How to Make Social Experiments Work for You (without Getting in Trouble)

0 Comments
Post Pic

Recently, Facebook raised some ethical concerns when it published its ‘emotional contagion‘ experiments. Facebook reported how it had manipulated the emotions of its platform users by skewing the news feed postings. Users shown predominantly positive or negative posts were then more likely to post positive or negative posts themselves, respectively.

While Facebook’s reaction to the social experimentation backlash was to reaffirm its commitment to user privacy, OKCupid decided to forgo any apologies for its social experiments. In a rather smug blog post, OKCupid’s cofounder Christian Rudder announced, “guess what, everybody: if you use the Internet, you’re the subject of hundreds of experiments at any given time, on every site. That’s how websites work.”

Granted, marketers take advantage of social experimentation every time they create an A/B test or analyze their site analytics. To a large extent, user implied consent is assumed- Facebook’s data use policy does state that “we may use the information we receive about you … for internal operations, including troubleshooting, data analysis, testing, research and service improvement.”

If this is the case, why was the recent research by Facebook and OKCupid deemed borderline unethical? And how can you run social experiments on your own affiliate marketing sites without being penalized for breaching privacy or informed consent rules?

1. Don’t manipulate emotions.

There are federal regulations and even human rights declarations in place that limit how much research subjects can be distressed or otherwise emotionally manipulated during the course of a psychological or social experiment. However, even if the long arm of national or international law doesn’t nab you, there is your audience’s reaction to consider if word ever leaks out about your experiments.

For example, consider what might happen if you purposely tout one of your least popular products as being a crowd favorite. You could write a great review of this unpopular product and even sing its praises in your email newsletter. This is akin to what OKCupid did when it tested the power of suggestion on its site and managed to match up bad date matches by touting how compatible they were with each other.

As OKCupid later reported and as you might even witness on your own site, people are swayed by “experts” telling them that something or someone is ideally suited to them. However, is conducting such a study ethical? And does conducting such a consumer psychology study inspire trust from your site’s users and subscribers? No.

2. Obtain informed consent.

Various marketing agencies recruit test subjects to study the effects of product color, scent, size, etc. on consumer purchasing decisions. The firms don’t always disclose the methods used to gauge participants’ reactions, or even what the objective of a particular study might be. However, what these firms do actually do, in order to cover their own legal behinds, is obtain informed consent from the test subjects first.

You can advertise that you’re conducting a consumer research study of your own and ask for recruits or volunteers from your audience. This is a great way to encourage audience participation and generate some brand recognition through customer testimonials and reviews. Your audience is also more likely to become involved with you and your affiliate products as a result of this study.

However, before you start recruiting people from your audience and organizing them into focus groups, be sure to have them consent to your study first. In this consent form, you need not outline everything you intend to measure, but you should at least mention that consumer behavior and consumer psychology will be analyzed. Also, if there are any physical or psychological risks inherent to the study, they need to be outlined. That way, should some kind of outcry erupt, you are legally covered.

3. Minimize harm.

Psychological studies are approved by institutional review boards all the time, provided they demonstrate that the value of the knowledge being gained is greater than the harm generated to test subjects. In the case of affiliate marketing, that harm is often monetary in nature.

Thus, if you are about to conduct a social experiment on your audience, you can minimize harm by reimbursing audience members for their time and effort or any purchases they eventually make. Alternately, you could make the study as theoretical as possible- for example, by conducting a survey of consumer purchasing decisions, you are less likely to induce your audience to spend money on affiliate products that they would normally not purchase.

4. Reveal your results- strategically.

If your audience is aware of your social experiment, they will eventually want to see the study results. To this end, you can release the information you have learned; however, it is well advised that you keep any far-reaching conclusions to yourself and publish only the base results. Don’t extrapolate or predict. And it goes (almost) without saying that you should never belittle individuals for making the choices they did.

You might also consider releasing some study results and keeping others to yourself. This way, you satisfy audience curiosity without insulting anyone or making people feel uncomfortable. Those unpublished results can be used to set up future consumer psychology experiments- or not.

Published results can go a long way towards generating additional audience participation and goodwill, so don’t hesitate to announce that you’re about to disclose a case study or some other social experiment. After all, it’s a psychological fact that people love to read about themselves. Just be sure to use tact when doing so.

Is ethical social experimentation possible?

Within the marketing arena, social experimentation is a necessity because it is nearly impossible to correctly predict consumer behavior through rational observation or algorithms. To this end, social experimentation fills the gap between theory and reality. Provided the audience is at least somewhat “in the know” that it being observed and analyzed, social experiments are tolerated. The more you can involve your audience in the experiment via informed consent, you more information- and even enthusiasm- you can garner.

Photo credit by Horia Varlan

Leave a Reply