Tech companies increasingly are partnering with research institutions. These partnerships include sharing data and project collaboration. What IRBs will want to know as this trend continues is what it means from a human research protection perspective.
“We’re seeing these ongoing and growing connections between academic researchers and industry,” says Elizabeth A. Buchanan, PhD, endowed chair in ethics and acting director of the office of research and sponsored programs at the University of Wisconsin-Stout in Menomonie.
“What’s happening is, more and more, research is sponsored by industry or a collaborative with industry,” Buchanan says. “We’re seeing these challenges, and we want to talk about these ethical and practical challenges that arise.”
Researchers can learn a great deal from social media data, says Brenda Curtis, PhD, MSPH, an IRB member and researcher, as well as an assistant professor in the department of psychiatry at Perelman School of Medicine at the University of Pennsylvania in Philadelphia.
Curtis has worked with Facebook and app developers on studies of populations of people who are seeking treatment for substance use disorder.
“I use language analysis to predict relapse,” Curtis says. “We’re using the social media language of people who consented to our study. There’s an opportunity for researchers to work with tech companies and big data companies like Facebook in order to help improve health,” Curtis says.
“More researchers are seeing the benefits of big data and they access data sets,” Buchanan says. “It’s changing how we do research, fundamentally, and IRBs need to know what to ask and how to think through a purchased data set.”
From a social-behavioral research perspective, researchers and tech companies share a curiosity and interest in learning more from social media data, including Facebook and Twitter.
The IRB’s role is to ask questions about the data set, including asking where it originated, Buchanan says.
Using tech companies’ big data can lead to important research. Buchanan has seen how social media data can better inform research into illegal drug use in urban areas. But it’s important for all investigators to outline to the IRB how they’ll protect data subjects’ privacy and safety.
“IRBs want to make sure we don’t have another Facebook emotional contagion problem,” she says. “So, we should review those projects.”
In 2014, Facebook and Cornell University of Ithaca, NY, were criticized for conducting a study not reviewed by an IRB. Called “Experimental Evidence of Massive-scale Emotional Contagion through Social Networks,” the study was published online June 2, 2014, in Proceedings of the National Academy of Sciences.
Cornell issued a statement at that time about how the research was conducted independently by Facebook and a Cornell professor only had access to results, which was why the Cornell University IRB concluded it was not subject to review. (The Cornell University statement can be found at: http://bit.ly/2hAAMuN.)
Facebook had changed its content on news feeds for users, chosen randomly. It adjusted the balance of positive and negative posts, with one group seeing mostly positive posts and another seeing mostly negative. The study looked at whether the emotional content of news feeds affected the emotional content of users’ own status updates. It found that it did affect news feeds, reflecting what users saw. No informed consent was obtained. (See story about the Facebook study in the December 2014 issue of IRB Advisor.)
“What have we learned, and has Facebook changed any internal policies since that experience?” Buchanan asks. “I don’t know the answer yet. Collectively, we’ve moved so quickly into accepting large data, big data sets, and accepting where they come from, whether Twitter or an aggregator site.”
The study had involved manipulation of multiple things — not just the negative versus positive words — and the effect was small, Curtis says. But the case did create caution among tech companies, researchers, and IRBs, she notes.
“Researchers will be more cautious in how we explain studies when we work with big data and are doing testing,” she explains. “Facebook has set up some internal review boards and policies, and published a couple of articles on it.”
Research institutions and tech companies can have a variety of arrangements. These include an institutional researcher partnering with the tech company or contracting for access to certain data, Buchanan says.
Tech companies might seek this collaboration to add academic integrity to the research and to reach a broader audience, she says.
IRBs can ask the following questions:
- How were these data collected originally, and for what purpose?
- How is the information now being used?
- Are there provisions in place that prohibit people from attempting to reidentify data, including provisions in data use management and agreements?
- What are the origins of the data, and how you are using them?
- Are there secondary or tertiary privacy concerns that come with reuse of data?
“We’ve moved away from a basic understanding of privacy into a complex place as data are used [in] more and more contexts, providing more opportunities for privacy to be challenged,” Buchanan says.
One way for an IRB to stay on top of these changes is to provide educational sessions on the institution’s rules regarding the use of social media data or collaborating with a tech company, Buchanan suggests.
“Make sure researchers, and maybe IRBs, too, are familiar with a particular social media platform,” she adds.
If they learn about Facebook, Twitter, Instagram, etc., they might come up with better ethical questions to ask before jumping into a research collaboration.
“And we’ve had interesting questions about the providence of data,” she adds. “For instance, if a data set is breached and a researcher has access to it, is it okay to use breached data, like data from WikiLeaks? And lots and lots of researchers are using those data.”
When data are from a dubious source or an unethical source, what does that mean for the scientific enterprise that plans to use the information?
Those types of questions drill down to the fundamental integrity of research and are more for an ethicist or philosopher to ponder than for an IRB to ask, she adds. “Just because data exist, are they there to be studied?” she asks.