How did Facebook try to manipulate my emotions? I will find out.
In January 2012, Facebook manipulated the news feeds of hundreds of thousands of people, actively attempting to influence their emotional state. I am trying to find out through legal efforts whether Facebook was giving me the positive (happy!) or negative (sad!) treatment. Below is the (long) letter I just sent to their Chief Privacy Officer.
Dear Ms Egan,
I am writing to you in your quality of Facebook Chief Privacy Officer.
I seek with my email to get access to some of my personal data, as envisioned by the US/CH Safe Harbor program, and more specifically personal data that relates to the Facebook Emotional Manipulation Experiment, conducted on January 11th-18th 2012. I trust that you are familiar with the experiment. Some of its results have been published in collaboration with Cornell researchers, in 2014. I quote here a selection of sentences from the article eventually published in the Proceedings of the National Academy of Sciences (PNAS), at http://www.pnas.org/content/111/24/8788.full.html
PNAS 01: In an experiment with people who use Facebook, we test whether emotional contagion occurs outside of in-person interaction between individuals by reducing the amount of emotional content in the News Feed.
PNAS 02: When positive expressions were reduced, people produced fewer positive posts and more negative posts; when negative expressions were reduced, the opposite pattern occurred.
PNAS 03: These results indicate that emotions expressed by others on Facebook influence our own emotions, constituting experimental evidence for massive-scale contagion via social networks.
PNAS 04: The experiments took place for 1 wk (January 11–18, 2012).
PNAS 05: Participants were randomly selected based on their User ID, resulting in a total of ∼155,000 participants per condition who posted at least one status update during the experimental period.
PNAS 06: The experiment manipulated the extent to which people (N = 689,003) were exposed to emotional expressions in their News Feed.
PNAS 07: Two parallel experiments were conducted for positive and negative emotion: One in which exposure to friends’ positive emotional content in their News Feed was reduced, and one in which exposure to negative emotional content in their News Feed was reduced.
PNAS 08: In these conditions, when a person loaded their News Feed, posts that contained emotional content of the relevant emotional valence, each emotional post had between a 10% and 90% chance (based on their User ID) of being omitted from their News Feed for that specific viewing.
PNAS 09: Both experiments had a control condition, in which a similar proportion of posts in their News Feed were omitted entirely at random (i.e., without respect to emotional content).
All told, this much is clear to me: Facebook conducted two experiments (one in the positive direction, one in the negative direction). Each experiment needed both a control group and an experimental group. This makes for four groups in total. Each participant in each experimental group was itself affected to a different degree: between 10% and 90% of posts with the relevant valence were withheld. PNAS 05 and 06 taken together make it clear that Facebook based its eventual analysis on having 155k people in each of those four groups, but introduces an additional condition: that the individual posted during that week.
It seems to me, but it is not completely clear, that Facebook would have needed to alter the newsfeeds of way more people than just those 600k, since it didn’t know when it started which would post during the week of the experiment.
I myself have not posted on Facebook during that week. I deduce from this that my profile was not part of the PNAS experimental analysis, but cannot exclude that:
- my newsfeed would still have been manipulated, with the intent of testing emotional manipulation;
- in anticipation that I would possibly post during that week and be included in the rest of the experimental protocol, I would have been assigned to one of two potential experiments, to one of two groups (experimental vs control), and if in the experimental group would have given one of several possible treatments (somewhere between 10% and 90%).
As part of Facebook’s obligations regarding personal data under the US/Swiss Safe Harbor program, I seek to know:
— whether my profile was eventually included in any of the categories documented in the research paper (I suspect not);
— whether my newsfeed was manipulated, later to be discarded as part of the experimental setup since I didn’t post during that period;
— whether I was part of the positive or the negative experiment;
— whether I was part of an experimental or control group;
— if part of an experimental group, what was the chance that posts would be withheld from my newsfeed.
I have obtained your email address through Facebook’s Safe Harbor registration on the FTC website, at
I couldn’t find appropriate links on Facebook’s data policy page, while TrustE arbitration procedures require that I make a direct attempt to contact Facebook before initiating legal action.
Facebook short name: paulolivier.dehaye
If you posted on Facebook during the week of January 11th-18th 2012, or are simply curious about your own treatment in this experiment, there might be easy ways for you to ask for the same information. Let’s talk. Here or here.
On April 26th 2016, I got the following response from Alex from Privacy Operations at Facebook:
Thank you for taking the time to contact us.
Facebook utilises the information we receive from our users to improve the services and features we are able to provide on our platform. From time to time we may also use this information to help improve our internal operations including troubleshooting, data analysis, testing, research and service improvement. These research initiatives not only allows us to provide Facebook as it exists today, but will also allow us to provide users with innovative features and services in the future.
Our research is carried out in consistence with our Data Policy https://www.facebook.com/policy.php
To learn more about the research conducted at Facebook please visit our dedicated website at https://research.facebook.com
I dutifully responded:
Dear Alex from Facebook Privacy Operations,
Thank you for taking the time to contact me, in response to my original request 17 days ago.
Unfortunately, your response completely fails to address the pointed questions I had concerning my Facebook personal data in the context of the Facebook Emotional Manipulation Experiment. I therefore invite you to reread my original request and reassess your response.
Update (June 29th 2016)
I have not received further response from Facebook, so I have initiated a TRUSTe dispute resolution with Facebook. On June 24th, TRUSTe has relayed my questions to Facebook.