Thirteen years ago, this court suggested that cross-examination may be required in school disciplinary proceedings where the case hinged on a question of credibility. Flaim v. Med. Coll. of Ohio , 418 F.3d 629 , 641 (6th Cir. 2005). Just last year, we encountered the credibility contest that we contemplated in Flaim and confirmed that when credibility is at issue, the Due Process Clause mandates that a university provide accused students a hearing with the opportunity to conduct cross-examination. Doe v. Univ. of Cincinnati , 872 F.3d 393 , 401-02 (6th Cir. 2017). Today, we reiterate that holding once again: if a public university has to choose between competing narratives to resolve a case, the university must give the accused student or his agent an opportunity to cross-examine the accuser and adverse witnesses in the presence of a neutral fact-finder. Because the University of Michigan failed to comply with this rule, we reverse.

I.

John Doe and Jane Roe were students at the University of Michigan. Halfway through Roe's freshman and Doe's junior year, the two crossed paths at a "Risky Business" themed fraternity party. While *579 there, they had a drink, danced, and eventually had sex. Two days later, Roe filed a sexual misconduct complaint with the university claiming that she was too drunk to consent. And since having sex with an incapacitated person (unsurprisingly) violates university policy, the administration immediately launched an investigation. Over the course of three months, the school's investigator collected evidence and interviewed Roe, Doe, and twenty-three other witnesses. Two stories emerged.

First, Doe told the investigator that Roe did not appear drunk and that she was an active participant in their sexual encounter. According to him, the night went something like this: after he and Roe had a drink, danced, and kissed at the party, the two decided to go upstairs to his bedroom. Once there, they kissed "vigorous[ly]" and eventually made their way onto his bed. R. 16, Pg. ID 332. After they jointly removed their clothing, he asked Roe if she wanted to have sex. She said, "Yeah," and the two proceeded to have intercourse followed by oral sex. Id. at Pg. ID 333-34. When they were done, they cuddled until Roe became sick and vomited into a trash can by Doe's bed. Doe rubbed her back for five or so minutes and then left to use the bathroom and talk with friends. By the time he returned, Roe was crying and another female student was helping her gather her things. He asked Roe if she was okay, but Roe's new companion told him to "[g]o away" and the two women walked out of the room. Id. at Pg. ID 335. At the time, he assumed that Roe was upset because he had left her alone after they had sex. He asserted that he had no reason to believe that she was drunk or that Roe thought any of his sexual advances were unwelcome.

Roe remembered the night differently. According to her, she was drunk and unaware of her surroundings when she and Doe went to his room. While kissing near the doorway, she told Doe "no sex" before she "flopped" onto his bed. Id. at Pg. ID 325-26. Without asking, Doe undressed her and had intercourse with her while she "laid there in a hazy state of black out." Id. at Pg. ID 326. And at some point, she passed out and woke up to Doe having oral sex with her. Afterwards, she felt a "spinning sensation" and fell back onto the bed. Id. at Pg. ID 327. Doe asked her if she was okay, and she told him that she was not. So Doe placed a trash can by the side of his bed and left the room. She proceeded to vomit into the trash can. Afterward, she attempted to find her clothes but could not get her bearings. Feeling a sense of "desperation and defeat," she tried to catch another female student's attention by making "vomit sounds." Id. It worked, and the female student ("Witness 2") helped Roe find her clothes, put them on, and get back to her dorm.

If Doe's and Roe's stories seem at odds, the twenty-three other witnesses did not offer much clarification. Almost all of the male witnesses corroborated Doe's story, and all of the female witnesses corroborated Roe's. For example, Doe's roommate said that Roe "didn't seem like she was hammered or that drunk," although he stated that he did not "want to speculate" about whether she had had some alcohol because he did not talk to her directly or "really interact with [her]" much. Id . at Pg. ID 339. Yet he mentioned that in his two interactions with her, he did not smell alcohol on her. Id. Doe's roommate further alleged that Roe and Witness 2 were just "rallying against a fraternity guy." Id. at Pg. ID 339-41. Another member of Doe's fraternity told the investigator that he saw Doe and Roe "making out" on the dance floor and there was no reason to suspect that either of them had too much to drink. Id. at Pg. ID 347. And two others stated that Roe did not appear drunk when she left Doe's room at the end of the night, *580 although they indicated they had limited observations of Roe.

Roe's sorority sisters, on the other hand, reported that Roe seemed "more than a little buzzed" at the party because her eyes were "open but unfocused" and she "trail[ed] off at the end of sentences." Id. at Pg. ID 345-46. The female student who helped Roe leave the party told the investigator that she found Roe crying and "very drunk" in Doe's bed. Id. at Pg. ID 342-43. And two other friends provided that when Roe returned to her dorm that night, she sobbed on the floor of her room and said "she thought she'd been raped." Id. at Pg. ID 352.

Given the students' conflicting statements, the investigator concluded that the evidence supporting a finding of sexual misconduct was not more convincing than the evidence offered in opposition to it. The investigator did note, however, that Witness 2 might have been a more credible witness because she had no prior connection to Doe, Roe, or their respective Greek organizations. But because Witness 2 only observed Roe after the sexual encounter had ended, the investigator concluded that she could not address the relevant question-Roe's level of intoxication during the encounter or what signs of intoxication she manifested at that time. So after three months of thorough fact-finding, the investigator was unable to say that Roe exhibited outward signs of incapacitation that Doe would have noticed before initiating sexual activity. Accordingly, the investigator recommended that the administration rule in Doe's favor and close the case.

