Tuesday, July 15, 2014

Do credentials scandals support the signaling model?

Every now and then, the world suddenly learns that a perfectly competent worker faked his credentials.  Consider the case of MIT's former head of admissions:

    "Marilee Jones, the dean of admissions at the Massachusetts Institute of Technology... admitted that she had fabricated her own educational credentials and resigned after nearly three decades at MIT. Officials of the institute said she did not have even an undergraduate degree.

    "I misrepresented my academic degrees when I first applied to MIT 28 years ago and did not have the courage to correct my résumé when I applied for my current job or at any time since," Jones said in a statement posted on the institute's Web site. "I am deeply sorry for this and for disappointing so many in the MIT community and beyond who supported me, believed in me, and who have given me extraordinary opportunities."


    Jones, 55, originally from Albany, New York, had on various occasions represented herself as having degrees from three upstate New York institutions: Albany Medical College, Union College and Rensselaer Polytechnic Institute. In fact, she had no degrees from any of those places, or anywhere else, MIT officials said.

    A spokesman for Rensselaer said Jones had not graduated from there, though she did attend as a part-time nonmatriculated student during the 1974-75 school year. The other colleges said they had no record of her."

When credential scandals break, fans of the signaling model of education often run a victory lap.  Should they?  On the surface, credential scandals seem inconsistent with both the standard human capital model and the signaling model.  Remember: Both models predict that the more-educated will be better workers than the less-educated.  In the human capital model, educated workers are better because school instills job skills.  In the signaling model, educated workers are better because school certifies job skills.  From both perspectives, then, the fact that Marilee Jones could fake her credentials, then do a fine job for 28 years, is weird.

On deeper reflection, however, the signaling model makes credential scandals less weird than the human capital model.  In the human capital model, there is no need to fake credentials; the market knows what you can do, and pays you accordingly.  In the signaling model, in contrast, the market has to guess what you can do.  If the initial guess is unfavorable, a qualified worker may fail to even land an interview, much less a job.

So what?  Ponder a job-seeker's incentive to lie in the following two scenarios.  In the first scenario, you lack both credentials and the skills those credentials normally indicate.  In the second, you're a diamond in the rough - someone whose credentials understate their true skills.

Scenario #1: In the absence of deception, the market will correctly guess that you're not good enough for the job.  What does lying get you?  If your employer fires you as soon as he discovers your incompetence, you still get paid above your true market value for a few weeks or months.  Given a typical firing-averse employer, you may get paid above your true market value until the next recession.

Scenario #2: In the absence of deception, the market will incorrectly guess that you're not good enough for the job.  What does lying get you?  Your big break!  Your lie wins you an opportunity, and that opportunity allows you to win an employer's esteem - and start permanently earning your true market value.

The incentive to lie in Scenario #1 is plain.  But in Scenario #2, the incentive to lie is overwhelming!  If you're really a diamond in the rough, one well-crafted lie can change your life for the rest of your life.  Or at least 28 years.

If all this is true, why do firms so often punish credential fraud with termination?  "Credibility" is the obvious answer.  Forgiving credential fraud committed in the past is a great way to encourage more credential fraud in the future.

But this answer dodges the more fundamental question: Why should firms detest credential fraud so much in the first place?  In the human capital model, it's unclear.  When you pay workers for their actual productivity, there's little need to police what workers claim about their productivity, because so little hinges on such claims.

In the signaling model, in contrast, perception is all-important, because you pay workers for their perceived productivity.  Harshly punishing credential scandals sends a valuable message to all your present and future workers: "Never mess with my perceptions."

So yes, credential scandals do support the signaling model.  But instead of running victory laps, fans of signaling should carefully explain why these scandals are weird, and how their preferred model makes the weirdness normal.


'Stop trying to make girls take science': It goes against their human nature, claims psychologist

Attempts to encourage more girls to study the sciences ‘completely deny human biology and nature’, an academic has claimed.

Schools should stop trying to close education gender gaps because innate differences between the sexes mean they will always be drawn to certain subjects and careers, according to a Glasgow University psychologist.

