Facebook's Emotion Experiment: Too Far, Or Social Network Norm?

Facebook's recently disclosed 2012 experiment in altering the tone of what its users saw in their newsfeeds has brought it plenty of negative opinions to chew on. Here's one, pointed out by an anonymous reader: Facebook's methodology raises serious ethical questions. The team may have bent research standards too far, possibly overstepping criteria enshrined in federal law and human rights declarations. "If you are exposing people to something that causes changes in psychological status, that's experimentation," says James Grimmelmann, a professor of technology and the law at the University of Maryland. "This is the kind of thing that would require informed consent." For a very different take on the Facebook experiment, consider this defense of it from Tal Yarkoni, who thinks the criticism it's drawn is "misplaced": Given that Facebook has over half a billion users, it’s a foregone conclusion that every tiny change Facebook makes to the news feed or any other part of its websites induces a change in millions of people’s emotions. Yet nobody seems to complain about this much–presumably because, when you put it this way, it seems kind of silly to suggest that a company whose business model is predicated on getting its users to use its product more would do anything other than try to manipulate its users into, you know, using its product more. ... [H]aranguing Facebook and other companies like it for publicly disclosing scientifically interesting results of experiments that it is already constantly conducting anyway–and that are directly responsible for many of the positive aspects of the user experience–is not likely to accomplish anything useful. If anything, it’ll only ensure that, going forward, all of Facebook’s societally relevant experimental research is done in the dark, where nobody outside the company can ever find out–or complain–about it."

Comments

One solution

[Anonymous Coward]

Just don't use social networking.

Re:One solution

[flyneye]

But, think of the implications for upcoming elections! How will the Repubmocrats keep 100% power against independents, tea party and other radical despots competing against the chosen ones? Control! The people obviously need controlled, they don't know what is good for them and the Repubmocrats always will.
Since there is a market, Facebook, who covers most demographics, can help by raising tensions toward sinister interlopers in our one party political system.
Look for upcoming "hour of hate" shows profiling Tea Party advocate Emanuel Goldstein, NRA talking heads, Farmers, Anti-war hippies and anyone without the Hillary Clinton seal of approval on their forehead or hand.....

Re:One solution

[rmdingler]

Some people have little interest in happiness and a good mood... you see it every day.

Within and without social media, people, events, results, and happenstance conspire to alter your mood each and every day... something that cannot happen without your tacit permission. Grow a thicker skin and remember that yelling at that jerk in traffic means you've allowed a complete stranger power over your behavior.

If giving up social media is too big a first step, don't go in with your eyes wide shut: you are the product, not the customer.

Re:One solution

[Impy the Impiuos Imp]

Wow that was an impressive misspelling of "surgical".

Re:One solution

[BobMcD]

How will the Repubmocrats keep 100% power against independents, tea party and other radical despots competing against the chosen ones?

https://www.youtube.com/watch?...

It's called 'First Past the Post', and Facebook has nothing to do with it.

Re:One solution

[s.petry]

Mostly this, but I believe it could have been worded differently for clarity.

1. Like all institutions that are impacting on the public, Facebook and Twitter are being, and have been used, for propaganda. Guiding opinion happens from the time a kid enters Public school though to graduating from a University. We don't like to admit it, because it's frightening when you investigate the scope with which people are being indoctrinated. Reality can be a bummer, but it is reality.

2. Just like with Television, there is a massive amount of social engineering with online media and content. How many "News" programs want you to "follow us on Twitter" and "Like us on Facebook"? All of them, and you will hear and see that phrase repeated over and over.

3. Use of institutions like Education and Social media for purposes of "Social Engineering" is Propaganda in it's purest form and insidious. This is much worse than bread and circuses alone, because it constantly provides an opinion that someone want's you to have without any dialogue or discussion on the impact or morality of the opinion. I.E. Iraq war, Syria, candidates for public offices, etc...

4. The fact that propaganda wars are waged against citizens happen is not new. It's been documented back as far as I can recall. Certain people owned Newspapers and provide an opinion. If a counter opinion is provided, it's generally squashed by ad hominem and when that doesn't work it simply won't be discussed on any mass media. This has been transferred to Radio, TV, and now Social media. The scale at which it's currently done is massive compared to when I was a kid.