Roe appealed. She argued that the evidence did not support the investigator's findings and asked the university to reconsider. The case went up to the university's Appeals Board, and a three-member panel reviewed the investigator's report. After two closed sessions (without considering new evidence or interviewing any students), the Board reversed. Although the Board found that the investigation was fair and thorough, it thought the investigator was wrong to conclude that the evidence was in equipoise. According to the Board, Roe's description of events was "more credible" than Doe's, and Roe's witnesses were more persuasive. R. 6-5, Pg. ID 274-75. As a result, the university set the investigator's recommendation aside and proceeded to the sanction phase. Facing the possibility of expulsion, Doe agreed to withdraw from the university. He was 13.5 credits short of graduating.

Since then, Doe filed a lawsuit claiming that the university's disciplinary proceedings violated the Due Process Clause and Title IX. He argues that because the university's decision turned on a credibility finding, the school was required to give him a hearing with an opportunity to cross-examine Roe and adverse witnesses. He also maintains that the Board violated Title IX by discriminating against him on account of his gender. The university filed a motion to dismiss, which the district court granted in full. Doe now appeals, and we review de novo. Kottmyer v. Maas , 436 F.3d 684 , 688 (6th Cir. 2006).

II.

To survive a motion to dismiss, a complaint must provide "a short and plain statement of the claim showing that the [plaintiff] is entitled to relief." Fed. R. Civ. P. 8(a)(2). A plaintiff shows that he is entitled to relief by "plausibly suggesting" that he can meet the elements of his claim. Bell Atl. Corp. v. Twombly , 550 U.S. 544 , 557, 127 S.Ct. 1955 , 167 L.Ed.2d 929 (2007). And a plaintiff's suggestion is plausible when it contains enough factual content that the court can reasonably infer that the defendant is liable. Ashcroft v. Iqbal , 556 U.S. 662 , 678, 129 S.Ct. 1937 , 173 L.Ed.2d 868 (2009). Legal conclusions, *581 "formulaic recitation[s]" of the claim's elements, and "naked assertion[s]" of liability are all insufficient. Id. (second alteration in original) (quoting Twombly , 550 U.S. at 557 , 127 S.Ct. 1955 ).

When evaluating a complaint's sufficiency, courts use a three-step process. First, the court must accept all of the plaintiff's factual allegations as true. Logsdon v. Hains , 492 F.3d 334 , 340 (6th Cir. 2007). Second, the court must draw all reasonable inferences in the plaintiff's favor. Id. And third, the court must take all of those facts and inferences and determine whether they plausibly give rise to an entitlement to relief. Iqbal , 556 U.S. at 679 , 129 S.Ct. 1937 . If it is at all plausible (beyond a wing and a prayer) that a plaintiff would succeed if he proved everything in his complaint, the case proceeds.

III.

Doe first argues that the university violated his due process rights during his disciplinary proceedings. He claims that because the university's decision ultimately turned on a credibility determination, the school was required to give him a hearing with an opportunity to cross-examine Roe and other adverse witnesses. The district court dismissed this claim, finding that even if credibility was at issue, the university's failure to allow for cross-examination was "immaterial" in Doe's case. R. 74, Pg. ID 2871. We disagree.

When it comes to due process, the "opportunity to be heard" is the constitutional minimum. Grannis v. Ordean , 234 U.S. 385 , 394, 34 S.Ct. 779 , 58 L.Ed. 1363 (1914). But determining what being "heard" looks like in each particular case is a harder question. The Supreme Court has declined to set out a universal rule and instead instructs lower courts to consider the parties' competing interests. See Mathews v. Eldridge , 424 U.S. 319 , 335, 96 S.Ct. 893 , 47 L.Ed.2d 18 (1976) ; Goss v. Lopez , 419 U.S. 565 , 579, 95 S.Ct. 729 , 42 L.Ed.2d 725 (1975). So, consistent with this command, our circuit has made two things clear: (1) if a student is accused of misconduct, the university must hold some sort of hearing before imposing a sanction as serious as expulsion or suspension, and (2) when the university's determination turns on the credibility of the accuser, the accused, or witnesses, that hearing must include an opportunity for cross-examination. Univ. of Cincinnati , 872 F.3d at 399-402 ; Flaim , 418 F.3d at 641 .

Due process requires cross-examination in circumstances like these because it is "the greatest legal engine ever invented" for uncovering the truth. Univ. of Cincinnati , 872 F.3d at 401-02 (citation omitted). 1 Not only does cross-examination allow the accused to identify inconsistencies in the other side's story, but it also gives the fact-finder an opportunity to assess a witness's demeanor and determine who can be trusted. Id. So if a university is faced with competing narratives about potential misconduct, the administration must facilitate some form of cross-examination in order to satisfy due process. Id. at 402 .

Doe claims that the university ran afoul of this well-established rule in his disciplinary proceedings. And the pleadings in his case suggest that he is right. The university's decision rested on a credibility determination: the Board found Doe responsible after concluding that Roe and *582 her witnesses were "more credible" than Doe and his. R. 6-5, Pg. ID 274-75. Nevertheless, Doe never received an opportunity to cross-examine Roe or her witnesses-not before the investigator, and not before the Board. As a result, there is a significant risk that the university erroneously deprived Doe of his protected interests. 2 See Mathews , 424 U.S. at 335 , 96 S.Ct. 893 .

This risk is all the more troubling considering the significance of Doe's interests and the minimal burden that the university would bear by allowing cross-examination in Doe's case. See id. at 334-35 , 96 S.Ct. 893 . Time and again, this circuit has reiterated that students have a substantial interest at stake when it comes to school disciplinary hearings for sexual misconduct. Doe v. Miami Univ. , 882 F.3d 579 , 600 (6th Cir. 2018) ; Univ. of Cincinnati , 872 F.3d at 400 ; Doe v. Cummins , 662 F. App'x 437 , 446 (6th Cir. 2016). Being labeled a sex offender by a university has both an immediate and lasting impact on a student's life. Miami Univ. , 882 F.3d at 600 . The student may be forced to withdraw from his classes and move out of his university housing. Id. His personal relationships might suffer. See id. And he could face difficulty obtaining educational and employment opportunities down the road, especially if he is expelled. Id.