Dr Gijsbert Stoet argued that Britain ‘probably needed to give up on the idea that we will get many female engineers or male nurses’.

Efforts to tackle gender inequalities were wasted because pupils with an aptitude for subjects such as physics, computing and engineering would choose them anyway, he claimed.

But his remarks, which fly in the face of campaigns by the Government and academic societies, drew a storm of controversy yesterday. The Institute of Physics attacked his position as ‘ridiculous’ and said it denied the strong influences of society and upbringing on women’s study choices.

Dr Stoet said it was ‘really hard’ to attract girls to subjects such as computing, telling the British Education Studies Association in Glasgow: ‘Girls will say, “Well, that’s boring, I’m just not interested in it”. ‘We need to have a national debate on why we find it so important to have equal numbers.

‘Do we really care that only 5 per cent of the programmers are women? … I don’t care who programs my computers. A wealthy, democratic society can afford to let people do what they want.

‘What is better? To have 50 per cent of female engineers who do not really like their work but say, “Yeah, well, I did it for the feminist cause”. Or do you want 3 per cent of female engineers who say “I really like my job”?’

His remarks were reported in the TES teachers’ journal. Dr Stoet went on to question the national focus given to girls’ struggles in subjects such as maths, when boys generally performed worse at school.

‘Nobody seems to be that interested that boys have problems. We have, as human beings, a natural tendency to see women as vulnerable and needing help. But if it’s a boy who needs help, he’s responsible for himself,’ he said.

Dr Stoet’s comments follow the launch of the Government’s Opening Doors campaign to ‘stamp out gender bias’ in schools, which it said is putting girls off taking maths and science A-levels.

Education Minister Elizabeth Truss warned that England’s school system was affected by ‘science deserts’, where too few pupils – particularly girls – are taking subjects such as chemistry and physics. At half of mixed state schools, not a single girl is taking physics A-level, she said.

In a report last December, the Institute of Physics suggested 49 per cent of state schools were ‘reinforcing gender stereotypes’ in the advice given to pupils about A-levels.

The institute’s Clare Thomson said Dr Stoet was implying differences in male and female brains drove subject choices. But evidence suggests that difference within the sexes are bigger than differences between them, she said.

‘Broadly you can say women are more interested in this and men are more interested in that but there’s a huge amount of crossover,’ she added.

‘Just to say, “We should give up on this because that is the way it is”, is ridiculous. It also denies the influences of society and upbringing which we know are incredibly strong.

‘Young people make choices based on the information they have and the advice they have taken.’


Proof and campus rape: Standards for campus disciplinary proceedings

By Hans Bader

Yesterday, the New York Times published my letter to the editor about evidence standards in college disciplinary proceedings over rape and sexual harassment, titled "Proof and Campus Rape." The letter was too short to provide a detailed discussion of the subject, so I am adding more discussion below for the benefit of readers.

As I noted in the Times,

It is unfortunate that the federal government now seeks to restrict the due process rights of students accused of sexual assault or harassment. As your editorial notes, “the Department of Education recommended that universities use ‘preponderance of the evidence’ as the standard of proof instead of the higher ‘clear and convincing’ standard.” But that higher standard protected due process.

As a Yale Law Journal article noted in 1987, “Courts, universities and student defendants all seem to agree that the appropriate standard of proof in student disciplinary cases is one of ‘clear and convincing’ evidence.”

The White House Task Force you cite demanded that students “not be allowed to personally cross-examine each other.” That contradicts court rulings like Donohue v. Baker (1997), which concluded that cross-examination must be allowed in certain campus disciplinary hearings.

The White House and Education Department argue that despite these facts, the Title IX statute requires a lower burden of proof and limits on cross-examination. But they are wrong, and they misconstrue court rulings to make this argument.

