Dismissed

Clark University officially kicked a graduate student out of her program last week, saying she had failed to find a new adviser within an initial, 30-day limit. But the student, Abby Nissenbaum, says the real story is that her department blackballed her after she complained about her first adviser’s research methods and behavior.

The university denies that. But Nissenbaum said Wednesday, "This whole thing is weird and in total violation of prescribed policies."

Nissenbaum’s case has attracted attention on social media, where she’s shared details about it. The primary concern among followers, including many faculty members, is that a student who is possibly a whistle-blower has not been afforded any special protections during the complaint process.

What happened?

Nissenbaum, who is interested in sexual violence interventions, started her graduate studies in Clark’s social psychology program in 2017. As sometimes happens, tensions soon emerged between her and her adviser, Andrew Stewart, an assistant professor. Nissenbaum believed that Stewart’s work was becoming increasingly qualitative, while she preferred quantitative work. And Stewart questioned the compatibility of their research agendas, writing in an email to Nissenbaum that she wasn’t interested in bystander outcomes -- even though her master’s thesis was on that subject. Their meetings together were infrequent, at best, Nissenbaum recalled. And by the end of the academic year, in mid-2018, the department advised Nissenbaum to better align her research with Stewart’s.

Stewart allegedly confided in Nissenbaum that he was having personal issues, and his overall attitude grew increasingly sour, she says. For instance, he accused her of being too “defensive” during informal lunch meetings to discuss research. Nissenbaum says the comment was gendered, in that it’s not a criticism that would be levied at a man.

Questions of Research Misconduct

In October 2018, in her second year of the program, Nissenbaum told Stewart that she was uncomfortable with how he was treating data from a study on sexual orientation and workplace harassment. Subjects were asked to rate their orientation on a scale of one to five, heterosexual to homosexual. The first time they ran the data, Nissenbaum says, she and Stewart considered one to be heterosexual, two to four to be bisexual, and five to be homosexual. But that analysis found no significant connection between sexual orientation and harassment, Nissenbaum says. So Stewart allegedly told her to re-run the analysis, with one to two being heterosexual, three being bisexual and four to five being homosexual. And this time, Nissenbaum says, the results were significant.

After attending a conference on open science, where she learned about p-hacking, or manipulating statistical significance, Nissenbaum asked Stewart if he thought what they’d done was unethical. That’s because study participants were told to select two or four if they weren’t totally straight or gay, respectively, reflecting the way she first ran the data, she says. 

Stewart allegedly told Nissenbaum that she was “intellectually underdeveloped.” She declined to continue working on the study.

At their meeting a month later, Stewart allegedly told Nissenbaum that she hadn’t grown intellectually since she entered her program. She says she asked for specific examples and an action plan to address the lack of growth but did not get either.

As a result of the “hostility,” Nissenbaum says, she applied to other graduate programs elsewhere that fall. She asked Stewart for a positive recommendation, but she says that a principal investigator at another institution she applied to tipped her off to his negative recommendation. Nissenbaum did not have a copy of the reference letter but included in her appeal to Clark an email from the unnamed PI. It recommends, delicately, given the confidentiality surrounding letter writing, that she seek another letter writer.

Nissenbaum did not get accepted elsewhere. So she resigned herself to finishing her Ph.D. at Clark, where at least she was guaranteed four years of funding.

Program Dismissal

In April, however, Stewart sent Nissenbaum emails saying that they had decided together that she would not continue on at Clark. Stewart also said he would not supervise Nissenbaum any longer as a graduate student, and that she had 30 days, as per university policy, to find another adviser.

During a meeting with the program chair, Johanna Vollhardt, Nissenbaum says she was asked to remove comments about the situation from her Twitter feed, as they made it harder for Vollhardt to help her.

In a separate email to Nissenbaum, Vollhardt that she was sorry Stewart's decision not to keep her on "hit out of the blue, I had assumed [Stewart] already discussed this with you. This was not a program decision behind your back, but rather a reflection of what I thought was a shared understanding that you both didn’t want to continue working with each other and that the program is not a good fit for you." She added, "I can assure you that none of this is vengeance for disloyalty by applying to other programs though! It seems there were several issues regarding fit and ability to work together that accumulated."

In her subsequent second-year portfolio review, Nissenbaum was faulted for not having made progress toward a first-author publication. But that’s in part because she refused to put her name on the first paper on sexual orientation, she says, and she was planning to submit her master’s thesis involving four studies with hundreds of participants instead. Additionally, she says, that criterion is applied unevenly across Ph.D. candidates.

In the next few weeks, Nissenbaum tried to find another adviser, including Vollhardt, but was unable to do so. One professor even told her via email that he was “reluctant to get involved in what I gather has been a difficult and conflictual process.” And by late May, she received an email from the department saying she’d been dismissed from the program.