In contrast, providing Doe a hearing with the opportunity for cross-examination would have cost the university very little. As it turns out, the university already provides for a hearing with cross-examination in all misconduct cases other than those involving sexual assault. So the administration already has all the resources it needs to facilitate cross-examination and knows how to oversee the process. See Univ. of Cincinnati , 872 F.3d at 406 (noting that a university does not bear a significant administrative burden when it already has procedures in place to accommodate cross-examination). And, importantly, the university identifies no substantial burden that would be imposed on it if it were required to provide an opportunity for cross-examination in this context.

Still, the university offers four reasons why Doe's claim is not as plausible as it seems. None do the trick. First, the university contends that even if Doe did not have a formal opportunity to question Roe, he was permitted to review her statement and submit a response identifying inconsistencies for the investigator. As such, the university claims that there would have been no added benefit to cross-examination. But this circuit has already flatly rejected that argument. In University of Cincinnati , we explained that an accused's ability "to draw attention to alleged inconsistencies" in the accuser's statements does not render cross-examination futile. Id. at 401-02 . That conclusion applies equally here, and we see no reason to doubt its wisdom. Cross-examination is essential in cases like Doe's because it does more than uncover inconsistencies-it "takes aim at credibility like no other procedural device." Id. Without the back-and-forth of adversarial questioning, the accused cannot probe the witness's story to test her memory, intelligence, or potential ulterior motives. Id. at 402 . Nor can the fact-finder observe the witness's demeanor under that questioning. Id. For that reason, written statements cannot substitute for cross-examination. See Brutus Essay XIII, in The Anti-Federalist 180 (Herbert J. Storing ed., 1985) ("It is of great importance in the distribution of justice that *583 witnesses should be examined face to face, that the parties should have the fairest opportunity of cross-examining them in order to bring out the whole truth; there is something in the manner in which a witness delivers his testimony which cannot be committed to paper, and which yet very frequently gives a complexion to his evidence, very different from what it would bear if committed to writing ...."). Instead, the university must allow for some form of live questioning in front of the fact-finder. See Univ. of Cincinnati , 872 F.3d at 402-03, 406 (noting that this requirement can be facilitated through modern technology, including, for example, by allowing a witness to be questioned via Skype "without physical presence").

That is not to say, however, that the accused student always has a right to personally confront his accuser and other witnesses. See Miami Univ. , 882 F.3d at 600 (noting that "even in the face of a sexual-assault accusation," the protections afforded to an accused "need not reach the same level ... that would be present in a criminal prosecution" (quoting Univ. of Cincinnati , 872 F.3d at 400 ) ). Universities have a legitimate interest in avoiding procedures that may subject an alleged victim to further harm or harassment. And in sexual misconduct cases, allowing the accused to cross-examine the accuser may do just that. See Univ. of Cincinnati , 872 F.3d at 403 . But in circumstances like these, the answer is not to deny cross-examination altogether. Instead, the university could allow the accused student's agent to conduct cross-examination on his behalf. After all, an individual aligned with the accused student can accomplish the benefits of cross-examination-its adversarial nature and the opportunity for follow-up-without subjecting the accuser to the emotional trauma of directly confronting her alleged attacker. Cf. Maryland v. Craig , 497 U.S. 836 , 857, 110 S.Ct. 3157 , 111 L.Ed.2d 666 (1990) (holding that where forcing the alleged victim to testify in the physical presence of the defendant may result in trauma, the court could use an alternative procedure that "ensures the reliability of the evidence by subjecting it to rigorous adversarial testing" through "full cross-examination" and ensuring that the alleged victim could be "observed by the judge, jury, and defendant as they testified"). The university's first argument is therefore unavailing. 3

Second, the university contends that Doe is not entitled to cross-examination because the university's decision did not depend entirely on a credibility contest between Roe and Doe. For support, the university brings our attention back to University of Cincinnati , where we emphasized the exclusively "he said/she said" nature of the investigation at issue in that case. 872 F.3d at 395, 402 . But the university reads far too much into that point. When we emphasized the exclusively "he said/she said" nature of the University of Cincinnati dispute, we were not implying that cross-examination would be less important in cases where the school's finding rested on the credibility of several witnesses instead of one or two. Rather, we merely distinguished that case from others holding that cross-examination was unnecessary when the university's decision did not rely on any testimonial evidence at *584 all . Id. at 401, 405 (distinguishing Plummer v. Houston , 860 F.3d 767 , 775-76 (5th Cir. 2017), which held that cross-examination was unnecessary when conduct depicted in videos and photos was sufficient to sustain a finding of misconduct without resorting to testimonial evidence); see also Flaim , 418 F.3d at 641 . Accordingly, University of Cincinnati does not stand for the proposition that cross-examination is required only if the university's decision depends solely on the accuser's statement. Instead, University of Cincinnati is consistent with our conclusion today: if credibility is in dispute and material to the outcome, due process requires cross-examination. See 872 F.3d at 406 (recognizing that credibility disputes might be more common in sexual misconduct proceedings than other university disciplinary investigations).

Third, the university claims that cross-examination was unnecessary in Doe's case because he admitted to the misconduct in a police interview the day after the incident in question. Here, the university is right about the law but wrong about the facts. This court has long held that cross-examination is unnecessary if a student admits to engaging in misconduct. Flaim , 418 F.3d at 641 . After all, there is little to be gained by subjecting witnesses to adversarial questioning when the accused student has already confessed. But at the motion-to-dismiss stage, we cannot conclude that Doe confessed to the misconduct in this case. To see why, a closer look at the police report is instructive.