In 2011, the Education Department ordered many colleges to change the burden of proof that they use in disciplinary proceedings over sexual harassment. (Under pressure from the Education Department's Office for Civil Rights, where I used to work, colleges are now routinely expelling or suspending students who are very likely innocent of sexual harassment or assault, see here, here, here, here, here, here, here, here, and here. They are also taking "interim measures" against accused students who have yet to be found guilty of anything, and the Education Department wants them to take such measures even against students who have never been formally charged with anything on campus.)

In doing so, the Education Department illegally legislated through administrative fiat, and undermined due-process safeguards. It also sought to eliminate cross-examination by the accused, which could reduce accuracy in campus adjudications, and could result in discipline based on constitutionally-protected speech in cases where verbal harassment is alleged.

Historically, most colleges used a “clear and convincing” evidence standard in student and faculty discipline cases, to safeguard due process. As James Picozzi noted in 1987 in the Yale Law Journal, “Courts, universities, and student defendants all seem to agree that the appropriate standard of proof in student disciplinary cases is one of ‘clear and convincing’ evidence.” (James M. Picozzi, University Disciplinary Process: What’s Fair, What’s Due, and What You Don’t Get, 96 Yale L. J. 2132, 2159 n. 17 (1987); see also Nicholas Trott Long, The Standard of Proof in Student Disciplinary Cases, 12 J. College & U.L. 71 (1985)). There was no exception for people accused of sexual harassment, who thus received the same due-process protections as everyone else.

But in an April 4, 2011 “Dear Colleague” letter, the Education Department’s Office for Civil Rights (OCR) rejected this traditional protection for students accused of harassment. It declared that schools cannot use a clear-and-convincing standard of proof typical in school disciplinary procedures for sexual harassment or sexual assault cases: “A school’s grievance procedures must use the preponderance of the evidence standard to resolve complaints of sex discrimination.” See Dear Colleague Letter: Sexual Violence Background, Summary and Fast Facts. “Preponderance of the evidence” means that if a school thinks there is as little as a 51 per chance that the accused is guilty, the accused must still be disciplined.

The Education Department’s position was based on a misunderstanding of who is subject to Title IX, the federal law banning schools from committing sex discrimination. Title IX’s requirements apply to schools, not individual students.

The Education Department’s reasoning for imposing a low “preponderance” standard on school disciplinary proceedings was that this “is the standard of proof established for violations of civil-rights laws” in lawsuits brought in federal court. (It cited court rulings under another civil-rights law, Title VII, which also bans discrimination). Therefore, it claimed, preponderance must also be “the appropriate standard for” schools to use in “investigating allegations of sexual harassment or violence.’”

It was completely true, and completely irrelevant, that the preponderance of the evidence standard applies in lawsuits in general, as well as civil-rights cases. But that burden of proof applies to whether the school violated Title IX by behaving inappropriately, not whether students or staff engaged in harassment. Students cannot violate Title IX; only schools can be sued under Title IX, not individuals. (See, e.g., Smith v. Metropolitan School District (1997).) Moreover, Students “are not agents of the school,” so their actions don’t count as the actions of the school.

The mere existence of harassment by students isn’t enough for liability under Title IX. More is required. The school’s own actions in response to harassment must be culpable. As the Education Department admitted in its 1997 “Sexual Harassment Guidance,” “Title IX does not make a school responsible for the actions of harassing students, but rather for its own discrimination in failing to remedy it once the school has notice.” (62 FR 12034 (1997)). So to violate Title IX, an institution’s own actions must be proven culpable under a “preponderance” standard — not the mere occurrence of harassment.

Since an institution itself must behave culpably, not just the accused harasser, federal courts have held that there is no violation of the civil rights laws even if harassment occurs, as long as the institution investigates in good faith in response to the allegation of harassment. That’s true even if the institution ultimately refuses to discipline a harasser based on the reasonable belief that he is innocent, after applying a firm presumption of innocence.

