October 16, 2017
Tweeting rage: How immigration policies can polarize public discourse
A UW study examined tweets at the height of debate over an anti-immigration bill in Arizona.
Before a border wall became a budget bargaining chip, before the presidential pardon of a controversial sheriff and before federal policies were announced on social media, there was Arizona Senate Bill 1070, the “show me your papers” law.
And of course, there was Twitter.
To René D. Flores, an assistant professor of sociology at the University of Washington, Twitter is a trove of insight into people’s beliefs and their willingness to express them. By analyzing tweets in the months before and after the 2010 passage of the controversial Arizona law, Flores found that the average tweet about Mexican immigrants and Hispanics, in general, became more negative. Social media data, Flores found, was useful in determining whether people had changed their attitudes about immigrants as a result of the law or whether they had begun behaving differently.
“The public discourse about immigrants became more negative, but this was not driven by people changing their mind about immigrants: It was driven by people changing the way they acted. Anti-immigrant users tweeted more as a result of the law,” Flores said. “Scholars and journalists were saying there’s more harassment and name-calling of immigrants, and they weren’t wrong about that. But the shift was behavioral, not attitudinal.”
Flores’ research was published in the September issue of the American Journal of Sociology. He recently spoke about his ongoing study of politics for No Jargon, a podcast of the Scholars Strategy Network.
In embarking on his study of SB 1070, Flores set out to determine whether punitive policies can shape people’s attitudes about immigrants. Since there were no available surveys to test this hypothesis, he turned to social media data. He sorted through 250,000 tweets from more than 24,000 users, starting from three months prior to the law’s passage and ending three months afterward. Officially named the Support our Law Enforcement and Safe Neighborhoods Act, the law required immigrants to carry their documents and allowed law enforcement officers to detain anyone suspected of lacking papers. (The U.S. Supreme Court has since struck down several parts of the law, though law enforcement officers still may check an individual’s immigration status during a traffic stop or other police action.)
As part of the UW study, Flores identified a “control” state — neighboring Nevada — with similar politics and demographics to Arizona, including a nearly 30 percent Hispanic population. But in Nevada, there was one key difference: No SB 1070. This allowed him to rule out general trends in public opinion unrelated to the law.
In Arizona, the number of tweets about immigrants jumped from about 3,000 in March 2010 to peak at nearly 18,000 in May, a month after then-Gov. Jan Brewer signed SB 1070 into law. But instead of demonstrating the ways SB 1070 changed what people in Arizona merely thought, Flores said, the tweets showed how it changed the way they acted. Arizonans who’d already expressed anti-immigrant beliefs posted more negative statements, more often.
To gauge the attitudes reflected in tweets, Flores used sentiment analysis, a technique that determines whether a piece of text is positive, negative or neutral based on the kinds of terms it contains. Flores added to the lexicon bank used in the analysis to capture nuances in negativity and positivity, establishing an intensity measurement for each word. The word hate, for example, would score higher on the negative scale than the word skeptical.
Examples of negative tweets identified in the study included “illegal aliens are criminals” and “deport illegal aliens,” among other, more stridently anti-immigrant statements. Positive statements included “in love with immigrants, I love soccer!”, but those were less prevalent.
“If anything, pro-immigrant people became somewhat more silent, while the anti-immigrant people became more mobilized,” Flores said. “It’s almost like the already-converted became more active. When you feel the center is moving, it can become harder to have a minority viewpoint and to express a view that’s different from the consensus. The pro-immigrant side may have felt increasingly uncomfortable.”
Another trend Flores noted: The anti-immigrant statements often extended to tweets about Hispanics, in general, but not about Asian Americans or African Americans. This may reflect a greater association with Hispanics as immigrants, Flores wrote in his study.
Flores’ paper also points out the value of Twitter as a sociological tool. Phone surveys, once the staple of social science research, have been met with more resistance in recent years as people avoid phone solicitations, the time commitment of a survey or a call from a number they don’t recognize. Twitter, on the other hand, is a completely voluntary sharing of one’s thoughts and is fast becoming a valuable historical archive of public opinion, Flores said, which can allow researchers to examine how the public reacted to specific events.
“People are providing data about how they think, which provides kind of a telescope for social scientists. It gives us a window into human expression and behavior we didn’t have in the past,” Flores said.
But it’s not necessarily representative of the population, Flores said. Twitter users — an estimated 328 million are considered monthly active users — are typically younger (often under 30), more likely to live in urban and suburban communities and more likely to be racial or ethnic minorities than Internet users overall. Therefore, it can be harder for researchers to generalize from tweets in the way that, statistically speaking, they can generalize from surveys, Flores said. Depending on the research question, then, tweets might serve more as a complement to, not a substitute for, surveys or other data sources.
The findings from this study, though, highlight a social consequence of a law that targets a specific group of people, Flores said. While such punitive laws may be an attempt to placate a vocal constituency, they also appear to spur mobilization against the targeted group.
Flores’ next project: analyzing whether Donald Trump’s rhetoric toward immigrants (he has described them as “criminals” and “rapists”) is affecting public opinion toward immigration.
“Political polarization is seemingly growing. Some political figures are increasingly using Twitter as a way to spread controversial ideas; fringe groups from both the left and the right are also using Twitter to recruit. It is important for social scientists, I think, to study these new developments both for theoretical but also civic reasons,” Flores said.
###
For more information, contact Flores at renedf@uw.edu or 206-543-9882.
Tag(s): College of Arts & Sciences • Department of Sociology • Rene Flores