During the police interview, a detective asked Doe to describe the previous night's sexual encounter. When doing so, Doe told the detective that Roe performed oral sex on him before they engaged in intercourse, and that when the pair began to have intercourse, Roe was on top. As it turns out, this story was different than the one Roe had reported to the detective earlier that day. According to the detective, Roe claimed that she told Doe "no sex" before making her way to the bed, and that she performed oral sex on Doe after the pair had intercourse. The detective thus relayed Roe's version of the story to Doe, and Doe immediately conceded that Roe was right and that he "got it all wrong." R. 16, Pg. ID 356. Even so, however, Doe reiterated that (1) he never heard Roe say "no sex," (2) he "didn't rape" Roe, and (3) he believed their sexual encounter was consensual. Id.

Because the district court made this report part of the pleadings, we must read it in the light most favorable to Doe. 4 See Logsdon , 492 F.3d at 340 . When we do, we cannot conclude that Doe admitted to any of the critical facts in his case-i.e., that Roe was too drunk to consent to sex, and that he knew or should have known as much. For one, we would have to ignore Doe's claim that the sex was "consensual." R. 16, Pg. ID 356. And for another, because Doe did not mention anything about Roe's level of intoxication in his own account of the night's events, his concession that Roe was correct and that he "got it all wrong" appears to relate only to the points on which the detective said their two accounts actually diverged-the order of the sexual acts. See id. , Pg. ID 354-56. This alleged confession thus does not sufficiently rebut the plausibility of Doe's claim.

The university offers one last ditch effort to avoid reversal. It points out that *585 although Doe did not have an opportunity to cross-examine Roe in the university disciplinary process, he recently deposed her in state civil proceedings. According to the university, because Roe's deposition is consistent with what she told the investigator, Doe's inability to cross-examine her during the disciplinary proceedings did not cause any prejudice. To start, Doe disputes whether Roe's deposition is, in fact, consistent with her earlier statements in the disciplinary process. But more importantly, Roe's later deposition has no bearing on this case. As discussed above, the value of cross-examination is tied to the fact-finder's ability to assess the witness's demeanor. Univ. of Cincinnati , 872 F.3d at 402 . So just as a written response insufficiently substitutes for cross-examination, so too does a written deposition transcript. And, critically, cross-examination for the sake of cross-examination is not what Doe seeks. Rather, Doe seeks cross-examination as part of the credibility assessment by the university . That a state court later allowed for cross-examination as a part of its fact-finding after the university had already made its decision is beside the point. If anything, the fact that the state court allowed cross-examination only goes to show just how far removed the university's fact-finding procedures are from the tried and true methods invoked by courts. See id. at 404-05 (noting that while classrooms are not courtrooms, at the very least a circumscribed version of cross-examination is required (citing Cummins , 662 F. App'x at 448 ) ).

In sum, accepting all of Doe's factual allegations as true and drawing all reasonable inferences in his favor, he has raised a plausible claim for relief under the Due Process Clause. We thus reverse the district court's decision to dismiss his claim. 5

IV.

Doe also sued under Title IX, which prohibits federally-funded universities from discriminating against students on the basis of sex. 20 U.S.C. § 1681 (a) ; Cannon v. Univ. of Chi. , 441 U.S. 677 , 717, 99 S.Ct. 1946 , 60 L.Ed.2d 560 (1979) (recognizing an implied private right of action under Title IX). He advances three theories of liability, claiming that the university (1) reached an erroneous outcome in his case because of his sex, (2) relied on archaic assumptions about the sexes when rendering a decision, and (3) exhibited deliberate indifference to sex discrimination in his disciplinary proceedings.

Erroneous Outcome . A university violates Title IX when it reaches an erroneous outcome in a student's disciplinary proceeding because of the student's sex. See Miami Univ. , 882 F.3d at 592 . To survive a motion to dismiss under the erroneous-outcome theory, a plaintiff must plead facts sufficient to (1) "cast some articulable doubt" on the accuracy of the disciplinary proceeding's outcome, and (2) demonstrate a "particularized ... causal connection between the flawed outcome and gender bias." Id. (alteration in original) (quoting Cummins , 662 F. App'x at 452 ). The district court held that Doe's complaint failed to meet either element and dismissed his claim. We reverse.

First, because Doe alleged that the university did not provide an opportunity for cross-examination even though credibility was at stake in his case, he has *586 pled facts sufficient to cast some articulable doubt on the accuracy of the disciplinary proceeding's outcome. See Yusuf v. Vassar Coll. , 35 F.3d 709 , 715 (2d Cir. 1994) (noting the "pleading burden in this regard is not heavy" and can be met by alleging "particular procedural flaws affecting the proof"); see also Univ. of Cincinnati , 872 F.3d at 401 ("Few procedures safeguard accuracy better than adversarial questioning."). Second, Doe has pointed to circumstances surrounding his disciplinary proceedings that, accepting all of his factual allegations as true and drawing all reasonable inferences in his favor, plausibly suggest the university acted with bias based on his sex. See Iqbal , 556 U.S. at 681-82 , 129 S.Ct. 1937 .

Around two years before Doe's disciplinary proceeding, the federal government launched an investigation to determine whether the university's process for responding to allegations of sexual misconduct discriminated against women. When news of the investigation broke, student groups and local media outlets sharply criticized the administration. The federal government's investigation and the negative media reports continued for years, throughout the Board's consideration of Doe's case.