For example, an appeals court reversed a jury verdict against the Postal Service for sexual harassment, even though harassment by Postal employees did occur, since the Postal Service had, after investigating the sexual harassment complaint, reasonably, but erroneously, failed to credit plaintiff’s allegations. As the court explained, “a good faith investigation of alleged harassment may satisfy the ‘prompt and appropriate response’ standard, even if the investigation turns up no evidence of harassment. . .[and] a jury later concludes that in fact harassment occurred.” See Swenson v. Potter, 271 F.3d 1184, 1196 (9th Cir. 2001), quoting Harris v. L & L Wings, 132 F.3d 978, 984 (4th Cir. 1997).

“Such an employer may avoid liability even if a jury later concludes that in fact harassment occurred” (see Harris v. L & L Wings, 132 F.3d 978, 984 (4th Cir. 1997)), and “an employer, in order to avoid liability for the discriminatory conduct of an employee, does not have to necessarily discipline or terminate the offending employee.” (See Knabe v. Boury Corp., 114 F.3d 407, 414 (3d Cir. 1997).)

For example, an appeals court held that an employer did not have to discipline an accused employee where the evidence did not convincingly prove the existence of harassment, citing the absence of a corroborating witness. (See Knabe v. Boury Corporation, 114 F.3d 407 (3rd Cir. 1997).) That employer escaped liability despite requiring more than a close case for discipline, as a preponderance of evidence would mandate. A corroborating witness is not needed to show proof under a mere preponderance-of-the-evidence standard.

Similarly, in another case, the appeals court rejected an attempt to hold an employer liable for harassment because it failed to discipline a harasser where it was genuinely unclear at the time whether he was guilty: “It would be unreasonable, and callous toward [the accused harasser’s] rights, for the law to require Wal-Mart to discipline [him] for events he denies, of which Wal-Mart could not find evidence.” (See Adler v. Wal-Mart, 144 F.3d 664 (10th Cir. 1998).) Thus, it can be perfectly reasonable, and thus legal, to give the accused a firm presumption of innocence, especially where the accused has no previous history of harassment.

Courts had said nothing since those decisions, or the Education Department’s 1997 harassment guidance, to suggest that schools violate Title IX just because students themselves are guilty of harassment under a mere preponderance-of-the-evidence standard. Accordingly, the Education Department had no legal basis for its demands.

Indeed, court decisions since the Education Department's 1997 harassment guidance had made it harder to hold schools liable than the Education Department wished: decisions like the Supreme Court’s 1998 Gebser decision, which dismissed a Title IX lawsuit for failure to show both a school’s knowledge of harassment and its “deliberate indifference” to it. OCR claims that “this deliberate indifference” requirement is irrelevant to its own interpretation of Title IX, and applies only to Title IX lawsuits against schools, in which plaintiffs must show both a Title IX violation and deliberate indifference to get damages for harassment.

But even if that claim by OCR were true, it wouldn’t matter. Giving someone a presumption of innocence isn’t a “violation” of the civil rights laws in the first place, as the cases discussed above illustrate. Under the Education Department’s own 1997 guidance, it’s the school, and its action in response to the harassment, that must be culpable to violate the statute, not just the harasser’s own conduct. And it’s not in any way culpable for a school to give someone a presumption of innocence. No one has a “right” to discipline just because harassment happens: they only have a right to a reasonable response by their school in response to their complaint, which may or may not lead to disciplinary action depending on the evidence.

The Supreme Court also noted in its 1999 decision in Davis v. Monroe County Board of Education that Title IX doesn’t give complainants a “right” to second-guess a school’s reasonable decisions about discipline, noting that there isn’t any “Title IX right” to “particular remedial demands” or “particular disciplinary action,” just because harassment has occurred:

“We stress that our conclusion here . . . does not mean that recipients can avoid liability only by purging their schools of actionable peer harassment or that administrators must engage in particular disciplinary action. . . the dissent erroneously imagines that victims of peer harassment now have a Title IX right to make particular remedial demands . . .courts should refrain from second guessing the disciplinary decisions made by school administrators,” who “must merely respond to known peer harassment in a manner that is not clearly unreasonable.”

