ROVNER, Circuit Judge.

Judy Norman-Nunnery applied for a job at Madison Area Technical College (“MATC”) in 2005. When she did not receive an interview, much less a job, she sued MATC and three employees involved in the Hiring process. She alleged that they discriminated against her because of her race- and retaliated against her because of her marriage to Willie Nunnery, a lawyer who had previously been involved in filing a frivolous lawsuit against these same defendants. The district court granted summary judgment in favor of the defendants, and we affirm.

Norman-Nunnery is an African-American woman who holds a doctorate in education from the University of Wisconsin-Madison. She has held a number of management positions in education and in state government, including as vocational education coordinator for the Milwaukee Public Schools, administrator for the Department of Health and Human Services, and administrator for the Workers’ Compensation Division and the Division of Vocational Rehabilitation in the State’s Department of Workforce Development. She is married to Willie Nunnery, a lawyer. In 2000, Mr. Nunnery represented Elvira Jimenez in a race discrimination suit against MATC, Carol Bassett, Jackie Thomas and William Stryker, who are all defendants in this case as well. The Jimenez case ended as badly as a case can end for a lawyer and his client. The court dismissed the Jimenez suit as frivolous and found that certain documents produced by Jimenez had been fraudulently created. Ultimately, Mr. Nunnery was sanctioned and his law license was suspended for a period of time as a result of his actions in the Jimenez case. Stryker, Thomas and Bassett all were questioned at the sanctions hearing. *425 Stryker and Bassett both testified regarding the harm they suffered as the result of being falsely accused of engaging in racial discrimination.

In 2002, Norman-Nunnery applied for a position with MATC as the Executive Dean of Learning. Thomas nominated four internal candidates for the position. Nothing in the record suggests that Thomas had any other involvement in the hiring process for the Executive Dean position. Although Norman-Nunnery passed the initial screening process for applicants, she was not interviewed for the job. As part of the hiring process, MATC compiled a list of information about the candidates that specified, among other things, their race. The list for the dean position identified Norman-Nunnery as “black.” There is no evidence in the record that any of the individual defendants here ever saw that list.

In 2005, Norman-Nunnery applied again for a position at MATC, this time for the position of Disability Resource Services Administrator (hereafter “DRS Administrator”). Norman-Nunnery learned of the position from Bob Wynn, a minority recruiter for MATC who contacted her because he thought she would be a good candidate for the job. With Wynn’s encouragement, Norman-Nunnery called Eugene Fujimoto, MATC’s Diversity Coordinator/Affirmative Action Officer, to discuss the job further. Fujimoto was responsible for monitoring MATC’s hiring process for fairness. He told Norman-Nunnery that her administrative experience would be helpful for this particular position. After discussing the job with Wynn and Fujimoto, Norman-Nunnery applied.

Seventy-six other persons also applied for the DRS Administrator position. Bassett, an administrator in the human resources department, conducted an initial screening of the applicants to weed out those who did not meet the minimum qualifications. The job posting specified that a minimally qualified applicant would have a master’s degree and at least 4000 hours of administrative experience. The job description set forth the basics of the position:

Direct the daily operation, activities and staff of Disability Resource Services across the MATC district. Plan, develop, implement, monitor and assess programs and services meeting the requirements of Section 504 of the Rehabilitation Act of 1973 and the Americans with Disabilities Act for eligible students with disabilities. Promote and support the success of students with disabilities emphasizing the development of the whole person with the student’s learning experience.

The initial screening reduced the applicant pool to 46 candidates. Bassett found that Norman-Nunnery met the standards for a minimally qualified applicant and Norman-Nunnery advanced to the next stage of the hiring process.

In order to select the ten most qualified candidates from the pool of forty-six to proceed to the interview phase, MATC used a selection committee consisting of Kevin Carini, Beth Bremer, Marilyn Fay-ram, Carol Higgans 1 and Jacquelyn Thomas. Thomas is a defendant in this *426 case, and of course had been a defendant in the Jimenez suit filed by Norman-Nunnery’s 'husband many years earlier. Thomas served as chair of the selection committee and was to be the direct supervisor of the DRS Administrator. Kristine Gebhardt, a MATC Human Resources Administrator, led a training session with the newly-formed committee to establish “depth and breadth” screening criteria for the DRS Administrator position. The committee developed a list of five criteria by whidh to assess the candidates and assigned a maximum number of points to each: (1) experience with higher education (two points); (2) experience with adult persons with disabilities (three points); (3) knowledge of current and emergent technologies for persons with disabilities (1 point); (4) supervisory experience (two points); and (5) experience with providing reasonable accommodations in an educational setting (three points).

