Hypothesis Based on TX-18 Demographics

Abrajano and Hajnal discussed in White Backlash (2015) that Whites who live in areas with higher immigrant populations are typically more conservative, vote Democrat less, are less supportive of general social welfare expansion, and more punitive of illegal immigrants. Based on this theory and the demographics of TX-18 – namely, that it has a large foreign born population (23% compared to the general US foreign born population of 13.5%) – I predict that TX-18 will be emblematic of this theory. To test this, I will compare the relevant views of Whites in TX-18 to those in another congressional district in Texas that has a low foreign born population. I choose to compare TX-18 (which comprises much of Houston) to TX-12 (which comprises much of Fort Worth); I made this decision because the districts were similar on covariates (besides the Fort Worth district having higher median incomes) yet had starkly different foreign born population sizes.

Data Collection and Methods

To test my prediction I need a big dataset that asks election/politics style questions to a large and representative audience; the CCES (Cooperative Congressional Election Study) dataverse is a great choice for this. From this dataset I pulled a total of eight variables that included relevant ID variables, the Respondent’s race, and four variables that directly related to the four public opinion focuses I am interested in exploring, given Abrajano and Hajnal’s theory, which includes the Respondent’s vote in the 2016 presidential election, their view on the expansion of social welfare, their view on finding and deporting illegal immigrants, and their political ideology. To compare TX-18 and TX-12 I will conduct a statistical contrasting of group means between the relevant focus variables of public opinion to try and detect any difference. The appropriate method for such a comparison is a Student’s t-test. This method requires quantitative data; the CCES variables I chose were coded, however, on qualitative scales. For example, the political ideology question has response options such as very conservative, slightly conservative, moderate, slightly liberal, and very liberal. I transformed this instead onto a numeric scale from -2 (very conservative) to 2 (very liberal). I similarly transformed the other variables onto numeric scales or changed them into dummy variables (for example, for the 2016 presidential election vote, I made it a binary variable for if the Respondent voted for the Democrats or not). After this data cleaning and transforming was done, I conducted a Student’s t-test in Rstudio for each of the opinion variables – i.e. I compared the average value for each opinion question between Whites in TX-18 and Whites in TX-12.

Findings

In my analysis I found multiple statistically significant differences between TX-18 and TX-12. Firstly, if we look at political ideology, we see that the average White person in TX-18 is more liberal than their counterpart in TX-12 (a difference of 0.5 points on a scale between -2 and 2), which is a statistically significant difference. Secondly, looking at punitive views concerning illegal immigrants, I found that 34% of Whites in TX-18 agree with finding and deporting illegal immigrants, compared to 56% of Whites in TX-12 (this difference is statistically significant). Thirdly, when observing 2016 voting patterns, I found that 52% of Whites in TX-18 voted Democrat compared to 38% of Whites in TX-12, however, these group values are not statistically distinguishable from one another. Finally, looking at views on social welfare (where the responses range from wishing to reduce welfare to wishing to increase it – on a scale from -1 to 1), I found that Whites in TX-18 were, on average, supportive of expanding welfare, while Whites in TX-12 were, on average, in favor of reducing it (a difference of 0.4 points on the scale, that is statistically significant). Overall, this suggests that the two districts are quite different with respect to public opinion among their White populations.

Discussion

My results are the opposite of what I had originally predicted. Based on Abrajano and Hajnal’s theory, TX-18 should have a White population that is more conservative, more punitive of illegal immigrants, more supportive of reducing social welfare, and votes Democrat less than the Whites in TX-12 – since TX-18 has a much higher foreign born population. However, as discussed, I found the opposite. Besides the voting habits between the two districts, which didn’t carry any statistical significance, TX-18 is exemplary of more liberal positions than TX-12. These findings suggest that perhaps Abrajano and Hajnal’s theory does not apply to TX-18, or perhaps even to urban congressional districts in the Southern, and more conservative, states in general. Further exploration into Southern locales would need to be done to see where Abrajano and Hajnal’s theorem holds regarding Whites’ views that are supposedly dependent on the size of the local foreign born population. One significant limitation of my study is that the CCES dataset only contained 32 observations after I restricted for Whites in TX-18 and 102 observations for those in TX-12. This calls into question the generalizability of my findings to the broader White population of TX-18 and TX-12 and, thus, the difference between the two. Further work would do well to find data sources with more observations in order to ensure the statistical rigor of further studies. In conclusion, White public opinion in  TX-18 appears to be inconsistent with extant scholarship – instead of being more conservative and anti-immigrant they actually demonstrate more liberal and pro-immigrant views, as compared to Whites in an area with a lower foreign born population. This finding bodes well for the state of immigrant affairs and immigration more broadly in TX-18.

