The IRB at the University of Illinois at Urbana-Champaign had a number of issues to consider when researchers submitted a study involving a social network analysis on sensitive issues. In this case, the sensitive issue was sexual assault on a college campus.
Social network analysis is new and growing in popularity, says Anita Balgopal, PhD, director of the university’s Office for Protection of Research Subjects (OPRS).
“Our board was mindful of keeping every aspect of this research in check,” Balgopal says. “It also was dealing with a highly sensitive topic.”
The research findings also have potential to inform the university’s administration about sexual assault on the campus, she adds.
“Everything about the study was treated with the utmost respect to participants, confidentiality of identifiable data and privacy of participation,” Balgopal adds.
“Sexual assault on college campuses is a hot topic right now,” says Emily Dworkin, MA, a doctoral candidate and co-investigator of the study, which is being led by associate professor Nicole Allen, PhD. Dworkin is a research assistant at the University of Illinois in Urbana-Champaign.
Dworkin’s and Allen’s study addresses gaps in research through the exploration of social networks with university sororities as contexts for help-seeking after sexual assault. Investigators examined the structure of friend networks within these sororities, the knowledge they contain about available resources, and the presence of social normal like rape myths as they recover from rape.1
“People experience the aftermath of sexual assault fundamentally in a social context,” Dworkin says. “They may disclose it to friends and family members, or they may choose to not disclose but just get information.”
Whether they disclose and the kind of responses they get could affect their later decisions and choices, so researchers wanted to find out more about characteristics of the social context and what influenced victims, Dworkin explains.
The two issues they are trying to understand are whether a victim seeks help and her ultimate sense of well-being. “Did the social contexts impact mental health?” Dworkin says. “We assessed the structure of the social contexts and the information they contained to understand these issues.”
The IRB had several ethical points to consider, Balgopal notes.
They included the following:
1. How would they enroll participants? Researchers planned to approach sororities and speak with their members, but the IRB questioned the feasibility of this. “We were not aware of how they were able to obtain a list of sorority members,” Balgopal says.
Typically, sorority membership lists are not pubic information, and a gatekeeper — such as a sorority president — is in charge of access to this information, she says.
2. Would the recruitment incentive create ethical concerns? Researchers proposed giving each sorority house a $500 recruitment incentive that they could donate to charity or keep for the sorority’s use. To receive the financial incentive, they would have to meet an 85% participation rate in study enrollment, Balgopal notes.
“The IRB was concerned about how a sorority might get to 85% — it could be coercive,” she says.
For instance, a sorority president might pressure members to take part in the study so her house would receive the extra funds, she says.
3. How would recruitment take place? “How were these sorority members going to be notified about the research study?” Balgopal adds. “It was difficult because sometimes the sorority chapter president was the gatekeeper, and the IRB wanted to make the distinction that recruitment should be from members of the research team and not members of the sorority house.”
Plus, sorority leaders were not researchers involved in the study and they wouldn’t be able to answer potential participants’ questions, she adds.
4. How would consent work with potential secondary subjects? The study’s goal was to understand links between victims and their friends and social support circle. If a woman, who agreed to participate in the study and acknowledged that she was a victim of sexual assault, were to name several women as people with whom she shared this information, it was entirely possible that one or more of the friends had not consented to participate in the study, Balgopal says.
“The IRB was concerned about whether it was appropriate to out one of those three members. Perhaps the participant went to this individual as a friend, for support or consult, or perhaps this was someone who was vocal about being a victim of assault, or not — we wouldn’t know,” she says.
Also, the IRB was concerned about the kind of questions the participants would be asked.
The IRB worked with researchers to develop procedures that would address the IRB’s concerns and reduce risk to participants.
Ultimately, all ethical questions and concerns were addressed to the IRB’s satisfaction, Balgopal notes.
For example, researchers handled recruitment by notifying all sorority presidents and inviting them and their members to participate, while specifying that presidents must not be involved directly in recruitment, Dworkin says.
“We set up protocols to have middle men let sorority presidents know we would be contacting them about the study,” she says. “We gave them full transparency about the study.”
In network surveys, standard practice is to list the names of people who have not directly consented to be part of the list but who are listed as members of an organization. Requiring consent before a name is listed would be the same as publishing the names of people who were participants in the study, eliminating all confidentiality, according to a presentation on the subject by Balgopal, Dworkin, and colleagues.1
Secondary subjects are considered to be the people about whom survey questions are asked about their behavior. The sexual assault study asked questions only about participants’ own behavior and their relationships to the people listed in the network.1
Researchers are keeping the sorority incentive from being coercive by not communicating participation rates to the sororities and by discouraging sorority leaders from conducting their own recruitment. So if a sorority reaches 80% participation, the sorority’s leadership will not know that with one or two more women participating they could receive the incentive money, Dworkin says.
Data were de-identified. All sorority member names were replaced with ID numbers, and researchers do not access data until it is replaced with the numbers, she adds.
“So we couldn’t know who was assaulted and who wasn’t,” Dworkin says. “What we found with the social network research is that someone doesn’t become a secondary subject until someone reports on their behavior, so we felt that our methods didn’t raise secondary subject concerns.”
Dworkin attended dinner announcements for the sororities. These were venues where outsiders were asked to be speakers. She told sorority members about the project and how it was up to them whether they wanted to participate.
“We sent a link about the study to their email addresses, and they could opt out,” she says.
The women who chose to participate would have to read a two-page, electronic informed consent document and check the box showing they were adults and volunteering to participate, she adds.
Participants were directed to a survey on a secure platform. They were asked whether they had an unwanted sexual experience and about their relationship with other women in their sorority. Specific names were given, but only first names and last initials. Participants would indicate whether a particular person was a good friend or acquaintance, Dworkin explains.
As the sensitive nature of the study required researchers to work closely with the IRB, the two groups decided to publish a case study on the process of social network analysis on sensitive issues. The collaboration was highly informative, Dworkin says.
“We’ve been thrilled with the process — given the care and concern and collaboration and having the IRB meet with us and discuss issues in a thoughtful way,” she says.
- Dworkin E, Balgopal A, Banks R, & Allen N. Social network analysis on sensitive issues: A human subjects research case study. Presented at the PRIM&R Advancing Ethical Research Conference, held Dec. 5-7, 2014, in Baltimore, MD.