Thread by @KlasfeldReports

Good morning from New York.

The #2020Census trial is about to begin a bit early. The attorneys are introducing themselves, and U.S. District Judge Jesse Furman has taken the bench.

As the SCOTUS denied a last-minute stay, trial ahead. courthousenews.com/taking-up-5-ca…

Note: This is a bench trial.

Furman indicated that he will watch the video of the depositions of John Gore and Earl Comstock, which means that the footage will become public.

Late last night, Gore's deposition became available in transcript form, and it was revealed that Gore confirmed a citizenship question was "not necessary" to enforce the Voting Rights Act, which was Wilbur Ross's justification.

Breakdown by @hansilowang

https://twitter.com/hansilowang/status/1059215777130262533

Furman: "I am a strong believer in public access to judicial documents."

The judge makes clear that he will docket the deposition designations along with the summaries.

The affidavits of the direct testimony will also be docketed.

Since this is a bench trial, we start with witness testimony and no opening statements.

The plaintiffs' first witness is Duke Professor D. Sunshine Hillygus, who has taught there for nine years. sites.duke.edu/hillygus/

Hillygus: "I served for six years on the scientific advisory committee for the Census Bureau."

She also wrote a book called "The Hard Count."

She says that she published dozens of academic articles and an upcoming book with Princeton University Press and testified in a VRA case in North Carolina.

"We have an election tomorrow. I'm a political scientist."

Asked why she chose to testify, Hillygus says she "has a foot" in the various issues involved in the census case.

"The purpose of my testimony is to evaluate the likely impact of a citizenship question on Hispanic and non-citizen households," she says.

Also evaluating Ross...

Hillygus said earlier there's "considerable evidence" that the decision to add the citizenship questions:

* will depress counts.
* was executed in a way against census guidelines.

Now explaining how census works. More on those conclusions later.

Hillygus: "With a very first census, George Washington said: 'We have a census number, but we think there's an undercount.'"

Asked about an overcount, Hillygus gives an example of college students and people with vacation homes being double-counted.

She says there historically has been an undercount of minority populations.

Hillygus describes a category of survey methodology called “burden of the survey.”

"That is the effort that is being asked of the respondent to fill it out," she says, describing such factors as "length," "complexity," "intrusiveness," and "sensitivity of what is being asked."

"When you have a sensitive item, people may be less likely to answer truthfully," she notes.

Hillygus differentiates "privacy" and "confidentiality."

"Privacy is like 'don't ask'; confidentiality is like 'don't tell,'" she says.

Asked about the language of Wilbur Ross's memo claiming the citizenship question poses "no additional imposition" on the survey taker, Hillygus says the statement "contradicts the scientific methodology research" on the issue.

Asked about whether the citizenship question qualifies as a "sensitive question," Hillygus replies: "The census bureau itself designates it as a sensitive question."

She adds there is "empirical evidence" that both non-citizens and Hispanics will be affected.

Hillygus cites public opinion polls, including by Pew, suggesting that Hispanics - regardless of citizenship status - have feared that they or their families will be deported.

The witness also cites "empirical research showing that even people who are not eligible for deportation are failing to use food stamps or ACA," using the shorthand for the Affordable Care Act.

The same is true for SNAP and the nutrition program WIC (Women, Infants, and Children), she says.

Questioning the witness is Arnold & Porter attorney John A. Freedman, who is introducing various exhibits into evidence now.

Freedman describes a focus groups finding that Hispanic respondents were "highly suspicious" about how data will be used against them and their families - regardless of citizenship status.

The study is called Census Barriers Attitudes and Motivators Survey (CBAM).

Hillygus: What’s so significant about this is post-the addition of the citizenship question, this offers one of the only attempts by the Census Bureau to evaluate the impact of the citizenship question.

Hillygus says that it offers "direct evidence" of what could be expected, and that the citizenship question could pose a "major barrier" to participation.

Freedman turns to analysis by John Abowd, the associate director of research and methodology and chief scientist at the Census Bureau.

Hispanics were 9x more likely from non-Hispanic whites to break off from the survey at the point of the citizenship question, Hillygus says.

Furman asks whether this would lead to an undercount.

Hillygus replies this would lead to "lower-quality data" for those households.

Referring to Abowd's research on the negative impact of the question, Hillygus says: "There are reasons that I think this is too conservative an estimate."

She emphasizes that this is Census Bureau research itself predicting depressed counts.

After Hillygus says the citizenship question "sticks out like a sore thumb," Freedman introduces the printed ACS (American Community Survey) form into the record.

10 minute recess.

We're back.

Hillygus are testifying about the Census Bureau's outreach campaign and NRFUO (non-response follow-up operations).

Hillygus: "I think it would be incredibly difficult for this outreach campaign to be successful" from respondents not participating in census over citizenship question.

She says that the "fragmented media environment" makes it more difficult today to get a message out.

Referring again to "non-response follow-up operations," Hillygus says that history shows that these will not make up for undercount if citizenship question added.

"NRFUO did not eliminate undercount in the past. This is an empirical pattern that we have seen in the past."

"This is going to be occurring during a presidential election," Hillygus says, referring to the follow-up operations to the survey.

"The potential for this to be politicized" complicates NRFUO even futher, she adds.

These include "proxy responses," from neighbors as opposed to self-responses.

“For noncitizen households and hispanics, proxy households are probably underestimating respondent size," Hillygus testifies.

Hillygus: The count imputation is contributing to the total population count.

Earlier, she defined imputation procedures as "a guess to fill in the numbers," and that these will undercount Hispanic and non-citizen households.

Furman asks Hillygus whether the imputation procedures are publicly available.

Hillygus responds that they are not.

Hillygus's expert report, via @HansiLoWang

https://twitter.com/hansilowang/status/1059490044111659008

Reading through her report, I'm correcting one of my acronyms.

Non-Response Follow-Up operations are known as NRFU, not NRFUO.

Hillygus notes that she has 175 footnotes in her report.

(Fact-check: She actually has 176.)

Hillygus: "The Census Bureau completely agrees that the addition of the citizenship question is going to reduce the accuracy of the data."

The witness describes the chain of events like this:

* Fewer people self-responding.
* Increase in NRFU
* Increase in proxy respondents
* Increase in imputations

All of which she says reduces accuracy.

Testimony turns to the integrity of the data - referring to the security of the information.

"I would argue that a citizenship question has an impact on the integrity of the data," Hillygus says. "Data breaches happen."

She adds the question increases the risk of "accidental or deliberate disclosure."

Testimony turns to Secretary Ross's memo.

Q: Does Secretary Ross talk about disclosure avoidance in his memo?
A: No.

One method of avoiding disclosure is infusing noise.

Asked if Ross mentions noise infusion, Hillygus responds: No.

Lunch break: Hillygus' testimony continues at 2 p.m.

"All rise."

Judge Furman takes the bench.

"You may be seated," he say.

We're back with more direct testimony from Hillygus. Furman addresses media requests for the deposition videos. He says they should be available to the press and public.

Translation for digital, radio and TV reporters: Roll tape on John Gore. 📺📻📺📻

(I will publish the video and edits once they become public.)

Now back to the testimony.

Hillygus called it "shocking" that there's no testing planned on the citizenship question before the surveys go out.

Hillygus: "It is surprising to say that you can add at the last minute a question without pretesting."

After Hillygus testifies the lack of pre-testing violates Census Bureau guidelines, a DOJ attorney goes up for cross-examination.

DOJ attorney says that Hillygus gave a presentation reacting to the citizenship question on March 19, 2018.

Hillygus' response: "WTH."

Asking if that communicated "What the hell," Hillygus said it communicated: "My shock, yes."

The DOJ attorney adds that Hillygus also said: "I want to say in no uncertain terms that this is an absolutely awful decision."

She confirms that she did say that.

The DOJ asks her whether she can rule out that they might still conduct RCT (randomized control testing) before the surveys go out.

Acknowledging she can't know, she adds: "I don't know if that's still feasible to conduct that before the 2020."

DOJ pressed Hillygus on cross on the fact that her data sources weren't citizenship question-specific. Hillygus kept agreeing that was true, but noted that this was the result of the Census Bureau foregoing random control testing.

Redirect:

Q: Could the Census Bureau have done additional question on the sensitivity of the question to Hispanics.
A: Not only could have. I believe they were obligated to.

* additional "testing"

Hillygus steps down from the witness stand. Next witness after a brief recess.

Next witness: Penn State scholar Jennifer Van Hook, whose bio can be found here. sociology.la.psu.edu/people/jxv21

Van Hook prepared an expert report for this case on Dec. 17, 2017 and she was deposed on Oct. 5.

Van Hook testifies that her hypothesis for that study was that questions about Hispanics and Asians were sensitive because a large share were non-citizens.

She acknowledges filing errata to her report giving incorrect statistics about what percentage were non-citizens.

Trial is over for the day. More scholars up as witnesses for the rest of the week.

Thanks for following and here's my recap of the opening day of #2020Census trial on @CourthouseNews here. courthousenews.com/scholar-warns-…