After the depth-and-breadth criteria were established, each committee member independently scored each of the forty-six candidates who had been found minimally qualified. The committee then met to discuss each candidate and the rationale for their scores until they reached a consensus on a score for each applicant. Thomas recorded the consensus score for each applicant on her copy of the screening form. The committee selected the top ten scoring applicants for interviews and Norman-Nunnery did not make the cut. Because the college set certain equal employment opportunity goals, three or more minority candidates were required to be in the interview pool. After selecting the original ten candidates, Higgans then retrieved the “affirmative action sheet” that identified the race of each applicant. On comparing that sheet to the list of ten potential interviewees, the committee learned that the list of ten candidates included only one minority applicant, an African-American woman. The committee then added to the interview pool the next two highest scoring minority applicants, an Asian man and a Hispanic woman. Two of the ten original interviewees removed themselves from consideration and the remaining candidates were interviewed. After the first round of interviews, the pool was reduced to three candidates. None of the minority candidates advanced to the second round of interviews.

At that point, Fujimoto became concerned that no minority candidates were advancing to the next round. He approached Stryker, the vice president of human resources, and Higgans to express his concerns. He asked why Norman-Nunnery had not been given an interview. Higgans told him that Norman-Nunnery had scored lower than the interviewed applicants in the depth-and-breadth categories for higher education and experience with persons with disabilities. Stryker then arranged a second meeting to address Fujimoto’s concerns. Present at the meeting were Stryker, Thomas and Bassett (the individual defendants here) as well as Higgans and Fujimoto. Fujimoto told the group that Norman-Nunnery had been specifically recruited for the job, had a high level of administrative experience and also had an excellent background. He felt that her level of relevant experience was not reflected in her application and asked the group to make certain assumptions about her experience from her résumé, asserting that they would learn during the interview process specific qualifications that she possessed that were not apparent on her paperwork. He also questioned the validity of the depth-and-breadth criteria. Higgans responded to these concerns by reiterating that Norman-Nunnery had scored lower than the other applicants in two areas of the depth-and-breadth criteria. When Stryker asked if anyone be *427 lieved that discrimination had played a part in the interview selection process, no one said it had. Stryker decided not to add Norman-Nunnery to the interview list, a decision in which Bassett and Higgans concurred. Fujimoto then sent a letter to Norman-Nunnery, explaining that she had not been selected for an interview because other candidates scored higher than she did in certain categories, including higher education experience and direct experience with persons with disabilities.

After the second round of interviews, MATC hired Sandra Hall, a white woman, for the DRS Administrator position. Hall, who had previously worked for Norman-Nunnery at the Division of Worker’s Compensation, had thirty years of experience in the disability field, a master’s degree in Rehabilitation Counseling, three years of administrative experience at the University of Wisconsin-Whitewater supervising disability services, five years of experience working as a Rehabilitation Counselor for adults with developmental disabilities or mental illness, and a good working knowledge of assistive technologies, particularly in the setting of students in higher education.

Before Norman-Nunnery applied for the DRS Administrator position, Bassett and Stryker did not know any “Nunnerys” in Madison, Wisconsin other than Willie Nunnery. Bassett conducted the initial screening of the applicants and so she was aware that there was a candidate who shared, in part, the same last name as the lawyer who had filed a frivolous suit against her. At some point, possibly during the meeting with Fujimoto to discuss Norman-Nunnery’s application, Bassett became aware that Norman-Nunnery is married to Willie Nunnery. She discussed this realization with Thomas and Stryker, although the record does not establish a definitive time frame for these discussions. All of the defendants were aware that Willie Nunnery is African-American. There is no other evidence in the record that anyone knew that Norman-Nunnery is African-American until Higgans retrieved the EEO sheet that listed the race of the applicants.

