Healthy Minds Study survey data informed 2016 Senior House decisions
Unbeknownst to participants, residence was associated with responses concerning drug use and mental health
The Healthy Minds Study (HMS) survey, administered to students in 2015, linked each participant’s responses to their dorm.
The survey included questions about illegal drug use and mental health. However, survey participants were not informed that their residence information would be linked to their responses.
The chancellor confirmed to The Tech in an email last night that the 2015 HMS survey and the 2013 National College Health Assessment “served to reinforce” concerns of illegal drug use in Senior House she announced in 2016. (See the surveys’ respective results here and here.)
This was the first time the chancellor confirmed that the HMS survey was used in decisions regarding Senior House, despite several opportunities to clarify the matter in 2016 and periodically since.
In her letter to the editor in this issue of The Tech, the chancellor writes that the HMS was not used in any decisions made “this year.” The Senior House turnaround announcement was made in 2016.
Concerns of illegal drug use had also stemmed from “information provided by faculty, staff, and students,” according to minutes sent to The Tech yesterday of a June 14, 2016 meeting between Senior House students and administrators. Emails about the minutes showed that the minutes had been written with input from the chancellor.
When the study was conducted in 2015, participants were presented with an informed consent section which was supposed to state exactly what information would be linked to the survey responses.
“Informed consent is designed to ensure study subjects understand what they are being asked to do,” according to MIT’s COUHES — Committee on the Use of Humans as Experimental Subjects.
MIT sent an excel file to HMS researchers to be linked with survey responses, Lydia Snover, Director of Institutional Research wrote to The Tech Tuesday. This file included: first name, last name, email address, whether the student was an undergraduate or graduate student, department, school, residence, gender, ethnicity, and birth year.
The first name, last name, and email address of each student was stripped from the file the HMS team at University of Michigan then returned to MIT, in order to preserve confidentiality.
“We strip any identifying information (e.g., name, email) and replace it with a non-identifying study ID (alphanumeric string),” Daniel Eisenberg, principal investigator for the survey at the University of Michigan, wrote to The Tech last Monday.
The contents of the excel file MIT sent to the University of Michigan differed somewhat from the variables listed in the consent form that students signed.
That consent section student received listed: name, email address, date of birth, gender, race/ethnicity, citizenship, degree program, year in program, and grade point average.
While GPA, date of birth, and year in program were listed as variables, MIT did not provide the HMS researchers with this information.
“We can infer citizenship and degree program from other variables provided,” Snover wrote in Tuesday’s email.
This means four additional variables: undergraduate or graduate status, department, school, and residence, were added to be linked to students’ responses, but not present on the consent form.
Elizabeth Glaser, a social scientist and parent of an MIT student who took the survey, concerned that MIT had data about her son he hadn’t consented to, reached out to Eisenberg.
“[O]ur protocol does allow for linking survey data to data elements provided by the institution (such as age and gender),” Eisenberg wrote. These data elements are then stated in the informed consent.
“We failed to customize that list of data elements in the informed consent to reflect the list used by MIT,” Eisenberg wrote in a July 14 correspondence to Glaser that Glaser then forwarded to The Tech.
“MIT could have used residence data, given that it was linked to the survey data. I don't know for sure if and how they actually did that analysis,” Eisenberg wrote in a July 17 email to The Tech.
MIT’s IR office received the data file containing survey results from the University of Michigan July 2015, Snover wrote to The Tech yesterday.
Snover previously wrote in Tuesday’s email that “Prior to knowing the error in the consent form, the Chancellor’s Office did conduct some limited analysis but they did not publish any data based on residence hall.”
“Once MIT learned of the error in the consent form, no offices conducted any analysis of the HMS data based on dorm/residence,” Snover added.
“When MIT contacted the HMS researchers about participating in the study, MIT requested that the survey be customized to ask survey respondents to identify their on campus address,” Snover wrote in the same email.
However, HMS researchers “declined to include a customized question regarding address, but informed MIT that it could provide the residence hall data, along with all other administrative data, in the data file that MIT sent to HMS researchers for the survey invitees.”
According to Eisenberg, the University of Michigan’s Institutional Review Board (IRB) found out about this discrepancy during the summer 2016 and notified MIT. The HMS team has since taken precautions to ensure this type of error does not occur in the future.
Asked if it is unethical to continue to use this data, Eisenberg wrote that he “would agree that it's not appropriate to continue using the data that wasn't mentioned in the consent” and said that once his team realized “the error, [they] removed the information from [their] copy of the data set.”
“I do not believe that the problem with the consent form invalidates the survey or rises to the level of any ethical breach as the error appears to be unintentional,” Leigh Firn, the COUHES chairman, wrote to The Tech Tuesday.
“MIT refrained from doing anything with the data once it discovered the problem; no identifying information was provided in connection with the survey responses; and most importantly, the individual survey responses have been treated as confidential and remain so,” Firn wrote.
Many students responded to this error in the informed consent with a reluctance to take MIT surveys in the future.
“Looks like they lied and it would be bad practice/bad character to use the data,” Travis Libsack ’18 wrote in an email to The Tech Tuesday. “Senior House 2k16 all over again.”
“I am, quite frankly, shocked and extremely angry that this happened, especially at MIT. The lack of informed consent on the part of students taking the survey, as well as the unethical way the results were used, is abhorrent,” said Madeleine Haworth ’19 in an email to The Tech Wednesday.
“To my knowledge MIT has made no official comment or apology related to this matter — which is, to me, a clear sign that research integrity isn't a priority when the administration wants to use questionably-obtained ‘data’ against students,” Haworth added.
“Suffice it to say that I won't be taking part in many Institute surveys or studies in the future if I can avoid it,” Haworth wrote.
Madison Evans ’18 wrote to The Tech Wednesday that she wasn’t sure if she would have thought that her survey results being linked to her house would have “been as concerning to [her] at the time.”
“But I do know in the future, I'd like to read those forms for future surveys. I may forget, like a lot of forms. But it's important to know what's happening with the information I give out,” Evans wrote.