JULIE A. ROBINSON, CHIEF UNITED STATES DISTRICT JUDGE

Dale E. Ho, Pro Hac Vice, R. Orion Danjuma, Pro Hac Vice, Sophia Lin Lakin, Pro Hac Vice, Emily Rong Zhang, American Civil Liberties Union Foundation, Neal A. Steiner, Pro Hac Vice, Daphne T. Ha, Rebecca Kahan Waldman, Dechert, LLP, New York, NY, Angela M. Liu, Dechert LLP, Chicago, IL, Lauren Bonds, American Civil Liberties Union of Kansas, Overland Park, KS, Stephen D. Bonney, Bonney Arbitration & Mediation, Kansas City, MO, for Plaintiffs.

Garrett Robert Roe, Kansas Secretary of State, Susan Becker, Topeka, KS, Kris Kobach, Lecompton, KS, for Defendant.

JULIE A. ROBINSON, CHIEF UNITED STATES DISTRICT JUDGE

CASES CONSOLIDATED FOR TRIAL

FINDINGS OF FACT AND CONCLUSIONS OF LAW

To register to vote, one must be a United States citizen. The Kansas legislature passed the Secure and Fair Elections ("SAFE") Act in 2011, which included a new requirement that Kansans must produce documentary proof of citizenship ("DPOC") when applying to register to vote. These cases were consolidated for trial because they both challenge the DPOC law as a method for enforcing the citizenship qualification. In Case No. 16–2105, the Fish Plaintiffs challenge the law as it applies to "motor voter" applicants—individuals who apply to register to vote at the same time they apply for or renew their driver's license online or at a Division of Motor Vehicles ("DOV") office. Plaintiffs include the Kansas League of Women Voters, as well as several Kansas residents who applied to register to vote when applying for a driver's license, but were denied voter registration for failure to submit DPOC. One claim remained for trial in that case alleging that under the Election Clause in Article 1 of the United States Constitution, the Kansas DPOC law is preempted by § 5 of the National Voter Registration Act ("NVRA"), which provides that voter registration applications may only require the minimum amount of information necessary for a State to determine applicants' eligibility to register to vote, and to perform its registration duties.

In Case No. 15–9300, Plaintiff Parker Bednasek challenges the DPOC law on constitutional grounds. His remaining claim for trial is brought under 42 U.S.C. § 1983, based on a violation of the right to vote under the Fourteenth Amendment's Equal Protection Clause. Mr. Bednasek's claim is not limited to motor-voter applicants.

The seven-day bench trial in these matters concluded on March 19, 2018. After hearing and carefully considering the evidence presented by the parties at trial, this Court first resolves the remaining motions by Plaintiffs to exclude expert testimony, and next issues its Findings of Fact and Conclusions of Law under Fed. R. Civ. P. 52(a). As explained more fully below, the Court grants in part and denies in part the motion to exclude Dr. Steven Camarota, and grants the motion to exclude Patrick McFerron. Under the test set forth by the Tenth Circuit Court of Appeals that governs whether the DPOC law violates § 5 of the NVRA, the Court finds in favor of Plaintiffs in the Fish case. The Court further finds in favor of Plaintiff Bednasek on his constitutional challenge to the law. Declaratory and injunctive relief is granted in both matters as set forth in this opinion. Further, the Court imposes specific compliance measures given Defendant's history of non-compliance with this Court's orders. And, the Court imposes sanctions responsive to Defendant's repeated and flagrant violations of discovery and disclosure rules.

I. Motions to Exclude Defense Experts Camarota and McFerron

The parties filed several motions to exclude expert testimony before trial. The Court orally ruled on all but two: Plaintiffs' written Motion to Exclude the Testimony and Report of Steven A. Camarota, and Plaintiffs' oral and written motion to exclude the expert testimony of Patrick McFerron under Rule 702, Daubert , and the rule against hearsay. These experts were offered by Defendant in both cases. The Court discusses each in turn after setting forth the appropriate legal standards.

A. Standards

The Court has broad discretion in deciding whether to admit expert testimony. The proponent of expert testimony must show "a grounding in the methods and procedures of science which must be based on actual knowledge and not subjective belief or unaccepted speculation." First, the Court must determine whether the expert is "qualified by 'knowledge, skill, experience, training, or education' to render an opinion." "[A] district court must [next] determine if the expert's proffered testimony...has 'a reliable basis in the knowledge and experience of his discipline.' " To determine reliability, the court must assess "whether the reasoning or methodology underlying the testimony is scientifically valid." The district court must further inquire into whether the proposed testimony is sufficiently "relevant to the task at hand."

It is within the discretion of the trial court to determine how to perform its gatekeeping function under Daubert . The most common method for fulfilling this function is a Daubert hearing, although such a process is not specifically mandated. In this case, the parties proffered each experts' testimony, which the Court provisionally admitted subject to later review under Rule 702 and Daubert . B. Steven A. Camarota

Defendant called Dr. Camarota to testify about the impact of the Kansas DPOC law on voter registration and participation rates. Specifically, Defendant offered Dr. Camarota "as an expert...in the fields of demography, census data, voter registration statistics, and voter participation statistics." Dr. Camarota earned a Ph.D. in American Government with a focus on policy analysis from the University of Virginia. He is currently the Director of Research at the Center for Immigration Studies ("CIS"), where his primary responsibility for the last nineteen years has been to analyze United States Census Bureau data. In this position, he helped construct the American Community Survey, which is a large annual survey conducted by the Census Bureau that includes questions about citizenship and voting. Dr. Camarota has also published peer-reviewed articles and book chapters about census data relating to immigration issues, but not on any issue related to voting. He has served as a peer reviewer for several scholarly journals. Dr. Camarota has published many non-peer-reviewed conference papers and reports for the Census Bureau and CIS, and he has testified before Congress several times about Census Bureau Data, mostly as it relates to immigration issues.

In his report and testimony, Dr. Camarota looked at Kansas administrative data provided by the SOS's Office, and data from the Current Population Survey ("CPS"), a large Census Bureau survey that asks about registration and voting in November of every other year when federal elections are held. Dr. Camarota observed that the administrative data showed an increase in registration and turnout between October 2010 and October 2014. Dr. Camarota then compared registration and voting rates in Kansas between November 2010 and November 2014, before and after the effective date of the DPOC law, based on the CPS data. Dr. Camarota also compared the Kansas registration and turnout rates to those rates nationally, and in neighboring states without DPOC laws, and found that there was no significant deviation. Dr. Camarota opined that because registration and turnout rates in Kansas increased between 2010 and 2014, the DPOC law did not unduly burden Kansans' ability to register and vote.

The Court finds that Dr. Camarota is qualified to testify as an expert in this case about Census Bureau data, including the CPS. His education and work experience qualify him to explain and present this Census data. However, the Court does not find him qualified to interpret these survey results as they relate to the DPOC law, particularly to the extent he challenges Professor Michael McDonald, whose expertise and scholarship in election law is extensive, and who more closely evaluated the administrative data. Dr. Camarota's experience at CIS is limited to scholarship and reports that generally deal with immigration and citizenship issues, not election issues such as voter registration. He has never published peer-reviewed research on the subjects relevant to this litigation, nor do his non-peer-reviewed articles contain analysis of the issues relevant to this case. To the extent Defendant offers Dr. Camarota as an expert on "voter registration statistics, and voter participation rates" beyond presenting Census Bureau data, that opinion is excluded. Dr. Camarota is qualified as an expert to explain the results of CPS data showing voter registration and turnout changes in Kansas between 2010 and 2014. And he is certainly qualified to explain how the CPS data was collected and whether it is reliable. But Dr. Camarota is not qualified to explain the reasons for the change in data between 2010 and 2014, or to insert assumptions into the record based on studies or academic literature regarding voter registration and turnout. These are not his areas of expertise.

The limitations of Dr. Camarota's expertise in this field were similarly evident in the recent NVRA case of Bellitto v. Snipes . There, the district court initially limited his testimony because he was not qualified to "offer testimony as to the degree of accuracy of...rates [of voter registration from the Census Bureau's American Community Survey]." That case went to trial and the district court issued its findings of fact and conclusions of law after this trial concluded, on March 30, 2018. In that order, the court found Dr. Camarota's population analysis to be misleading and inaccurate by comparing mismatched data.

Plaintiffs further challenge the reliability of Dr. Camarota's opinions in this matter on several grounds: (1) he fails to control for confounding factors, such as general interest in the election, whether an advocacy group took an interest in the election, get-out-the vote efforts, competitiveness of the election, laws governing registration, education levels, ethnicity, age, and natural population growth; (2) the choice to compare non-presidential election year data fails to account for any change that may be due to the DPOC law as opposed to other factors; (3) he relied on and cherry-picked flawed statistical data from the SOS's Office; and (4) he relied on conclusory assumptions, such as that some noncitizens mistakenly believe they are citizens when they register to vote. Plaintiffs also point to evidence of Dr. Camarota's bias based on public positions taken by CIS, and based on statements made by executives at CIS. Defendant maintains that these issues go to the weight and not the admissibility of the evidence.

The Court agrees with Defendant that Plaintiffs' reliability challenges largely go to the weight and not the admissibility of Dr. Camarota's report and testimony, to the extent the testimony relates to his area of expertise. As described below in its findings of fact and conclusions of law, the Court does not credit Dr. Camarota's opinion that: (1) CPS data about registration and turnout is a better measure of registration and turnout than the actual numbers maintained by the SOS's Office; (2) comparing election years 2010 and 2014 is an accurate measure for determining how the SAFE Act impacted registration and voting rates; (3) that election years 2010 and 2014 in Kansas are comparable to one another, or to other states; and (4) registration rates and voter turnout is the best measure of how burdensome the DPOC law is. The Court therefore grants in part and denies in part Plaintiffs' motion to exclude Dr. Camarota's testimony. As to the Census Bureau data described in Dr. Camarota's report, the Court gives it little weight in determining the overall burdensomeness of the DPOC law, as described in the Court's findings of fact and conclusions of law. C. Patrick McFerron

In May 2016, Patrick McFerron conducted a telephone survey of 500 Kansans by CHS & Associates to help determine the rates of possession of DPOC. The survey purports to control "for gender, age, and geographic region in order to replicate US Census information." The executive summary of the survey concludes that it "reveals requiring proof of citizenship in order to register to vote is not a concern for residents and is not hampering voter registration." The study surveyed a sample of 500 Kansans, and found 83% are registered to vote. Of those not registered, only one reported that lack of DPOC was the reason.

Mr. McFerron drafted the survey results, but did not complete his own expert report, nor was he ever designated as an expert in this case. Instead, Hans von Spakovsky, one of Defendant's other experts, attached it to his expert report. Mr. McFerron is listed on the executive summary as the President of CHS & Associates, which is in Oklahoma City, Oklahoma. Plaintiffs deposed him on June 7, 2016, during which Plaintiffs' counsel asked him whether he purported to testify as an expert in this case, and he testified that he did not believe so. Mr. McFerron testified about the survey, its methodology, and its results. On January 30, 2018, Defendant filed his final witness disclosures, listing Mr. McFerron as a fact witness by written deposition.

Plaintiffs moved in limine to exclude McFerron's testimony, strike and exclude from trial his deposition designations, and exclude his survey. They argued that his testimony is inadmissible lay opinion, that it should be excluded as expert opinion because it was not disclosed under Rule 26(a)(2), and Defendant's failure to disclose was neither harmless nor substantially justified. The Court ruled that McFerron's testimony was not lay opinion, and as a sanction for failing to designate him as an expert witness, the Court required him to testify at trial as a live witness instead of by deposition. The Court took under advisement Plaintiffs' motion to exclude as inadmissible hearsay, and at trial, Plaintiffs also moved to exclude the survey under Rule 702 and Daubert . The Court provisionally admitted his testimony, subject to a later admissibility ruling.

Defendant vacillated at trial between offering Mr. McFerron as a fact and expert witness, despite the Court's ruling that his testimony was not admissible lay opinion. Mr. McFerron testified for the first time on cross-examination that he was paid an hourly rate of $100 per hour for research, and $150 per hour for his testimony based on an agreement reached with Defendant two weeks before trial that was not previously disclosed to Plaintiffs. This fact, in conjunction with the nature of Mr. McFerron's substantive testimony, reinforces this Court's previous ruling that he is clearly offered as an expert witness. He testified not only about the methodology of his survey, but about its accuracy and conclusions, including that the DPOC law is not burdensome because most Kansans possess DPOC, or can obtain it easily. Mr. McFerron's testimony illustrates the prejudice involved in allowing an expert to testify without first meeting the requirements of Rule 26(a)(2)(B). Mr. McFerron's "Report" contains a two-page summary of the survey's results; it does not contain a statement of his compensation, qualifications, or a list of all publications he has either authored or co-authored in the last ten years. He did not sign the report. In fact, Mr. McFerron admitted during his testimony that he was not sure whether he was being paid to testify as a fact or expert witness.

The Court has already excluded Mr. McFerron's testimony to the extent Defendant offers it as lay opinion. With respect to the admissibility of Mr. McFerron's expert testimony, the Court grants Plaintiffs' motion to exclude because he is not qualified to render the opinion contained in the report's summary, and because his survey is unreliable and not relevant. Also, because it fails to adhere to generally accepted survey principles, the survey lacks the indicia of trustworthiness required for survey evidence to meet an exception to the hearsay rule.

1. Qualifications

Plaintiffs argue that Mr. McFerron is not qualified to provide expert testimony on the subject matter of his survey because he is a pollster, and not a trained statistician. The Court agrees. It is true that McFerron's qualifications are based on his experience, and not an academic background in statistics. It is also true that Mr. McFerron has spoken to numerous university classes regarding polling, and that the firm he works for conducts approximately 50 to 70 public opinion surveys in any given year. Mr. McFerron has conducted approximately 15 to 20 polls in Kansas since 1993. But, as Plaintiffs point out, Mr. McFerron only took one statistics course as an undergraduate and one while he was studying for his master's degree, though he cannot recall the name of the graduate statistics course. Mr. McFerron has never written a peer-reviewed article, nor has he ever served as a peer reviewer for a journal. Mr. McFerron has not published anything on survey methodology or polling methodology. At the time of deposition, Mr. McFerron was not familiar with the American Association of Public Opinion Research, nor other standard survey research principles described by Dr. Matthew Barreto, Plaintiffs' rebuttal expert. He is unfamiliar with the basic concept of social desirability bias leading to overreporting in survey research, a concept that applies to surveys concerning voter registration and voting, or that asks if one possesses an underlying document deemed socially important. Notably, Mr. McFerron has never previously testified as an expert witness.

While an academic background is not required to testify as an expert witness, the expert testimony in this case requires a background in survey methodology that Mr. McFerron does not have. In sum, while the Court finds that Mr. McFerron obviously is an experienced pollster, particularly in the Midwest, he is not qualified to render an expert opinion about the accuracy of the results of this study about DPOC possession under well-accepted survey principles.

2. Reliability and Trustworthiness

Survey evidence is admissible in this circuit as an exception to the hearsay rule "if the survey is material, more probative on the issue than other evidence and if it has guarantees of trustworthiness." The Court will find a survey trustworthy "if it is shown to have been conducted according to generally accepted survey principles." Therefore, the survey standards for reliability under Daubert and trustworthiness under the hearsay exception are parallel. Assuming Mr. McFerron is qualified to render an expert opinion about the survey's methodology and the accuracy of the conclusions stated in the report's summary, the survey must be excluded because Plaintiffs established during Mr. McFerron's cross-examination, and with their rebuttal expert Dr. Matthew Barreto, that the survey relies on flawed methodology and is thus unreliable and untrustworthy for several reasons.

The Court finds Dr. Barreto credible and qualified to discuss accepted survey methodology. He explained the myriad flaws with the McFerron Survey that render it inadmissible under Rule 702, Daubert , and the rule against hearsay. First, the McFerron Survey does not contain a large enough sample for reliable estimates about individuals who might be burdened by the DPOC requirement. The survey targeted eligible Kansas voters generally, rather than the pool of individuals who are subject to the DPOC requirement: eligible Kansas voters who are not yet registered to vote. The McFerron Survey contained a sample of only 65 individuals who were not yet registered to vote. This is substantially less than the sample size of 300–500 people that Mr. McFerron himself testified would be necessary for reliable statistical results at the statewide level.

Second, the sample of 500 Kansas adults in the McFerron Survey was not a representative sample of the entire eligible voting population. The "most important and single first principle" one considers in a survey is whether the survey sample is representative of the population as a whole. But the McFerron Survey did not look at respondents' educational attainment, household income categories, and homeownership or renter status to ensure representativeness. Moreover, as Mr. McFerron admitted during his testimony, surveys typically use weights to achieve a representative sample, yet he relied on a quota-based approach. He acknowledged that academic literature for decades has discredited quotas, but believes that criticism is outdated because it was based on the prevalent use of landlines to conduct surveys, which does not pose a concern today. He could provide no citation to authority that contradicted Dr. Barreto's strongly-cited opinion that generally accepted survey methodology relies on weighting, and not quotas. Indeed, the results of Mr. McFerron's survey, which substantially differ from the Census Bureau data relied on by Dr. Camarota, illustrates the problems with Mr. McFerron's approach. For example, Mr. McFerron reported in his survey that 39% of households had incomes below $50,000, while the Census data shows that this figure is 48%.

Third, the McFerron Survey was only conducted over a three-day period in the evening hours between a Monday and a Wednesday. This sampling schedule precluded participation from individuals who, due to their work schedule, may not have been available during those limited days and hours. This practice violated the norms of survey research.

Fourth, when reporting his survey results, Mr. McFerron did not include a response rate, which makes it impossible to assess the reliability and the generalizability of the data collected. As another district court explained, "[n]on-response bias typically becomes a concern when the response rate falls below eighty percent. Response rates below that point—even far below that point—are commonplace and do not necessarily invalidate a survey, but they do require an analysis as to the reasons for the nonresponses and the effect they may have on the results." Here, Mr. McFerron has provided no response rate to evaluate.

Fifth, the questions on the McFerron Survey about DPOC possession are contaminated by bias because their wording primed respondents to state that they possessed DPOC even if they did not. Before any questions about possession of DPOC were asked, respondents were asked a series of nine questions prefaced by the following:

Now I want to read you a short list of documents. Only one of these documents is needed in order to register to vote in Kansas. For each of these, please let me know if you have that document at your home, office, or other location or if someone else keeps the document for you and could get it to you if necessary, or if the document does not exist.

The following nine questions asked about the types of documents that can be used to meet the DPOC requirement (e.g., birth certificates). The prefatory statement to that series of nine questions, "primed" the respondent that one of the documents on a list he/she would hear was needed to register to vote in Kansas. Extensive political science research suggests that such priming will lead to overreporting of access to documents.

Respondents were also asked in Question 18: "In 2011 because of evidence that aliens were registering and voting in Kansas elections, the Kansas legislature passed a law requiring that people who register to vote for the first time must prove that they are United States citizens before they can become registered. Do you support or oppose this law?" The Court easily finds that this question primed the respondent to answer that they support the law. Indeed, the Defendant himself drafted this loaded question and demanded that Mr. McFerron include it. Mr. McFerron "had reservations about" the question, so much so that he decided to place it toward the end of the survey so that it would not impact the earlier questions.

For these reasons, the Court finds the McFerron Survey is neither reliable nor trustworthy.

3. Relevance

Finally, even assuming the reliability of Mr. McFerron's methodology, the relevance of the survey is nominal at best. As already discussed, only 65 of the survey's 500 respondents were unregistered voters. Because the law does not apply to registered voters, there is no constitutional burden to assess for these individuals as a matter of law. Setting aside the fact that this percentage of the sample does not match the Census data touted by Defendant's other expert, Dr. Camarota, it is simply not relevant how burdensome the law is on individuals who need not comply with the law because they were registered before the law's effective date.

The survey also failed to ask several relevant questions. The survey's possession questions were compound, so it is impossible to know whether each respondent did not have the particular document addressed in the question, or whether someone else keeps the document for them. Respondents were not asked how long it would take for them to get a copy of their DPOC if they did not personally possess it. Nor were they asked how much it would cost them to obtain a birth certificate or other form of DPOC. Respondents were also not asked whether the name on any document that could have been used to meet the DPOC requirement matches their current name. Mr. McFerron acknowledged that people sometimes change their names, and thus, a person who answered that they do possess DPOC might still be unable to satisfy the DPOC requirement because of a name mismatch. For these reasons, the Court does not find the McFerron Survey is helpful to the trier of fact.