This public attention and the ongoing investigation put pressure on the university to prove that it took complaints of sexual misconduct seriously. The university stood to lose millions in federal aid if the Department found it non-compliant with Title IX. The university also knew that a female student had triggered the federal investigation and that the news media consistently highlighted the university's poor response to female complainants. Of course, all of this external pressure alone is not enough to state a claim that the university acted with bias in this particular case. Rather, it provides a backdrop that, when combined with other circumstantial evidence of bias in Doe's specific proceeding, gives rise to a plausible claim. See Twombly , 550 U.S. at 570 , 127 S.Ct. 1955 .

Specifically, the Board credited exclusively female testimony (from Roe and her witnesses) and rejected all of the male testimony (from Doe and his witnesses). In doing so, the Board explained that Doe's witnesses lacked credibility because "many of them were fraternity brothers of [Doe]." But the Board did not similarly note that several of Roe's witnesses were her sorority sisters, nor did it note that they were female. This is all the more telling in that the initial investigator who actually interviewed all of these witnesses found in favor of Doe. The Board, by contrast, made all of these credibility findings on a cold record.

When viewing this evidence in the light most favorable to Doe, as we must, one plausible explanation is that the Board discredited all males, including Doe, and credited all females, including Roe, because of gender bias. And so this specific allegation of adjudicator bias, combined with the external pressure facing the university, makes Doe's claim plausible. Indeed, other courts facing similar allegations have reached the same result. See, e.g. , Miami Univ. , 882 F.3d at 594 (plaintiff's complaint was sufficient where allegations included that the university faced "pressure from the government to combat vigorously sexual assault on college campuses and the severe potential punishment-loss of all federal funds-if it failed to comply"); Doe v. Columbia Univ. , 831 F.3d 46 , 56-57 (2d Cir. 2016) (plaintiff's complaint pointing to student group criticism and university statements was sufficient to raise a plausible inference of bias under a "minimal plausible inference" standard); Doe v. Amherst Coll. , 238 F.Supp.3d 195 , 223 (D. Mass. 2017) (plaintiff's complaint was sufficient where allegations *587 suggested that university was trying to "appease" a biased, student-led movement); Doe v. Lynn Univ., Inc. , 235 F.Supp.3d 1336 , 1340-42 (S.D. Fla. 2017) (plaintiff's complaint was sufficient where allegations suggested that university was reacting to "pressure from the public and the parents of female students" to punish males accused of sexual misconduct); Wells v. Xavier Univ. , 7 F.Supp.3d 746 , 751 (S.D. Ohio 2014) (plaintiff's complaint was sufficient where, taken together, his allegations suggested that university was "reacting against him[ ] as a male" in response to a Department of Education investigation).

The dissent disagrees, taking a deep and thoughtful dive into the factual record to conclude that there is "no basis to reasonably infer" that Doe was a victim of gender discrimination. But when viewed against the backdrop of external pressure, the Board's decision to discredit Doe's fraternity brothers in part because they were fraternity brothers, while not holding Roe's witnesses to the same standard, is basis enough at the motion-to-dismiss stage. Of course, anti-male bias is not the only plausible explanation for the university's conduct, or even the most plausible. The university might have been unaffected by the federal investigation or the media's criticism, and the significance of the Board's decision to disregard Doe's witnesses' statements might be overblown. And as the dissent points out, the Board might have ruled the way it did because it believed Witness 2's testimony was more credible. But alternative explanations are not fatal to Doe's ability to survive a Rule 12(b)(6) motion to dismiss. See 16630 Southfield Ltd. P'ship v. Flagstar Bank, F.S.B. , 727 F.3d 502 , 505 (6th Cir. 2013) ("[T]he mere existence of more likely alternative explanations does not automatically entitle a defendant to dismissal."); Watson Carpet & Floor Covering, Inc. v. Mohawk Indus., Inc. , 648 F.3d 452 , 458 (6th Cir. 2011) (noting that there are often several plausible explanations for a defendant's conduct, but "[f]erreting out the most likely reason for the defendants' actions is not appropriate at the pleadings stage"). Doe's allegations do not have to give rise to the most plausible explanation-they just have to give rise to one of them. See Iqbal , 556 U.S. at 678 , 129 S.Ct. 1937 (stating that there is no "probability requirement" at the pleading stage (quoting Twombly , 550 U.S. at 556 , 127 S.Ct. 1955 ) ).

As this case proceeds and a record is developed, evidence might very well come to show that today's inference is the least plausible of the bunch. Certain allegations that we must assume are true might be proven false. And with the benefit of exhibits, testimony, and cross-examination, a fact-finder may conclude that the inferences we were required to draw in Doe's favor are simply untenable. But these possibilities cannot affect this court's evaluation of Doe's complaint. Our job is simply to ensure that Doe is not deprived of an opportunity to prove what he has alleged unless he would lose regardless. Because Doe has alleged facts that state a plausible claim for relief, we reverse the district court's decision to dismiss his complaint. Whether he will ultimately succeed is a question for another day.

Archaic Assumptions and Deliberate Indifference . Doe advances two more theories of Title IX liability. First, he maintains that the university relied on archaic assumptions about the sexes when resolving his case. And second, he claims that the university was deliberately indifferent to the Board's sex discrimination. The problem for Doe, however, is that neither of these theories applies in the context of university disciplinary proceedings.

Title IX plaintiffs use the archaic-assumptions theory to show that a *588 school denied a student an equal opportunity to participate in an athletic program because of historical assumptions about boys' and girls' physical capabilities. See Mallory v. Ohio Univ. , 76 F. App'x 634 , 638-39 (6th Cir. 2003). This court has never applied the theory outside of the athletic context, and, indeed, we have repeatedly refused litigants' requests to do so. See Cummins , 662 F. App'x at 451 n.9. Since Doe has not offered any reason why we should change course and take that step today, we affirm the district court's decision to dismiss on this ground.

