SOURCE ALERT: The questionable ethics (and conclusions) of Facebook’s emotion manipulation study

July 1, 2014

Experts can comment on the challenges of doing research in the digital age, as well as raising questions about the conclusions of Facebook’s study.

Contact: Andrew Good at (213) 740-8606 or gooda@usc.edu.

Conclusions go “way beyond the data”

Karen North, Director of the USC Annenberg School’s Program on Online Communities, is an expert in social media, digital privacy, and psychological research. She said that the conclusions of Facebook’s study go “way beyond the data. When they said people’s emotions changed – there’s no evidence whatsoever of that.”

The study, she said, found evidence of behavior changing – but that presupposes knowledge of the sample group’s emotional state. Users seeing positive posts may be influenced to post more positive things themselves, but that does not mean Facebook knew their emotional state at the time, North said.

North, who was involved in creating the the National Bioethics Advisory Commission under the Clinton Administration, said the study was also questionable from an ethical perspective. “The chances of a negative effect are pretty small, but if you’re doing a random sample, you don’t want to pull in someone vulnerable. You don’t know if there’s already someone suicidal or schizophrenic in that group.”

Contact: (310) 650-5689 or knorth@usc.edu.

Is privacy a polite fiction?

Mark Marino, Associate Professor in the Teaching of Writing at the USC Dornsife College of Letters, Arts & Sciences, is an expert in new media, online storytelling and netprov. He developed an online project spoofing terms of service agreements that readers could click through. Most didn’t read the fine print explaining they were agreeing to sell their soul.

“It’s interesting to see people posting their outrage with Facebook through their Blogger and Twitter feeds. The cognitive dissonance between the excessive claims of the terms of service documents and users’ sense of their inherent rights to Internet privacy is startling. What we’ve learned from the NSA and this recent news is simple: if symbolic exchange occurs over an electronic network, it’s in the public sphere. Any expectation of privacy is at best a polite fiction.”

Contact: (213) 740-1980 or markcmarino@gmail.com.

Was Facebook’s terms of service enough for informed consent?

Alex Capron, Distinguished Professor of Law and Medicine at the USC Gould School of Law, is an expert on bioethics and informed consent. He can discuss the issue of informed consent when conducting research and having a transparent terms of service agreement.

“I think the potential to harm at least some of the 600,000 people included in the experiment cannot be dismissed. When such risk exists, and when a ‘private’ space is being entered, informed consent is the usual requirement.”

Contact: (213) 740-2557 or acapron@law.usc.edu.

What are the ground rules for conducting research when using private data?

Jennifer Swift, Associate Professor & Director of Graduate Studies at USC’s Spatial Sciences Institute, is an expert in geospatial information sciences (GIS), app development and data modeling. She can discuss the challenges and potential pitfalls of conducting research using sensitive, private data in the digital age.

“In terms of general ground rules for researchers at accredited academic institutions considering studies using private or sensitive data, an Institutional Review Board (IRB) is available to evaluate any research proposal involving human subjects, to make sure that all human subject research is conducted in accordance with all federal, institutional, and ethical guidelines.”

Contact: (213) 740-5841 or jswift@dornsife.usc.edu.