For all of these reasons, the Court grants Plaintiffs' motion to entirely exclude the McFerron Survey and his expert testimony. He is not qualified to render an expert opinion on the survey's methodology or conclusions. Moreover, the survey is unreliable and untrustworthy because it fails to follow accepted survey methods and practices. Finally, the survey is not helpful to the trier of fact. Even if the Court admitted Mr. McFerron's testimony and report, for the same reasons identified above, the Court would give it little to no weight.

II. Findings of Fact

A. Kansas Law Governing Citizenship Eligibility

Under Kansas law, legally qualified voters must register to be eligible to vote, and only United States citizens over the age of 18 may register to vote. Before January 1, 2013, Kansas voter registration applicants met these eligibility requirements by signing an attestation of eligibility on the registration application. The attestation states: "I swear or affirm that I am a citizen of the United States and a Kansas resident, that I will be 18 years old before the next election, that if convicted of a felony, I have had my civil rights restored, that I have abandoned my former residence and/or other name, and that I have told the truth on this application." Kansans may apply to register to vote in person, by mail, through a voter registration agency, in conjunction with applying for a Kansas driver's license, or "by delivery to a county election officer to be registered."

Defendant Kansas Secretary of State ("SOS") Kris Kobach does business in and is an elected official of the State of Kansas. In his capacity as SOS, he is the Chief Election Officer for the State of Kansas. During his campaign to become SOS, news stories about the problem of noncitizen voting fraud began to increase. Defendant campaigned on that issue, asserting it was a pervasive problem. After becoming SOS, he helped craft the SAFE Act, which became law in April 2011. In addition to an attestation of eligibility, the SAFE Act requires that voter registration applicants submit DPOC at the time they apply to register to vote. The law provides thirteen forms of acceptable documentation:

(1) The applicant's driver's license or nondriver's identification card issued by the division of vehicles or the equivalent governmental agency of another state within the United States if the agency indicates on the applicant's driver's license or nondriver's identification card that the person has provided satisfactory proof of United States citizenship;

(2) the applicant's birth certificate that verifies United States citizenship to the satisfaction of the county election officer or SOS;

(3) pertinent pages of the applicant's United States valid or expired passport identifying the applicant and the applicant's passport number, or presentation to the county election officer of the applicant's United States passport;

(4) the applicant's United States naturalization documents or the number of the certificate of naturalization. If only the number of the certificate of naturalization is provided, the applicant shall not be included in the registration rolls until the number of the certificate of naturalization is verified with the United States bureau of citizenship and immigration services by the county election officer or the SOS, pursuant to 8 U.S.C. § 1373(c) ;

(5) other documents or methods of proof of United States citizenship issued by the federal government pursuant to the immigration and nationality act of 1952, and amendments thereto;

(6) the applicant's bureau of Indian affairs card number, tribal treaty card number or tribal enrollment number;

(7) the applicant's consular report of birth abroad of a citizen of the United States of America;

(8) the applicant's certificate of citizenship issued by the United States citizenship and immigration services;

(9) the applicant's certification of report of birth issued by the United States department of state;

(10) the applicant's American Indian card, with KIC classification, issued by the United States department of homeland security;

(11) the applicant's final adoption decree showing the applicant's name and United States birthplace;

(12) the applicant's official United States military record of service showing the applicant's place of birth in the United States; or

(13) an extract from a United States hospital record of birth created at the time of the applicant's birth indicating the applicant's place of birth in the United States.

The DPOC requirement became effective on January 1, 2013.

If an applicant is a United States citizen but unable to provide one of the thirteen forms of identification listed in subsection (l), the statute allows that applicant to submit another form of citizenship documentation by directly contacting the SOS's Office. Although information about the subsection (m) hearing alternative has been available on the SOS's website, it is not publicized to applicants at the time they apply to register to vote. To avail oneself of this option, an applicant must submit a "RCD" form with the SOS's Office, and schedule a hearing. The form requires a declaration under penalty of perjury that the applicant does "not possess any of the documents...that may be used for proof of citizenship according to Kansas law." The form also states that a false statement on the affirmation is a severity level 9 nonperson felony.

The hearing must be before the State Election Board, which will assess the alternative evidence of citizenship to determine whether it is satisfactory. The State Election Board is comprised of three high-ranking State officials: the SOS, the Attorney General, and the Lieutenant Governor. The RCD form states that the Board will give the applicant five days' notice of the date, time, and location of the hearing. In practice, a hearing may be held with two out of the three members of the Board, and one representative of the third member. Personal attendance by the applicant is not required.

There is no statute, regulation, or list maintained by the SOS of specific documents that would satisfy the State Election Board. Bryan Caskey, the Director of Elections at the SOS's Office, testified and Defendant argued that an applicant's own declaration explaining his or her circumstances and why he or she does not possess a proof of citizenship document would satisfy the board. Five individuals have completed this hearing process since the law became effective, and all had their citizenship approved.

If a voter registration applicant fails to submit the requisite DPOC before the registration deadline in Kansas, that applicant can still submit DPOC to the county election office in person, by mail, or electronically (including by text message) before midnight on the day before an election.

On June 25, 2015, Defendant Kobach promulgated K.A.R. § 7–23–15, which became effective on October 2, 2015. The regulation applies to registration applications that have been deemed "incomplete" and therefore held "in suspense." Such applications are "canceled" if they do not produce DPOC, or otherwise cure the deficiency in the application, within 90 days of application. The applicant must submit a new, compliant voter registration application in order to register to vote.

The Bednasek case was filed on September 30, 2015, just before K.A.R. § 7–23–15 became effective. The Fish case was filed on February 18, 2016. On May 17, 2016, this Court issued an extensive Memorandum and Order granting in part the Fish Plaintiffs' motion for a preliminary injunction barring enforcement of the Kansas DPOC law until the case could be decided on the merits. It was effective on June 14, 2016. The Tenth Circuit affirmed that ruling on October 19, 2016, providing significant guidance on Plaintiffs' preemption claim that § 5 of the NVRA displaces the Kansas DPOC law. On remand, the Court reopened discovery in Fish as to evidence relevant to the Tenth Circuit's guidance.

B. DOV Policies and Procedures

Driver's license applicants in Kansas must provide proof of lawful presence when they apply for the first time. As part of this requirement, the Kansas Division of Vehicles ("DOV")

shall require valid documentary evidence that the applicant: (A) Is a citizen or national of the United States; (B) is an alien lawfully admitted for permanent or temporary residence in the United States; (C) has conditional permanent resident status in the United States; (D) has an approved application for asylum in the United States or has entered into the United States in refugee status; (E) has a valid, unexpired nonimmigrant visa or nonimmigrant visa status for entry into the United States; (F) has a pending application for asylum in the United States; (G) has a pending or approved application for temporary protected status in the United States; (H) has approved deferred action status; or (I) has a pending application for adjustment of status to that of an alien lawfully admitted for permanent residence in the United States or conditional permanent resident status in the United States.

The DOV website identifies five documents that purportedly "show your date of birth, identity, and lawful status as a U.S. citizen" when applying for an original Kansas driver's license or non-driver identification card: a certified U.S. birth certificate, an unexpired United States Passport or Passport Card, a U.S. Consular Report of Birth Abroad, a Certificate of Naturalization, and a Certificate of Citizenship. These documents also meet the DPOC requirement for voter registration. In order to renew a Kansas driver's license, the applicant must also provide the DOV with proof of identity (such as an expiring Kansas driver's license), a Social Security number, and proof of Kansas residency.

After reviewing an applicant's documentation, a DOV employee enters the applicant's name and date of birth into the DOV database and takes the applicant's photo as well as captures their signature. Currently, DOV procedure and training provides that driver's license examiners are to scan all documents an applicant provides during a driver's license renewal. If a proof of citizenship document was scanned into the DOV system during a prior transaction and a voter applies to register to vote during a renewal, the DOV is to inform the SOS's Office that such document is on file. The DOV only has documents scanned into the system since 2013.

As part of the driver's license application and renewal processes, the driver's license examiner is to ask each applicant if they want to register to vote. The DOV currently has a policy of not offering voter registration to driver's license applicants who self–identify as noncitizens, such as TDL applicants or driver's license applicants who show a green card to demonstrate lawful presence while applying for a driver's license. The examiners are trained to enter a "Y" in the appropriate field of the computer interface if a customer answers "yes" to the voter registration question. The examiner then directs the applicant to read a voter oath located on the counter in front of the applicant and to ask the applicant to read that oath. Next, the examiner is to ask the applicant if he/she affirms the voter oath. Applicants are not required to provide a signature after reading the voter oath. The signature occurs during the photo and signature portion of driver's licensing process before the voter registration part of the process begins.

The examiners are to ask applicants who affirm the voter oath a series of questions including whether they are citizens of the United States, whether they will be 18 years of age before the next election, whether they want to register with a political party, and whether they want to provide their telephone numbers. The examiners are to record the customers' answers to these questions in the computer interface.

Noncitizens who apply for a driver's license may receive a temporary driver's license ("TDL"), the duration of which is tied to the length of time that the documentation they provided to the DOV permits their presence in the United States. Noncitizen lawful permanent residents who apply for a driver's license receive a standard six-year license. Lawful permanent residents are not required to provide a lawful presence document when they renew their driver's license. The DOV does not keep statistics on the number of driver's licenses issued to permanent residents.

A voter registration receipt prints automatically when someone applies to register to vote at the DOV. DOV procedure requires examiners to provide the applicant with the voter registration receipt. The receipt is on a small piece of paper that resembles a fast-food receipt, according to one witness, and it contains the applicant's picture. The following language appears on the receipt in small font:

Thank you for your voter registration application. Your application will be sent to your county election office for processing.

Unless you already submitted to the division of vehicles a document proving U.S. citizenship, you need to submit one to your county election office before you will be added to the voter registration list. Visit www.gotvoterid.com for a list of acceptable documents. If you were a registered voter in Kansas before 2013 and are still registered, you do not need to provide a citizenship document.

A notice will be mailed to you when processing is completed. If you have questions about your application, please call the county election office...or call the Kansas SOS....

C. Impact of the DPOC Law on Kansas Applicants

1. Administrative Data

The Kansas Election Voter Information System ("ELVIS") is a statewide voter registration database, maintained by Defendant; ELVIS assigns a unique identification number to all voters. Each county election office is responsible for maintaining the voter lists for its county, so this central database reflects data that is entered by the counties. When a voter registration application is received by the relevant county election office, a record is created in the ELVIS database. County election officers have been instructed to enter into ELVIS all voter registration applications regardless of whether the applicant provided proof of citizenship. When a person applies for a driver's license or a renewal at the DOV but does not apply to register to vote at that time, an ELVIS file is not created and the SOS is not notified.

ELVIS contains codes for "source of information description," showing how the applicant registered to vote. "MV" is the code recorded in ELVIS to indicate that an applicant has applied to register to vote at the DOV in conjunction with a driver's license application. ELVIS contains status codes, including "A" for Active, "R" for Canceled, and "S" for Suspense. ELVIS contains voter registration reason codes, which explain the reason an applicant is or is not registered to vote. "CITZ" is the code recorded in ELVIS to indicate that an applicant has failed to provide DPOC.

Defendant and county election officers may accept DPOC at a different time or in a different manner than an application for voter registration, as provided in (l), "as long as the applicant's eligibility can be adequately assessed by the SOS or county election officer as required by this section." Under this authority, Defendant has established interagency agreements with two Kansas agencies to verify whether one of the thirteen forms of DPOC listed in § 25–2309(l) may be on file.

First, on January 7, 2014, Defendant and Robert Moser, MD, Secretary of the Kansas Department of Health and Environment ("KDHE"), entered into an Interagency Agreement called the "Birth/Voter Registration Data Link," whereby the KDHE agreed to crosscheck the names of incomplete voter registration applicants with the database of birth certificates and marriage licenses on file with the Kansas Department of Vital Statistics ("OVS"), and provide Defendant with the results. Defendant sends a list of new voter registration applicants on the suspense list to the KDHE on approximately a monthly basis. The agreement makes clear that "The Kansas OVS maintains records only on Kansas vital events occurring in the State of Kansas. The voter registration form does not collect State of birth for the voter." The SOS' Office does not currently check with any agencies outside of Kansas to verify citizenship of voter registration applicants.

Second, in May 2016, after the preliminary injunction hearing in the Fish case, Defendant implemented an interagency policy for coordinating with the Kansas Department of Revenue ("KDOR") to verify citizenship documents that may have been provided by voter registration applicants when they applied for a Kansas driver's license. Defendant and the county clerks were given access to a secure internet portal whereby they may check the DOV database for records of any registration applicant on the suspense list to determine if the DOV possesses DPOC for that resident. Defendant has instructed the counties to check for every applicant on their suspense list to determine whether incomplete voter registration applicants may have provided acceptable DPOC to the DOV when applying for a driver's license.

The SOS's Office has instructed the counties to contact each voter registration applicant on the suspense list at least three times before the 90–day period under K.A.R. § 7–23–15 expires. The notices from Douglas County in evidence at trial list the various acceptable forms of citizenship under § 25–2309(l), and state the applicant can send copies to the county election office by regular mail or e-mail. The notices do not reference the hearing procedure in § 25–2309(m).

According to ELVIS records, as of January 1, 2013, there were 1,762,330 registered voters in Kansas. As of October 2016, there were 1,817,927 registered voters. As of March 28, 2016, before the preliminary injunction was issued requiring Defendant to register to vote applicants suspended or canceled for failure to provide DPOC, there were 14,770 applicants on the suspense list. Of these, 5,655 were motor voter applicants. As of March 28, 2016, 16,319 individuals had their applications canceled under K.A.R. § 7–23–15 due to lack of DPOC. Of these, 11,147 were motor voter applicants. These figures amount to 31,089 total applicants who were denied registration for failure to provide DPOC, 16,802 of whom applied through the DOV.

Professor Michael McDonald testified as an expert witness for Plaintiffs about the composition of the suspense and cancellation lists. Dr. McDonald is an Associate Professor of Political Science at the University of Florida and a leading scholar on American elections, voter registration, and factors affecting voter behavior and turnout. He has received numerous research grants and honors for his academic work. Dr. McDonald has offered expert testimony in numerous election law cases, including cases involving voter registration and the NVRA. He has written numerous peer-reviewed books, book chapters, and articles about elections and voter registration.

Dr. McDonald examined data extracts from the ELVIS database to evaluate the individuals whose applications were canceled or suspended for lack of DPOC, and he offered opinions about the effect of the law based on that analysis. He looked at three sources of information: (1) a list of suspended applicants as of September 24, 2015, provided to him by the Plaintiffs; (2) the electronic voter registration file dated December 11, 2015, including the list of suspended applications as of that date, provided by Defendant; and (3) a list of canceled and suspended applicants as of March 31, 2016, disclosed by Defendant.

Dr. McDonald examined the voter registration data and determined that most of the individuals on the suspense list as of September 25, 2015 did not become registered by December 11, 2015. 22,814, or 70.9% of the applicants on the September 2015 list, remained on the December 2015 list. Canceled or suspended applicants represented 12.4% of new voter registrations between January 1, 2013 and December 11, 2015. Dr. McDonald acknowledges that a few of these suspended or canceled applicants may in fact be noncitizens, however given the individual-level data he reviewed, he believes that the majority are eligible citizens.

As of March 31, 2016, the confidential ELVIS data provided to Dr. McDonald pursuant to the protective order in this case showed a total of 30,732 voter registration applications were either held in suspense or canceled due to the DPOC requirement—16,749 applications were canceled, and 13,983 applications were suspended. These 30,732 unregistered applicants represented approximately 12% of the total voter registration applications submitted since the law was implemented in 2013. Of the 30,732 applicants whose applications were, as of March 31, 2016, suspended or canceled due to failure to provide DPOC, approximately 75% were motor-voter applicants. Dr. McDonald opined that these numbers would have increased further before the 2016 presidential election but for the Court's preliminary injunction order, in part because voter registration activity typically increases in the months leading up to a presidential election. Indeed, Mr. Caskey's testimony and Defendant's own statements during the contempt hearing that followed this trial support Dr. McDonald's opinion. They suggested that problems coordinating certificates of registration to those affected by the preliminary injunction were tied to their increased activity and workload associated with the runup to that election.

Dr. McDonald further credibly opined that the DPOC law disproportionately affects the young and those who are not politically affiliated. He testified that 43.2% of motor voter applicants held in suspense or canceled were between the ages of 18–29, and 53.4% of suspended and canceled motor voter applicants were unaffiliated. To be sure, the law only applies to new voter registration applicants—those registering for the first time in Kansas after January 1, 2013. Those voters tend to be young and unaffiliated with a political party. But that is the point: the fact that the law affects only new applicants means that it disproportionately affects certain demographic groups. Dr. McDonald explained that there is a consensus in social science that barriers to voter registration increase the cost of voting and dissuade individuals from participating in the political process. Moreover, these groups—the young and unaffiliated—already have a lower propensity to participate in the political process and are less inclined to shoulder the costs associated with voter registration. This opinion is borne out by Ms. Marge Ahrens' testimony, discussed infra, which provided examples of how difficult it has been for the Kansas League of Women Voters to help register young voters due to the DPOC law.

2. Current Population Survey Data

As described in the Court's Daubert ruling, Dr. Camarota disagrees with Dr. McDonald about whether the DPOC law poses a burden on voter registration and voting. He primarily relies on the Census Bureau's Current Population Survey to opine that the burden must be low because voter registration and turnout rates in Kansas increased between 2010 and 2014. The Court has already ruled that Dr. Camarota's qualifications limit his expert opinion to explaining the CPS data; he is not qualified as an expert in voter registration, voting trends, or election issues, so he is not qualified to opine on issues of causation.

Even if Dr. Camarota is deemed qualified, the Court gives little weight to his ultimate opinion for several reasons. Primarily, the Court finds that the best evidence about the DPOC law's burden is the actual data from the suspense and cancellation lists, evaluated by Dr. McDonald. This data demonstrates a concrete burden for thousands of voter registration applicants, many of whom were not registered in time to vote in the 2014 election by operation of the DPOC law. Because this data was presented to the Court, it need not look at indirect survey data that is based on sampling, nor must the Court look at how Kansas compares in terms of Census data to neighboring states. As Dr. McDonald explained, the individual-level data that he analyzed is the "gold standard," so there is no need to rely on statistical sampling. In the same vein, the Court gives no weight to Dr. Camarota's bare observations about the uptick in new registration and voter turnout numbers between 2010 and 2014, as shown in the administrative data. As the Court discussed in its Daubert ruling, Dr. Camarota is not qualified to opine about this administrative data. He did not verify this data, as Dr. McDonald did, with the individual data. Importantly, Dr. Camarota's observation that the stipulated registration and turnout numbers are larger in 2014 is not helpful to the trier of fact—the Court has accepted the parties' stipulations as to these numbers and can glean for itself that the 2014 figures are higher. For the reasons described below, such a comparison tells the Court little about the impact of the DPOC law.

Moreover, comparing 2010 and 2014 election data is not a reliable way to measure the impact of the DPOC law. To make a valid comparison between the voter registration and turnout statistics between these two election years, one would have to assume that the only difference in Kansas between 2010 and 2014 is the DPOC law. But as Plaintiffs submitted, this isn't true. First, the 2014 election in Kansas was highly competitive compared to 2010. The Gubernatorial and U.S. Senate races were close elections. Sam Brownback won the race for governor by only 3.7 points; the Democratic candidate for U.S. Senate withdrew and consolidated support behind an independent candidate. Also, there were several Kansas Supreme Court justices on the ballot and a strong advertising effort had been made by groups urging Kansans to vote against retention. The states with which Dr. Camarota compared, Oklahoma and Nebraska, did not have similarly competitive races. The competitiveness of these high-profile races could easily account for the increased registration and turnout between 2010 and 2014. Dr. Camarota conceded that he did not take these facts into account when comparing 2010 to 2014, nor when comparing Kansas rates to those of other states. Similarly, he did not control for differences in state laws between 2010 and 2014 that may have explained his observation that Kansas "bucked the national trend" of a decline in voter registration.