The same problem dooms Doe's deliberate-indifference theory. The deliberate-indifference theory was designed for plaintiffs alleging sexual harassment . See Horner v. Ky. High Sch. Athletic Ass'n , 206 F.3d 685 , 693 (6th Cir. 2000) (explaining that the deliberate-indifference test arose from Supreme Court cases that "all address deliberate indifference to sexual harassment"). And though sexual harassment is a form of discrimination for purposes of Title IX, Davis v. Monroe Cty. Bd. of Educ. , 526 U.S. 629 , 649-50, 119 S.Ct. 1661 , 143 L.Ed.2d 839 (1999), we have held that to plead a Title IX deliberate-indifference claim, "the misconduct alleged must be sexual harassment," not just a biased disciplinary process. Miami Univ. , 882 F.3d at 591 . Because Doe did not allege that actionable sexual harassment occurred during his disciplinary proceedings, he failed to state a claim under Title IX by way of deliberate indifference.

V.

Accordingly, we REVERSE the district court's dismissal of John Doe's procedural due process claim insofar as it is based on the university's failure to provide a hearing with the opportunity for cross-examination, we REVERSE the district court's dismissal of John Doe's Title IX claim insofar as it is based on erroneous outcome, and we REMAND for further proceedings consistent with this opinion.

JULIA SMITH GIBBONS, Circuit Judge, concurring.

CONCURRENCE

I write separately to make one discrete point with respect to the Title IX Claim. I agree that Doe has plausibly alleged a claim of gender bias. The inclusion of materials, other than the complaint, in the record makes a summary judgment standard tempting. The dissent avoids summary judgment language, but its analytical approach is analogous to the process by which a judge determines the existence of a genuine issue of material fact. Yet Doe is entitled to the full benefit of the standard for considering a motion to dismiss. Under that standard, my view is that Doe's complaint survives.

RONALD LEE GILMAN, Circuit Judge, concurring in part and dissenting in part.

CONCURRING IN PART AND DISSENTING IN PART

I concur in the majority's judgment (but not in its discussion) with regard to the disposition of Doe's due process claim. As to Doe's Title IX claim, I would affirm the judgment of the district court because of Doe's failure to plausibly state a claim under Title IX.

I. Due process claim

Although I agree that Doe's due process rights were violated when he was not permitted the opportunity to engage in any form of cross-examination of the witnesses against him, I disagree with the majority about the scope of cross-examination mandated by the United States Constitution in *589 this context. I particularly believe that the majority has traveled "a bridge too far" in mandating that "if the university does not want the accused to cross examine the accuser under any scenario, then it must allow a representative to do so." Maj. Op. at 583 n.3.

This court has found that when witness credibility is at issue, the accused must have an opportunity for at least a "circumscribed form" of cross-examination where he or she is allowed to submit questions to the trier of fact, who will then directly pose those questions to the witnesses. Doe v. Cummins , 662 F. App'x 437 , 446 (6th Cir. 2016). Cummins held that this requirement was met even where the trier of fact did not ask all the questions submitted or allow an opportunity to submit follow-up questions. Id. at 448 ; see also Doe v. Univ. of Cincinnati , 872 F.3d 393 , 406 (6th Cir. 2017) (emphasizing that the university has only a narrow obligation to provide the trier of fact with an opportunity "to evaluate an alleged victim's credibility, not [to allow] the accused to physically confront his accuser," and that "what matters" is that the trier of fact has the "ability to assess the demeanor of both the accused and his accuser"); Nash v. Auburn University , 812 F.2d 655 , 664 (11th Cir. 1987) (finding that the due process rights of a student suspended for academic dishonesty were not violated where he was given the opportunity to submit questions to the trier of fact, who would then direct the questions to the witnesses, and that the student did not have a right to "question the adverse witnesses in the usual, adversarial manner").

Although Cummins is factually distinguishable because the two accused students faced only suspension and probation rather than expulsion, the majority cites no case that would support its expansion of Doe's cross-examination rights beyond those set forth in Cummins . Nor does the majority explain why the Eldridge balancing factors would require the added protection of unfettered cross-examination by a representative whenever expulsion from a university is a potential penalty. See Mathews v. Eldridge , 424 U.S. 319 , 335, 96 S.Ct. 893 , 47 L.Ed.2d 18 (1976).

And this expansion, in the absence of a focused and caselaw-supported analysis, leaves many questions unanswered. For example, who is the "representative" that will be allowed to question witnesses on the accused's behalf? Is it an attorney? If so, then this expanded right of cross-examination conflicts with our caselaw making clear that a student has no constitutional right to have an attorney actively participate in his disciplinary hearings, except in very limited circumstances. See Flaim v. Med. Coll. of Ohio , 418 F.3d 629 , 640 (6th Cir. 2005) (noting that a student has no recognized right to have counsel participate in school disciplinary proceedings except, possibly, where the proceedings are complex or where the university itself utilizes an attorney); Cummins , 662 F. App'x at 448-49 (same); Gorman v. Univ. of Rhode Island , 837 F.2d 7 , 16 (1st Cir. 1988) (noting that "the weight of authority is against [recognizing the right to] representation by counsel at [school] disciplinary hearings, unless the student is also facing criminal charges stemming from the incident in question"); Donohue v. Baker , 976 F.Supp. 136 , 146 (N.D.N.Y. 1997) (noting that a student in a school disciplinary hearing has a right to counsel only to protect his Fifth Amendment right against self-incrimination, not to affect the outcome of the hearing through cross-examination).

Should the representative instead be a teacher or an administrator? Such an individual would undoubtedly need to be paid for his or her work, imposing an additional burden on the university. Could the representative *590 be a friend or family member of the accused? And would the rules of evidence apply to the cross-examination? Cf. Flaim , 418 F.3d at 635 (observing that "[c]ourts have generally been unanimous ... in concluding ... that neither rules of evidence nor rules of civil or criminal procedure need be applied" in school disciplinary hearings). Assuming they would not, who would decide what limits to impose on the representative's questioning and using what criteria?