Fujimoto, as the college’s Diversity Coordinator and Affirmative Action Officer, was responsible for reporting on MATC’s affirmative action hiring and retention efforts. In his research for the report, Fujimoto determined that, for the ten-year period between 1995 and 2005, six to seven percent of the administrators at the college were persons of color. This represented a decrease from the prior ten-year period. After conducting a “non-experimental study,” Fujimoto concluded that race affected the college’s hiring decisions on a structural level because the depth-and-breadth system favored insiders. 2 In the area of faculty hiring, for example, 65% of full-time positions were filled from the pool of part-time staff. This had a disparate impact on minority candidates because 95% of the part-time faculty was white.

After she was not offered an interview for the DRS Administrator position, Norman-Nunnery filed a charge of race discrimination in the Equal Rights Division of the Wisconsin Department of Workforce Development (hereafter “ERD”). After filing her claim, MATC discovered that *428 certain documents pertaining to the hiring process for the DRS Administrator position were missing. Twenty-five of the original seventy-seven applications had been lost, including Norman-Nunnery’s. Thomas’ screening form, which included the consensus-based depth-and-breadth scores for the minimally qualified candidates was also lost. The depth-and-breadth scoring forms of the other four committee members were found and produced, as was Bassett’s screening form for minimal qualifications.

Ultimately, Norman-Nunnery filed suit in federal court against MATC, Bassett, Stryker and Thomas. She alleged that the defendants violated Title VII and the Fourteenth Amendment when they precluded her from entering into an employment contract because of her race or because of her association with her husband, that they violated 42 U.S.C. § 1981 and the Fourteenth Amendment by discriminating against her in the terms and conditions of pursuing the DRS Administrator job, and that they violated her rights under the First and Fourteenth Amendments by retaliating against her because of her marital association with Willie Nunnery. The district court granted summary judgment in favor of the defendants because Norman-Nunnery offered no evidence from which a rational jury could conclude that race or marital status motivated the defendants’ decision not to hire her. She appeals.

II.

On appeal, Norman-Nunnery argues that she is entitled to an inference under the spoliation doctrine that the missing documents would have favored her claims of race discrimination and marital association retaliation. She contends that she produced adequate evidence that the defendants knew her race from the beginning of the hiring process, that they were motivated by race and dislike for her husband in deciding not to interview or hire her, and that the record is replete with evidence that the defendants’ stated reasons for not hiring her were pretextual. Our review of the district court’s grant of summary judgment is de-novo. Gunville v. Walker, 583 F.3d 979, 985 (7th Cir.2009); George v. Walker, 535 F.3d 535, 538 (7th Cir.2008).

The first issue is whether Norman-Nunnery is entitled to an inference that the missing documents would support her claims for discrimination and marital association retaliation. “An employer’s destruction of or inability to produce a document, standing alone, does not warrant an inference that the document, if produced, would have contained information adverse to the employer’s case.” Park v. City of Chicago, 297 F.3d 606, 615 (7th Cir.2002). In order to draw an inference that the missing documents contained information adverse to the defendants, Norman-Nunnery must demonstrate that the defendants intentionally destroyed the documents in bad faith. Faas v. Sears, Roebuck & Co., 532 F.3d 633, 644 (7th Cir.2008); Park, 297 F.3d at 615. The crucial element in a spoliation claim is not the fact that the documents were destroyed but that they were destroyed for the purpose of hiding adverse information. Faas, 532 F.3d at 644; Trask-Morton v. Motel 6 Operating L.P., 534 F.3d 672, 681 (7th Cir.2008). Some courts have found a spoliation sanction to be appropriate only where a party has a duty to preserve evidence because it knew, or should have known, that litigation was imminent. Trask-Morton, 534 F.3d at 681. See also Park, 297 F.3d at 615 (a violation of a record retention policy creates a rebutta *429 ble presumption that the missing record contained evidence adverse to the violator).