Importantly, comparing 2010 and 2014 registration data does not provide a reliable measure of the impact of the DPOC law because there is no way to know when the increased registration occurred—the 2014 data represents an increase from 2010, but the DPOC law did not become effective until January 1, 2013. Dr. Camarota's analysis does not demonstrate when the increased registrations occurred, before or after the law was passed. Similarly, as Dr. McDonald testified, because the DPOC law only applies to new registrants, it makes sense that the law would not have a large impact on the overall registration numbers and turnout rates, as measured by survey data. Most registered voters surveyed in Kansas in 2014 were registered before January 1, 2013, before the law became effective, and were thus exempt from the DPOC requirement.

3. Kansas League of Women Voters

Margaret Ahrens, the immediate past co-president of Plaintiff League of Women Voters of Kansas (the "Kansas League") and an advisor and mentor to the current leadership, testified on behalf of the Kansas League. The Kansas League is a nonpartisan, nonprofit volunteer organization that encourages informed and active participation of citizens in government and works to influence public policy through education and advocacy. Founded almost 100 years ago, the Kansas League is active throughout Kansas, with nine local affiliates and more than 800 members. The Kansas League was established to encourage and assist voters to access the vote, register, and "participate in the vote" in an informed manner. As Ms. Ahrens testified, "[t]he biggest passion of the [Kansas L]eague is to engage every possible citizen in the vote."

To accomplish this mission, the Kansas League provides educational resources and holds voter registration drives at various locations including schools, libraries, grocery stores, nursing homes, naturalization ceremonies and community events. The Kansas League also performs studies on a variety of public policy issues to inform membership action and advocacy efforts as well as to educate its members and the public on these issues. The Kansas League assists all prospective voters, but it is particularly committed to engaging individuals who are "underrepresented in the vote," including the first-time voter, the elderly, and individuals with limited resources and time.

Ms. Ahrens was President of the Kansas League from 2015–2017, after the DPOC law became effective. She explained that the Kansas League has opposed the SAFE Act since before its passage because it "saw [the law] as a complex network of hoops and jumps for the average Kansas citizen" that would "create barriers to the vote."

Once it went into effect, the DPOC requirement substantially affected the Kansas League's work in at least three respects. First, the DPOC requirement significantly hampered the Kansas League's voter registration work. Ms. Ahrens described the impact of the DPOC Law on the Kansas League's ability to fulfill its mission as "huge. It was a dead hit. It was absolutely a blow and I found the word shock to be appropriate in thinking about this." When the law came into effect, the Kansas League initially stopped all registration activity in every county but one, to protect volunteer members from any liability that could arise from handling or copying applicants' personal documents. The Kansas League leadership spent considerable resources on developing a copying policy to mitigate the risks associated with handling DPOC.

Once the copying policy was in place, the Kansas League re-initiated registration efforts, but the number of individuals the Kansas League could successfully register declined significantly. Ms. Ahrens provided several examples during her testimony. In Wichita, the Kansas League estimated that it helped register 4,000 individuals the year before the DPOC became effective. In 2013, after the law became effective, the Kansas League estimated it registered 400. Ms. Ahrens explained that this decline was because many individuals do not have the necessary documents at hand, or are not willing to provide such documents to League volunteers, to satisfy the DPOC requirement. She estimated that before the law passed, it took the League 3–4 minutes to assist a voter registration applicant, but after the DPOC law, it would take an hour per applicant.

In one registration effort, Kansas League volunteers in Douglas County went to high schools to register voters but returned with such "large numbers of incomplete voter registrations" due to the fact that the students did not have DPOC at hand that the volunteers called the students' families and schools and went back three times "to try to get as many young people registered [as possible]." During another voter registration effort at Washburn University, Kansas League volunteers provided multiple opportunities for students to complete their voter registration applications and to provide DPOC, by maintaining a voter registration table at the university over multiple weeks. Despite this concerted effort, out of approximately 400 students who attempted to register to vote, only about 75 students successfully completed their registration applications.

Second, the DPOC requirement forced the Kansas League to devote substantial resources to assist voters whose applications are in suspense due to the failure to provide DPOC. To reach these suspended voters, the Kansas League purchased from the SOS both the suspense list as well as the full voter file several times. The Kansas League has published the suspense list on the Kansas League website and circulated the list to local newspapers to do the same to notify applicants that their registrations are not complete. Kansas League volunteers also spent considerable time and effort to reach individuals on the suspense list directly to assist them in completing their registration applications. Ms. Ahrens provided the notable example of efforts by Kansas League volunteers in Douglas County who, after unsuccessful attempts to reach individuals on the suspense list by phone and email, visited the residences of 115 people whose voter registration applications were on the suspense list with a mobile copy machine. Of those 115 people, only 30 ultimately registered. At least half of these 30 individuals who completed their registrations did not personally possess or were not able to provide DPOC to the Kansas League volunteers and were unable to complete their registrations immediately onsite. Since the DPOC Law went into effect, the Kansas League has devoted thousands of hours to contacting the tens of thousands of voters on the suspense list and attempting to help them satisfy the DPOC requirement.

Third, the DPOC requirement has forced the Kansas League to spend a considerable amount of member resources—including volunteer time—and money to educate the public about registering under the DPOC law. The Kansas League created thousands of informational trifolds "to help people understand the changes in the law and how to participate in the vote" that volunteers distributed to community colleges, public libraries, and high schools across the state. The Kansas League had in the past developed written educational materials to assist voters in registering but "not to this extent." The Kansas League also developed a teaching module and an accompanying instructional video to distribute on its website and to universities, community colleges, vocational and technical schools, and high schools throughout the state in order to educate new voters about how to register to vote under the SAFE Act.

Following the Court's preliminary injunction in this case, the Kansas League again obtained a copy of the suspense list from the SOS. This list included the names of voters who were registered under court orders, including this Court's preliminary injunction ruling. The SOS refused the Kansas League's request for a list of suspended applicants that did not include voters registered under court orders. As a result, the Kansas League is no longer able to effectively use the suspense list to inform and reach voters who are unable to vote because their registration applications are on the suspense list because it lacks confidence that the list is accurate.

4. Access to DPOC by Suspended and Canceled Applicants

There was little admissible evidence presented at trial about the rate of DPOC possession by suspended and canceled applicants. As already discussed, the McFerron Survey is inadmissible, but even if admissible, the Court gives its findings no weight due to its many methodological flaws. There is no evidence about how many canceled and suspended applicants in fact lack DPOC, although the Court can reasonably infer from the suspense and cancelation numbers that either (1) these applicants lack immediate access to such documents because they were repeatedly notified of the need to produce DPOC in order to register, yet they did not complete the registration process; or (2) these applicants were not well enough informed about the DPOC requirement to locate their DPOC and provide it to the county election office in order to become registered; or (3) these applicants were otherwise unable or unwilling to go through the steps to produce DPOC.

Dr. Jesse Richman estimates that only 2.2% of the applicants on the suspense list lack access to DPOC, based on a survey he conducted of individuals on the suspense list. Yet Dr. Richman's results are not statistically distinguishable from zero, as the margin of error is 2.7%. Furthermore, Dr. Richman concludes that 97.8% of citizens on the suspense list have what he describes as "immediate access" to DPOC, but his estimate includes individuals who do not personally possess DPOC, but have someone who "keeps" such a document for them. Obtaining a document from another person constitutes an additional step in the voter registration process, which increases the costs of voting. As Dr. Richman himself has written in published articles, "electoral rules that increase the costs of voting are expected to diminish voter participation."

Although Dr. Richman speculated that it would be relatively easy for a registration applicant to obtain a citizenship document from another person who "keeps" the document for them, his survey provides no support for this conclusory statement. As such, Dr. Richman conceded during his trial testimony that it was an overstatement to say that a respondent has "immediate access" to DPOC when answering yes to his survey question.

Defendant argues that the suspense list is dynamic and constantly in flux, therefore it does not represent the universe of applicants prevented from registering to vote—many are ultimately registered under the State's interagency agreements, or because they later submit DPOC. There are several problems with this argument. First, while the suspense list may be dynamic, the cancelation list (before the preliminary injunction) is not. More than 16,000 voter registration applicants had been canceled under K.A.R. § 7–23–15 at the time of the Court's preliminary injunction. Moreover, at the time of the Court's preliminary injunction, more than 13,000 individuals were on the suspense list. To be sure, the evidence established that some portion of this number may come off the list due to the State's interagency agreement with the DOV, but as of March 2016, the KDHE agreement had been in place for three years. Yet, each time Dr. McDonald took a snapshot of the suspense list between September 2015 and March 2016, the combined number of suspended and canceled applicants represented about 12% of all new voter registration applications. Dr. McDonald found that 22,814, or 70.9% of the applicants on the September 2015 list, remained on the December 2015 list. While that number certainly was lower by March 2016, that is undoubtedly because many of those on the suspense list were canceled under the regulation by that point, given that in December 2015, the 90–day rule had not yet been effective for 90 days. Defendant, by contrast, provided no data about the number of those on the suspense list who have come off because DPOC was ultimately verified, or provided, as opposed to cancellation. The Court finds that the majority of those on the suspense list ultimately did not become registered.

5. Lay Testimony by Individuals Lacking DPOC

Dr. McDonald's analysis demonstrates that tens of thousands of individuals who applied to register to vote after the DPOC law became effective were held in suspense or canceled for failure to submit DPOC. He further credibly opined that the clear majority of those suspended or canceled are in fact United States citizens. Ms. Ahrens' testimony demonstrates that the DPOC law made the Kansas League's mission of helping register voters difficult, by substantially reducing the number of individuals it could assist in registering to vote, particularly within the groups it targets: first time voters, the elderly, and individuals with limited resources and time. This evidence leads the Court to the conclusion that tens of thousands of eligible citizens were blocked from registration before this Court's preliminary injunction, and that the process of completing the registration process was burdensome for them.

The experiences of several lay witnesses, including the individual Plaintiffs in both cases, illustrate Dr. McDonald's findings and Ms. Ahrens' concerns about the barriers to registration after the DPOC law became effective. Plaintiff Steven Wayne Fish is a U.S. citizen, a resident of Kansas, and over 18 years old. He works the overnight shift at an American Eagle distributor. In August 2014, he applied to register to vote while renewing a Kansas driver's license at the DOV in Lawrence, Kansas. Mr. Fish brought documents to fulfill the Kansas residency requirement for obtaining a driver's license. The driver's license examiner did not inform him that he needed a citizenship document to register to vote; when he left the DOV, he believed he had registered to vote. Subsequently, he received notices in the mail from the Douglas County election office telling him that he needed to provide DPOC in order to become registered. Those notices listed the 13 acceptable forms of DPOC under the K.S.A. § 25–2309(l). They make no mention of the alternative hearing process under subsection (m). He searched for his birth certificate but could not find it. He attempted to obtain a replacement birth certificate but could not determine how to do so—he was born on a decommissioned Air Force base in Illinois. Mr. Fish was unable to vote in the 2014 general election, and his voter registration application was subsequently canceled for failure to provide DPOC under the 90–day rule.

Later, in May 2016, Mr. Fish's sister located a copy of his birth certificate that had apparently been placed in a safe by Mr. Fish's mother, who passed away in 2013. Although the birth certificate was ultimately located, it took nearly two years to find it. Due to the preliminary injunction in this case, Mr. Fish became registered to vote in June 2016. In September or October 2016, Mr. Fish relocated within Douglas County and changed his address with the DOV in person. At that time, Mr. Fish filled out a second voter registration application and provided his birth certificate. He is now registered to vote based on this October 2016 application, having provided DPOC. He voted in the 2016 general election.

Plaintiff Donna Bucci is a U.S. citizen, a resident of Kansas and over 18 years old. She was born in Baltimore, Maryland. Ms. Bucci has been employed at the Kansas Department of Corrections for the last six years. She is a cook in the prison kitchen on the 3:00 a.m. to 12:00 p.m. shift. She is provided with limited time off, and must provide two-weeks' notice to use it. In 2013, Ms. Bucci applied to register to vote while renewing a Kansas driver's license at the DOV in Sedgwick County, Kansas. The driver's license examiner did not tell Ms. Bucci that she needed to provide proof of citizenship, and did not indicate that she lacked any necessary documentation. When she left the DOV, she believed she had registered to vote. Later, she received a notice in the mail informing her that she needed to show a birth certificate or a passport to become registered to vote. It did not include information about how to pursue the hearing process in K.S.A. § 25–2309(m). Ms. Bucci does not possess a copy of her birth certificate or a passport. She cannot afford the cost of a replacement birth certificate from Maryland and she credibly testified that spending money to obtain one would impact whether she could pay rent. Ms. Bucci's voter registration application was canceled for failure to provide DPOC. She could not vote in the 2014 election, but was able to vote in the 2016 election by operation of the preliminary injunction. Ms. Bucci first learned of the alternative hearing procedure when defense counsel informed her of it during her deposition in this case. She testified that it would be hard for her to even participate in a telephonic hearing because she is not allowed to use her cell phone on a work break.

Plaintiff Charles Stricker is a U.S. citizen, a resident of Kansas, and over 18 years old. He was born in Missouri and has lived in Kansas since late 2013, after a period of living in Chicago. Prior to living in Chicago, Mr. Stricker lived in Kansas and was registered to vote in Kansas during that time. He works as a hotel manager in downtown Wichita. Mr. Stricker applied to register to vote while renewing a Kansas driver's license at the Sedgwick County DOV in October 2014. He was told that he had insufficient documentation, and a clerk provided him with a list of documents he needed. Mr. Stricker was attempting to register on the last day of registration before an election, it was so important to him to become registered that he took the day off work to accomplish it. Mr. Stricker rushed home and "grabbed every single document that I could and started shoving them into a file folder to try to get back before the DMV closed," including his birth certificate. He made it back to the DOV in time to complete his application, and recalls telling the clerk that he wanted to register to vote. The DOV clerk did not tell him that he needed any further documentation to register. The clerk printed a small receipt for Mr. Stricker and explained to him that it would be his temporary driver's license until he received his license in the mail. He asked the clerk if there was anything else he needed to do, including whether he needed a voting card. The clerk told him nothing more was necessary. He believed that he was registered to vote.

Mr. Stricker attempted to vote in the 2014 midterm election. He presented his driver's license to the poll worker, but she could not find a record of his registration. He was given a provisional ballot to fill out at an open table with another voter. Mr. Stricker testified that he was confused and embarrassed by the experience. Election day was the first time Mr. Stricker learned that he was not registered to vote. He testified that he learned about the DPOC law sometime later through a press report and wondered if it could explain why he was not allowed to vote. He does not recall receiving any notices from Sedgwick County asking him to provide proof of citizenship.

In 2015, Mr. Stricker's voter registration application was canceled in the ELVIS system. His registration was reinstated by operation of the preliminary injunction on June 22, 2016. At some point in advance of the November 2016 election, Mr. Stricker attempted to check his registration status online and by calling the Sedgwick County election office. The person with whom he spoke told him that it was unclear whether he would be able to vote in the upcoming election because there were legal issues that were still up in the air. When he checked online, there was no record of his registration. On October 26, 2016, the Sedgwick County Election Office sent Mr. Stricker a "Notice of Voter Registration Status." It states:

This notice is to inform you that you have been granted full voter registration status in Kansas and that you are qualified to vote in all official elections in which voters in your precinct are eligible to participate.

According to Kansas Statutes Annotated 25, 2309(l), any person registering to vote for the first time in Kansas on or after January 1, 2013, must provide evidence of United States citizenship along with the registration application in order to be granted full registration status.

Our records indicate that you submitted a voter registration application during the above-mentioned time period, but you did not provide evidence of your U.S. citizenship. We have since received information from the Kansas Department of Health and Environment's Office of Vital Statistics indicating that you have a Kansas birth certificate on file. Based on that determination, your registration status is deemed complete, and we have granted you full voter registration status.

This notice was signed by Tabitha Lehman, the Sedgwick County Election Commissioner. Ms. Lehman testified that, despite Fed. R. Evid. 615 being invoked at the beginning of trial, she read media reports about the trial, including reports of Mr. Stricker's testimony. She testified that ELVIS records indicate Mr. Stricker is active and "fully registered," and that after reviewing his file prior to her testimony, she believes that the notice he received erroneously referenced his Kansas birth certificate, when in fact his citizenship document was in the DOV's database. She testified that in October 2016, just prior to the election, her office had not updated the generic notice sent to applicants whose DPOC was verified by the county to include the DOV database check, a policy that had changed in May 2016. Therefore, between May 2016 when the DOV policy went into effect, and the 2016 election, Sedgwick County—the second largest county in the State—was apparently sending out erroneous and confusing notices to individuals stating that citizenship was confirmed through the department that maintains Kansas birth certificates, when in fact that was not true.

Plaintiff Thomas Boynton is a U.S. citizen, a resident of Kansas, and over 18 years old. He was given a code of "suspense" in ELVIS for failure to provide DPOC. He moved to Kansas for the first time in July 2014 to begin teaching as a professor of English at Wichita State University. In August 2014, Prof. Boynton attempted to register to vote at a DOV in Wichita. He recalls being asked if he would like to register to vote, and responded that he did. Prof. Boynton brought several documents with him that he suspected he might need to obtain a driver's license, including his Illinois birth certificate. He does not recall which documents he specifically showed the clerk, but he produced the documents the clerked requested during the transaction. The DOV clerk did not tell him that he did not have the necessary documentation to register to vote, and when he left the DOV, Prof. Boynton understood he was registered to vote. He went to his polling place in November 2014, but the poll worker told him that his name was not on the rolls and offered him a provisional ballot. He was surprised to learn for the first time that his registration had not been completed at the DOV.

In December 2014 or January 2015, Prof. Boynton received a notice in the mail informing him that he would need to submit DPOC to complete the voter registration process. It did not advise him about the hearing process under K.S.A. § 25–2309(m). Prof. Boynton was frustrated upon receiving this notice, believing that he had registered before the November 2014 election, and understanding that his provisional ballot had not been counted. He was disappointed and irritated upon learning that his vote did not count. Voting is important to him and he had regularly voted in federal elections up until that point.

Prof. Boynton visited the DOV two times in 2015 to obtain replacement driver's licenses. Both times he declined when the clerk asked him if he wanted to register to vote. He testified that he was dissuaded from registering after his 2014 attempt failed. "I thought to myself, this doesn't seem to be the kind of process that leads to me being successfully registered, so I might as well just save myself the effort and say no this time...."

Prof. Boynton's ELVIS file shows that a certified United States birth certificate was submitted on August 4, 2014, contradicting Mr. Caskey's testimony that Prof. Boynton did not apply to register to vote until November 4, 2014, Election Day. On November 5, 2015, Prof. Boynton's voter registration application was canceled. According to Mr. Caskey, the SOS found a citizenship document through the DOV web portal after access had been granted in 2016. The fact that the web portal located a citizenship document—likely the birth certificate he took to the DOV that day—supports Prof. Boynton's testimony that he in fact applied to register to vote in August 2014. His ELVIS record now shows that he is active based on the citizenship document the SOS office located on June 20, 2016, after this Court issued its preliminary injunction.

Plaintiff Douglas Hutchinson is a U.S. citizen, a resident of Kansas, and over 18 years old. Mr. Hutchinson applied to register to vote while renewing a Kansas driver's license at a Johnson County DOV in 2013. His application was suspended for failure to provide DPOC and it was ultimately canceled on January 27, 2016. He was later registered by operation of this Court's preliminary injunction order. On July 30, 2016, his status changed to active after he submitted DPOC at the Johnson County Election Office and is now considered "fully registered."

Plaintiff Parker Blake Bednasek is a United States citizen over the age of 18, who moved to Kansas in August 2014 to attend school at the University of Kansas ("KU"). Plaintiff was born in Oklahoma. His parents, who live in Texas, possess his Oklahoma birth certificate. Prior to moving to Kansas, Mr. Bednasek was registered to vote in Tarrant County, Texas. In the fall of 2015, Mr. Bednasek volunteered with the Kansas Democratic Party. Through this work, he discussed the issue of voter registration, including the DPOC law, with the party's field and political director. He canceled his Texas voter registration on December 3, 2015. On December 4, 2015, Mr. Bednasek applied to register to vote in person at the Douglas County Election Office. He did not provide DPOC for two reasons: (1) he did not physically possess DPOC at the time of application; and (2) he does not agree with the law requiring DPOC. The Douglas County Clerk's Office accepted Plaintiff's application, but deemed it incomplete for failure to submit DPOC. Mr. Bednasek received two or three letters from the Douglas County election office, informing him that he needed to provide his DPOC and advising him that he had been placed on a 90–day waiting list. Plaintiff's voter registration application was canceled on March 4, 2016 under K.A.R. § 7–23–15.