Many reasonable school officials with expertise in civil-rights law have long supported using a higher burden of proof than a mere preponderance of the evidence, contrary to the Education Department’s recent demand. For example, Harvard Law School’s “Policy and Guidelines Related to Sexual Harassment,” drafted by a committee that included a professor who taught gender-discrimination law, and adopted by Harvard Law’s faculty in April 1995, until recently contained the following provision: “Burden of proof: Formal disciplinary sanctions shall be imposed only upon clear and convincing evidence.” The Education Department’s rule also conflicted with faculty collective bargaining agreements and university regulations mandating a clear-and-convincing standard. In 2011, Inside Higher Education gave one example, citing Ohio State University, “where the OCR began a compliance review in June after students alleged too high a burden of proof when considering whether to discipline those accused of sexual harassment, lowering that standard could require approval from the institution’s governing board.”

In addition to warning against “second-guessing” of school officials’ disciplinary decisions, the Supreme Court in Davis also emphasized that to successfully sue a school district for damages, a complainant alleging sexual harassment must also show that school officials were “deliberately indifferent to sexual harassment, of which they have actual knowledge.”

Applying that “deliberate indifference” standard, appellate judges ruled that where a school district does not discipline a harasser because it failed to conclusively determine that harassment occurred, it was not liable even where that failure to find guilt was “flawed,” and led to future harassment. (See Doe v. Dallas Independent School District, 220 F.3d 380 (5th Cir. 2000).) Thus, it is clear that giving someone a strong presumption of innocence is not the basis for a Title IX lawsuit, regardless of what the Education Department may think.

Even in the absence of these court rulings, the Education Department’s demands would be invalid, because they impose new legal obligations without complying with the Administrative Procedure Act, which requires agencies to publish rules and regulations in proposed form before imposing them on the public, and give the public an opportunity to comment before adopting them. Legal scholars have criticized the Education Department’s Dear Colleague letter for that very reason. For example, Cornell law professor Cynthia Bowman argued that Cornell University should not change its procedures in response to the Education Department’s demands, because the lower standard of proof mandated by the Department in its letter “has not been subjected to notice and comment.” Bowman observed that the Education Department’s directive was “not an administrative regulation, has not been subjected to notice and comment, and thus does not have the status of law.” “The consequences for someone expelled for sexual assault are enormous and will follow him throughout his life, leading to rejection by other schools, inability to qualify for the bar and a great deal of stigma,” Professor Bowman noted. “To impose those consequences on someone requires a rigorous standard of proof and many due process protections to ensure fairness.” “Indeed, there is general agreement among faculty at the Law School that the procedures being proposed [by the Education Department] are Orwellian,” she added.

Leading law professor Richard Epstein also challenged the legality of the Education Department’s demands, declaring that “the Department of Education is on a collision course with the Bill of Rights.” The Education Department’s position was criticized by lawyers such as former U.S. Civil Rights Commissioner Jennifer C. Braceras; Wendy Kaminer; Robert Smith; former Massachusetts ACLU leader Harvey Silverglate; Ilya Shapiro; and Greg Lukianoff; groups like the American Association of University Professors and the Foundation for Individual Rights in Education; and many civil libertarians and journalists.

As former ACLU board member Wendy Kaminer noted, the Education Department is also discouraging colleges from allowing accused students to cross-examine their accusers: “students accused of harassment should not be allowed to confront (or directly question) their accusers, according to OCR, because cross-examination of a complainant ‘may be traumatic or intimidating.’” As OCR puts it, “OCR strongly discourages schools from allowing the parties personally to question or cross-examine each other during the hearing.” This is perverse, since the Supreme Court has described cross-examination as the “greatest legal engine ever invented for the discovery of truth.” (See Lilly v. Virginia, 527 U.S. 116, 124 (1999)). Moreover, courts themselves invariably permit cross-examination of complainants in sexual harassment lawsuits, so OCR can hardly claim that courts view cross-examination as intrinsically unfair to complainants, much less claim it is illegal. Yet OCR ludicrously claims to be merely restating existing legal obligations under federal court rulings in its “Dear Colleague” letter.