Norman-Nunnery fails on every element of the test for the spoliation inference. First, the documents disappeared before Norman-Nunnery filed any claim against MATC and the individual defendants. Thus, the documents were lost before MATC knew or should have known that litigation was imminent. Second, Norman-Nunnery has offered no evidence that the defendants lost or destroyed the documents for the purpose of hiding adverse information. Between the time Norman-Nunnery applied for the job and the filing of her claim of discrimination, MATC moved its human resources department twice due to office remodeling. Each time, all of the records in the office were boxed up and moved. At the time Norman-Nunnery applied for the DRS Administrator position, MATC still used paper applications. MATC’s paper filing system left something to be desired. The stack of papers awaiting filing in the human resources office sometimes reached three feet in height. No single individual was responsible for filing; various human resources employees filed applications and other paperwork, including student employees. After the moves, MATC was unable to locate twenty-five of the original seventy-seven applications, including Norman-Nunnery’s. Although Thomas’s depth-and-breadth screening form, which contained the consensus scores, was missing, the scoring forms of the other four committee members were located and produced, along with Bassett’s screening for minimal qualifications. MATC also located and produced the written applications of fifty-two candidates, including the application submitted by Hall, the person hired for the job. It is unclear from the record whether Norman-Nunnery kept a copy of the application she submitted, but the last person at MATC who could account for the whereabouts of Norman-Nunnery’s application was Fujimoto. He had Norman-Nunnery’s application with him when he met with Stryker, Bassett and Thomas. At the end of that June 2005 meeting, he returned Norman-Nunnery’s application to the human resources department. Between June 2005 and December 2005, the human resources files were boxed up and moved twice. Norman-Nunnery filed her claim against MATC and the individual defendants in April 2006.

As evidence of the defendants’ bad faith, Norman-Nunnery points out that two witnesses could not recall hiring documents being lost prior to this occasion. She also asserts that MATC constantly shifted its story about what happened to the missing documents, first claiming that Fujimoto had them, then simply claiming that her application was missing, then claiming to have lost the documents in an office move and then claiming to have lost the documents in two office moves. That documents had not been lost before implies nothing in and of itself. And MATC never changed its explanation but simply supplied more information over time. For example, MATC never stated that Fujimoto had the missing documents. Rather, a lawyer for MATC stated in a letter:

The complainant’s application materials were with Eugene Fujimoto when he investigated her concerns. Mr. Fujimoto has left the College and her material is not available.

R. 42, Ex. S, at 2. Although one reading of the letter implies that Fujimoto took the papers with him when he left the College, each part of the statement is literally true. Fujimoto was the last person known to have the papers. Although he stated that he returned them to the human resources department, no one has been able to find *430 them since that time. Thus, Norman-Nunnery’s material is not available. Stating simply that'the documents were missing is also true and implies nothing. Because Norman-Nunnery did not file her claims until after the documents had been moved twice in the office remodeling project, and because no one had any reason to look for the documents between the two moves, it is not surprising that the defendants are not able to pinpoint when the documents were lost. In light of the undisputed circumstances of the office moves, there is nothing suspicious about the defendants’ inability to say When or how the documents disappeared. Norman-Nunnery’s conclusory statements to the contrary, there is simply no evidence supporting a bad faith motive on the part of the defendants in the disappearance of the documents.

Moreover, although the consensus scoring sheet was not found, the scoring sheets of four of the five committee members were located. These documents supplied a good part of the information used in reaching the consensus scores. If the scoring sheets of four out of five committee members do not support Norman-Nunnery’s claims, it is highly unlikely that the missing fifth document contains information adverse to the defendants. Presumably Norman-Nunnery knew what was on her own application and Fujimoto also could have testified about the contents of her application as he was the last person in possession of it. Norman-Nunnery has presented no evidence suggesting an intent on the part of the defendants to hide adverse information. Because there is no evidence that any of the defendants destroyed the documents in bad faith, Norman-Nunnery is not entitled to a favorable inference under the spoliation doctrine.