This court has repeatedly held that "[f]ull-scale adversarial hearings in school disciplinary proceedings have never been required by the Due Process Clause." Univ. of Cincinnati , 173 F.Supp.3d at 603 (quoting Flaim , 418 F.3d at 640 ). And the burden of allowing a representative to participate by cross-examining witnesses "in every disciplinary hearing would be significant due to the added time, expense, and increased procedural complexity." Cummins , 662 F. App'x at 449 ; see also Flaim , 418 F.3d at 640-41 ("[C]onducting [full adversarial hearings] with professional counsel would entail significant expense and additional procedural complexity."). This would be especially true if the university were also required to provide representation for a student who could not provide his or her own.

Although a university may choose to allow an agent or representative of an accused student to cross-examine the complainant and his or her witnesses, no court has previously held that this is constitutionally required. This court has instead held that the university must provide at least the "circumscribed form" of cross-examination set out in Cummins , 662 F. App'x at 446 . And because Doe was not provided with even this level of cross-examination, I agree that his due process rights were violated.

I recognize that a case might arise where the Constitution requires more than the procedures that this court approved of in Cummins , but we should address that issue only if and when it arises. We need not-and should not-resolve it today because we have been given neither the facts nor the arguments necessary to conduct an adequate analysis. I therefore believe that we should refrain from imposing on all universities a rigid requirement to provide students facing expulsion with an opportunity to have a representative cross-examine adverse witnesses. See Kremer v. Chem. Constr. Corp. , 456 U.S. 461 , 482, 102 S.Ct. 1883 , 72 L.Ed.2d 262 (1982) ("We must bear in mind that no single model of procedural fairness, let alone a particular form of procedure, is dictated by the Due Process Clause. ... 'The very nature of due process negates any concept of inflexible procedures universally applicable to every imaginable situation.' " (quoting Mitchell v. W.T. Grant Co. , 416 U.S. 600 , 609, 94 S.Ct. 1895 , 40 L.Ed.2d 406 (1974) ) ); Gorman v. Univ. of Rhode Island , 837 F.2d 7 , 16 (1st Cir. 1988) ("[O]n review, the courts ought not to extol form over substance, and impose on educational institutions all the procedural requirements of a common law criminal trial. ... In all cases the inquiry is whether, under the particular circumstances presented, the hearing was fair, and accorded the individual the essential elements of due process.").

II. Title IX claim

I now turn to Doe's claim under Title IX. Doe contends that the University faced pressure from the United States Department of Education, the general public, and student groups to adequately address sexual-assault complaints made against male students on campus and that, as a consequence, the University erroneously found him responsible for sexual misconduct because of his gender. But no circuit court has ever held that a student plausibly states a claim that deficiencies in his disciplinary *591 proceedings were motivated by gender bias where the only fact that he alleges to show such bias is general pressure on the university to adequately address allegations of sexual assault. Cf. Doe v. Miami Univ. , 882 F.3d 579 , 593 (6th Cir. 2018) (noting that to survive a motion to dismiss, a plaintiff must show a "causal connection between the flawed [disciplinary] outcome and gender bias" by alleging, "inter alia, statements by members of the disciplinary tribunal, statements by pertinent university officials, or patterns of decision-making that also tend to show the influence of gender").

In Miami University , this court found the complaint sufficient where it alleged facts showing a pattern of gender-based decision-making, in addition to general pressure on the university to take sexual-assault claims seriously. Id. at 593 . This evidence included "an affidavit from an attorney who represents many students in Miami University's disciplinary proceedings, which describes a pattern of the University pursuing investigations concerning male students, but not female students." Id. It also included an allegation that the university investigated the complaint of sexual misconduct made against the male plaintiff but did not investigate his allegation that his female accuser actually perpetrated sexual misconduct against him. Id. at 590-91, 593 .

The Miami University court further noted that the university "was facing pressure to increase the zealousness of its 'prosecution' of sexual assault and the harshness of the sanctions it imposed because it was a defendant in a lawsuit brought by a student who alleged that she would not have been assaulted if the University had expelled her attacker for prior offenses." Id. at 594 . This alleged pattern of activity relating to sexual-assault matters, combined with the general pressure on the university, was deemed sufficient to "support a reasonable inference of gender discrimination" and therefore to survive a motion to dismiss. Id.

Similarly, in Doe v. Columbia University , 831 F.3d 46 , 57 (2d Cir. 2016), the Second Circuit held that a complaint plausibly alleged gender discrimination when it contended that, "during the period preceding the disciplinary hearing, there was substantial criticism of the University[ ] both in the student body and in the public media [that] accus[ed] the University of not taking seriously complaints of female students alleging sexual assault by male students." Id. The complaint further alleged that "the University's administration was cognizant of, and sensitive to, these criticisms, to the point that the President called a University-wide open meeting with the Dean to discuss the issue." Id. Moreover, the investigator in that case had been subjected to "personal criticism" by the student body and in news articles "for her role in prior cases in which the University was seen as not taking seriously the complaints of female students." Id. at 51, 58 .

The investigator in Columbia University was thus aware that the university "had been criticized for ... conducting the investigations in a manner that favored male athletes and that was insufficiently protective of sexually assaulted females." Id. Finally, the plaintiff in Columbia University alleged that the investigator failed to interview key witnesses identified by the plaintiff, that she was hostile to him during interviews, and that she failed to inform him of his right to make a statement at the hearing. Id. at 49, 52 .