Since the DPOC law was passed, 6 individuals have applied for a hearing under § 25–2309(m) with the State Election Board. One of these individuals, Ms. Jo French, lost her birth certificate after moving several times. She testified about the lengthy and burdensome process of registering to vote without a citizenship document. Ms. French's many encounters with the SOS's office led her to characterize her relationship with former-Deputy SOS Eric Rucker as a friendship. She testified that she hoped her testimony would make Defendant "look good." But her testimony contradicted Defendant's position that the DPOC requirement is not burdensome. As she testified, Ms. French's first of many hurdles was to pay $8 for the State of Arkansas to search for her birth certificate to prove that it did not exist, even though she already knew did not exist because she had requested it twice before. Second, she had to collect documents with the help of several other people—her baptismal record through an old friend in Arkansas and school records from her old school district in Arkansas. Third, she spoke with Mr. Rucker, who in turn reached out to her friends and cousin to vouch for her citizenship. Fourth, Ms. French relied on a friend to drive her 40 miles to the hearing; it was difficult for her to drive because she had recently had knee replacement surgery.

Ms. French's hearing before the State Election Board lasted 30 to 35 minutes and was attended by Defendant, the Lieutenant Governor, and a representative from the Kansas Attorney General's office. Also present were members of the media. The entire process from application to the date of her hearing took more than five months. After the hearing, Ms. French was interviewed, and stated: "I just thought it was strange that I had to go through this procedure to be able to vote. And any other state, you go in, throw down your driver's license and that gives you the right to vote. So this was totally off the wall for me....I don't look funny. I don't talk funny, I've been here all my life."

The hearing records contain information on the other four individuals who availed themselves of the hearing process. One established citizenship through a hearing and was represented by retained counsel. Another individual, Mr. Dale Weber, stated that he did not possess DPOC and that procuring such a document would be cost-prohibitive. The State Election Board ultimately accepted an affidavit that Mr. Weber executed on his own behalf as proof of his citizenship, attesting that he had been born on a military base and was a U.S. citizen. The State Election Board apparently found that Mr. Weber's mere attestation was sufficient to establish his citizenship.

D. Noncitizen Registration in Kansas Before and After the DPOC Law

1. Empirical Cases of Noncitizen Registration or Attempted Registration

Pretrial, Defendant stipulated that it had confirmed instances of 127 noncitizens who either registered to vote, or attempted to register to vote since 1999, based on data collected from Mr. Caskey and Tabitha Lehman, the Sedgwick County Election Officer. Of the 127 individuals identified by Mr. Caskey and Ms. Lehman, 43 had successfully registered to vote, and 11 have voted. 88 are motor-voter applicants, 25 of whom successfully registered to vote; 5 have voted.

At trial, Mr. Caskey asserted that his office had uncovered 129 instances where noncitizens had registered or attempted to register to vote. But the documentary evidence does not fully support this testimony. The underlying ELVIS records reveal that many of these instances are a result of false positive matches, confusion, or administrative error by either the county election office or a driver's license examiner. At trial, Defendant submitted evidence of: (1) 38 incidents of noncitizen registration or attempted registration in Sedgwick County; (2) 79 possible incidents of noncitizen registration by comparing the voter rolls with the DOV's list of TDL holders; and (3) 3 noncitizens who were found because they stated on juror questionnaires that they were noncitizens.

Ms. Lehman testified about the first category, based on a spreadsheet she helped maintain for several years, reflecting incidents of noncitizen registration in Sedgwick County. She did not create the spreadsheet; it was created by and is now maintained by the SOS's Office. The spreadsheet was last updated in January 2018, reflecting 38 incidents of noncitizen registration or attempted registrations going back to 1999, 18 of whom successfully registered to vote between 1999 and 2011, 5 of whom voted. The other 13 were on the voter rolls for extended periods of time, but never voted. Between 2013 and November 2016, the spreadsheet reflects 16 noncitizens who attempted to register to vote. And the spreadsheet reflects 4 individuals who are now citizens, but had been held in suspense because they applied to register to vote before becoming naturalized citizens. Because these 4 applied to register after the DPOC law passed, they were registered to vote pursuant to the Court's preliminary injunction order. Most of the individuals on this spreadsheet were discovered during naturalization ceremonies held in Wichita, Kansas, which members of Ms. Lehman's office regularly attend to help register to vote newly-naturalized citizens.

The ELVIS records for many of the individuals on the Lehman spreadsheet demonstrate instances of applicant confusion and administrative error. For example, one individual listed an "A-number" in the field for "Naturalization number (if applicable)" on the voter registration form. Another individual voted four times between 2004 and 2008, but stated that she "was a permanent resident of the U.S. and did not know she wasn't allowed to vote until after 2008 when one of her friends told her she couldn't, she then stopped voting."

The records for several "attempted registrations" on this spreadsheet are in fact instances where a noncitizen applicant did not intend to register to vote. One person Ms. Lehman lists as an attempted registrant on January 1, 2014, indicated to the DOV that she was not a citizen, but the DOV processed the voter registration application anyway. In fact, Ms. Lehman had an e-mail exchange about this applicant with former election director Brad Bryant, who stated, "I just wish DMV would not register people who they know to be noncitizens." This applicant also wrote to the DOV "Please put in the record that I am not a citizen. I cannot vote." She underlined "am not" and "I cannot vote."

Another applicant indicated to the DOV that she was not a citizen, but the DOV nonetheless processed the application. This applicant "came into the office w/a POC notification letter and stated that her registration was a mistake on the part of the DMV when she renewed her license. She is not a U.S. citizen. She filled out a [c]ancellation form." Another applicant replied in the negative when the DOV clerk asked if she was a United States citizen and produced a "Resident Alien" card, yet the DOV submitted an application, prompting the applicant to request cancellation. There are also two examples of voter registration forms being submitted to the county election office despite the applicants' failure to answer the question, "Are you a citizen of the United States of America" on the form.

Ms. Lehman personally transmitted to Defendant an application that had checked "no" to the citizenship question on the form. Ms. Lehman disingenuously testified that this "would be a case where it would be something anomalous to report up....I think it's a dicey one so I send it on." The Court does not find this testimony credible. Ms. Lehman testified that she is one of four county election commissioners directly appointed by Defendant, and that she reports directly to him. She stated that her office was charged with helping determine instances of noncitizen registration, and that when they "see something they suspect" could be noncitizen registration, her staff reports that to her, and she in turn reports it to Defendant. It appears that Ms. Lehman was either instructed or took it upon herself to pass along to Defendant even noncolorable attempts at noncitizen registration caused by State employee oversight or lack of training, rather than deliberate attempts to register to vote.

For his second category of empirical evidence, Defendant identified 79 instances of purported noncitizen voter registration by comparing a list of TDL holders generated by the DOV, with the voter rolls. Mr. Caskey testified that he compared these lists 4 times, in 2009, 2011, 2016, and 2017. Plaintiffs' expert Eitan Hersh was retained to conduct his own matching analysis of these two lists. Dr. Hersh is an expert in voter registration records and matching analysis. He is a tenured professor of political science at Tufts University, whose academic research is focused on studying large-scale individual databases, such as the ELVIS system, and matching those databases to other sources of individual-level data. The Court finds Dr. Hersh qualified, and that his testimony was credible as to the significance of Mr. Caskey's TDL list matches. Dr. Hersh conducted an extensive and thorough matching analysis, which is fully set forth in his report. He found that of the 79 individuals matched by Mr. Caskey, only 14 successfully registered to vote, and 12 had the "CITZ" code in their ELVIS records at some point, indicating they applied to register to vote and were suspended or canceled for lack of DPOC. Therefore, of these 79 individuals, 26 either successfully registered to vote or were stopped from registering under the DPOC requirement. One of these 26 individuals voted. Nine of the 79 individuals successfully registered to vote at a DOV, 1 of whom had a "CITZ" code at some point.

Moreover, as Dr. Hersh testified, Mr. Caskey's TDL matches to the voter file do not necessarily represent cases of noncitizen registration or attempted registration. Even where there are correct matches, as with the 79 individuals identified by both Mr. Caskey and Dr. Hersh, it is possible a person could obtain a TDL and later naturalize prior to registering to vote. Dr. Richman agreed with Dr. Hersh, testifying that a person is not necessarily a noncitizen simply by virtue of appearing in the TDL file. Thus, Defendant has not demonstrated that all 79 individuals matched on the TDL list were noncitizens at the time they registered to vote.

Giving full credit to Mr. Caskey's evidence that 3 noncitizens were discovered to be registered voters through juror questionnaires, and ignoring evidence that several of the Lehman spreadsheet applicants were confused about whether they had the right to register to vote, and/or State employees submitted their applications despite having knowledge that they were noncitizens, the evidence shows that 67 noncitizen individuals registered to vote under the attestation regime, or attempted to register after the DPOC law was passed. Of these, 39 successfully registered to vote despite the attestation requirement, and 28 noncitizens attempted to register to vote after the DPOC law was passed but were thwarted by operation of that law. Extrapolating percentages based on these numbers, the total number of confirmed noncitizens who successfully registered to vote between 1999 and 2013 is .002% of all registered voters in Kansas as of January 1, 2013. Of the estimated 115,500 adult noncitizens in Kansas, .06% have successfully registered or attempted to register to vote since 1999. And, the number of attempted noncitizen registrations since the DPOC law became effective in 2013 is .09% of the total number of individuals canceled or suspended as of March 31, 2016, for failure to provide DPOC.

2. Expert Testimony Regarding Incidents of Noncitizen Registration in Kansas

The Court admitted in part the expert opinion and testimony of Defendant's expert Hans von Spakovsky in the areas of elections, election administration, and voter fraud. Mr. von Spakovsky is a senior legal fellow at The Heritage Foundation, "a think tank whose mission [is to] formulate and promote conservative public policies." He is an adjunct, non-tenured professor at the Law School of George Mason University. He has never testified as an expert witness before and has published no peer-reviewed research on any subject. Notably, Mr. von Spakovsky could not identify any expert on the subject of noncitizen voter registration. The methodology Mr. von Spakovsky utilized in his expert report entailed collecting information on prosecutions, and various other reports of noncitizens voting and summarizing that information.

Mr. von Spakovsky opined that there is a problem with noncitizen voter registration and that attestation is not sufficient to prevent noncitizens from registering to vote. These opinions are premised on his assertion that any time a noncitizen registers to vote, regardless of that person's intent, it defrauds the votes of legitimate citizens. In his view, the numbers of individuals held in suspense or canceled under the Kansas DPOC law is irrelevant because those individuals could become fully registered with effort. Although he could provide no example of a noncitizen vote affecting the outcome of a close election, he opines that the mere possibility of that happening justifies the DPOC law. He based his opinions on the summary of noncitizen voting in his expert report.

The Court gives little weight to Mr. von Spakovsky's opinion and report because they are premised on several misleading and unsupported examples of noncitizen voter registration, mostly outside the State of Kansas. His myriad misleading statements, coupled with his publicly stated preordained opinions about this subject matter, convinces the Court that Mr. von Spakovsky testified as an advocate and not as an objective expert witness.

As to Kansas noncitizen registration, Mr. von Spakovsky is aware of only 30 instances of noncitizen registration or attempted registration provided to him on the Lehman spreadsheet at the time he prepared his report. He did nothing to verify this information. Yet, he opined that "[c]learly aliens [in Sedgwick County] who applied to register at the DMV were not dissuaded from falsely asserting U.S. citizenship by the oath requirement." He later admitted during cross-examination that he had no personal knowledge as to whether or not any of these individuals had in fact falsely asserted U.S. citizenship when they became registered to vote and that he did not examine the facts of these individual cases. As the Court has already discussed, several of the individual ELVIS records for those on the Lehman spreadsheet include noncitizens who disclosed their noncitizen status to the DOV clerk, so his statement that they "were not dissuaded from falsely asserting U.S. citizenship" is not supported by the record.

Mr. von Spakovsky stated in his report that a local NBC television station in Florida identified 100 individuals excused from jury duty who were possible noncitizens on the voter rolls; but on cross-examination, he admitted that he failed to include a follow-up story by the same NBC station that determined that at least 35 of those 100 individuals had documentation to prove that they were, in fact, U.S. citizens. He claimed that, at the time of his expert report, he was unaware of the NBC follow-up report, and only learned about it at his deposition. Yet after his deposition Mr. von Spakovsky never submitted a supplement or correction to his expert report to acknowledge this omission.

Mr. von Spakovsky also cited a U.S. GAO study for the proposition that the GAO "found that up to 3 percent of the 30,000 individuals called for jury duty from voter registration rolls over a two-year period in just one U.S. district court were not U.S. citizens." On cross-examination, however, he acknowledged that he omitted the following facts: the GAO study contained information on a total of 8 district courts; 4 of the 8 reported that there was not a single-noncitizen who had been called for jury duty; and the 3 remaining district courts reported that less than 1% of those called for jury duty from voter rolls were noncitizens. Therefore, his report misleadingly described only the district court with the highest percentage of people reporting that they were noncitizens, while omitting any mention of the 7 other courts described in the GAO report, including 4 that had no incidents of noncitizens on the rolls.

Mr. von Spakovsky wrote an editorial in 2011, alleging that 50 noncitizens from Somalia voted in an election in Missouri. Yet, nearly one year earlier, the Missouri Court of Appeals issued an opinion, Royster v. Rizzo , affirming the trial court's finding that no fraud had taken place in that Missouri election. While he testified that he was not aware of the court opinion at the time he wrote the op-ed, Mr. von Spakovsky admitted that he never published a written retraction of his assertion about Somalian voters illegally participating in that election.

The record is replete with further evidence of Mr. von Spakovsky's bias. Dr. Minnite testified to, and Mr. von Spakovsky's CV demonstrates, his longtime advocacy of voting restrictions. He admitted during his testimony that, at least as early as 2012, he was already an advocate for DPOC requirements like the one at issue in this case. Moreover, as early as 2001, Mr. von Spakovsky was already of the view that the NVRA was "a universal failure," and "was so flawed as to actually undermine our registration system." Mr. von Spakovsky also contributed to Defendant's first campaign for SOS and wrote an email promoting a fundraiser for that campaign. He did not disclose these facts in his report or on direct examination. When Mr. von Spakovsky opined in his report that the incidence of noncitizen registration collected by Ms. Lehman in Sedgwick County is likely just the "tip of the iceberg," he used the exact same phrase employed by Defendant to describe the same 30 incidents of noncitizen registration in Sedgwick County in a press release issued just a few months earlier. Indeed, that phrase has been Defendant's refrain in this case in describing the problem of noncitizen voter registration.

As stated above, the Court gives little weight to Mr. von Spakovsky's opinions. While his lack of academic background is not fatal to his credibility in this matter, the lack of academic rigor in his report, in conjunction with his clear agenda and misleading statements, render his opinions unpersuasive. In contrast, Plaintiffs offered Dr. Lorraine Minnite, an objective expert witness, who provided compelling testimony about Defendant's claims of noncitizen registration. Dr. Minnite is an associate professor at Rutgers University–Camden, a tenured position, where her research focuses on American politics and elections. Dr. Minnite has extensively researched and studied the incidence and effect of voter fraud in American elections. Her published research on the topic spans over a decade and includes her full-length, peer reviewed book, The Myth of Voter Fraud , for which Dr. Minnite has received grants and professional distinction, and numerous articles and chapters in edited volumes. This topic has been the focus of her work and research for the past seventeen years. Dr. Minnite has been offered and accepted as an expert on the incidence and effect of voter fraud in numerous cases.

Notably, Dr. Minnite testified that when she began researching the issue of voter fraud, which includes noncitizen voter fraud, she began with a "blank slate" about the conclusions she would ultimately draw from the research. This stands in stark contrast to Mr. von Spakovsky's starting point as an advocate. In forming her opinions on the incidence of voter fraud and noncitizen registration in Kansas, Dr. Minnite relied on numerous quantitative, qualitative and archival sources. These include, among other sources, thousands of news reports, publicly available reports, court opinions, as well as various documents relied on by Defendant as evidence of noncitizen registration, including various iterations of Ms. Lehman's spreadsheet and underlying voter registration records. To evaluate these sources, Dr. Minnite employed a "mixed methods" research approach, in which different data sources are triangulated in order to identify patterns across the sources. This Court found on the record at trial Dr. Minnite's methodology is reliable under Daubert .

Although she admits that noncitizen registration and voting does at times occur, Dr. Minnite credibly testified that there is no empirical evidence to support Defendant's claims in the in this case that noncitizen registration and voting in Kansas are largescale problems. Of the nominal number of noncitizens who have registered and voted, many of these cases reflect isolated instances of avoidable administrative errors on the part of government employees and/or misunderstanding on the part of applicants.

This testimony is supported by the ELVIS records underlying the Lehman spreadsheet, discussed supra . Although the ELVIS records at times are less than pellucid, mostly due to the many codes insisted upon by Defendant's office to continue to track individuals registered by operation of law pursuant to this Court's order, Dr. Minnite could comment and interpret them without undergoing training as "an election administrator or DMV clerk." In fact, both Dr. Minnite and this Court can draw the reasonable inference from an ELVIS record where the applicant replied "No" they were not a United States citizen, that a State employee erroneously completed the voter registration application in the face of clear evidence that the applicant was not qualified.

Dr. Minnite's testimony is further supported by Dr. Hersh's expert testimony. Dr. Hersh explained that the number of purported incidents of noncitizen registration found by Defendant is consistent with the quantity of other low-incidence idiosyncrasies in ELVIS and in voter files more generally, and is suggestive of administrative errors. For example, 100 individuals in ELVIS have birth dates in the 1800s, indicating that they are older than 118. And 400 individuals have birth dates after their date of registration, indicating that they registered to vote before they were born. In a state with 1.8 million registered voters, issues of this magnitude are generally understood as administrative mistakes, rather than as efforts to corrupt the electoral process. Accidental registrations could have occurred as a result of clerks' administrative errors in inputting handwritten data from paper forms. Moreover, the very low incidence of voting among purported noncitizen registrants suggests that those individuals ended up in ELVIS due to accidents, as opposed to intentional unlawful registrations. The voting rate among purported noncitizen registrations on Mr. Caskey's TDL match list is around 1%, whereas the voting rate among registrants in Kansas more generally is around 70%. If these purported noncitizen registrations were intentional, one would expect these individuals to vote more frequently; the fact that they do not suggests that these registrations are the product of administrative mistakes by State employees or by the applicants themselves.

In short, the Court gives more weight to the careful, documented, and nonmisleading testimony of Dr. Minnite and Dr. Hersh on the issue of the significance of noncitizen voter registration in Kansas.

3. Statistical Estimates of Noncitizen Registration in Kansas

The remaining category of evidence offered by Defendant to demonstrate the degree of noncitizen registration in Kansas is statistical estimates generated by his expert, Dr. Jesse Richman. The Court admitted Dr. Richman's expert report and testimony, finding him qualified as an expert in the fields of elections, voter registration, survey construction and analysis, and political methodology.

Dr. Richman holds a M.A. and a Ph. D. in Political Science from Carnegie Mellon University. He is an associate professor at Old Dominion University and was the Director of University Social Science Research Center there for 3 years. Dr. Richman teaches research, research design, and advanced statistics, including statistical analysis. His academic research includes, among other topics, voting and participation, and he has published 12 or 13 peer-reviewed articles, several of which involve elections or voting.

Dr. Richman has published one peer-reviewed article on noncitizen registration, in the British journal, Electoral Studies . The article was based on data collected through a survey known as the Cooperative Congressional Elections Study ("CCES"), a large online survey concerning American voting behavior. Dr. Richman is not and has never been involved in designing or implementing the CCES. He has published one peer-reviewed paper based on a survey that he designed, and he has never published any peer-reviewed research addressing the accuracy of survey responses for a survey that he designed, or any peer-reviewed research involving his own efforts to compare survey responses to government records to assess the validity of those survey responses. He has never, other than in this case, designed or implemented any survey to measure citizenship rates of survey respondents.