Investigating and learning about the psychology being used is like taking the red pill. A whole new world opens up to you, and you can see how much propaganda is being generated.

Re:One solution

[AthanasiusKircher]

How will the Repubmocrats keep 100% power against independents, tea party and other radical despots competing against the chosen ones?

It's called 'First Past the Post', and Facebook has nothing to do with it.

You're right that "first past the post" is a big part of the problem. But that's far from the whole story.

For one thing, "Repubmocrats" do NOT have 100% of the power. It may be near 100%, but it isn't 100% -- for example currently two U.S. senators, four major city mayors, and hundreds of state and local officials across the U.S. are elected independents or members of 3rd parties.

That means that quite a few voters across the U.S. have actual experience in ELECTING someone who is not a member of the two major parties. Those sorts of successes can make it more easier for some voters to justify taking a chance on a 3rd-party candidate.

Also, the Tea Party (for another example) has made huge inroads into mostly Republican territory in recent elections. One can argue that this has created huge problems for the traditional Republican leadership that disagrees with Tea Party ideology. In some elections, this has resulted in long-standing prominent members of the Republican party losing primaries and/or general elections.

The video you posted labels this the "spoiler effect" when it leads to, say, a Democrat being elected. But the very term "spoiler effect" is not a neutral term -- it's a propaganda term used by the 2 major parties to convince people to keep them in power.

We could, instead, call this the "disenfranchisement effect," where the two-party system focus on candidates A and B might disenfranchise voters who hate both and would prefer to vote for C. According to the spoiler effect logic (and your video), the C voters supposedly should learn between elections that their votes were "wasted" by voting for C, and thus come back and vote for A or B in the next election.

But that's not the only choice the C voters have. They could also (1) not vote at all, because they believe they've been effectively disenfranchised by the two parties or (2) continue to vote for a candidate like C, as the only candidate who represents their views. (One could certainly argue that at least condition (1) is incredibly important in the U.S. today, given the "get out the vote" efforts both parties employ, and also the fact that most adults don't even bother to vote in most U.S. elections.)

The point of all this is that while "first past the post" gives a mathematical tendency toward two party domination, it does NOT guarantee that the two parties will always stay the same -- one could eventually be replaced by another, or (more often) one or both parties could shift ideological ground significantly to retain members in the face of increasing defections in major elections.

This in fact has happened many places with the Republicans for example, where Tea Party affiliates have forced the Republican party candidates to change their ideologies or else lose elections. Rather than seeing it as a "spoiler" effect to lose an election, many voters instead see it as effective "disenfranchisement" and refuse to continue voting for candidates they hate. The only choice of the Republicans is to nominate candidates that have a chance of winning... and thus, to retain power, they must actually respond to those voters who cast votes for 3rd parties (or didn't even vote at all).

So, it's not quite as dire as people make it out to be. Yes, "first past the post" sets up some unhelpful constraints, but significant change is still possible. We can start by refusing to use 2-party propaganda terms like "spoiler effect," since that implies that one of the top two candidates somehow "deserves" to win. Untrue. If a 3rd-party candidate draws enough votes away from one of the top 2 candidates, the major party candidate FAILED to attract enough votes to win. That's not a 3rd-party candidate "spoiling" an election -- that's a major party "losing" an election by failing to satisfy enough of the electorate.

more interessting,..

[Selur]

it doesn't sound like this is the first experiment done by the facebook crowd -> What other experiments happened? Were the participants informed about it later? Who takes the blame if such an experiment results in someone getting hurt?

Re:more interessting,..

[N1AK]

Where do you draw the line? If Facebook realised that showing more negative stories (by monitoring what people already see) makes people more likely to click adverts is that really any better/worse than them artificially increasing/decreasing the amount of positive stories a user sees?

If Google was having a hard time deciding if a page was junk or not, would it be unethical to put it in the results for some users and see how they react? Clearly that's an experiment without user knowledge, but it certainly doesn't sound like it's unethical to me and stopping that kind of experimentation or flooding sites with notices about them would make things better for users.

Obviously there are experiments they could run that would be unethical if users weren't informed and monitored; discussing where the lines are and agreeing some best practices would therefore make sense.

