Unknowing test subjects: The ethics of social media sites conducting experiments on users

It has recently come to light that social media companies including, Facebook and the popular dating website OKCupid, have been conducting experiments on users of the platforms.  In what Facebook calls a “massive experiment” they tweaked the news feeds of almost 700,000 users. They deliberately manipulated users’ news feeds so that some users saw more depressing statuses (determined based on keywords) and other users saw more happy statues.They found that emotions are contagious and that people tend to experience similar emotions as those of their news feeds. In a separate incident OKCupid lied to some users and told them they were a good match with someone when in fact their algorithms determined they would be a bad match. They found that being told you were a good match did lead to more interactions between users. In both cases the users were unaware that the information they were given had been deliberately manipulated for the purpose of the experiment, nor were they informed even after the fact. Additionally, they were not given the option to opt-out of such experiments. While the Facebook researchers apologized for how the experiment was described (i.e. did a little damage control), the CEO of OKCupid was very unapologetic (see image above).

Technically this is not illegal, but as I always tell my students, “just because you can do something, doesn’t mean you should.” In this case, I feel the experiments were not ethically conducted. Let’s consider other institutes that conduct research on humans. According to the National Research Act, any federally funded research must adhere to basic standards of ethics when it comes to experiments involving human subjects. Likewise, virtually all universities require that human subject research meets ethical standards set by Institutional Review Boards. And finally, many reputable journals require scholars to follow proper ethical procedures prior to publishing the work. These ethics ensure that human subjects are properly informed and can consent to the research. It also covers issues of confidentiality and anonymity. The review boards acknowledge that informed consent could sometimes alter results and thus they do provide a way for experiments to be conducted without explicit consent. However the stipulation is that after the conclusion of the experiment the researcher must still inform and debrief those involved. This was not the case with OKCupid or Facebook- there was no consent before or after the experiments.

Some are defending the experiments, arguing that you agree to the terms and services when you sign up for each platform. Others are pointing out that these services track data and run tests all the time. However, there is a big difference in anonymized data gathering that can be used to make the sites better, track trends, etc. versus this kind of blatant manipulation of content intended to affect users’ emotions and trust. As a user of a site I know they are tracking my interactions and I agree to that. However, it’s because I trust that the site will use my information in ways that do not harm me. I assume the tests are to better predict user preferences and to make the site better.  Additionally, when it comes to tracking, behavioral targeting, and third-party advertising the sites do give me some options and control over what is collected, how it is used, and who it is shared with – if I am uncomfortable with it, I may choose not to use the site at all. This was not the case with the experiments – users were not informed nor were they given opt-out options.

Now to a certain extent I think some of the outrage was disproportionate. Reading people’s comments I discovered that a lot of people falsely assume Facebook doesn’t already manipulate our news feeds. That simply isn’t the case, they manipulate our news feed all the time. The posts I see are not randomly generated nor are they even the most recent posts (depending on your settings). Rather than being chronological, Facebook determines which posts to show you based on complicated algorithms based on your interactions, your network, interests, trends, advertisers, etc. The placement of ads is already manipulation because I have not chosen to add the companies to my network, yet they appear in my news feed. While I find this annoying at times, I am aware of it and do not feel it harms me. In fact, it can be beneficial when I see an older post that is actually important or relevant to me (e.g. a friend made a big announcement and it shows up at the top of my news feed because so many of my other friends have interacted with that post). However, deliberately altering the mood of my feed could have harmful effects – for example, making a depressed person feel more depressed. This is something most ethical review boards would be reluctant to approve without serious stipulations and protection mechanisms in place.

Likewise, I understand that OKCupid manipulates users’ exposure to profiles. However, I’m assuming most users trust that OKCupid really is trying to match them with the right people. I have less of a problem with the OKCupid experiment because I can see why it might be important for them to understand how effective their rating system is and what effect it has on users’ interactions. But again, this comes down to consent and control. These kinds of overt manipulations should not be done without user consent. And the disregard the CEO has for people’s uneasiness and outrage speaks to larger issues of privacy, trust, and control.

What upsets me the most though is that these companies are not regulated in the same way that scientists and academics are (you know, the individuals that are actually highly trained to design and conduct experiments and to interpret the data and findings in ethical ways). While I do get frustrated with university review boards at times, I believe they serve a very necessary purpose that protects me as a researcher, the people I research, as well as the university. It has always bothered me that marketers and journalists are afforded greater opportunities to speak with youth than researchers. I find this particularly problematic because they are more likely to have exploitative goals than researchers. I know this is not always the case, but it is likely. For example, marketers conduct research for the sole purpose of selling a product versus bettering young people’s lives. However, as of now social media companies are protected and afforded the ability to conduct actual experiments on users without consent. This just doesn’t seem right. Again, there is a difference in observing, collecting, tracking, and analyzing (often anonymous) data versus deliberate manipulation of people’s emotions and moods.

Good research benefits society, but it should never be at the cost of people’s rights to consent, privacy, and control. I hope the government (some have suggested the FTC) or another governing body consider the implications of these recent experiments. While neither probably caused any long term consequences (I hope!), it does bring to light the need for ethical guidelines. As more and more private companies collect data about us (which is growing exponentially all the time), the potential for harm, misuse, and abuse is a real threat to the integrity and rights of citizens. Private companies have always and will always construct us solely as consumers, but we need basic ethical guidelines that protect us first and foremost as citizens.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s