Emigrants and the Body Politic Left Behind: Results from the Latino National Survey

Abstract

There is a duality at the heart of the migration phenomenon, as the very same people who are immigrants are also emigrants, making a living and possibly setting down roots in the receiving society, but still connected to and oriented toward the home society where their significant others often reside. While research has shown that home-country political conditions and experiences affect immigrant political behaviour in the receiving society, scholarship has yet to ask how those same factors affect the ways in which emigrants relate to the body politic left behind. This paper seeks to fill that lacuna. We find that pre-migration political experiences impart a lasting post-migration interest in home-country politics and that such effects are substantial compared with the impacts associated with other cross-border connections, such as remittance sending or return travel.

Keywords:

• Emigrants
• Transnationalism
• Diaspora
• Cross-Border Politics
• Migrant Political Engagement

Acknowledgements

This research was supported by grants from the Russell Sage (#88-08-05) and National Science (SES-0751944) Foundations and a fellowship to Roger Waldinger from the John Simon Guggenheim Memorial Foundation. Support for Thomas Soehl was provided by the UCLA Graduate Division, a pre-doctoral advanced quantitative methodology training grant (#R305B080016) awarded to UCLA by the Institute of Education Sciences of the US Department of Education, and the ZEIT-Stiftung Ebelin und Gerd Bucerius. Responsibility for conclusions reported in this paper is that of the authors alone.

Notes

1. To ensure that our results were not an artifact of the estimation technique, we also ran the regressions using a three-category variable, collapsing each of the two agree and disagree responses into separate categories each, leaving ‘Haven't you thought about it?’ as the third response option. A multinomial regression produced results substantively similar to those produced by the binomial logistic regression reported on in the paper.

2. Entering the variable as a set of dummies does not substantively alter our results but produces a more unwieldy table.

3. To assess the fit of the models we calculated predicted values for each observation and compared them to the observed dependent variable. For the logistic regression model on having an opinion about expatriate voting we follow convention and use p=0.5 as the cut-off, meaning each individual with a produced probability greater than 0.5 is classified as having an opinion and those with p < 0.5 are classified as not having an opinion. By this measure our model correctly classifies 76 per cent of the observations. For ordered logistic regression models there is no straightforward way to assess model fit. Calculating predicted probabilities for each outcome and classifying individuals by their model predicted probability gives us a correct classification rate of 43 per cent. The polychoric correlation between the categorised and observed outcomes is 0.47 for the ordered logistic model and 0.41 for the logistic regression model, indicating a good fit.