“It could be 3 million, it could be 30 million”: Quantitative misperceptions about undocumented immigration and immigration attitudes in the Trump era

Abstract

Recent changes in the sociopolitical US landscape calls for the examination of the level of quantitative misperception about undocumented immigration and its connection with immigration attitudes. Nationally representative survey data are used to analyze whether being misinformed about the proportion of US immigrants that are undocumented in 2015 is linked with abstract immigration attitudes and four immigration policy options in 2016. The results reveal that people who overestimated undocumented immigration—a common misperception—are more likely to report that all immigrants present symbolic threats to the country than are their accurately informed peers. Consistent with the especially high salience of the US–Mexico wall in this period, overestimators also place more importance on building the wall but not on other policy options. These findings have important theoretical and real-world implications, given the current social and political context and spillover effects on Latinx and other racialized communities.

Resumen

Cambios recientes en el panorama sociopolítico de los Estados Unidos exigen un análisis del nivel de percepciones cuantitativas erróneas sobre la inmigración indocumentada y su relación con las actitudes hacia la migración. Se utilizaron datos de encuestas representativas de la nación para analizar si el hecho de estar mal informado sobre la proporción de los inmigrantes de EE.UU. con estatus indocumentado en 2015 se vincula a las actitudes migratorias abstractas y a cuatro opciones de política migratoria en 2016. Los resultados revelan que las personas que sobreestimaron la inmigración indocumentada (una percepción errónea común) tienen más probabilidades que sus pares mejor informados de reportar que todos los inmigrantes representan riesgos simbólicos para el país. Las personas que sobrestimaron también dan mayor importancia a la construcción del muro entre EE.UU. y México pero no a otras opciones de políticas, lo cual es congruente con la prominencia especialmente del muro durante ese periodo. Estos hallazgos tienen implicaciones importantes teóricas y del mundo real, dado el contexto sociopolítico actual y los efectos colaterales en la comunidad latina y otras poblaciones racializadas.

Appendix

Analyses of Trump’s rhetoric and attention to immigration policy options

A research assistant and the author carried out keyword searches of Trump’s speeches and Trump’s Twitter to examine whether Trump focused different levels of attention on the four immigration policy issues examined in this study. Factba.se provides a searchable database of Trump’s video recorded speeches, providing both video and transcripts.¹⁹ The corpus was created by limiting searches to those dated between January 1, 2015 and December 31, 2016, and to those using “video” as the media format (rather than text or social media) and “speech” as the type (rather than interviews, press conferences, or video logs). The database does not yield individual video recorded speeches matching the keyword search; what is reported are what Factba.se labels “references,” a several-sentence-long paragraph-length segment of Trump’s speech, with a transcript and a link to the recorded video. As these are segments of a longer speech, a term might be mentioned several times, leading to Factba.se yielding multiple segments that are part of the same speech. It is not possible to easily download the keyword searches from Factba.se into Excel to count only the number of speeches rather than speech segments. Although Factba.se does not indicate how many individual speeches or speech segments are in the database for 2015–2016, it is possible to approximate a denominator for the “references”/speech segments. For example, a keyword search of Factba.se for how many speech segments occur in which Trump uses the word “I” shows 37,578 references in 2015–2016, with 5578 speech segments in 2015 and 32,000 segments in 2016.²⁰

Table 4 identifies the search terms and common examples of segments from video recorded speeches in 2015–2016, the same years of the ATP survey data analyzed in the study. The search terms tap into the same four specific immigration policy priorities covered in the ATP survey and the outcome variables in the multivariate analyses.²¹ The counts regarding the “Wall” and deport* (hereafter deportations), were relatively straightforward to collect. For more extensive searches and keywords used to identify Trump’s attention on and rhetoric about establishing a way for undocumented immigrants to legally remain and allowing undocumented youth to stay and apply for legal status, see Table 4. The term “amnesty” was one of the more common ways that he referred to whether undocumented immigrants would be allowed to stay, with keywords such as “dreamers” used to refer to undocumented youth much less commonly, as identified in an endnote in the manuscript. Speech segments identified in the keyword searches for undocumented immigrants/youth allowed to remain were read to examine whether Trump indicated support for these options.

Table 4 Description of keyword search of recorded video speech segments in Factba.se, January 1, 2015–December 31, 2016

The Trump Twitter Archive provides a searchable list of Trump’s tweets and retweets.²² The corpus was created by limiting searches to those items dated between 1 January 2015 and 31 December 2016 using the same keywords as used for the analysis of Factba.se. The archive indicates 11,749 tweets or retweets on his account in 2015–2016. Table 5 provides more information about the search terms and examples from the analysis of Trump’s Twitter.

Table 5 Description of keyword search of Trump’s tweets and retweets in the Trump Twitter Archive, January 1, 2015–December 31, 2016

Figures 1 and 2 report the raw results of these keyword counts from Trump’s recorded speeches and from his Twitter in 2015 and 2016. To identify whether keywords such as the “wall” were discussed more frequently than “deportations” or “amnesty,” χ² tests were carried out that compare the proportions from Factba.se and the Trump Twitter Archive separately. The numerators for the proportions are the absolute size number of segments or tweets/retweets and the denominators are approximated for Factba.se (37,578 speech segments in 2015–2016) and are exact for Trump’s Twitter (11,749 tweets/retweets in 2015–2016). The results for the χ² tests are provided in the manuscript.

Fig. 1

Keyword search results in Trump speech segments, 2015–2016

Fig. 2

Keyword search results in Trump’s Twitter, 2015–2016

Analyses of the relationship between quantitative misperceptions about undocumented immigration and immigration attitudes

Tables 6 and 7 provide the full set of regression results for the analyses presented in Tables 2 and 3.²³ The full set of results are consistent with the previous literature on immigration attitudes. Aside from the variables tapping into misperceptions about the proportion of immigrants that are undocumented younger respondents, those with more education, women, and less conservative people are less likely to report that immigrants threaten American culture (Table 7) and to place more importance on building a wall or increasing deportations than their counterparts without these characteristics. Such people also are more likely to place more importance on finding ways for undocumented immigrants and undocumented youth to legally remain (Table 7). Note that the models for building a wall and increasing deportations are coded in the direction of reflecting more exclusionary policy positions (Models A and B), while the opposite is true for the models regarding undocumented immigrants and undocumented immigrant youth able to legally remain (Models C and D). The results also point to residual racial/ethnic differences between Hispanics and others, such as non-Hispanic Whites and, in some cases, African Americans and Asians, as well. Finally, respondents with higher values on the racial prejudice indicator are more likely to report that immigrants threaten American customs and values and attribute higher importance to building a wall and increasing deportations. National sociotropic concern about immigrants’ negative impacts on the economy is strongly associated with all five immigration attitudes.

Table 6 Logistic regression of immigrants threaten traditional American customs and values in 2016 on perceptions about the relative size of undocumented immigration and other variables in 2015

Table 7 Ordinary least squares regression of immigration policy preferences in 2016 on perceptions about the relative size of undocumented immigration and other variables in 2015