgo the multiple imputation approach. Instead, we created an income dummy variable for the highest household income category ($50,000 or more) and ran the analysis to address the issue (see Appendix Table A.7). The results are the same as those reported in the main analysis of this study.

8 The electoral participation variables included in the survey are voting in the 2016 presidential election and voting in the 2018 primary elections. As this article examines how the increased deportation threat under the Trump administration after his election in 2016 affected Latino electoral participation, the 2016 presidential election variable loses its validity. For this reason, participation in the 2018 primary elections was used in the analysis. However, we re-estimated the voting models with the 2016 election variable for a robustness check. The reanalysis does not change the results.

9 Those responses include “Not a U.S. citizen”, “Permanent resident”, “Not allowed to vote”, “Felony”, “Not qualified”, “Immigration status”, etc.

10 The survey did not include a citizenship question due to a concern that this question could be intrusive for the RGV population. We acknowledge that this is not as ideal as having a citizenship question to filter out ineligible voters from the analysis. As a result, our sample used in the analysis might still include some unidentifiable non-citizens and the results we found could have been driven by this group of respondents. To address this potential concern, we ran a separate analysis with a subsample of those who were born in the U.S. See Appendix Table A.9. Also, we ran the entire models again using a subsample of eligible voters that we identified to see if ineligible voters (non-citizens) drive the regression analysis for paying attention to politics and having political discussion besides voting. See Appendix Table A.8 for the results. They do not change the findings reported in the main analysis of this study.

11 Although the RGV is a predominantly Hispanic region, there are variations among Latinos in the community in terms of their Spanish- versus English-fluency. Therefore, it is crucial to control for the effect of the use of Spanish. In our sample, 39 per cent of the respondents were interviewed in Spanish.

12 It is possible that political engagement variables might create a collinearity problem. We re-estimate the models by omitting other engagement variable from each equation. The results do not change. See the Appendix A.5.

13 Not on voting because it is not intuitive to assume that voting would increase deportation threat.

Additional information

Funding

No funding was reported by the authors