The district court found that Norman-Nunnery’s claims for race and marital association discrimination failed at the outset because she produced no evidence that any of the relevant decision-makers knew her race or marital status at the time she was eliminated from the' pool of candidates to be interviewed. The court also found that Norman-Nunnery failed to demonstrate that the defendants’ stated non-discriminatory reason for not hiring her — that she scored lower than the other candidates in the depth-and-breadth criteria — was a pretext. We think the question of whether the defendants knew.Norman-Nunnery’s race or marital association is a closer call than the district court’s opinion reflects. For example, Bassett admits she drew a connection between Willie Nunnery and Norman-Nunnery at some point and discussed this connection with the other defendants, although the record does not tell us when she made the connection or when she discussed it with the other defendants. Thomas chaired the depth-and-breadth committee and was to be the direct supervisor of the new DRS Administrator. Bassett, Thomas and Stryker had all been sued by Willie Nunnery in the Jimenez case. A person is unlikely to quickly forget being falsely sued for racial discrimination, and any mention of Nunnery’s fairly unusual name would probably have attracted the attention of these three defendants. On summary judgment, when we must construe the facts in favor of Norman-Nunnery, we are reluctant to find that there is no genuine issue regarding whether and when the defendants were aware of her race and marital association. A reasonable fact-finder could infer from these highly unusual facts that the defendants knew Norman-Nunnery is African-American and that she is married to Willie Nunnery. In the end, though, this dispute is irrelevant because Norman-Nunnery cannot show that the defendants’ nondiscrimi *431 natory reason for refusing to hire her is pretext. Nor can she demonstrate causation on her claim for marital association retaliation.

Norman-Nunnery contends that she has direct evidence of race discrimination and that she can also defeat summary judgment using the McDonnell Douglas burden-shifting analysis. We begin with her direct evidence. Norman-Nunnery contends that Fujimoto’s study of MATC’s hiring practices and the suspicious circumstances surrounding the missing documents both supply direct evidence of discriminatory intent. As we discussed above, there is nothing suspicious about the loss of certain documents in the two office moves that took place between the time the hiring process was underway and the time Norman-Nunnery first complained of discrimination. Nor does Fujimoto’s non-experimental study aid her case. Fujimoto’s study concluded only that MATC’s use of the depth-and-breadth criteria tended to favor inside candidates, which had the effect of favoring non-minority candidates. But Norman-Nunnery must show that it is more likely than not that she was harmed by discriminatory acts. A study showing that certain hiring practices correlated with a negative effect on minority hiring, although relevant, will not by itself meet the “more likely than not” standard for a claim of discrimination against an individual plaintiff. Baylie v. Federal Reserve Bank of Chicago, 476 F.3d 522, 524 (7th Cir.2007).

Statistical analysis is relevant in the technical sense that it “has a tendency to make the existence of [a material] fact ... more probable or less probable than it would be without the evidence.” Fed. R.Evid. 401. But data showing a small increase in the probability of discrimination cannot by itself get a plaintiff over the more-likely-than-not threshold; it must be coupled with other evidence, which does most of the work.

Baylie, 476 F.3d at 524. See also Nichols v. Southern Illinois University-Edwardsville, 510 F.3d 772, 782 (7th Cir.2007) (“a plaintiff may use pattern evidence of disparate treatment even if that evidence is not rigorously statistical, although, standing alone, it is insufficient evidence to withstand summary judgment”); Barricks v. Eli Lilly & Co., 481 F.3d 556, 559 (7th Cir.2007) (noting the importance of further analysis and context when relying on raw data in employment discrimination disputes). Norman-Nunnery has no other direct evidence of discrimination. Therefore, she must make her claim, if at all, under the McDonnell Douglas burden shifting analysis. See McDonnell Douglas Corp. v. Green, 411 U.S. 792, 802-02, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973).

Under the burden shifting analysis, a plaintiff must first establish a prima facie case of hiring discrimination by demonstrating:

(i) that he belongs to a racial minority;

(ii) that he applied and was qualified for a job for which the employer was seeking applicants; (iii) that, despite his qualifications, he was rejected; and (iv) that, after his rejection, the position remained open and the employer continued to seek applicants from persons of complainant’s qualifications.