TX-18 slides (see these for corresponding slides to below descriptions)

Slide 1: Demographics

Between 2007 and 2017, there has been little, virtually unrecognizable, change in the demographic composition of Texas’ 18th Congressional District. The racial breakdown of the area has remained constant with it being predominantly made up of Hispanic/Latinos and African Americans. The foreign born population of the district has grown as the overall population has grown, but its proportion of the whole has remained largely unchanged. The foreign born population has been consistently made up of roughly 80% Hispanic/Latino origin individuals.

Slide 2: Existing research predicting outcomes

Extant research suggests that sudden growth of the immigrant or Hispanic population is associated with anti-immigrant views among current residents (this includes Whites as well as other racial minorities like African Americans) particularly when coupled with the issue of immigration being one that is salient at the national level which, in turn, causes people to notice and politicize changes in their area (Hopkins, 2010; Craig and Richeson, 2017). Enos (2014) goes further to suggest that even minor changes in immigrant demographics lead to exclusionary attitudes among residents. Large populations of immigrants or Latinos in general (not necessarily spikes in their numbers) are associated with Whites holding more punitive views of immigrants, being less supportive of social welfare programs, identifying as more conservative, and voting for Democrats less (Abrajano and Hajnal, 2015). Such conservative views are correlated with places than have large foreign born populations and with respondents who are White and identify as Republican, are older, and typically have lower education levels (Dunaway et al, 2010). The presence of an immigrant population is also associated with the existence of immigrant advocacy organizations who do extensive work to integrate immigrants into political and civic life, provide legal assistance, and positively influence local policy and programs concerning immigrants in local government (de Graauw, 2008; Andersen, 2008). In a similarly positive vein, immigrant populations promote Spanish news outlets which are more positive and informative about immigration, which engenders more positive views among their audience and is beneficial to the community of Spanish-speaking immigrants and Latinos (Abrajano and Singh, 2008). Conversely, mainstream news outlets typically house much anti-immigrant sentiment, particularly those in border states, which spurs similar tones in their audience and is ultimately bad for immigrants (Branton, 2009). Ultimately, these outcomes present a mixed bag of positive and negative outcomes regarding immigrant experiences, media coverage, and public opinion.

Slide 3: Mapping outcomes onto TX-18

Given the presented research on the impacts of immigrant demographics, and the anticipation that these would present a varied assortment of outcomes, I predict a few contrasting results for TX-18. Regarding immigrant experiences, there are a wide assortment of advocacy organizations in Houston that lead me to believe that immigrants will integrate well politically in TX-18. Texas, and thus Houston, is also home to one of the largest assortments of Spanish news-media in the country (Martinez, 2016); this will lead to a high amount of more pro-immigrant sentiment which is good in and of itself but also supportive for the existing immigrant population. Regarding media coverage specifically, beyond Spanish news, since Houston is only 350 miles away from the Southern border there will likely be larger amounts of immigrant-related stories, which will be mainly negative. Regarding public opinion, since there has been no influxes of foreign born or Latino individuals, the sentiment around immigrants should have been fairly consistent over the last 10 years or so. Further, due to the continuously large Latino and foreign born populations (far above US average), these opinions are likely to be conservative and somewhat hostile towards immigrants among Whites and African Americans in TX-18. These unsympathetic views and the negative mainstream media coverage will also feed into a negative immigrant experience. These predictions, overall, stack up to paint a depressing situation for immigrants in TX-18 with their main solace being found in advocacy groups, Spanish media, and other immigrants or residents of Latino origin.

Slide 4: Research plan

I will test my hypothesis about public opinion. To do this I shall look at public opinion data to explore the prediction stemming from Hopkins work that, since the foreign born and Latino/Hispanic population has remained relatively unchanged over the past 10 years or so (2007-2017), negative attitudes towards immigrants and Latinos/Hispanics should have stayed relatively constant in that time period too. To do this I will gather public opinion data from the reputed ANES Time Series data (a longitudinal study of public opinion) from 2008, 2012, and 2016, for TX-18 congressional district. The relevant questions in this survey include: a feelings thermometer towards illegal immigrants; a feelings thermometer towards Hispanics; and the respondent’s views on immigration policy regarding citizenship, government checks, and overall policy preferences. I will then explore the bi-variate relationships between responses to these questions and the demographics of the respondents (race, age, party, education level). Following this initial analysis I will regress the question responses on the demographics aforementioned to identify any significant relationships. This will allow me to conclude whether or not the expected respondents hold conservative views, as the prior research suggests, and if these views have changed over time.