Unlike the plaintiffs in Miami University and Columbia University , Doe crucially fails to link general pressure on the University of Michigan to the particular proceedings that he faced. See Doe v. Cummins , 662 F. App'x 437 , 452 (6th Cir. 2016) (noting that "to state an erroneous-outcome *592 claim, a plaintiff must plead ... a ' particularized ... causal connection between the flawed outcome and gender bias' " (emphasis and first ellipsis added; second ellipsis in original) (quoting Yusuf v. Vassar College , 35 F.3d 709 , 715 (2d Cir. 1994) ) ). Nor does Doe allege any facts suggesting a pattern of discriminatory behavior by the University in its response to sexual-assault allegations, or that he made any sexual-misconduct complaints himself that the University ignored, in contrast to the plaintiff's allegations in Miami University , 882 F.3d at 590-91, 593 . There is also no allegation here that the investigator faced individualized criticism for her handling of previous sexual-assault claims and subsequently manifested hostility toward Doe, in contrast to the plaintiff's contentions in Columbia University , 831 F.3d at 49, 51-52, 58 . In fact, the investigator here found in favor of Doe, and Doe acknowledged that her investigation was "thorough."

Doe also fails to show how general pressure on the University's administration to pursue and effectively address sexual-assault complaints led the Appeals Board-a board made up of an assistant dean from the law school, a retired professor from the dentistry school, and a student-to take actions against him based on gender bias. He also fails to identify any practice or policy adopted by the University in response to either external or internal pressure that would reflect bias against males. Moreover, the media reports published after the Appeals Board decision (which reports allege that the University was continuing to inadequately address sexual-misconduct complaints) would appear to belie any contention that the University had overcorrected by adopting policies or practices biased against male students accused of sexual misconduct.

The majority in fact recognizes that the alleged external pressure on the University alone is not sufficient to plausibly show that a university acted based on gender bias in Doe's particular case. Maj. Op. at 586. But it concludes that this pressure is sufficient when combined with Doe's allegation that the Appeals Board adopted all of the statements made by the female witnesses and rejected all of the statements made by the male witnesses. Maj. Op. at 586-87. More specifically, the majority reasons that "when viewed against the backdrop of external pressure, the Board's decision to discredit Doe's fraternity brothers in part because they were fraternity brothers, while not holding Roe's witnesses to the same standard, is basis enough at the motion-to-dismiss-stage." Maj. Op. at 587. But the majority's observation about the Appeals' Board's alleged disparate treatment of the witnesses is not borne out by the record. (I recognize that the record would not normally be considered at the motion to dismiss stage of the case. But as acknowledged in footnote 4 of the majority opinion, the administrative record was referenced in the complaint and, without objection by either party, considered as part of the pleadings.)

To start with, the Appeals Board discussed statements from only two of Roe's sorority sisters, although additional sorority sisters provided statements that were contained in the investigator's report. The record reflects the following evaluation by the Appeals Board:

Two witnesses who know [Roe] reported that they observed [Roe] drinking from the wine bag at [Doe's] fraternity and also reported that they perceived she was intoxicated for a variety of reasons (very energetic when she's drunk; inhibitions were lowered; and speech that was 'not completely clear,' contained 'occasional slurs,' and occasionally trailed off at the end of sentences).

*593 The Appeals Board provided no further discussion of these statements that would suggest that it was relying on them beyond its observation that Roe's statements were "corroborated by other witnesses, particularly by Witness 2's observations of [Roe's] behavior and physical condition immediately after the sexual encounter." And this observation by the Appeals Board leads directly to the biggest weakness in both Doe's and the majority's position: the Appeals Board's decision to credit the testimony of Roe and Witness 2 (and subsequently to find Doe responsible for sexual misconduct) was based on the considerations that (1) Witness 2 spent significant time with Roe following Roe's sexual encounter with Doe, and (2) Witness 2 had no connection to Doe, Roe, or their respective Greek institutions.

The Appeals Board explained:

Although there were other witnesses who observed Complainant both prior to and after the sexual encounter with Respondent, many of them were fraternity brothers of Respondent, and all of them only observed Complainant briefly and/or at a distance. For these reasons, we find their statements to be significantly less persuasive than the statements of Complainant and Witness 2. Complainant knew that she consumed an excessive amount of alcohol and recognized that she was not mentally or physically in control. Witness 2 had no previous connection to Complainant and observed her for a lengthy period of time, spanning from a few minutes after Complainant's sexual encounter with the Respondent, through the time she got Complainant into bed at her dorm.

Whether the statements made by Roe's sorority sisters were credible was not discussed. The Appeals Board's decision instead shows that the statements by Doe and his witnesses were disfavored only as compared to the statements of Roe and Witness 2, and that there was no categorical preference shown for or against statements by fraternity brothers versus sorority sisters, or for or against statements by men versus women as such. The Appeals Board also noted that Witness 2's observations were further corroborated by two witnesses who helped Roe into a vehicle outside the fraternity house and who, like Witness 2, had no connection to Doe or Roe.

I therefore find no basis to reasonably infer that the Appeals Board declined to rely on the statements made by Doe and his witnesses simply because they were men . This leaves us with only one fact from which to infer that gender bias caused the procedural defects in Doe's disciplinary proceedings-the general pressure on the University to adequately address sexual-assault claims. But as discussed above, this is not sufficient to show the "particularized ... casual connection" required to plausibly allege a claim of gender bias under Title IX. See Doe v. Miami Univ. , 882 F.3d 579 , 592 (6th Cir. 2018) (quoting Doe v. Cummins , 662 F. App'x 437 , 452 (6th Cir. 2016) ); Cummins , 662 F. App'x at 453 (noting that a complaint is insufficient if it shows at most "a disciplinary system that is biased in favor of alleged victims and against those accused of misconduct"). Absent an allegation of some particularized facts linking gender bias to the University's disciplinary practices or proceedings, I respectfully dissent as to the viability of Doe's Title IX claim.