McDonnell Douglas, 411 U.S. at 802, 93 S.Ct. 1817; O’Neal v. City of New Albany, 293 F.3d 998, 1003 (7th Cir.2002); Senner v. Northcentral Technical College, 113 F.3d 750, 754-55 (7th Cir.1997). To withstand summary judgment on the prima facie case, the plaintiff need only show that there is a genuine issue of material fact regarding these elements. O’Neal, 293 F.3d at 1003. If the plaintiff establishes a *432 prima facie case, the burden shifts to the defendant to articulate a legitimate, nondiscriminatory reason for not hiring the plaintiff. McDonnell Douglas, 411 U.S. at 802, 93 S.Ct. 1817; O’Neal, 293 F.3d at 1005; Senner, 113 F.3d at 755. Normally, we first determine whether a plaintiff has established a prima facie case before putting the employer to the burden of demonstrating a non-discriminatory reason for not hiring the plaintiff and engaging in the pretext analysis. Everroad v. Scott Truck Systems, Inc., 604 F.3d 471, 477 (7th Cir.2010).

In some cases, though, the issue of qualifications for the job and the question of pretext overlap. When the employer asserts as the non-discriminatory reason for not hiring the plaintiff that she was not as qualified as other candidates for the position, the credibility of the employer’s assertion is at issue for both the second and fourth elements of the plaintiffs prima facie case and the pretext analysis. Everroad, 604 F.3d at 477-78. There is a good deal of overlap in this case between these key issues and we therefore analyze them together.

There is no dispute that Norman-Nunnery belongs to a racial minority or that she met the minimal qualifications for the job. Nor is there any dispute that she was not chosen for the job, and MATC then continued to seek applicants from the pool of other qualified candidates. But MATC had a large number of minimally qualified candidates and sought to select from that larger pool the candidates who were most qualified for the job to proceed to the next phase of the hiring process. MATC has proffered as its legitimate, nondiscriminatory reason for not interviewing or hiring Norman-Nunnery that she scored lower on the depth-and-breadth criteria than the applicants who advanced to the interview stage. She also had lower scores than the person ultimately hired for the position. The numerical evaluation system the college developed to screen applicants meets the employer’s burden to state a non-discriminatory reason for refusing to hire Norman-Nunnery. Senner, 113 F.3d at 755. The burden therefore shifts back to Norman-Nunnery to produce evidence from which a reasonable fact-finder could conclude that MATC’s stated reason was a pretext. Senner, 113 F.3d at 755.

The district court was correct that Norman-Nunnery has presented no evidence calling into question the sincerity of the non-discriminatory reason the defendants gave for not interviewing or hiring Norman-Nunnery. As evidence of pretext, Norman-Nunnery first cites MATC’s shifting explanation for the disappearance of important documents. As we discussed above, there is nothing suspicious about MATC’s explanation for the disappearance of the documents. Norman-Nunnery next cites Fujimoto’s study. But we have already determined that statistical data, although relevant, cannot alone meet the “more likely than not” standard for a specific hiring decision about an individual. Without context, Fujimoto’s study sheds little to no light on whether race played a part in this particular employment decision. In context, the undisputed facts demonstrate that MATC purposely sought to increase the number of minority candidates who would advance to the interview phase of the hiring process. In addition to the African-American woman who scored high enough on the depth-and-breadth criteria to be included in the original pool of ten interview candidates, the committee also selected the next two highest scoring minority candidates to be interviewed.

At every stage of the process, MATC applied the same criteria to each candidate, and Norman-Nunnery has no evi *433 dence to the contrary. The depth-and-breadth criteria were established by a committee before any applications were reviewed, so any suggestion that the criteria were rigged to exclude Norman-Nunnery fails for lack of evidence. That Wynn and Fujimoto thought Norman-Nunnery was qualified for the job does not advance her cause. Each spoke to Norman-Nunnery before the depth-and-breadth criteria were established and neither man was involved in the process of comparing the applications of all of the minimally qualified applicants. 3 Thus neither could say whether Norman-Nunnery’s application was comparatively better (or worse) than the candidates who advanced to the interview process. Nor is it significant that the defendants refused to stray from the depth-and-breadth criteria when asked to do so by Fujimoto. Refusing to give Norman-Nunnery an advantage that no other candidate had is not evidence of discrimination. In short, after reviewing the evidence cited by Norman-Nunnery, we find no evidence supporting her claim of pretext. The district court correctly granted judgment in favor of the defendants on her claim of racial discrimination in hiring.