I think it's fine

[Threni]

I love how overblown the coverage of this has been..as if it's driven people to suicide. It's their site, they can do what they want; people are free to leave if they want. Nothing to see here.

Re:I think it's fine

[DarkOx]

That is kinda my reaction as well. It seems this issue people have here is that facebook sought to manipulate peoples emotional state. The thing is that is exactly what just about every advertiser does all the time.

Home Security System ads: clearly designed to make you feel vulnerable and threatened.

Cosmetic surgery ads: clearly designed to make you feel inadequate.

Beer ads: very often designed to make you feel less accepted, you need their product to be preceived as cool, ditto for clothing, and personal care products

Political ads: feelings of security and family (at least if you pick their candidate)

This list goes on...

It might not have the same rigor as the academic world but they absolutely do focus group this stuff and find out how people 'feel' the marketers have researched what words, phrases, and imagery can best evoke these feelings. If what facebook did is illegal or even just unethical than so is pretty much everything the modern advertising industry has been up to for the past 70 years.

I am sure many people would actually agree with that, but I don't see why its suddenly so shocking and deserving of attention just because facespace does it.

Re:I think it's fine

[oh_my_080980980]

For the cognitively impaired: Advertising is is identified as advertising. The experiment that was conducted was not identified, people did not know this was occurring. The "researches" did not get informed consent - which you are required. The "researches" claimed the EULA was the informed consent.

Re:I think it's fine

[Tim the Gecko]

Facebook uses psychology to make minor changes in our happiness... Something must be done!

Soda companies use psychology to sell huge buckets of sugar water... Hands off our soda, Mayor Bloomberg!

Re:I think it's fine

[danudwary]

More PUBLISHED experiments, though, please. Let's know what they're doing, and what the outcomes are.

Re:I think it's fine

[Darinbob]

But Facebook has never provided 100% of all posts of all friends/non-friends/acquaintances in your feed. There's a wide variety of stuff there and they pick and choose. I don't use facebook so I don't know all of what's there, but on G+ the list of "what's hot" clearly doesn't include everything, if I reload the page I get a different list of items appearing, and sometimes I do have to go directly to a friend's page to see some recent posts.

There's no evidence given here that Facebook decided not to show you your friend's bad news.

Shock and awe

[Stumbles]

Not really, not in the least bit should anyone be surprised. Some years ago when Zuckerburg was asked about facebook users data, his reply; they are fucking idiots to trust him.

Re:Shock and awe

[s.petry]

This is probably the most important consideration for people to understand. Do you really think that the people pulling the strings give a rats ass about individuals? To people like Zuckerberg, you are something to be exploited for whatever purpose they wish.

So now that we know that they have done this once, I simply wonder how many more experiments they have done without disclosure. I have a hard time believing this is the first, or the last.

As a first guess, the whole "timeline" feature that was forced on people a while back was probably a result of this experimentation. It is impossible to hold or read rational dialogue on Facebook main pages now, because they reorder by "popularity" and not by chronological order, or even relevancy.

Re:This news piece has been greatly exagerated

[astro]

Bullshit. How do you know that you don't know anyone that was affected by it? Do you know which week in 2012 the experiment was conducted? Do you know which of the ~billion FB accounts were the 700k experimented upon? I find it pretty shocking that so many people are having difficulty understanding the difference between A/B testing and intentional emotional manipulation where a significant negative (or positive) result was the data point the study strove to measure.

I can quite imagine that a significant number of offline lives were impacted by this experiment. People exposed to negative content presumably don't limit their negative reactions to behavior only in the venue where they were exposed to the negative content.

Re:This news piece has been greatly exagerated

[Anonymous Coward]

Seriously, come on. Do you PERSONALLY know ANYONE who was affected by this? Neither do I.

Do you PERSONALLY know anyone who was affected by warrantless wire tapping? Neither do I.

As long as they never admit who it happened to, so that nobody can know whether it happened to them, then we're good? Look, there probably isn't anyone alive today (and certainly not on thus website) who knew Little Albert but that doesn't make the experiment that was done to him any less unethical.

Facebook's TOS can obtain consent, but it can never obtain informed consent.

A/B-Testing

[bickerdyke]

I understand why this should be considered wrong and fully understand users who don't want to have someone (less some company!) playing with their feelings.

But on the other hand, considering that creating an emotional response has been a standard marketing tool for the last 20 years, how is this different from regular A/B-Testing? 50% of your website users will see a slightly altered version of your website, and you compare response rates to the users receiving the "old" or "original" website.

Advertisers are manipulating our feelings for decades.News outlets have been doing it to an extent it became part of the news format itself (I guess anyone who was watching tv news last night saw that light-hearted, cozy, human-intrest or slightly oddball or cute item concluding the broadcast, right?) While creating negative feelings toward someone else has always been used in political campaigns.

It even becomes less spectacular if you consider, that on facebook, there always has been a selection algorithm in place, that tried to select those items from all your facebook-sources, that might keep your intrest focused onto facebook. Without selection, your facebook would scroll past like the Star Wars end titles. Only the parameters of the selection have been fine tuned, as they probably are at each facebook server update. It would be some new quality if that selection had been "objective" before, but being "personal" and emotional instead, is what kept us at facebook already.

So this is old news. But it should be a wake-up call: WAKE UP, THIS IS OLD NEWS! PEOPLE ARE TRYING TO MANIPULATE YOUR FEELINGS FOR AGES!

Just in case you haven't noticed. I'm surprised about the number of people who are surprised.

Re:A/B-Testing

[TheRaven64]

A significant amount of marketing is intended to cause harm to 100% of the user-base, so being ethically unsound doesn't appear to be a problem.

Re:A/B-Testing

[mr100percent]

The issue is clear; if a doctor or psychologist tried this, they would have to get IRB approval. You need informed consent; such laws were passed after psychologists had tried a LOT of experiments on the unwitting public; simluating muggings, imminent death scenarios, etc.

I know people say "it's just manipulating feeds, what's the harm?" There can be plenty of harm if you manipulate the feeds. Where is the line? What if facebook had decided to see what happens if you try showing depressing posts and bad news for a year? Or a feed where you were always ignored? No IRB would allow something like that if it risked permanent mental scarring or created a suicide risk.

Bad move, Facebook. Experiments are definitely cool (I'm a researcher), but we go through proper channels and regulation for a darned good reason.

Re:A/B-Testing

[Trepidity]

Apparently they did actually get IRB approval, oddly enough. The study was jointly done with two universities, and from what other researchers have told me, the two universities' IRBs approved the protocol. I'm surprised myself that they would. Would be curious to see what their reasoning was.

Re:A/B-Testing

[frank_adrian314159]

The reasoning was the same as what many are saying (IMO, incorrectly) here - that FB was already manipulating feeds so it was OK. I find this reasoning specious because, normally, FB modifies what it shows to attempt to change a narrow behavior with relatively finite consequences - whether a user clicks on an ad or not - while with this experiment the researchers were trying to alter something much more broad - a person's entire emotional state - a change with much broader implications. Given what we know from online bullying episodes, targeting a broad population in this manner could possibly exacerbate underlying emotional problems in some of the subjects. I can't see how these review boards let this study through - especially with a bullshit EULA for "informed consent".

The grounds for and circumstances surrounding these approvals should be investigated thoroughly.

Re:This news piece has been greatly exagerated

[bickerdyke]

I find it pretty shocking that so many people are having difficulty understanding the difference between A/B testing and intentional emotional manipulation where a significant negative (or positive) result was the data point the study strove to measure.

Creating an emotional response is part of marketing and therefore webdesign.

Of course you're not directly monitoring emotions as a data point during A/B-Tests. You measure e.g. the clicks, pages read or the time spent on the website. But every marketing guy worth its money could tell you that you can increase all of that by "making the user feel at home".

Too far?

[ebonum]

What about what advertisers do every day?
Our government (for us Americans) runs campaigns to alter opinions in other countries.
I'd like to everyone in the business of "caus[ing] changes in psychological status" get "require informed consent" first.
Beer companies anyone?

"Victims" received positive or negative newsfeeds?

[by (1706743)]

According to the WSJ's coverage http://online.wsj.com/articles... ,

The impetus for the study was an age-old complaint of some Facebook users: That going on Facebook and seeing all the great and wonderful things other people are doing makes people feel bad about their own lives.

So although conventional wisdom might say that seeing positive things makes you happier, here there have been accusations to the contrary -- positive things about other people makes you feel lousy about yourself. This study ostensibly looked at that (and I think it found something along the lines of conventional wisdom: happy posts make you post happy stuff, a [dubious!] proxy for your own happines...).

If Facebook knew (and how would they?) that X makes you depressed, then yes...there might be some moral issues with that. But it seems that Facebook asked a legitimate question -- especially so given that it was published in PNAS.

That said, yeah...it feels a little shady. But then, when I log onto Facebook, I am certainly not expecting any aspect of the website to be designed with my best interests in mind!

Natural vs randomized experiments

[sjbe]

Given that Facebook has over half a billion users, it’s a foregone conclusion that every tiny change Facebook makes to the news feed or any other part of its websites induces a change in millions of people’s emotions. Yet nobody seems to complain about this much...

If this guy actually thinks nobody complains about this much then he isn't paying attention. However putting that aside his argument is a straw man. There is a VERY significant difference between changing a service and that change having an emotional impact versus actually experimenting on the emotions of your customers directly and without their permission without even so much as review by an independent review board. Anyone who can't comprehend the difference between the two has a pretty big ethical blind spot. The fact that Facebook seems to be genuinely surprised by this response tells me everything I need to know about how they regard their users. They see them the same way an entomologist sees bugs - something to be cataloged and experimented on but not worthy of the respect one normally gives other human beings.

–presumably because, when you put it this way, it seems kind of silly to suggest that a company whose business model is predicated on getting its users to use its product more would do anything other than try to manipulate its users into, you know, using its product more

There is a big and fairly bright line between observing users behavior given certain stimuli as a natural experiment and the experimental investigators manipulating those users directly without their permission in a designed experiment. The later generally requires informed consent for a variety of very sensible reasons relating to ethics. The fact that emotional manipulation is done in other contexts is utterly irrelevant. That's the same argument children make when they claim that "...but all my friends are doing it too". I suppose since Facebook is owned and run by an immature child billionaire that I shouldn't be surprised.

And no, the Facebook terms of use does NOT rise to the level of informed consent.

Yarkoni misses the point

[Registered Coward v2]

Facebook didn't simply set out to make tweaks and see how users responded; they setup a controlled experiment on subjects without their consent; a practice that appears to violate ethical and possibly legal guidelines for behavioral research. I agree it could push them to continue to do such research and not reveal it; but when it inevitably leaks that they are doing that it will create a PR nightmare. Facebook could have simply asked people to opt in to the study and provide the standard information regarding the study and this would be a non-issue. For those looking for info on humane research protection guidelines in the US google Office of Human Research Protection.

Re:Yarkoni misses the point

[Registered Coward v2]

without their consent

What's actually more problematic to me is that the paper explicitly claimed they asked for and received "informed consent". But their justification is that users agreed to the Facebook EULA. That is a serious misunderstanding of what constitutes informed consent in research ethics; it does not just mean that someone agreed to some fine print, possibly months ago, in a transaction unrelated to the current study.

If they want to argue that this doesn't require informed consent at all, because it's e.g. just data mining of effectively existing data, that would be less problematic imo than watering down the standard for informed consent to include EULAs.

I agree, with an added thought. It wasn't just data mining but a controlled experiment that altered the data they received. That, IMHO, cross the line between "let's look at the existing data" to "let's conduct an experiment."

Another part to his argument seem to be "the impact was so small as too be negligible and thus it was OK." However, the researchers did not know the results would be negligible so using that as an excuse after the fact doesn't fly.

Messaging versus manipulation of content

[sjbe]

But on the other hand, considering that creating an emotional response has been a standard marketing tool for the last 20 years, how is this different from regular A/B-Testing?

Because they aren't just throwing messages at people to see how they react. They were actively changing the messages and how they were received. HUGE difference and one that crosses an ethical line. If you are a beer company, you can try to promote your product to me in a way that you think might make me more inclined to buy it and that is fine as long as you aren't overly intrusive about it (think telemarketers). What is NOT fine is for them to take what I say and manipulate that to try to convince me (or others) to buy their product.

Just in case you haven't noticed. I'm surprised about the number of people who are surprised.

Then you do not understand what is going on. Facebook stepped over an ethical line in their "research". No, nobody got (badly) hurt but that doesn't make it acceptable. Screwing around with people's emotions in a controlled experiment should require at minimum review by a genuinely independent ethical review board and probably genuine informed consent. Facebook could be bothered with neither one. They seem to regard their users as insects to be manipulated and dissected.

Re:Messaging versus manipulation of content

[bickerdyke]

Because they aren't just throwing messages at people to see how they react. They were actively changing the messages and how they were received. HUGE difference and one that crosses an ethical line.

But according to /., not what happend here.

According to this article here, no messages were changed:

Facebook briefly conducted an experiment on a subset of its users, altering the mix of content shown to them to emphasize content sorted by tone

(emphasis mine).

I agree with you that changing the actual messages would not be acceptable by any standard.

Just in case you haven't noticed. I'm surprised about the number of people who are surprised.

Then you do not understand what is going on. Facebook stepped over an ethical line in their "research". No, nobody got (badly) hurt but that doesn't make it acceptable. Screwing around with people's emotions in a controlled experiment should require at minimum review by a genuinely independent ethical review board and probably genuine informed consent. Facebook could be bothered with neither one. They seem to regard their users as insects to be manipulated and dissected.

And again I agree with you that you're stepping over a line when you're consciously manipulating people's feelings for economic reasons. But this line is crossed thousandfold already. The type of environment is secondary. A/B-tests take place in controlled environments, too.

Re:This news piece has been greatly exagerated

[PolygamousRanchKid]

1. Look up what wacky crimes were committed in January 2012.
2. Blame them on Facebook.
3. Sue.
4. Profit!.

In January 2012 a bunch of kids formed the Islamic Caliphate of the Rusted Chevy on Cinder Blocks on my front lawn, despite their parents instructing them: "You best be staying away from Mr. Kid, he ain't right in the head. "

Obviously Facebook manipulation caused this.

Creating emotional response is not the issue

[CodyRazor]

The problem is not that they attempted to create an emotional response or manipulate people's emotions. As people are constantly pointing out advertisers so that all the time. People don't seem to grasp that there is a large difference between this and advertising.

The problem is the way it was done. People use facebook with the expectation that they are seeing a (reasonably) objective representation of what their friends are trying to express or convey. Facebook is the equivalent of the telephone in a telephone call. If the telephone somehow manipulated what you heard to make your friend sound more negative or positive without changing their core meaning that would be unethical without informed consent, just as this is.

A more extreme version would be facebook subtly modifying the content of what your friends post as it appears to you without anyone knowing it was doing this. That would be even more unethical. The problem is mirepresentation, the method by which they attempt to manipulate emotions.

Not exaggerated at all

[sjbe]

Seriously, come on. Do you PERSONALLY know ANYONE who was affected by this? Neither do I.

Nobody knows who was affected or exactly how. That's part of the problem. They did it without knowledge or consent. They did not inform people of what they were doing or the fact that they did it after the fact. They did not have their design of experiment reviewed by an independent ethics board. They violated the (misplaced) trust their users had to deliver their messages as the users intended.

This isn't legal documents we're talking about here, anyway. I'm also pretty sure this is covered under Facebook's EULA/TOS you didn't read.

NOTHING in Facebook's TOS remotely qualifies as informed consent to be experimented upon. I don't even have to read it to know that. It's not THAT they did this experiment, it is HOW they did this experiment. It's not hard to check the experiment proposal in front of an ethics panel. It's not hard to get informed consent if that is deemed appropriate by the ethics panel. It is standard practice to do those things for some very very good reasons. Facebook couldn't be bothered.

This is not advertising

[sjbe]

The thing is that is exactly what just about every advertiser does all the time.

No it is NOT the same thing. The beer company does not have any control over what *I* say and they do not get to (legally) change what I say or how it is delivered to others. There is a HUGE difference between putting a message out there and seeing how people react to it versus actually changing what you or I say and how it is delivered to someone else without my consent. The former is advertising which is fine as long as it isn't too intrusive. The later is a violation of personal sovereignty unless you obtain informed consent beforehand.

Furthermore even if every advertiser actually did this (which they do not) and you have an ethical blind spot so large that you can't actually see what Facebook did wrong, two wrongs don't make a right. "Everyone else is doing it" is a juvenile argument that little kids make to justify behaviors that they shouldn't be engaging in.

Re:This is not advertising

[nine-times]

As far as I could tell from reading about this, they didn't change what people said.

Here's the thing, Facebook already filters what you see with the default setup. Your 500 friends each post 10 posts today, and when you load up your page on a social networking site, the page only displays 15. So how are those 15 chosen? (I'm making up numbers here, obviously)

The obvious choice would be to show the 15 most recent posts, but that means there's a good chance you'll miss posts that are important and that you'd like to see, since you're only getting a brief snapshot of what's going on in that social networking site. Facebook instead has an algorithm that tries to determine which of those 5,000 posts you'll care most about. I don't know the specifics, but it includes things like favoring the people who you interact with most on Facebook.

So what Facebook did in this study is they tweaked that algorithm to also favor posts that included negative words. The posts were still from that 5,000 post pool and the contents of the posts were unedited, but they subjected you to a different selection in order to conduct the research.

It's still an open question as to whether this sort of thing is appropriate, but it's important to note that this is something Facebook does all the time anyway. I think where is gets creepy is that Facebook is also an ad-driven company, so you have to wonder what the eventual goal of this research is. I can imagine Facebook favoring posts that include pictures of food to go along with an ad campaign for Seamless. Maybe they'll make a deal with pharmaceutical companies to adjust your feed to make you depressed, while at the same time plastering your feed with ads for antidepressants.

Outrage due to Censorship, not the test

[Anonymous Coward]

I talked to several (non-tech) friends about this, and they were more upset about Facebook "censoring" out posts than the emotional manipulation. In their minds, Facebook allows everything to be shown, but certain topics gain preference due to likes or dislikes. However, they will show you everything if you scroll far enough.

Their outrage came from the thought that FB was removing "happy" content from their feed. (That it was no longer a "dumb" pipe for social data).

Re:This news piece has been greatly exagerated

[mr100percent]

I don't know anyone who was affected by the Tuskeegee syphilis study, but that doesn't mean it was right or we shouldn't be outraged.

Re:This news piece has been greatly exagerated

[bickerdyke]

So you don't want to create websites that people enjoy using?

That may explain the design of the average Linux user group website, but wold also explains why websites like facebook or even lolcats, that target emotions, have more commercial success.

Isn't the FB Newsfeed a giant experiment anyway?

[swb]

I quit using Facebook six months ago, but for a couple of years was a regular user.

The "newsfeed" always struck me as enormously manipulated, with Facebook constantly altering the algorithm that determines what you're shown. Even nontechnical users would comment about this, wondering why they didn't see some posts from some people some times.

Some of this may have been benign, trying to figure out what order to display posts relative to relationships, posting frequency, sort of ordinary attempts to sort out "importance".

But I'm sure there was commercial manipulation -- ranking user comments with links to advertising-affiliated sites higher than non-affiliated sites, downranking links to sites likely to lead a person to shorten their Facebook session, etc.

All of this could be considered "manipulation" even though there might not be one single motivation behind it and not all the factors may be even focused on a specific outcome.

Human experimentation needs close supervision

[Opportunist]

There is no fine line here. There's only a bold one. Does it involve humans? If so, not only is tight ethic supervision required (to avoid a Milgram scenario) but, and that's the even more important part, the active and willing consent of the participating people is required.

Anything else, no matter how "trivial" it may be seen, is simply and plainly wrong. And no, some clause somewhere hidden between another few billion lines of legalese in an EULA does NOT constitute consent to being a guinea pig!

Re:This news piece has been greatly exagerated

[Darinbob]

All advertising is about manipulating emotional responses. And advertising has always involved experimentation, they just don't call it that normally.

Re:This news piece has been greatly exagerated

[Darinbob]

Which is exactly what broadcasters do. They want to know if they should show the car crash story of the story with the kitten stuck in the tree, and they want to know which will grab the viewer's attention and increase ratings, etc. So they actually do the research, monitoring viewership over time as the channel's style changes, though often this is done by aggregating word of mouth from other stations, following the trade magazines, etc. Facebook was just being more efficient in this regard.