Nissenbaum notes that Clark policies allow for a graduate student to leave a lab voluntarily, but not for an adviser to unilaterally dismiss a student. And even when a student receives an unsatisfactory review, she says, university policies call for a remediation process, in which the student is put on probation. Nissenbaum was never put on probation.

As for her earlier tweets about being subjected to gender discrimination -- the ones Vollhardt mentioned -- Nissenbaum wrote in her appeal to Clark that she has "always been outspoken about discrimination and harassment, and a vocal supporter of the #MeToo movement." And when Stewart said he wouldn’t work with her, she continued, “I objected publicly to the gender-based poor treatment I had received as his advisee.”

If those tweets played into her dismissal, she says, then that is retaliation.

Final Word

Even though Nissenbaum had only 30 days to find a new adviser, the university took longer than that month to adjudicate her ultimate appeal of dismissal. The final expulsion letter, written by Yuko Aoyama, associate provost and dean of research and graduate studies, and dated Oct. 24, says that the research misconduct claim is not valid. According to additional information from the university, Nissenbaum's sexual harassment claim is still in the initial assessment stage.

“Please rest assured that the department of psychology and university have taken your case very seriously and spent considerable time carefully evaluating the details of your claims through departmental grievance and university appeal processes. Taken as a whole, the evidence you provided does not support your many claims,” the letter said.

Stewart said this week that he could not comment on specific allegations that would violate federal laws governing student privacy. But in a lengthy email, he said that what he did with the sexuality and harassment study was not data falsification.

Stewart recalled the coding scheme somewhat differently than Nissenbaum, with one being “completely homosexual” on the scale, and so forth. In any case, the first analysis demonstrated that only bisexual participants reported statistically significantly different levels of heterosexist harassment than heterosexual participants, he said. But at a weekly program meeting, someone suggested that the “robustness” of findings could be tested by recoding the sexual orientation scale, so that “completely” or “mostly” heterosexual would be a category, along with bisexual and “mostly” or “completely” homosexual. 

He took up the suggestion, he said, as it’s “useful” in psychological research to “identify whether the findings are robust to changes in measurement.”  And in any manuscript submitted for publication, Stewart added, “we would report tests for robustness to evaluate whether the results (and consequent conclusions) would change if the variables were measured differently.” 

In the new analysis, homosexual participants reported statistically greater levels of harassment than did heterosexual participants, but the effect was “very small," Stewart said. And the greater theoretical conclusions remained the same as before.

Clark said in a statement that it takes any allegation of research misconduct “extremely seriously” and that it “respects federal law, which protects student privacy rights, and thus the university cannot and will not comment upon specific details of this student’s case.”

Clark can, however, “affirm that the department reviewed the individual’s appeal of the dismissal decision. Then, upon the individual’s appeal of that decision, the university conducted its own review.” Both “thorough reviews” concluded that there is “no basis for this individual’s allegations, that the claim that there was research misconduct is without merit, and that the dismissal was supported by the facts and appropriately followed department and university policies and procedures.”

Who's at Fault?

Aoyama's letter also blames Nissenbaum for her failure to secure a new adviser.

“The role of the faculty adviser is absolutely fundamental to doctoral education,” the letter says. “In a Ph.D. program, the faculty adviser serves not only as a mentor, but also as a research collaborator; as such, the relationship may be terminated by either party if it is found that collaborations are not functioning or if research interests diverge.” And when things became “untenable” with Stewart, Aoyama wrote, “it was your obligation to identify other faculty whose research interests aligned with your own.” The letter notes that Nissenbaum was given a three-week extension, or 21 more days beyond the initial 30, to find a new adviser. 

Nissenbaum said she was “shocked” by the Aoyama’s letter, as she didn’t necessarily have faith in the department, but she did trust that the university would follow its own policies. She called it an instance of “corruption” that leaves her without funding or a place to study, nearing the middle of what would have been her third year in graduate school. If Clark can’t properly mentor the students it commits to, she said, it probably shouldn’t be a research university.

Steven Piantadosi, an assistant professor of psychology at the University of California, Berkeley, who advocated for graduate students involved in a harassment case at his former institution, the University of Rochester, said that it seems “unusual for a student to be unable to find a new adviser, especially just a second one.” Many programs even require multiple possible advisers to agree before admitting an applicant, "just in case research with one adviser doesn't work out."

The short timeline for finding a new adviser is also unusual, Piantadosi said. “Good faculty should jump at the chance to work with someone who has actually stood up for scientific integrity and risked -- as you can see -- their own career by making a good-faith complaint” about a “serious issue” in the field.

He added, “If a program kicks someone out because one adviser doesn't work, that's a sign the program and faculty don't care at all about students.” And that should send a message to others “about whose careers the institution is there to protect.”