While there is no blanket right to cross-examine in campus disciplinary proceedings, the right to cross-examine was often afforded by education codes, collective bargaining agreements, or other contracts or regulations. Title IX did not require that these be disregarded. As the Supreme Court observed in the Davis case, it is a school is entitled “to refrain from a form of disciplinary action that would expose it to constitutional or statutory claims.” Moreover, in a few campus disciplinary cases, such as Donohue v. Baker (1997), judges ruled that cross-examination was constitutionally required on due-process grounds to test the credibility of the accuser.

The Education Department’s attack on cross-examination undermines accuracy in disciplinary proceedings. The subjective nature of the legal definition of harassment means that there is no category of cases in which cross-examination is more useful or essential to ensure due process. To legally qualify as sexual harassment under Title IX, or racial harassment under Title VI, speech must be severe and pervasive enough to create a hostile learning environment for the listener, and interfere with the listener’s education, both in subjective and objective terms, according to court rulings like the Supreme Court’s 1999 Davis decision. Transitory offense is not enough. If the accuser admits on questioning that she did not really view the offensive speech as being a “big deal,” or was not shocked or surprised by it, that probably rules out the existence of a subjectively hostile environment. Indeed, a federal appeals court dismissed a racial harassment claim for just that reason in Newman v. Federal Express Corp., 266 F.3d 401 (6th Cir. 2001).

But a wrongly-accused person can’t establish that lack of a subjectively-hostile atmosphere without questioning the accuser, and may not be able to show that the accuser wasn’t greatly impacted by the speech without cross-examining the accuser about its alleged effect on her and her studies, such as whether she continued to enjoy her college experience after overhearing the allegedly “harassing” remarks.

There is a fine line between protected speech about unpleasant sexual topics and unprotected sexual harassment, and it is crucial that accused people be able to prove that their speech did not amount to sexual harassment. Even sexually vulgar speech on political issues is protected on college campuses, as the Supreme Court’s Papish decision illustrates. And perfectly civil, non-vulgar students have been subjected to disciplinary proceedings for sexual and racial harassment, in violation of the First Amendment, merely for expressing commonplace opinions about sexual and racial issues, like criticizing feminism or affirmative action, or discussing racial implications of the death penalty. (See the examples cited in the Amicus brief of Students for Individual Liberty in Davis v. Monroe County Board of Education, available at 1998 WL 847365.)

To fully defend themselves against sexual harassment charges over speech on sexual topics that doesn’t really amount to sexual harassment, people who are wrongly accused of sexual harassment will sometimes need to cross-examine their accuser to show that their speech did not really have any sexually harassing effect, and thus did not legally amount to sexual harassment, despite their accuser’s attempt to make a mountain out of a molehill.

The Education Department’s attack on cross-examination will lead to free speech violations, by resulting in students being convicted of harassment even when their speech did not create a subjectively-hostile environment, much less interfere with the accuser’s educational opportunities. If the speech has not created such an environment, it has not caused tangible harm, and cannot be banned merely because a hypothetical listener might have objected to it. One case illustrates this principle. In Meltebeke v. Bureau of Labor and Industries, 903 P.2d 351 (Or. 1995), the Oregon Supreme Court struck down a religious business owner’s fine for religious harassment because the state agency’s harassment rule violated religious-freedom guarantees. Justice Unis, in his concurrence, noted the rule also violated free speech, and was unconstitutionally overbroad, because it only required that the speech create a hostile environment for a hypothetical reasonable person — not for the actual complainant, who did not need to experience a subjectively-hostile environment.

In another disturbing move, the Education Department’s letter also forbids colleges to allow students found guilty to appeal their punishment or finding of guilt, unless the college also allows complainants to appeal findings of innocence, and to challenge supposedly inadequate punishments: “If the school provides for an appeal of the findings or remedy, it must do so for both parties.” That conflicted with the practices of many campus disciplinary systems. There is nothing nefarious about allowing only the accused to appeal. Courts generally permit only defendants — not the prosecution or the complainant — to appeal the verdict in a criminal case.


No comments: