SAN FRANCISCO (CN) - President Donald Trump's fiery comments and immigration policies took center stage Wednesday on the second full day of a federal bench trial over the push to add a citizenship question to the 2020 Census.
Speaking from the witness stand, UCLA professor Matthew Barreto explained that comments like those made in Trump's Tuesday night speech serve only to intensify a climate of fear and distrust among immigrants and Latinos toward the federal government.
"We saw this reflected in the statements made last night," Barreto said.
In his border wall speech Tuesday, Trump highlighted the number of undocumented immigrants who committed "violent killings" and "sex crimes" while stressing the need to stop an influx of illegal drugs from crossing the border, even though most seized narcotics are found at official ports of entry.
The combination of incendiary words and policies perceived as anti-immigrant, including terminating protected status for Dreamers and other groups, creates an unwelcoming "macro environment" that makes non-citizens, their families, friends and neighbors less likely to participate in the 2020 Census, Barreto insisted.
The UCLA professor, who specializes in survey research and polling with a focus on Latinos, concluded based on a review of academic research that adding a citizenship question in a hostile environment will inflict maximum harm on participation rates and accuracy for the decennial count.
"When the question that is sensitive is directly related to a perceived threatening environment, that is when participation suffers the most," Barreto said.
Barreto was the second witness called by California and six cities, including Los Angeles and San Jose, who are challenging U.S. Commerce Secretary Wilbur Ross's March 2018 decision to add a citizenship question to the 2020 Census.
California claims the question will discourage participation in a state where Latinos make up nearly 40 percent of the population. An undercount could lead to a reduction in federal funds and congressional representation for the state and its cities.
The Trump administration argues the citizenship question is needed to provide essential data for voting rights enforcement.
Barreto designed a nationwide survey for more than 6,000 randomly selected people, which resulted in 3,609 completed phone interviews. The UCLA professor had interviewers ask if people planned to participate in the census and if a citizenship question would change their mind. The survey found a higher rate of people in California said they would not participate if a citizenship question were asked.
People were also asked if they trusted the Trump administration to protect their personal information. The survey found 56.9 percent of Latinos in California and 47.3 percent of Latinos nationwide said they would not trust the administration to keep their information confidential.
On cross examination, Justice Department lawyer Kate Bailey asked why Barreto used "Trump administration" and "federal government" instead of "Census Bureau" in his survey questions. She also asked why he didn't test to see if responses would vary without the words "Trump administration."
Barreto replied that "Census Bureau" was referenced in the first question of the survey, and that the follow-up questions were worded precisely as intended.
"At the time of your survey, wasn't the president's approval rating in the low 40s, and wasn't his approval rating 10 percent lower among Latinos," Bailey asked rhetorically.
Bailey also pointed out that Barreto counted those who did not respond to certain questions or replied "I don't know" or "maybe" as negative responses, suggesting that the number of people who said they would not participate due to the citizenship question was inflated.
Barreto said counting a non-response as "no" is standard practice and interviewers were trained to press respondents to answer before counting them as "not willing to participate."
Finally, Bailey pressed Barreto on whether the hostile "macro environment" created by Trump that he described would by itself suppress census responses even without a citizenship question.
Barreto acknowledged that the current political climate would present challenges, but he insisted that combining the unfriendly atmosphere with a citizenship question was the most effective way to suppress participation and harm the accuracy of the decennial count.
The trial is expected to continue through early next week.
This is the second federal trial on legal challenges against the census citizenship question. Another bench trial ended in New York this past November and a decision is still pending in that case.