Norman-Nunnery also claims that the defendants failed to hire her because of their animosity toward her husband, Willie Nunnery. She frames this as a claim for “marital association retaliation.” We take Norman-Nunnery’s claim to be not that the defendants discriminated against her because she was married in general but because she was married to Willie Nunnery in particular. Norman-Nunnery relies in part on Christensen v. County of Boone, Illinois, 483 F.3d 454 (7th Cir.2007). There, we held that the Due Process Clause of the Fourteenth Amendment protects a non-marital, romantic relationship against direct and substantial interference by the government. 483 F.3d at 463. We emphasized that, although the “Constitution prevents fundamental rights from being aimed at,” it does not “prevent side effects that may occur if the government is aiming at some other objective.” 483 F.3d at 463. Moreover, we noted that “[ojfficial conduct that represents an abuse of office (as opposed to, say, the implementation of a statutory duty) violates the substantive component of the due process clause only if it ‘shocks the conscience.’ ” Christensen, 483 F.3d at 464.

Norman-Nunnery also cites Adler v. Pataki, 185 F.3d 35 (2d Cir.1999), where a public employee contended that he was fired from his job in retaliation for a lawsuit his wife filed against state officials. The Second Circuit found that a claim that adverse action was taken against a person in retaliation for the conduct of that person’s spouse should be analyzed under the First Amendment, as a violation of the right of intimate association. 185 F.3d at 44. The court noted that the plaintiff would be required to demonstrate that his wife’s lawsuit was the motivation for his discharge. 185 F.3d at 45-46.

Under either case, Norman-Nunnery’s claim cannot survive summary judgment because she has failed to provide any evidence that the defendants refused to hire her because of her marriage to Willie Nunnery. Arguably she has presented ev *434 idenee that they were aware she was married to Willie Nunnery and that they harbored ill feelings toward him. But she has presented no evidence that they were motivated by their animosity toward Willie Nunnery in passing her over for the job. To the contrary, all of the uncontroverted evidence demonstrates that the candidates who advanced to the interview stage and the candidate who ultimately was hired all had depth-and-breadth scores that exceeded those of Norman-Nunnery. The depth-and-breadth scores were a legitimate, non-discriminatory criteria for distinguishing among the minimally qualified applicants. Her claim thus fails for lack of evidence. The defendants would prefer that we find that Norman-Nunnery’s claim fails as a matter of law, that a public employer’s refusal to hire a person because of animosity toward that person’s spouse can never be actionable as a Constitutional claim. Because Norman-Nunnery’s claim fails for a lack of evidence, we need not decide whether the claim is viable as a matter of law. We i’eserve that question for the case in which it is clearly presented.

III.

Norman-Nunnery has no evidence that the defendants’ legitimate, non-discriminatory reason for not hiring her is pretext. She is unable to demonstrate that the decision not to interview or hire her was due to her race or her marital association with Willie Nunnery. The judgment of the district court is therefore

. There appears to be some disagreement in the record over when Higgans joined the committee. According to the defendants, Stryker originally filled the role of EEO representative on the committee and he was present for the initial training session at which the depth-and-breadth criteria were determined. Higgans took his place on the committee after receiving training on the criteria, prior to scoring any of the candidates. The plaintiff does not allege that Stryker took any wrongful actions during his brief time on the committee and so the dispute is immaterial to the issues presented.

. Norman-Nunnery characterizes Fujimoto's study as "non-experimental.” The record does not tell us what this term means to the parties. We take it to mean that Fujimoto's study was conducted without the use of controls and was instead based on observation of past events. Such a study typically can establish a correlation between two factors but not necessarily a cause-and-effect relationship. The analysis of the claim is the same regardless of how the study was conducted.

. Fujimoto testified that he compared Norman-Nunnery's application to those of the applicants chosen for interviews. After that review, Fujimoto argued to Stryker, Thomas and Bassett that Norman-Nunnery’s qualifications were not obvious from the face of her application but would become apparent if they interviewed her. This contention demonstrates a belief by Fujimoto that Norman-Nunnery’s application compared unfavorably on its face to those of the candidates who advanced.