Plaintiffs offered the testimony of their rebuttal expert, Dr. Stephen Ansolabehere to assess Dr. Richman's various statistical estimates. Dr. Ansolabehere is the Frank G. Thompson Chair at Harvard University in the Department of Government. He has been on the board of American National Election Studies for 12 years, which is the longest running political science research project in the country, was the founding director of the Caltech/MIT voting technology project, and has worked for CBS News since 2006 on the election night decision desk that designs the surveys used and the data collection process. He has published a substantial body of peer-reviewed work: 5 books and approximately 80 articles on a variety of topics, including survey research methods, statistics for analyzing large sample data, and for matching large surveys. He has received a variety of research grants. Dr. Ansolabehere has testified in numerous voting rights cases, which all cite to his testimony favorably.

Dr. Ansolabehere is the creator and principal investigator of the CCES, the survey on which Dr. Richman relied in his Electoral Studies article on noncitizen registration. Dr. Richman considers Dr. Ansolabehere to be knowledgeable about survey research, and believes that Dr. Ansolabehere has a good reputation as a political scientist among other political scientists. Indeed, the Court found his testimony and report to be persuasive, consistent, supportable, and methodologically sound. Dr. Ansolabehere opined that Dr. Richman's various estimates—collectively and individually—did not provide statistically valid evidence of noncitizen registration in Kansas. For the following reasons, the Court credits Dr. Ansolabehere's testimony and finds that Dr. Richman's estimates are not statistically valid.

a. Estimates of Noncitizen Registration or Attempted Registration in Kansas

Dr. Richman offers four different estimates based on four different data sources of noncitizen registration or attempted registration in Kansas: (1) 14 Kansas respondents in the 2006–2012 CCES who stated that they were noncitizens, out of which 4 stated that they were registered to vote, (2) records of approximately 800 newly-naturalized citizens in Sedgwick County, 8 of whom had records of pre-existing registration applications; (3) a survey of 37 TDL holders, 6 of whom stated that they were registered or had attempted to register to vote; and (4) 19 survey responses from a group of "incidentally-contacted" noncitizens, 1 of whom stated that they were registered or had attempted to register to vote. Dr. Richman failed to provide margins of error with his original report for the first three estimates he discussed, and he admitted during his testimony that such failure does not conform to peer-review standards for statistical estimates. Dr. Richman issued a supplemental report that included margin of error calculations, under various alternative methods. Nonetheless, all four of these estimates, taken individually or as a whole, are flawed.

i. CCES Survey Results—4 of 14 respondents

After extrapolating the CCES survey results of 4 out of 14 noncitizen registrations to an estimated noncitizen adult population in Kansas of 114,000, Dr. Richman estimated that 28.5%, or 32,000 noncitizens, were registered to vote, with a confidence interval of between 11.7% and 54.6%. The first problem with Dr. Richman's estimate is that the sample size is too small. As both Dr. Richman and Dr. Ansolabehere testified, estimates based on such small samples have large margins of error, and do not amount to reliable or probative statistical evidence. To be sure, Defendant discounted this estimate in his opening statement due to this flaw.

Second, Dr. Richman failed to demonstrate that the 14 CCES respondents were in fact noncitizens. Dr. Ansolabehere credibly testified, as the creator and principle investigator of that survey, that individuals who are U.S. citizens sometimes mistakenly respond that they are noncitizens, and published a peer reviewed article explaining this error. He explained that while this "citizenship misreporting" error occurred relatively infrequently, the number of errors is large when compared to the number of individuals who identify themselves as noncitizens on the CCES, and thus fatally contaminates any attempt to use the CCES to make statistical estimates about noncitizens. Indeed, Dr. Richman's published findings about noncitizen voting can be accounted for entirely by citizenship misreporting. In fact, a group of approximately 200 political scientists signed an open letter criticizing Richman's work on essentially the same grounds.

Third, Dr. Richman's CCES estimate suffers from registration overreporting. Dr. Hersh testified that social desirability bias sometimes causes individuals to respond to survey questions that they are registered to vote when they are not. In a peer-reviewed article, Drs. Ansolabehere and Hersh have documented overreporting of registration in the CCES. Dr. Richman testified that he has no reason to doubt Dr. Ansolabehere's findings on this issue. In fact, Dr. Ansolabehere explained that registration overreporting in Dr. Richman's survey sample is supported by the survey results—of the 4 respondents in the CCES sample who stated that they were registered to vote, only 1 can be validated to an actual voter file.

Finally, Dr. Richman did not weight the CCES sample to accurately reflect the population of Kansas. Even though Dr. Richman weighted his national estimates of noncitizen registration using CCES data in his Electoral Studies article, among other things, by race and Hispanic ethnicity, he did not conduct any weighting for his estimates of noncitizen registration based on the same underlying data in his expert reports.For all of these reasons, the Court gives no weight to Dr. Richman's estimate that 32,000 noncitizens registered or attempted to register to vote based on responses to the CCES.

ii. Records of Newly–Naturalized Citizens in Sedgwick County—8 of 791

To reach a total estimate based on the Sedgwick County naturalization records, Dr. Richman observed that in that county, "roughly 1 percent of newly naturalized citizens since January 1 2016 (8/791) turned out to have previously registered to vote while non-citizens." In his supplemental report, Dr. Richman applied 1.01% to the updated statewide estimate of noncitizens in Kansas, estimating that 1,169 noncitizens registered to vote across the state.

This estimate contains the largest sample size of Dr. Richman's various estimates, and therefore should offer the greatest statistical certainty. But as Dr. Ansolabehere explained, his results are not statistically distinguishable from zero, meaning there is so much uncertainty about this estimate that the number could be zero or close to it. "[I]t's just an expression of how much uncertainty there is and whether we accept the hypothesis that this is any more than a—a minimal or de minimus amount of non-citizens in the state attempting or registering to vote."

Dr. Ansolabehere calculated a theoretical margin of error of plus or minus 3.6%, meaning that Dr. Richman's 1% estimate is within the margin of error. Dr. Richman vehemently disagreed with Dr. Ansolabehere's calculation of the theoretical margin of error, which assumed that "the only source of variation in estimates is due to random sampling." He contends that his estimate is within the bounds of several alternative confidence intervals he calculated in the supplemental report, although he claims the "Wilson Score" method is recommended in the field for the types of estimates he utilizes in this case. Dr. Richman also points to the fact that his sample includes confirmed examples of individuals who registered and then re-registered upon naturalization, so to suggest that the true figure is zero or below is plainly erroneous.

But Dr. Ansolabehere explained that he applied a margin of error that uses a "P" of .5 because that is "the standard approach to calculating the standard errors" where, as here, there is no information about the assumptions of the survey researcher. Dr. Ansolabehere testified that is the conventional method used by political scientists. The "Exact," "Agresti", "Jefferys" and "Wilson Score" methods used in Dr. Richman's supplemental report "are all under specific assumptions and there's no reporting of any of the assumptions for the sample data collection, so I have no reason to believe that those were appropriate methods as opposed to just applying a bunch of methods that are in a toolbox." Dr. Richman points to no assumption contained in his report that would justify a different method. The Court is more persuaded by Dr. Ansolabehere's testimony about the appropriate margin of error for Dr. Richman's estimates and therefore finds that Dr. Richman's estimate based on Sedgwick County data from naturalization ceremonies is not statistically distinguishable from zero.

Additionally, the estimate of noncitizen registration based on Sedgwick County naturalization data is not based on a representative sample of noncitizen adults in Kansas. This sample includes only newly-naturalized citizens, and categorically excludes undocumented immigrants and legally present noncitizens who have not naturalized. Dr. Ansolabehere testified that noncitizens who naturalize tend to be older, more stable in their living situations and better-educated than noncitizens who do not. All of these factors tend to correlate with higher registration rates. As a result, an estimate of noncitizens based on naturalized citizens is likely to overestimate the number of noncitizens who are registered to vote in Kansas. Dr. Richman speculated that registration rates among those about to naturalize are likely to be lower than registration rates among other noncitizens, but he acknowledged that he has never done any research that attempts to quantify or compare registration rates among those noncitizens who naturalize and those who do not, and the only authority he cites in support of that proposition is a news interview with a single former Immigration and Customs Enforcement officer.

Dr. Richman did not weight the sample of newly-naturalized citizens to accurately reflect the noncitizen population of Kansas. Even though he weighted his estimates of noncitizen registration in his published research by, among other things, race and Hispanic ethnicity, and he weighted his suspense list survey by age and gender, he did not conduct any weighting for his estimates of noncitizen registration. Dr. Richman did not collect information from the voter registration applications of these individuals that, for example, could have enabled him to weight his sample by age. Finally, Defendant did not establish that this estimate is based on noncitizens who had successfully registered to vote prior to naturalizing. In his initial report in this case, quoting Ms. Lehman, Dr. Richman noted that the 8 identified noncitizen registrants "were already in ELVIS," but he admitted that individuals who merely attempt to register to vote can be found in ELVIS, even if they never successfully registered.

For these reasons, the Court gives no weight to Dr. Richman's estimate that 1,169 noncitizens registered or attempted to register to vote based on responses to the CCES.

iii. TDL List Survey—0 or 6 of 37

Extrapolating the TDL list survey results statewide, Dr. Richman estimated in his initial report that 16.5%, or up to 18,000 noncitizens, could be registered to vote. In his opening statement, Defendant cited this figure of 18,000, and described it as the "best estimate" of noncitizen registration in the State of Kansas. Yet, Dr. Richman ultimately weighted his estimate based on the TDL sample to match the overall noncitizen population in Kansas, which reduced his original estimate to about 13,000 noncitizen registrations. Dr. Richman testified that he considers this weighted estimate to be more reliable than his original 18,000 estimate.

Like the previous two estimates, Dr. Richman's estimate based on the TDL list suffers from flaws that give it little probative value. First, the sample size of 37 is too small to draw credible estimates. Dr. Richman himself admitted that the TDL sample has "a very modest sample size," and that the estimate therefore has "substantial uncertainty." Assuming the validity of Dr. Richman's confidence interval using the Wilson Score method, it is still a large interval of over 20 percentage points. As the Court has already explained, this is too uncertain to produce a reliable estimate of noncitizen registration.

Second, Dr. Richman's estimate of noncitizen registration based on the TDL list survey is attributable entirely to registration overreporting, i.e., individuals who said that they had registered or had attempted to register, but who in fact had done neither. Dr. Hersh looked for the 6 individuals who self-reported being registered to vote or having attempted to register to vote in the ELVIS database, but was unable to find them. This indicates that none of them even attempted to register to vote. Dr. Richman did not conduct a similar analysis, and did not dispute Dr. Hersh's findings in this regard.

Third, as with the newly-naturalized citizen list, Dr. Richman's calculations do not provide information about the number of noncitizens who successfully registered to vote. The survey instrument he used asked what Dr. Ansolabehere referred to as a "double-barreled" question: did the person register to vote or attempt to register to vote.

Finally, Dr. Richman did not provide a response rate for the TDL survey. Although he provided an estimate for the overall response rate for all of the surveys performed by Issues and Answers of 16%, he was unable to provide a response rate for his survey of TDL holders. It is therefore impossible to assess the statistical reliability of the TDL survey. He explained on direct examination that this response rate compares well to the response rate for national polling, but fails to consider whether non-response bias could have affected the results of this survey. Dr. Richman's failure to consider the response rate for the TDL survey and whether it was affected by non-response bias further reduces the value in this estimate of noncitizen registration.

iv. Incidentally–Contacted Individuals—1 of 19

Finally, as to the 19 incidentally contacted noncitizens from his survey, Dr. Richman estimates that 5.3%, or 6,000 noncitizens, registered or attempted to register to vote, with a confidence level of between .9% and 24.6%.

Again, there are several methodological flaws with this estimate. First, using the conventional method for calculating the margin of error, the estimate is not statistically distinguishable from zero. Second, the sample size is "extremely small," as Dr. Richman conceded in his report, and therefore has low statistical power. Using any of the methods of calculating margins of error that Dr. Richman employs in his supplemental report, the confidence interval for this estimate is more than 20 percentage points, which is too large to form a probative estimate of noncitizen registration in Kansas. Third, the estimate is based on the same faulty survey question as the TDL survey, so there is no way to distinguish between a respondent who registered or attempted to register to vote. And finally, the sample was not weighted to match the noncitizen population in Kansas. For all of these reasons, the Court agrees with Dr. Richman's assessment in his report that his extrapolation based on this single survey response is "very uncertain."

v. Survey of Registered Voters—0 of 576

Included in Dr. Richman's report, but discounted by him during his testimony, are survey results from another component of the Issues and Answers survey. This survey contacted more than 500 registered voters in 4 Kansas counties: Ford, Finney, Grant, and Seward. Dr. Richman explained in his report that these counties were selected due to their "large non-citizen populations." Zero respondents indicated that they were noncitizens, yet Dr. Richman did not calculate an estimate of noncitizen registration based on this particular survey, and gave the results short shrift. He testified that if he had estimated the rate of noncitizen registration based on this survey, it would be zero.

In his supplemental report, Dr. Richman argued that because this sampling included citizens, "it is inappropriate to include as analogous to the other items," as it "is an estimate of the percentage of 2008 through 2012 voter registrants who are still on the voter rolls and still at the phone number provided when the [sic] registered, and also non-citizens."

Given that the Court's task in this matter is to determine the number of noncitizens who successfully registered to vote during the attestation regime, i.e., before 2013, the Court is not persuaded by Dr. Richman's conclusory legal assertion about the relevance of this data. To the contrary, this data appears to be highly relevant to the Court's endeavor. If Defendant's contention is true that substantial numbers of noncitizens registered to vote before the DPOC law was passed, one would expect this number to be higher than zero, particularly given the robust sample size as compared to Dr. Richman's other estimates, and even assuming some level of social desirability bias. These results do not support Dr. Richman's overall opinion, and the Court is not persuaded by his attempt to eliminate it from his analysis.

vi. Meta–Analysis

Because Dr. Richman failed to identify a "best estimate" among his myriad calculations, Dr. Ansolabehere performed a "meta-analysis" of Dr. Richman's four estimates along with the results of the registered voters survey, weighting them based on sample size. Assuming all this data is accurate, taken together Dr. Ansolabehere estimates a 1.3% rate of noncitizen registration with a wide margin of error of 7.6%. Given this margin of error, a collective estimate of Dr. Richman's results is not statistically significant—there is so much uncertainty associated with them that "there can be no confidence that the number of non-citizen registration is more than the nominal cases in the sample. It is not possible to reject the hypothesis that the rate of non-citizen registration in the State of Kansas is different from zero." Dr. Richman responded to these calculations in his supplemental report, producing his own meta-analysis for the first time with his own confidence interval calculations, ranging from 1.1% to 1.8%. As already described, the Court was more persuaded by Dr. Ansolabehere's explanation of the appropriate margin of error that applies to Dr. Richman's calculations.

The Court finds Dr. Richman's testimony and report about the methodology and basis for concluding that a statistically significant number of noncitizens have registered to vote in Kansas, are confusing, inconsistent, and methodologically flawed. Most importantly, his refusal to opine as to the accuracy of any one estimate undercuts this Court's ability to determine that any one of his wildly varying estimates is correct. The extrapolations included in his report and testimony range from 0 to 32,000 noncitizen registrations. Given this range of estimates, most of which are based on sample sizes that cannot produce reliable results, this Court finds none of them represents an accurate estimate of the numbers of noncitizens registered to vote in Kansas.

b. Survey Results

Dr. Richman further testified about his survey of over 1,300 individuals on the suspense list. Seven of these respondents reported that they were noncitizens. After weighting, he estimated that these results demonstrate that .7% of the suspense list are noncitizens, although his report concedes that "given the small sample size, any inference about the precise magnitude of the non-citizen presence on the suspense list is fraught with substantial uncertainty." Beyond Dr. Richman's admission that the sample size renders this estimate substantially uncertain, there are other problems with the estimate. First, although Dr. Richman did not provide a margin of error, Dr. Ansolabehere did, and determined that Dr. Richman's estimate that 0.7% of the people on the suspense list are noncitizens is statistically indistinguishable from zero. Dr. Richman did not dispute Dr. Ansolabehere's finding in this regard. Moreover, taken at face value, the corollary to Dr. Richman's estimate that 0.7% of people on the Suspense List are noncitizens, is that more than 99% of the individuals on the suspense list are United States citizens.

Second, Dr. Richman weighted his sample by various characteristics that, in his view, could correlate with citizenship status, including age, gender, party identification, geographic region, year of registration, and "foreign name," in order to account for differences between the sample and the suspense list. After weighting the sample, he concluded that 117 individuals on the suspense list are noncitizens. Dr. Richman and a graduate student assistant went through the suspense list and determined which names were, in their view, foreign. Neither Dr. Richman nor his assistant had any experience in identifying so-called foreign names. By his own admission, their determinations were subjective and based primarily on whether the name was "anglophone," meaning originating in the British Isles. Dr. Richman also testified that their work was performed quickly, and that they made many mistakes along the way. A review of their coding revealed inconsistencies; for example, of five individuals with the last name of "Lopez," two were coded as foreign and three were coded as non-foreign. On cross examination, Dr. Richman admitted that he would have coded Carlos Murguia, a United States District Judge sitting in this Court, as foreign.

E. Alternative Methods of Enforcing Citizenship Eligibility

The parties presented evidence about several methods of enforcing the State's citizenship eligibility requirement for voter registration, other than the DPOC law.

1. DOV List Comparisons

The DOV has compared the list of individuals on the suspense list to information in the DOV database concerning driver's license holders who presented proof of permanent residency (or "green cards") in the course of applying for a driver's license, and identified possible non-citizens. There is no evidence that the SOS's Office has conducted vigorous follow-up investigations on these individuals to determine if they were still noncitizens at the time they applied to register to vote.

Similarly, as the Court has already found, Defendant's office has compared the TDL list to the ELVIS database four times in the last decade. As of January 30, 2017, Defendant has identified 79 individuals through matching these lists. While showing some false positives, this method has allowed Defendant to identify some noncitizens on the voter rolls.

2. DOV Training

Brad Bryant, the former Elections Director for the SOS's Office, testified by deposition that in the mid-1990's into the early 2000's, the DOV believed that under the NVRA it was required to offer voter registration to any driver's license applicant, regardless of citizenship. "[T]here was nobody telling them be careful if somebody is not a citizen." Mr. von Spakovsky testified that after the NVRA was passed, there was a nationwide problem with State motor vehicle offices offering voter registration to noncitizens. He opined that motor vehicle officials did not want their clerks making judgment calls about whether an applicant should be offered the right to register to vote. He testified that several states took the position that voter registration should be offered to every applicant.

Bryant recalls a greater effort in 2010 when Chris Biggs was SOS, to make clear to the DOV that its employees need not offer voter registration forms to noncitizens. The SOS's Office relied primarily on posters sent to each DOV office throughout the State, to clarify that noncitizens could not register to vote. Nonetheless, since the time of this effort in 2010, there is evidence that DOV employees sometimes mistakenly offered noncitizens voter registration applications, and that even when applicants denied U.S. citizenship, the application was completed by the clerk, creating an ELVIS file. This evidence included the email between Ms. Lehman and Mr. Bryant lamenting this problem with DOV registration of obvious noncitizens. Since Mr. Caskey has been Elections Director, it has been the SOS's Office's policy that noncitizens should not be offered the opportunity to register to vote at the DOV, yet his office has not drafted a written instruction to DOV clerks to not offer voter registration applications to noncitizens.

Another problem revealed by this record, and confirmed by Mr. Caskey, is that at least for some period the DOV did not accept DPOC offered by applicants if it was unnecessary to fulfill the proof of residency requirement to obtain a driver's license. Mr. Caskey testified, "I'm aware of many cases where a person brought a documentary proof of citizenship document and it was not needed as part of their driver's license application and was not scanned in the system and as part of their voter registration record."

Mr. Caskey testified as follows about the scope of the SOS's role in training DOV staff:

Our office routinely talks with the Division of Motor Vehicles executive staff as well as their trainers concerning requirements that the Division of Motor Vehicles has to comply with the National Voter Registration Act as it relates to offering the opportunity to register to vote to anyone who conducts a transaction at the Division of Motor Vehicles.

We also periodically review any changes in the state and federal law regarding elections. Previously we have reviewed the training materials that their trainers use. DMV routinely uses kind of a train-a-trainer approach where they have a group of trainers who then train their individual field offices. So we have provided—in years past, we provided posters and reviewed their training materials, you know, over the last 10 to 15 years.

DOV clerks in Kansas receive, on average, no more than 30 minutes of training regarding motor voter registration laws during their two-day in-classroom training. They were provided updated training after the SAFE Act became effective in 2013. Between February 2015 and June 2016, the SOS's Office did not provide any new written training materials to the DOV concerning motor voter registration laws. And Mr. Caskey testified that there have been no recent changes to DOV training or procedures.

Mr. Caskey is also charged with providing instruction and training to the 105 county election officials on the rules and regulations governing elections. There is an electronic training manual maintained by Mr. Caskey that used to be available online. It was last revised in 2012 to reflect the DPOC law. The office does not update the document as procedures change. Instead, Mr. Caskey testified that his office provides training to the counties by e-mail and phone, and that they attend regional and statewide meetings. Mr. Caskey and Defendant insist, however, that they have no authority to force the counties to comply, and that it is impractical to monitor whether they are implementing the SOS's policies.

3. Department of Homeland Security's Systematic Alien Verification for Entitlements Program

The Systematic Alien Verification for Entitlements ("SAVE") program is overseen by the United States Department of Homeland Security ("DHS"). In a letter dated August 20, 2012, DHS notified Kansas that

States will be able to access SAVE to verify the citizenship status of individuals who are registered to vote in that state provided that the requesting state has a signed information sharing agreement with the Department of Homeland Security and that each state be able to supply for each individual it seeks to verify (1) a specific type of unique identifier like an alien number or certificate number that appears on immigration-related documentation, and (2) a copy of the immigration-related documentation in question to complete the verification process.

On this basis, Defendant maintains that "A-numbers" or Alien Verification Numbers ("AVNs") are required to run SAVE searches, such that this is not a viable option because this information is not required on the registration application.

But the trial record demonstrates two instances where the SOS's Office has confirmed noncitizenship status of registrants through DHS: (1) it confirmed noncitizenship of three individuals who stated they were noncitizens on juror questionnaires in 2017; and (2) it confirmed noncitizenship of 6 respondents to Dr. Richman's TDL survey. Moreover, Mr. Caskey acknowledged in his testimony that other states such as Florida, Virginia, and Colorado use, or have attempted to use, SAVE for voter registration purposes. Mr. Caskey has not contacted any of these states' election officials to learn how they utilized SAVE.

Other agencies in Kansas have access to noncitizen documentation that could be used for SAVE searches. For example, DOV collects noncitizen documents when it issues TDLs. While the SOS obtains information from the DOV to confirm whether a driver's license applicant has provided DPOC, it does not obtain A-number information or copies of noncitizen documents from the DOV. The SOS has not requested from the Legislature a law that would enable it to obtain A-number information or copies of noncitizen documents from the DOV, and has not investigated whether other agencies in Kansas (such as the KDHE) have AVN information that could be used for SAVE searches.

4. Prosecutions

Former Deputy SOS Eric Rucker, an experienced prosecutor, testified by deposition that criminal prosecutions can prevent and deter criminal conduct. Since July 1, 2015, Defendant has had the independent authority to prosecute any person who has committed or attempted to commit any act that constitutes a Kansas elections crime. Since July 1, 2015, the SOS's Office has become aware of multiple instances of noncitizens registering to vote in Kansas. Since obtaining prosecutorial authority over Kansas elections crimes, the SOS has filed zero criminal complaints against a noncitizen for allegedly registering to vote. As of June 20, 2017, Defendant has filed one criminal complaint and obtained one conviction against an individual who actually voted while being a noncitizen.

5. Juror Questionnaires

In Kansas, people who are called for jury service are sent jury duty questionnaires that include a question about U.S. citizenship. District Courts send to the SOS's Office on at least a monthly basis the lists of individuals who requested to be excused from jury service based on their claims of noncitizenship. The SOS's Office has compared lists of individuals who answered on their jury questionnaires that they were not citizens with the voter registration roll.

As of March 24, 2017, Defendant identified 3 individuals who were on the voter rolls but who had self-identified as noncitizens on their jury questionnaires. An investigator employed by Defendant provided the full names and dates of birth of the 3 individuals who had self-identified as noncitizens on their jury questionnaires to the DHS via email. After receiving the full names and dates of birth of the 3 individuals who had self-identified as noncitizens on their jury questionnaires from the investigator, DHS responded with an email describing information known to DHS about the immigration and citizenship status of these 3 individuals. The response from DHS was the first time in Mr. Caskey's experience that DHS has aided the SOS, despite seeking assistance in the past. Only noncitizens called for jury duty can be identified by comparing ELVIS records with jury questionnaires.

III. Justiciability Challenges

Defendant has challenged Plaintiffs in both cases repeatedly on standing and mootness grounds. Defendant challenges Plaintiff Bednasek's standing because he has access to DPOC yet fails to produce it, and because he is a Texas resident. Defendant asserts that the claims of the individual Fish Plaintiffs are moot. The Court has addressed these challenges in detail in prior orders, but briefly addresses them again to the extent they are raised in Defendant's proposed findings of fact and conclusions of law, and addresses the mootness of Mr. Fish's claim sua sponte.

A. Parker Bednasek

Defendant challenges Bednasek's standing on the basis that he is a Texas, not Kansas, resident, and because he has access to DPOC. The Court has twice addressed and rejected Defendant's standing arguments in prior, lengthy orders. The Court incorporates by reference those rulings, and finds no further evidence presented at trial changes those decisions.

B. William Stricker, III, Thomas Boynton, Douglas Hutchinson, and Steven Wayne Fish

Defendant challenges the claims asserted by Plaintiffs Stricker, Boynton, and Hutchinson in Case No. 16–2105 on mootness grounds because they have become fully registered while this case has been pending and a favorable decision would not change their status. These challenges have been raised and rejected. The record shows that all 3 Plaintiffs' voter registration applications were canceled under the DPOC law before they were resurrected by operation of the Court's preliminary injunction order. But for that order, which required Defendant to register motor voter applicants whose applications had been suspended or canceled, these Plaintiffs would have been required to file new applications for registration to become registered.

Mr. Stricker's voter registration application was canceled in the ELVIS system in 2015. His registration was reinstated by operation of the preliminary injunction on June 22, 2016. Only after his registration was reinstated pursuant to the Court's preliminary injunction order did Defendant investigate behind the scenes and locate a citizenship document. Unfortunately for Mr. Stricker, this was well after he was denied the right to vote in the 2014 election.

On November 5, 2015, Prof. Boynton's voter registration application was canceled. According to Mr. Caskey, the SOS's Office later found a citizenship document through the DOV web portal after that access was granted in May 2016, after Prof. Boynton's registration was reinstated by operation of the preliminary injunction. Prof. Boynton's ELVIS file shows that a certified United States birth certificate was found and added on August 4, 2014 notwithstanding the fact that Mr. Caskey testified that Mr. Boynton did not apply to register to vote until November 4, 2014, Election Day.

Plaintiff Hutchinson's 2013 motor voter application was suspended for failure to provide DPOC, and it was ultimately canceled on January 27, 2016. He was registered by operation of this Court's preliminary injunction order. On July 30, 2016, his status changed to active after he submitted DPOC at the Johnson County Election Office and is now considered "fully registered."

But for this lawsuit and the preliminary injunction, these applications would not have been reinstated and they each would have been required to reapply to register to vote. Under these circumstances, the Court finds Defendant has not met his burden of demonstrating mootness for the same reasons explained in the Court's May 4, 2017 Memorandum and Order. Defendant's unilateral enforcement actions of the statute, which have been a moving target since this case's inception, and which were possible due to this Court's Order, do not render these Plaintiffs' claims moot. In fact, this Court previously warned Defendant that if he "continues his pattern of picking off Plaintiffs through targeted back-end verifications in an attempt to avoid reaching the merits of this case, the Court may be inclined to revisit its previous decision denying Plaintiffs' motion for class certification."

There is one exception to the Court's mootness ruling, which this Court addresses sua sponte because it is jurisdictional. "If an intervening circumstance deprives the plaintiff of a 'personal stake in the outcome of the lawsuit,' at any point during litigation, the action can no longer proceed and must be dismissed as moot." Mr. Fish testified that in September or October 2016, he relocated within Douglas County and changed his address with the DOV in person. At that time, Mr. Fish filled out a second voter registration application and provided his birth certificate. He is now registered to vote, having provided DPOC. The evidence demonstrates that Mr. Fish's active registration status is therefore not due to the Court's preliminary injunction, but his voluntary action of reapplying to register to vote at which time he provided DPOC. Unlike the other 3 Plaintiffs, whose applications were canceled but resurrected due to the unilateral efforts of Defendant, Mr. Fish filed a new application in full compliance with the DPOC law while this action was pending. Therefore, although he suffered an injury at the time the Complaint was filed, he no longer suffers an injury that can be redressed in this case and his remaining claim must be dismissed as moot.

IV. Conclusions of Law in Fish v. Kobach , 16–2105

The parties in the Fish case went to trial on the only remaining claim in this case—Count 1, which alleges a violation of § 5 of the National Voter Registration Act ("NVRA") based on preemption under the Election Clause in Article 1 of the United States Constitution. Section 5 of the NVRA requires that every application for a driver's license, "shall serve as an application for voter registration with respect to elections for Federal office." Subsection (c)(2)(B)–(C) of § 5 provides:

(2) The voter registration application portion of an application for a State motor vehicle driver's license— .... (B) may require only the minimum amount of information necessary to

(i) prevent duplicate voter registrations; and

(ii) enable State election officials to assess the eligibility of the applicant and to administer voter registration and other parts of the election process;

(C) shall include a statement that—

(i) states each eligibility requirement (including citizenship);

(ii) contains an attestation that the applicant meets each such requirement; and

(iii) requires the signature of the applicant, under penalty of perjury.

Under the NVRA, Defendant is "responsible for coordination of State responsibilities" under the NVRA.

Shortly after this case was filed, Plaintiffs successfully moved for a preliminary injunction based on their likelihood of success on the merits of their § 5 claim. Defendant appealed, and the Tenth Circuit affirmed, providing detailed guidance on whether the Kansas DPOC law is preempted by § 5's mandate that a motor-voter registration application contain the minimum-amount of information necessary for the state to exercise its eligibility-assessment and registration duties.

In its opinion, the Tenth Circuit set forth the applicable rules of statutory interpretation and preemption under the Elections Clause, interpreted the NVRA's requirements under § 5, and applied that interpretation to the facts as found by this Court in its preliminary injunction order. In the course of its detailed analysis, the Tenth Circuit "rejected Secretary Kobach's readings of the NVRA." As the Court previously explained on summary judgment, under both the law of the case doctrine, and the mandate rule, the Tenth Circuit's opinion with regard to issues of law governs at all subsequent stages of the litigation. The Court therefore proceeds to apply the standards announced by the Tenth Circuit in its October 19, 2016 published opinion in this case to the trial record. The Court once again declines to revisit Defendant's arguments that were resolved by that opinion. The Tenth Circuit held that the attestation requirement in § 5(c)(2)(C) presumptively satisfies the minimum-information requirement for motor voter registration in subsection (c)(2)(B). However, this presumption is rebuttable if the state can demonstrate "that the attestation requirement is insufficient for it to carry out its eligibility-assessment and registration duties." The court went on:

More specifically, in order to rebut the presumption as it relates to the citizenship criterion, we interpret the NVRA as obliging a state to show that "a substantial number of noncitizens have successfully registered" notwithstanding the attestation requirement. In EAC , we held that the EAC was not under a nondiscretionary duty to add state-specific DPOC instructions to the Federal Form at two states' behest. We reached this conclusion because "[t]he states have failed to meet their evidentiary burden of proving that they cannot enforce their voter qualifications because a substantial number of noncitizens have successfully registered using the Federal Form." The failure to make such an evidentiary showing was seemingly dispositive there of Secretary Kobach's Qualifications Clause challenge.

...

This results in the preemption analysis here being quite straightforward: if Kansas fails to rebut this presumption that attends the attestation regime, then DPOC necessarily requires more information than federal law presumes necessary for state officials to meet their eligibility-assessment and registration duties (that is, the attestation requirement). Consequently, Kansas's DPOC law would be preempted.

In a footnote, the court explained that if a state could show that attestation does not satisfy the minimum-information standard by demonstrating that substantial noncitizens are able to register to vote notwithstanding attestation of citizenship, then this Court would need to consider whether DPOC should be deemed "adequate to satisfy" the minimum-information standard. This second inquiry would require the state to "show that nothing less than DPOC is sufficient to meet those duties."

In its preliminary injunction order, this Court found that between 2003 and the effective date of the DPOC law, 14 noncitizens had registered or attempted to register to vote in Sedgwick County, Kansas. The Tenth Circuit found that this number "fall[s] well short of the showing necessary to rebut the presumption that attestation constitutes the minimum amount of information necessary for Kansas to carry out its eligibility-assessment and registration duties." In addressing this evidence, the court considered and rejected Defendant's argument that "even if one noncitizen successfully registers under the attestation regime, then DPOC is necessary to ensure applicant eligibility." This is because in adopting the NVRA registration procedures, Congress intended "to ensure that whatever else the states do, 'simple means of registering to vote in federal elections will be available.' " If 1 vote by a noncitizen is too many, then states would be able to justify even harsher means of verifying citizenship. The court explained, "[t]he NVRA does not require the least amount of information necessary to prevent even a single noncitizen from voting."

After remand, the Court reopened discovery in the Fish matter related to the Tenth Circuit's test. Defendant was given the opportunity to retain experts and marshal evidence to meet his burden of demonstrating that "a substantial number of noncitizens have successfully registered to vote under the attestation requirement" in order to rebut the presumption that attestation meets the minimum-information requirement of § 5 and that nothing less than DPOC is sufficient to meet his eligibility-assessment and registration duties under the NVRA. As described below, the Court finds that on the trial record Defendant has failed to make a sufficient showing on the first inquiry. Moreover, even if Defendant could demonstrate a substantial number of noncitizen registrations, he has not demonstrated that nothing less than the DPOC law is sufficient to enforce the State's citizenship eligibility requirement.

A. Substantial Number of Successful Noncitizen Registrations Under Attestation

Under the first part of this test, the parties dispute the meaning of "substantial." Defendant has argued that "substantial" should be interpreted to mean any number that can change the outcome an election. Plaintiffs argue that "substantial" must be evaluated in comparison to the number of total registered voters. In its summary judgment order, the Court provided the parties with guidance as to how it would apply the standard. After reviewing the Tenth Circuit's opinion in this case, and the decisions upon which it rested, this Court found that Defendant's showing must go beyond the number of registrations that would impact the outcome of an election to be substantial. Instead, the Court considers the number of noncitizen registrations in relation to the number of registered voters in Kansas as of January 1, 2013, when the DPOC law was passed, and is otherwise guided by legal authority cited with approval by the Tenth Circuit in its October 2016 ruling when determining whether Defendant's evidence meets the threshold of "substantial."

During his opening statement at trial, Defendant invited the Court to adopt a third approach to substantiality he deemed a "functional failure test." Under Defendant's test, the Court would determine whether a reasonable person would find that the attestation requirement failed to perform the function of preventing noncitizens from registering to vote. Defendant offers no explanation about how the Court is to apply this test, nor any authority for using a "reasonable person" test. Moreover, Defendant's proposed approach effectively modifies the Tenth Circuit's test by removing the words "substantial number." The plain meaning of the word "substantial" when describing an amount, means "considerable in quantity; significantly great." By asking this Court to consider whether attestation "functions" a certain way, instead of quantifying the noncitizens allowed to vote under attestation—a straightforward question—this inquiry runs afoul of the test formulated by the Tenth Circuit.

Defendant makes no mention of his "functional failure test" in his proposed findings of fact and conclusions of law. Instead, he argues for the first time that the Court should look to social security cases construing "substantial" in the context of reviewing administrative law judges' ("ALJ") decisions about whether "substantial gainful activity exists in significant numbers in the national economy," an inquiry relevant to a disability determination under the Social Security Act. The Court does not find this line of cases helpful or relevant in determining the meaning of substantial in this case. First, the social security cases review ALJ decisions to determine if they are supported by substantial evidence in the record, a different standard of review than this Court employs during a civil bench trial. Second, those cases construe the term "significant number," not "substantial number," in an entirely different context, guided by such factors as "the level of claimant's disability; the reliability of the vocational expert's testimony; the distance claimant is capable of travelling to engage in the assigned work; the isolated nature of the jobs; the types and availability of such work, and so on." Obviously, none of these factors are relevant to the question before the Court in this case. Instead, the Court considers whether Defendant's evidence of noncitizen registration is substantial according to the guidance provided at summary judgment, which relied on cases determining this question in the context of noncitizen voter registration.

For the reasons already explained, the Court finds no credible evidence that a substantial number of noncitizens registered to vote under the attestation regime. The only information about Kansas registration rates relied upon by Mr. von Spakovsky was provided to him by Mr. Caskey, and the Court has already evaluated that underlying data in more detail than Mr. von Spakovsky, who simply accepted the numbers as true. His generalized opinions about the rates of noncitizen registration were likewise based on misleading evidence, and largely based on his preconceived beliefs about this issue, which has led to his aggressive public advocacy of stricter proof of citizenship laws. The Court likewise does not find Dr. Richman's opinion as to the numbers of noncitizen registration carry weight given the numerous methodological flaws set forth in the Court's findings of fact.

That leaves Defendant's empirical evidence of noncitizen registration. He has submitted evidence of 129 instances of noncitizen registration or attempted registration since 1999, but looking closely at those records reduces that number to 67 at most. Even these 67 instances are a liberal estimate because it includes attempted registrations after the DPOC law was passed, a larger universe than what the Tenth Circuit asked the Court to evaluate. Only 39 successfully registered to vote. And several of the individual records of those who registered or attempted to register show errors on the part of State employees, and/or confusion on the part of applicants. They do not evidence intentional fraud. As discussed below, in determining whether nothing less than requiring DPOC is sufficient to enforce the citizenship requirement, it matters whether noncitizens are intentional registrants or not.

Moreover, the Court is unable to find empirical evidence that a substantial number of noncitizens successfully registered to vote under the attestation regime. As stated in the Court's findings of fact, there are only 39 confirmed noncitizens who successfully registered to vote between 1999 and 2013 when the DPOC law became effective. This is but .002% of all registered voters in Kansas as of January 1, 2013 (1,762,330). Furthermore, the 67 confirmed registrations and attempted registrations between 1999 and 2018 amounts to only .004% of registered voters. Of the estimated 115,500 adult noncitizens in Kansas, .06% have successfully registered or attempted to register to vote since 1999. The number of attempted noncitizen registrations since the DPOC law became effective is .09% of the total number of individuals canceled or suspended as of March 31, 2016, for failure to provide DPOC. The Court finds none of these numbers are substantial when compared to the total number of registered voters, the total number of noncitizens in Kansas, or the number of applicants on the suspense/cancellation list as of March 2016. They all fall below 1%. Instead, these numbers support the opinions of Drs. Minnite, McDonald, and Hersh that while there is evidence of a small number of noncitizen registrations in Kansas, it is largely explained by administrative error, confusion, or mistake.

Defendant insists that these numbers are just "the tip of the iceberg." This trial was his opportunity to produce credible evidence of that iceberg, but he failed to do so. The Court will not rely on extrapolated numbers from tiny sample sizes and otherwise flawed data. Dr. Richman's estimates were not only individually flawed and wildly varied, but his refusal to opine as to the best method of estimating the iceberg renders them all suspect. Mr. von Spakovsky's opinions fare no better. His advocacy led him to cherry pick evidence in support of his opinion, and he failed to demonstrate knowledge of Kansas noncitizen registration that would allow him to reliably quantify the iceberg beyond Mr. Caskey and Ms. Lehman's testimony. While the Court acknowledges that Defendant has limited tools at his disposal to quantify the statewide numbers of noncitizen registrations, the Court does not assume as Defendant does that this means there must be additional, substantial cases of noncitizen registration. Instead, the Court draws the more obvious conclusion that there is no iceberg; only an icicle, largely created by confusion and administrative error.

B. Alternatives to DPOC

Although the Court finds Defendant has not met his burden of showing a substantial number of successful noncitizen registrations under attestation in Kansas, out of an abundance of caution it proceeds to consider whether "nothing less than DPOC is sufficient to meet" Kansas's NVRA eligibility-assessment and registration duties. The parties presented evidence about the following alternatives to enforcing the Kansas citizenship requirement: (1) better training of State employees, particularly at the DOV; (2) list matching; (3) reviewing juror questionnaires; (4) the SAVE program; and (5) prosecution and enforcement of perjury for false attestations.

The Court begins its analysis of this part of the test by finding that no system for detecting noncitizen registration is perfect because there is no way to completely eliminate human error. The experts for both sides agree on this point. But the evidence at trial showed that a greater effort to pursue several of these alternatives, taken together or individually, would be sufficient to meet Kansas's NVRA duties. The testimony of Drs. Minnite and Hersh established that many confirmed instances of noncitizen registration or attempted registration in Kansas were due to either applicant confusion or mistake, or errors by DOV and county employees, not intentional voter fraud. Lack of intent matters not as a means of determining legal liability, but because it frames the acceptable alternative approaches that would allow Defendant to better enforce the State's citizenship requirement while imposing a less burdensome process on Kansans who apply to register to vote. If most noncitizen registrations are due to mistake or administrative error, as opposed to intentional fraud, that fact shapes the best method for enforcing the citizenship requirement.

1. Training

The evidence made clear that several noncitizens who registered or "attempted" to register, according to Defendant, either did not intend to register to vote or did not understand that they were prohibited from registering to vote. Some applicants told the DOV clerk that they were not citizens, yet the clerk completed a voter registration application. For some period of time prior to 2010, the evidence established that the DOV had been offering voter registration to all applicants as a matter of course, even if the clerk knew that the applicant was a noncitizen. It is not difficult to understand why many noncitizens registered to vote during this period if they were offered a voter registration application notwithstanding their disclosure to the DOV clerk of noncitizen status. These applications are part of the universe forwarded to the county clerks that were flagged as attempted registrations instead of mistakes. There was also evidence of DOV and county clerk error in implementing the DPOC law prior to and after the preliminary injunction order became effective in June 2016. There was evidence that DOV clerks continued to offer voter registration to noncitizens after 2010. And there was evidence that DOV clerks did not accept DPOC presented by driver's license applicants, and therefore applicants were suspended or canceled despite fully complying with the law.

The SOS's Office could make better, more meaningful efforts toward training DOV employees charged with completing motor voter applications. While it is true that an effort was made back in 2010 to train DOV clerks that noncitizens should not be offered voter registration, Mr. Caskey could not quantify a sustained effort by his office to train these workers not to fill out applications for those who disclose they are noncitizens either by providing evidence of noncitizenship during the application process, or by answering "no" to the question about U.S. citizenship. The evidence suggests a shift in policy around 2010 when the SOS's Office came up with training posters for the DOV on this issue. Yet, the errors identified during Ms. Lehman's testimony involved applications after 2013, suggesting that the problems continued after 2010. Mr. Caskey's testimony provided no indication of how often he or the SOS's Office "talks with" DOV trainers about these policies, nor what "periodically review[ing]" their training materials entails. Although drivers' license clerks received updated training between February 2015 and June 2016, there have been no new written training materials since that time. Despite the interagency agreement between the SOS's Office and the DOV that became effective in May 2016, and the Court's preliminary injunction that became effective in June 2016, the record does not establish that robust, updated training was provided to the DOV. The Court is convinced that a greater effort at training DOV staff would reduce the amount of inadvertent noncitizen registrations. The Court is further convinced that without the burdensome DPOC law to enforce, Mr. Caskey and his staff would have far more resources to devote to this endeavor.

Mr. Caskey testified that there is a training manual for the counties that was last updated in 2012, which used to be available online. Instead of updating the manual, he sends out emails and holds regular conference calls with the counties when items must be updated. Therefore, there is no centralized, updated training manual at the counties' disposal containing policy guidance from the SOS. Instead, the counties apparently must maintain their own index of emails and conference call notes to determine current policy on voter registration under the SAFE Act, the enforcement of which has been a moving target for the last 5 years given the many changes in internal policy since the law was passed. Since the law was passed, Defendant implemented the cancellation regulation under K.A.R. § 7–23–15, and it has established interagency agreements with the DOV and KDHE. In addition, the preliminary injunction created additional duties for the counties. There is no record of exactly how many updates to the training manual have been implemented through informal emails and phone calls, but the Court can reasonably infer that there have been many. More consistent, centralized training for county clerks would be another less burdensome way to ensure they understand the State's eligibility requirement, and how to ensure to the best of their abilities that noncitizens do not inadvertently end up on the voter rolls.

2. DOV List Matching and Juror Questionnaires.

Defendant has relied on list matching between the TDL list and the ELVIS database to produce evidence of some noncitizen registrations in this case. He could certainly continue to compare these lists, and confirm noncitizenship through either the individual records, or through DHS's help, as he has done in previous cases. Moreover, there is evidence that Defendant could compare and investigate those applicants in the DOV database that presented green cards during the driver's license application process, with the ELVIS database.

Defendant also demonstrated that he can examine juror questionnaires for self-identified noncitizens who are called to jury duty from the voter rolls. Although these methods may generate false positives, as the Court discussed in its findings of fact, they at least offer a less burdensome starting point for investigation and confirmation. Given that Defendant currently uses these alternative means for detecting and addressing noncitizen registration, Defendant has failed to establish that nothing less than a DPOC requirement is sufficient to address the problem of noncitizen registration.

3. SAVE Database

Defendant maintains that he cannot rely on the SAVE database to determine noncitizen status of Kansas voter registration applicants, relying on a 2012 letter from DHS requiring A–numbers and a copy of immigration documentation in order to share information with Kansas. Yet, the evidence at trial demonstrated that DHS has confirmed noncitizenship status in the past without this information, and that other states without DPOC laws use SAVE under agreements with DHS. Mr. Caskey admitted that he has not attempted to contact these states to determine how they utilize SAVE, and whether it might be an acceptable alternative to the DPOC law in Kansas. And, it is not clear that the SOS's Office has leveraged information from other state agencies to access the information needed by DHS to confirm citizenship status of voter registration applicants. The Court finds that such an approach would be less burdensome than the DPOC law.

4. Prosecutions

Defendant already has prosecutorial authority over Kansas election crimes. Yet, since obtaining this authority, and despite claiming to have located 129 instances of noncitizen registration in Kansas, Defendant has filed zero criminal complaints against noncitizens for registering to vote. To the extent Defendant takes the position that these are all cases that meet the definition of perjury, or otherwise involve fraudulently registering to vote, his own office has taken the position that prosecuting such individuals should act as a deterrent to future registrations by noncitizens. However, as the Court has already found, the evidence at trial demonstrates that many of the 129 cited instances of noncitizen registration were mistakes or the result of administrative error, which may not be prosecutable and which may undermine the deterrent effect of future prosecutions. Given that Defendant has not meaningfully sought to utilize criminal prosecutions, at least when he detects intentional cases of noncitizen registration, he has failed to establish that nothing less than DPOC is sufficient to address the problem of noncitizen registration.

The second prong of the Tenth Circuit's test does not require Defendant to establish that an alternative to DPOC would eliminate noncitizen registration; indeed, all the experts agree that may not be possible given the component of human error involved. The test instead requires that Defendant demonstrate that nothing less than DPOC would be a sufficient alternative. He does not satisfy this test. Several alternatives exist, especially when taken together, that would be sufficient to reduce the nominal amount of noncitizen registration that occurs through an attestation regime. Thus, the Court finds that Defendant has failed to rebut the presumption that the attestation clause meets the minimum information principle in § 5 of the NVRA, and therefore orders judgment in favor of Plaintiffs on this remaining claim.

V. Conclusions of Law: Bednasek v. Kobach

Plaintiff Bednasek claims that the DPOC law unconstitutionally burdens his right to vote under the Fourteenth Amendment. The Supreme Court has made clear that there is no "litmus test" for considering a constitutional challenge to a State's election laws. Instead, the Court is to "first, consider the character and magnitude of the asserted injury to the rights protected by the First and Fourteenth Amendments that the plaintiff seeks to vindicate." Second, the court "must identify and evaluate the precise interests put forward by the State as justifications for the burden imposed by its rule." In considering the State's interest, the Court is to both "determine the legitimacy and strength of each" State interest, and also "consider the extent to which those interests make it necessary to burden the plaintiff's rights." The Court has explained the balancing test as follows:

Under this standard, the rigorousness of our inquiry into the propriety of a state election law depends upon the extent to which a challenged regulation burdens First and Fourteenth Amendment rights. Thus, as we have recognized when those rights are subjected to "severe" restrictions, the regulation must be "narrowly drawn to advance a state interest of compelling importance." But when a state election law provision imposes only "reasonable, nondiscriminatory restrictions" upon the First and Fourteenth Amendment rights of voters, "the State's important regulatory interests are generally sufficient to justify" the restrictions.

In 2008, the Supreme Court decided Crawford v. Marion County Election Board , which considered a challenge to an Indiana law requiring its citizens to present photo identification ("photo–ID") when voting in-person. Indiana identified the following interests to justify the law's burden on voters: (1) deterring and detecting voter fraud; (2) election modernization; and (3) safeguarding voter confidence. As to voter fraud, the Court acknowledged no record evidence of in-person voter fraud (the only kind of fraud the statute could address) at any time in Indiana. However, the Court found that "flagrant examples of such fraud in other parts of the country," "occasional examples [that] have surfaced in recent years" in other places, and "Indiana's own experience with fraudulent voting in the 2003 Democratic primary for East Chicago Mayor" involving the use of absentee ballots, "demonstrate that not only is the risk of voter fraud real but that it could affect the outcome of a close election." The Court found that the State's interest in preventing voter fraud was legitimate and proper. The Court also found that the State has an interest in modernizing elections, pointing to the NVRA and the Help America Vote Act ("HAVA"), which "indicate that Congress believes that photo identification is one effective method of establishing a voter's qualification to vote and that the integrity of elections is enhanced through improved technology." Finally, the Court acknowledged the "independent significance" of the State's interest in public confidence in the integrity of the electoral process.

In considering the burdens imposed by Indiana's photo-ID law, the Supreme Court distinguished between the types of burdens it imposes on voters. Burdens "arising from life's vagaries," such as a lost or stolen wallet, are not constitutionally significant because "the availability of the right to cast a provisional ballot provides an adequate remedy." Instead, the Court considered burdens imposed on those who are eligible to vote, but who do not possess a photo ID that complies with Indiana law. The Court found that the burden on this subgroup was low because Indiana issued free photo-ID cards to these individuals, and: "For most voters who need them, the inconvenience of making a trip to the BMV, gathering the required documents, and posing for a photograph surely does not qualify as a substantial burden on the right to vote, or even represent a significant increase over the usual burdens of voting." The Court found that the evidence demonstrated a heavier burden was placed on elderly persons born outside of Indiana, persons who have difficulty obtaining a birth certificate required to obtain a photo-ID, homeless persons, and persons with religious objections to being photographed. However, the severity of the burden on these groups is "mitigated by the fact that, if eligible, voters without photo identification may cast provisional ballots that will ultimately be counted. To do so, however, they must travel to the circuit court clerk's office within 10 days to execute the required affidavit." The burden would only be a constitutional problem if it was "wholly unjustified," and even then, the burden on "just a few voters" would be insufficient to facially invalidate the statute.

In balancing the State's interests against the burden on voters, the Court stressed that instead of weighing the burden that the law imposes on all voters, the plaintiffs asked the Court to look at only a narrow group of voters that experienced a special burden. The Court found that the evidence in the record was insufficient to quantify "either the magnitude of the burden on this narrow class of voters or the portion of the burden imposed on them that is fully justified." The petitioners presented no evidence of the number of registered voters lacking a photo ID, or of the specific burdens felt by the categories of burdened voters identified by the Court. Moreover, those with difficulty obtaining a photo-ID, such as the elderly, could vote absentee without presenting photo-ID. Thus, "on the basis of the record that has been made in this litigation, we cannot conclude that the statute imposes 'excessively burdensome requirements' on any class of voters." The Court declined to invalidate the entire statute on this showing. Given this guidance, the Court proceeds to consider the balancing test as it applies to the evidence in this trial record.

A. The State's Interests

Defendant submits that the DPOC law furthers the State's legitimate interests in preventing noncitizen registration, maintaining accurate voter rolls of only qualified U.S. citizens, and maintaining confidence in the electoral process. The Court finds that although all 3 are legitimate interests, Defendant failed to produce evidence that they are strong enough to outweigh the tangible and quantifiable burden on eligible voter registration applicants in Kansas who were not registered to vote before January 1, 2013.

The Court has already determined that at most, 67 noncitizens registered or attempted to register in Kansas over the last 19 years. Even looking beyond Kansas, Defendant's evidence of noncitizen registration at trial was weak. Dr. Richman's Electoral Studies article concluding that millions of noncitizens registered and voted was credibly dismantled by Dr. Ansolabehere, the architect of the survey upon which Dr. Richman's conclusions were based. He explained that Dr. Richman's findings in that article are based on a flawed data analysis, and over 200 political scientists wrote an open letter criticizing its methodology and conclusion. Similarly, the Court does not fully credit Mr. von Spakovsky's summary of reported incidents of noncitizen registration, given its inclusion of misleading and false assertions. While there is certainly some evidence of noncitizen voter registration nationally, the trial evidence did not demonstrate the largescale problem urged by Defendant.

Maintaining accurate voter rolls of only U.S. citizens is again a legitimate State interest, however, for the same reasons already described, evidence that the voter rolls include ineligible citizens is weak. At most, 39 citizens have found their way onto the Kansas voter rolls in the last 19 years. And, as Dr. Hersh explained, given the almost 2 million individuals on the Kansas voter rolls, some administrative anomalies are expected. In the case of Kansas, this includes 100 individuals in ELVIS with birth dates in the 1800s, and 400 individuals with birth dates after their date of registration.

Finally, the Court concludes that maintaining confidence in the electoral process has independent significance, as found in Crawford. B. Burdens Imposed by the DPOC Law

This Court previously denied summary judgment based on genuine issues of material fact as to the burdens imposed by the DPOC law in this case. The Court explained then that unlike the photo-ID cases largely relied on by Defendant that deal with requirements for casting an in-person ballot, the DPOC law is distinguishable because it applies to registration. There is no safety valve such as a provisional ballot that can serve to mitigate the burden on voters. Therefore, unlike the Indiana law in Crawford , an eligible Kansas applicant on the suspense or cancellation list does not have the option to fill out a provisional ballot, produce DPOC after the election, and have their ballot counted.

This distinction is illustrated by Mr. Stricker's experience—the burden imposed on him by the DPOC law disenfranchised him in 2014. He was offered a provisional ballot, but because he was not registered before the election, there was no way for him to cure and have his ballot counted after the election. The only way to cure a violation of the DPOC law is to submit DPOC before midnight on the day before the election. Of course, in Mr. Stricker's case, he did not know he was not registered the day before the election because he had provided a birth certificate at the time he applied to register, he told the driver's license clerk he wished to register, he signed the attestation of eligibility, and he was told that there was nothing more he needed to do, including that he did not need a "voting card." Similarly, Prof. Boynton brought a birth certificate to the DOV when he sought to register to vote, and he too believed that he had registered when he left the DOV. He learned for the first time that he was not in fact registered when he was offered a provisional ballot in the 2014 general election. Prof. Boynton was so frustrated by the experience that he declined to register to vote the next 2 times he applied for a driver's license, lamenting that he "might as well save [him]self the effort and say no this time."

The DPOC law's deterrent effect on registration is further supported by Ms. Ahrens's and Dr. McDonald's testimony. The DPOC requirement fundamentally changed the Kansas League's ability to accomplish its mission of encouraging and assisting citizens in registering to vote and voting. Ms. Ahrens explained that the law was a "dead hit" on this mission, stopping all its registration efforts until it could conceive of a safe copying policy—previously, the League could assist registrants by having them fill out an application and sign an attestation of eligibility. Now, they were required to ask for a sensitive, personal document and maintain a copy of that document with each registration application. In 2013 in Wichita, for example, the Kansas League helped register only 400 individuals, compared to 4000 the year before. The Kansas League was also hampered in helping young people register to vote for the first time—less than one-quarter of all students who tried to register to vote at one Kansas university could successfully complete their applications due to lack of immediate access to DPOC.

This testimony supports Dr. McDonald's expert opinion formed after analyzing the individual ELVIS records for those on the suspense and cancellation lists, that the DPOC law disproportionately affects the young and unaffiliated. Dr. McDonald explained that the consensus in social science is that barriers to voter registration increase the cost of voting and dissuade individuals from participating in the political process. Dr. Richman agreed with this proposition. And these groups—the young and unaffiliated—already have a lower propensity to participate in the political process and are less inclined to shoulder the costs associated with voter registration. As compared to photo-ID laws, the Kansas DPOC law is an absolute bar to registration for any applicant lacking access to such DPOC. And even for those that have access the additional steps to possess such a document, such as locating it from a family member or separately obtaining an underlying document, increase the "costs" of voting that Dr. McDonald testified dissuade individuals from participating in the political process.

As the Court explained in its summary judgment order, another distinguishing feature between this case and Crawford , is that the number of incomplete and canceled registration applications for failure to submit DPOC provides concrete evidence of the magnitude of the harm. These individuals all sought to register to vote but were blocked by the DPOC requirement. This evidence contrasts with the photo-ID cases, where courts were unable to determine how many people were unable to vote based on the photo-ID requirement, and therefore found the burden to be speculative. The administrative data presented in this case, coupled with the expert opinions of Dr. McDonald, Minnite, and Hersh, all support the burdensomeness of the law.

Of the tens of thousands of individuals whose voter registration applications have been suspended or canceled due to lack of DPOC, less than 1% have been confirmed to be noncitizens. There is no evidence to suggest that the rest are not U.S. citizens, and thus these applicants are all eligible to register to vote but have been unable to produce DPOC. Defendant argues that any burden imposed by the law only applies to individuals on the list who do not have access to DPOC, and only a small number of those prevented from registering under the DPOC law do not either possess DPOC, or have immediate access to it. He points to Dr. Richman's estimate that only 2.2% of people on the suspense list lack immediate access to DPOC. He also points to the McFerron survey. As discussed earlier in this opinion, the McFerron Survey is inadmissible. And even if it was admissible, it is riddled with so many methodological errors that the Court would give it no weight. Likewise, Dr. Richman's estimate that only 2.2% of the individuals on the suspense list lack DPOC list is flawed for the many reasons discussed in the findings of fact. Moreover, that survey tells the Court nothing about those whose applications were canceled, nor does it provide evidence about the universe of eligible unregistered Kansas citizens subject to this law.

Defendant argues that the law is easy to comply with, pointing to the ability to submit DPOC electronically, and the State's attempts to verify citizenship through the KDOR and KDHE. Moreover, Defendant argues that the suspense list is dynamic, and that most of the applicants on the list eventually come off the list either through the State locating a citizenship document, or because the applicant eventually submits a compliant document. First, there is no clear evidence about the number of applicants that have been cleared through the KDHE and KDOR process since it went into effect; Defendant opted not to update discovery requests with new suspense list figures before trial. But the suspense list before this Court's preliminary injunction order is a credible snapshot of the overall burdensomeness of the law—those figures represent the number of applicants impacted during the first 3 years it was enforced and subsumes the KDHE policy that was effective long before 2016. While the Court acknowledges that the KDOR agreement likely lowers the number of individuals on the suspense list somewhat, it could not resurrect applications that were canceled before the agreement became effective.

Ms. Lehman testified that her office received bounce-back notices from about one-third of the individuals on the suspense list, and surmised that because many of them moved, they should not be counted. First, there is no evidence of statewide bounce-back notices on this scale. Second, even if that rate of notices bounced back, it says nothing about the citizenship status of the recipients. Instead, it shows how burdensome the notice process is, and the fact that many of those impacted by the law are not receiving notice of (1) the fact that they are not registered to vote, after in at least some cases being told that they were at the time of application; and (2) what they need to do to cure the problem. In sum, unlike in the in-person voting photo-ID cases, there is evidence here that "substantial numbers of persons eligible to vote have tried" to register but have been unable to do so.

Defendant suggests that the hearing procedure in § 25–2309(m) mitigates the burden imposed by the DPOC law, because if a person lacks the ability to obtain one of the 13 forms of DPOC in the statute, the hearing procedure allows them to submit some alternative proof of citizenship. He claims that this procedure is easy to comply with, but the evidence does not support that statement. First, the hearing procedure in subsection (m) is not explained to applicants when they apply to register, nor to applicants who were suspended for lack of DPOC. Neither the small DOV receipt, nor the example notices sent by the counties, contain any language explaining the hearing option to applicants. None of the named Plaintiffs in either case recall this option being mentioned to them. This explains why only 5 individuals, out of the more than 30,000 individuals on the suspense and cancellation list in March 2016, availed themselves of this option in the 5 years that the law has been in effect.

Second, even for those individuals who were registered after going through the alternative hearing process, their experiences were burdensome. One such individual, Ms. French, testified at trial. Ms. French admitted to a newspaper after the hearing that she "thought it was strange that I had to go through this procedure to be able to vote." Although she whole-heartedly agreed with the law as a policy matter, her experience illustrates the many steps required to comply under this procedure, which took five months to accomplish in the spring and summer of 2016, incidentally during the very timeframe when the Fish case was filed and the preliminary injunction in that case was being heard and decided. The Court does not find it to be coincidental that Mr. Rucker became Ms. French's "friend" during this time period. The Court would not normally expect a high-level government official to invest the time and attention Ms. French described during her testimony to help an applicant locate citizenship records and navigate the hearing procedure, much less keep in close personal contact after the process is complete. Based on this level of individual attention, and the fact that the media was present at the hearing, it appears the State was motivated to help this applicant navigate the system and become registered through the hearing process. The Court therefore does not find that Ms. French is a typical example of how an applicant would expect to experience this process, assuming the average applicant was aware that this process was available.

The hearing records reveal that another applicant was represented by retained counsel at the hearing, and yet another was required to execute his own affidavit explaining that he had been born on a military base and was therefore a U.S. citizen. The Court finds that this alternative procedure adds, not subtracts, from the burdensomeness of the law.

The Court determines that the magnitude of potentially disenfranchised voters impacted by the DPOC law and its enforcement scheme cannot be justified by the scant evidence of noncitizen voter fraud before and after the law was passed, by the need to ensure the voter rolls are accurate, or by the State's interest in promoting public confidence in elections. Unlike in Crawford , Plaintiff has presented evidence of the number of voters who were unable to register to vote due to lack of DPOC, and the specific burdens felt by those who lack DPOC. Also, there is no mitigating provision comparable to the provisional ballot in Indiana that would cure the failure to register before an election. Given the evidence in the trial record that before the Court's preliminary injunction about 12% of all new voter registration applicants were either suspended or canceled, the Court finds that the burden imposed on Kansans by this law outweighs the state's interest in preventing noncitizen voter fraud, keeping accurate voter rolls, and maintaining confidence in elections. The burden is not just on a "few" voters, but on tens of thousands of voters, many of whom were disenfranchised in 2014. At least one voter, Prof. Boynton, was deterred from registering again after the burdensome process he endured in 2014, a result supported by the testimony of several election experts in this case that increased "costs," or steps to registration, decrease the likelihood of registration and voting. This deterrent effect on young voters is particularly acute.

Moreover, the evidence does not support a fit between the DPOC law and the State's interest in ensuring only qualified citizens are included on the State's voter rolls. The experts agree that several of the nominal cases of noncitizen registrations identified by Defendant can be explained by administrative error and confusion. Indeed, the evidence showed that other far less burdensome methods are available to the State to maintain accurate voter rolls of eligible Kansans by utilizing tools such as matching DOV lists, investigating self-reported noncitizens who are called for jury service, and utilizing the SAVE database. Moreover, better training of DOV staff could help ensure that voter registration applications are not offered to noncitizens.

And while maintaining confidence in the electoral process has independent significance as Crawford held, the evidence in this case does not show that the DPOC law furthers this significant interest. Instead, the law has acted as a deterrent to registration and voting for substantially more eligible Kansans than it has prevented ineligible voters from registering to vote. At least one applicant testified that he opted not to apply to register to vote again, despite possessing DPOC, because of the burdensome experience of being held in suspense and prevented from voting in 2014 due to the law. There has also been evidence of incorrect notices sent to applicants, incorrect information about registration status communicated over the phone by State employees, failure to accept DPOC by State employees, failure to meaningfully inform applicants of their responsibilities under the law, and evolving internal efforts to verify citizenship, that have all caused confusion during the 5 years this law has been effective. If Kansans who try to register to vote cannot be sure if they are in fact registered, particularly after they have been led to believe they complied with all registration laws, it erodes confidence in the electoral system. If Kansans receive misinformation from State officials about whether they are registered to vote, it erodes confidence in the electoral system. If eligible Kansans' votes are not counted despite believing they are registered to vote, it erodes confidence in the electoral system.

In sum, the type of burden and the quality of the evidence in support of that burden is distinguishable from Crawford , which the Supreme Court was careful to limit to the record in that case. Based on this record, the magnitude of the burden on unregistered eligible Kansas voters cannot be justified by the State interests relied on by Defendant. The evidence at trial demonstrated that those interests, while legitimate, are not furthered by the DPOC law. Instead, the DPOC law disproportionately impacts duly qualified registration applicants, while only nominally preventing noncitizen voter registration. It also may have the inadvertent effect of eroding, instead of maintaining, confidence in the electoral system given the confusing, evolving, and inconsistent enforcement of the DPOC laws since 2013. For all of these reasons, the Court finds in favor of Plaintiff Bednasek on his § 1983 claim alleging a Fourteenth Amendment violation of his right to vote.

VI. Remedies

Plaintiffs' requests in both cases for declaratory and injunctive relief is granted. As already stated, K.S.A. § 25–2309(l) and K.A.R. 7–23–15, violate § 5 of the NVRA and infringe on the right to vote under the Fourteenth Amendment. Defendant shall not enforce the DPOC law and accompanying regulation against voter registration applicants in Kansas. As the Court stated in an earlier opinion finding Defendant in contempt, Defendant's well-documented history of avoiding this Court's Orders, and providing confusing notices and information on the State's websites in conjunction with this Court's rulings, warrant specific compliance measures with this injunction as spelled out below.

Defendant argues that Plaintiffs lack standing to seek the specific compliance measures requested by the Fish Plaintiffs, and that the NVRA does not require any specific educational materials or ballot types. The Court disagrees. First, these items are part and parcel of registering eligible Kansans to vote, and ensuring that they are not any more confused than necessary by the change in policy. If Defendant takes the position that he is entitled under the NVRA to continue to falsely assert the status of the law on his website, and that he may require registered voters to complete provisional ballots, he is invited to file a brief to the Court, not to exceed ten pages in length, citing authority for this proposition. Moreover, the Court rejects Defendant's standing argument. These measures would not be required but for past enforcement problems necessitated by Defendant's claims of failing to understand the confines of the Court's preliminary injunction order. These specific compliance measures attempt to address Defendant's past complaints that the Court's directives were not specific enough, and to avoid the need for further compliance directives going forward. They allow Plaintiffs' claims under § 1983 and the NVRA to be fully redressed.

(1) To the extent he has not already done so, Defendant shall provide all registrants covered by the permanent injunction with the same information provided to other registrants (including but not limited to certificates of registration); and must ensure that all elections-related public education materials (including but not limited to voter-aimed notices and websites, in all languages in which those documents are available, including English and Spanish) make clear that voter registration applicants need not provide DPOC in order to become registered to vote, and need not provide any additional information in order to complete their voter registration applications.

(2) Defendant shall instruct all state and county elections officers, and must ensure that all training and reference materials for elections officials in Kansas (including but not limited to the SOS's County Elections Manual) make clear, that voter registration applicants need not provide DPOC in order to be registered to vote, and need not provide any additional information in order to complete their voter registration applications.

(3) Defendant shall maintain the "Voter View" website to accurately reflect covered voters' registration status.

(4) Defendant shall ensure that, in counties that use paper poll books, the names of all registrants lacking DPOC appear in the same manner and in the same list as all other registered voters' names, and that all registrants covered by this Order shall be entitled to vote using standard ballots rather than provisional ballots at polling places on Election Day or when they request advance mail-in ballots.

The parties shall meet, confer, and file a joint status report 30 days before the next primary election scheduled in Kansas to verify compliance with the permanent injunction. Following this joint status report, the Court may determine whether modification of its final order is warranted or whether any additional steps may be necessary to ensure that effective relief for covered voters is not denied or otherwise undermined by Defendant.

VII. Sanctions

Throughout this opinion, the Court referenced several instances when Defendant failed to disclose evidence under Fed. R. Civ. P. 26(a), or to supplement discovery under Rule 26(e). In many of these instances, the Court excluded the evidence. In others, Plaintiffs either withdrew the objection, or the Court allowed the evidence with some limitation. At least once, Defendant attempted to introduce such evidence despite the Court's ruling excluding it. These violations led to objections throughout trial despite the Court's repeated efforts to educate Defendant about his Rule 26 disclosure obligations.

The Court's rulings were governed by Fed. R. Civ. P. 37(c)(1), which states that when a party fails to produce information or identify a witness in violation of Rule 26(a) or (e),

the party is not allowed to use that information or witness to supply evidence on a motion, at a hearing, or at a trial, unless the failure was substantially justified or is harmless. In addition to or instead of this sanction, the court, on motion and after giving an opportunity to be heard:

(A) may order payment of the reasonable expenses, including attorney's fees, caused by the failure;

(B) may inform the jury of the party's failure; and

(C) may impose other appropriate sanctions, including any of the orders listed in Rule 37(b)(2)(A)(i)-(vi).

In determining whether a failure to disclose is harmless or substantially justified, the Court looks to several factors: "(1) the prejudice or surprise to the party against whom the testimony is offered; (2) the ability of the party to cure the prejudice; (3) the extent to which introducing such testimony would disrupt the trial; and (4) the moving party's bad faith or willfulness." The burden to demonstrate that the failure to disclose is harmless or substantially justified is on the party who failed to properly disclose.

There were several violations that justified sanctions under Rule 37(c)(1). First, the Court issued a pretrial ruling on disclosure issues involving Mr. McFerron and Dr. Richman. The Court found that Defendant failed to designate Mr. McFerron as an expert witness after determining that his proffered testimony was not admissible lay opinion. As a sanction for failing to properly disclose him as an expert, the Court required that he testify live, and not by deposition as urged by Defendant. As for Dr. Richman, Defendant attempted to submit an untimely supplemental report by attaching it to his final witness and exhibit list for trial. This report contained estimates and extrapolations based on new data that was available to him in July 2017, yet Defendant failed to supplement at that time, and failed to disclose to Plaintiffs that Dr. Richman would be issuing a supplemental report when the parties informally exchanged their expert witness lists in January 2018 to determine whether to file pretrial Daubert motions. As the Court explained in its previous Order, this last-minute attempt at supplementation with new data prevented Plaintiffs' rebuttal experts from evaluating the report and updating their own opinions. It also prevented Plaintiffs from filing a Daubert motion by the deadline challenging the new figures.

At trial, Defendant again attempted to introduce updated extrapolations by Dr. Richman not included in his original or supplemental report. For example, Defendant tried to introduce a new estimate by Dr. Richman of 3,813 noncitizen registrations in Kansas through a demonstrative exhibit. Dr. Richman explained that this figure is derived from multiplying 3.3%—the percentage of noncitizens identified in the national 2008 CCES dataset with a voter match file and self-reported registration—by an updated adult noncitizen population estimate for the State of Kansas of 115,500, included in his supplemental report. Dr. Richman had not previously, in his supplemental report or otherwise, opined that this represented an accurate estimate of noncitizen registration in Kansas. Dr. Richman attempted to explain the omission from his previous reports by testifying that "this initial report came out just before the source I was getting that number [the noncitizen population in Kansas] from updated. And so in the supplemental report, partly in response to the prompting of one of the experts for the plaintiffs, I updated the number to the more current census estimate of the number of adult non-citizens in the state of Kansas."

Dr. Richman's explanation for his late-disclosed extrapolation was misleading and conflated Dr. Ansolabehere's criticisms of his national estimates with his Kansas estimates. They are both based on the CCES, but on different datasets. Page 3 of Dr. Richman's original report, which he pointed the Court to during his testimony, recites the findings of his 2014 Electoral Studies article, that 3.3% of noncitizens nationally were registered to vote based on 2008 CCES data. Dr. Richman did not extrapolate this 3.3% figure to Kansas by applying it to the American Community Survey's estimate of the number of adult noncitizens in Kansas at the time of his original report (114,000). Instead, he presented a different extrapolation based on self-reported noncitizens who resided in Kansas (14) from the 2006 through 2012 CCES. That extrapolation was 32,000 noncitizen registrations.

Dr. Ansolabehere's report extensively criticized Dr. Richman's Electoral Studies article and its findings about the national rate of noncitizen registration, including that the CCES is not representative of the noncitizen population, and that classification error "can completely account for Dr. Richman's findings." Dr. Ansolabehere separately criticized Dr. Richman's estimate that 32,000 noncitizens in Kansas (an estimate abandoned by Defendant during his opening statement given the small sample size) was flawed, in part because it did not use the 2014 CCES, which included only 4 respondents who stated they were noncitizens, none of whom had a matching voter record. This criticism targeted the CCES sample he used, not the number of noncitizens in Kansas.

Therefore, it was disingenuous to suggest during his testimony that his new estimate of 3,183 noncitizen registrants in Kansas was merely an updated figure from his original report to answer Dr. Ansolabehere's criticisms. Notwithstanding the fact that Dr. Richman had an opportunity to answer such criticisms in his supplemental report, he never opined that the appropriate way to estimate noncitizen registration in Kansas would be to multiply 3.3% by the estimated number of adult noncitizens in Kansas. As the Court explained at trial, it is troubled by Dr. Richman's attempt to insert yet another noncitizen registration estimate into the record during the trial that had not been previously disclosed. The Court is even more troubled by his misleading testimony upon closer inspection of the reports post trial.

Second, Defendant failed to disclose documents underlying the subsection (m) hearings that have taken place since the DPOC law was passed. Exhibit 150 was provided to Plaintiffs for the first time during the trial. These records include the RCD forms for 6 applicants, and the State Election Board orders and copies of supporting documentation for 5 registrants, including Ms. French. As the Court found on the record, this late-produced discovery violated Rule 26(a) because Plaintiffs had requested Defendant provide "correspondence between Defendant Kobach and any other person concerning the purpose or implementation of the DPOC....This Request 1 is not intended to include uniform letters sent to individual registration applicants regarding their voter registration applications." Had Defendant properly disclosed these documents to Plaintiffs, they would have been on notice of Ms. French, a witness Defendant convinced the Court to allow as a rebuttal witness mid-trial, despite his failure to disclose her as a witness before trial. Plaintiffs would have also been aware of the identities of the other individuals who availed themselves of a subsection (m) hearing who could have potentially held discoverable information or testified as witnesses at trial.

Third, Defendant attempted to elicit testimony from Mr. von Spakovsky about updated information and opinions not included in his original report, which he completed in 2016 and never supplemented.

Fourth, Defendant repeatedly attempted to introduce updated numbers of suspended and canceled voter registration applicants based on reports that Mr. Caskey generated the weekend before trial. These figures had not been disclosed to Plaintiffs during discovery. Defendant first attempted to introduce these figures during his opening statement in demonstrative exhibits. The Court excluded the demonstrative exhibits. He then tried to elicit the information through Mr. Caskey. When that did not work, he asked for the Court to take judicial notice of these figures. The Court excluded the evidence under Rule 37(c)(1) for failure to supplement discovery under Rule 26(e), and because different numbers had been stipulated to in the Pretrial Order. The Court also found that they were not appropriate facts for judicial notice because they could not "be accurately and readily determined from sources whose accuracy cannot reasonably be questioned." Defendant again attempts to introduce these numbers, despite their exclusion at trial, into his proposed findings of fact and conclusions of law.

In responding to Plaintiffs' disclosure objections, Defendant was incredulous that the Court would not allow him to introduce into the record the most recent statistics on canceled and suspended applications. But these numbers were provided to Plaintiffs for the first time a few days before trial in the form of a demonstrative exhibit. In fact, Plaintiffs argued that Defendant had not disclosed any updated information about the suspense and cancellation lists after March 2016. Because they were not previously provided, Plaintiffs had no way to verify their accuracy with the underlying ELVIS records. And their experts were not provided an opportunity to supplement their opinions based on these new numbers. Both Drs. McDonald and Hersh relied on the March 2016 records provided to them by Defendant to render their detailed analysis about the contents of the suspense and cancellation lists, and to perform certain list matching. Had these lists been updated, Plaintiffs could have asked these experts to reevaluate their original reports and supplement them based on the new data.

Because the updated information was provided in a demonstrative exhibit, Defendant suggested that this Court's courtesy rule requiring disclosure of demonstrative exhibits 24 hours in advance of their use should excuse his failure to abide by Rule 26(e)'s duty to supplement discovery. This argument was and is unavailing. Because Defendant failed to provide updated figures about the number of DOV applicants suspended or canceled after March 2016, he may not ambush the Plaintiffs and this Court with updated information on the eve of, or during, trial. This was a straightforward violation of Defendant's ongoing duty to supplement discovery disclosures under Rule 26(e), and under Rule 37(c)(1) this Court finds that the violation was not harmless or substantially justified. The prejudice described above demonstrates the failure was not harmless. And, given Mr. Caskey's testimony about the ease of running reports to capture this data, the failure to disclose was not substantially justified. Indeed, Ms. Ahrens described the Kansas League's ability to purchase the suspense list several times after the DPOC law was passed, as recently as the summer of 2017. Given this evidence, Defendant's failure to supplement his March 2016 disclosure about the contents of the suspense and cancelation lists was not substantially justified.

The disclosure violations set forth above document a pattern and practice by Defendant of flaunting disclosure and discovery rules that are designed to prevent prejudice and surprise at trial. The Court ruled on each disclosure issue as it arose, but given the repeated instances involved, and the fact that Defendant resisted the Court's rulings by continuing to try to introduce such evidence after exclusion, the Court finds that further sanctions are appropriate under Rule 37(c)(1), which permits, in addition to exclusion of the evidence, "other appropriate sanctions." It is not clear to the Court whether Defendant repeatedly failed to meet his disclosure obligations intentionally or due to his unfamiliarity with the federal rules. Therefore, the Court finds that an additional sanction is appropriate in the form of Continuing Legal Education. Defendant chose to represent his own office in this matter, and as such, had a duty to familiarize himself with the governing rules of procedure, and to ensure as the lead attorney on this case that his discovery obligations were satisfied despite his many duties as a busy public servant. The Court therefore imposes a CLE requirement of 6 hours for the 2018–2019 reporting year in addition to any other CLE education required by his law license. These 6 additional hours must pertain to federal or Kansas civil rules of procedure or evidence. Defendant shall file a certification with this Court before the end of the reporting period on June 30, 2019, certifying that this CLE requirement has been met.

IT IS THEREFORE ORDERED BY THE COURT that Defendant's Motion to Exclude the Testimony and Report of Dr. Steven Camarota (Doc. 429) is granted in part and denied in part , and Plaintiffs' Motion to Exclude Patrick McFerron (Doc. 460, Bednasek Doc. 183) is granted .

IT IS FURTHER ORDERED that in Case No. 16–2105, the Court finds in favor of Plaintiffs Donna Bucci, Charles Stricker, III, Thomas Boynton, and Douglas Hutchinson on their remaining claim under § 5 of the NVRA. Plaintiff Steven Wayne Fish's remaining claim is dismissed as moot.

IT IS FURTHER ORDERED that in Case No. 15–9300, the Court finds in favor of Plaintiff Parker Bednasek on his remaining claim under 42 U.S.C. § 1983.

IT IS FURTHER ORDERED Plaintiffs' requests for declaratory and injunctive relief are granted as set forth in this Order. Defendant shall strictly comply with the directives in this Order meant to enforce the Court's permanent injunction of the DPOC law and K.A.R. § 7–23–15.

IT IS FURTHER ORDERED that Defendant shall attend 6 hours in addition to any other CLE education required by his law license for the 2018–2019 reporting year. The additional CLE must pertain to federal or Kansas civil rules of procedure or evidence. Defendant shall file a certification with this Court before the end of the reporting period on June 30, 2019, certifying that this CLE requirement has been met. This sanction is imposed under Fed. R. Civ. P. 37(c)(1) for the disclosure violations identified in this Order under Rule 26(a) and (e).

IT IS SO ORDERED.