Journal that published Facebook emoto-furtle study: Proper boffins get CONSENT Facebook's ethical standards do not meet those of most researchers who conduct studies on human subjects, the journal which published the "secret", emotion-manipulative research on nearly 700,000 of the social network's users has said. The journal of the Proceedings of the National Academy of Sciences (PDF), has now made a …
In psychological research how is informed consent even possible?
Leaving aside the basic problem that research is meant to discover new effects, psychology is a wholly imprecise science subject and the average person's understanding of it is pretty much zero. As a consequence a researcher is not qualified to either list all the possible effects (on the test subject), nor to quantify the risks of taking part - or any possible benefits, either. Nor is a lab-specimen, sorry: test subject in a position to assess whether the value they get from the experiment is worth whatever might, or might not, happen to them: either as a result or just naturally, from life.
Without that basic level of information, any consent is worth no more than "we'd like to do something to you. It probably won't (physically) hurt, but apart from that we can't make any promises". Which is probably little different from the level of personal protection that FB commits to in the Ts & Cs that nobody ever bothers to read, before they accept.
You ask them if they mind taking part in an experiment based on their fb news feed. You tell them when it will start and finish. You offer them an opt out at any time, you offer them a final opt out after the experiment finishes when you also explain what it was about.
I'm not defending Facebook, but that definitely wouldn't work, because if you told them you were doing an experiment with their News Feed, they would start paying close attention to their News Feed and try to work out what their experiment was. And they might deliberately try to act in the way they think the researchers are looking for - or if they're the contrary type, do exactly the opposite to pollute the data.
I'm not saying that Facebook went about it the right way, but the way you suggest would be totally worthless as research.
i see your point, but i'm not sure i agree. The way round that problem (which is a fair point) would be to start the experiment but not make the changes to the algorithm until a period of time into the expeirment (e.g. get their consent, enroll them and start making the changes a week later).
the important part is also to de-brief them afterwards - so if you turn the algorithm off agin after a month, make sure that they know what you did, how long you did it for and where they can find out more about the study.
I don't think that it is hard to do well, its just easier to be arrogant about it...
All the psychological research I've conducted or supervised has been quite specific about potential risks. Psychology works incrementally so there is almost always some precedent for the research from which you can make reasonable inferences about the potential risks. Furthermore, my colleagues and I often do pilot studies on ourselves to check for unforeseen consequences. Ethics committees normally require VERY detailed consideration of possible risks that are weighed up in a cost/benefit analysis. We can't anticipate every outcome with 100% accuracy, but we can give potential participants a very good idea of what they are signing up for. Furthermore, we always design studies so participants can withdraw at any time during the study and afterwards request their data be withdrawn if they aren't happy.
Fortunately the Facebook study hasn't particularly damaged the participants. However, it violates many core principles of how psychological research should be conducted and brings the subject into disrepute. Psychologists rely upon trust between researchers and participants in order for people to volunteer. This has the potential to seriously undermine that trust.
Psychology is imprecise if your benchmark is one of the natural sciences. The problem for psychology is that the thing it investigates is the same as the tool used to do the investigating (the human mind). I don't think that invalidates its importance as a subject area, but it does present a major challenge that can't easily be designed out of the research. So whilst I would say research should be rigorous, objectivity isn't always possible or necessarily desirable. I'm often dubious of research that implies objectivity and ignores the role of the researcher.
N.B. The term "subjects", when referring to people is historic in psychology. Generally they are referred to as "participants". That isn't just playing with words, it is genuinely intended to reflect a different relationship between researchers and volunteers.
Anonymous coward because I vaguely know one of the journal editors, although I don't think she was involved in publishing this.
Very good post, pity I can't upvote it more than once.
"Psychology is imprecise if your benchmark is one of the natural sciences."
Not necessarily true. Experiments in physics and chemistry are conducted with equipment made from the same things that are being studied. A good example is the discovery of the cosmic microwave background, where the detectors had their own internal noise and this had to be painstakingly eliminated before it became clear that there really was a signal there. The history of chemistry is littered with experimental problems, like, assuming you can make ultrapure water, what do you keep it in?
This is before you get involved in the question of what constitutes an "observation" in quantum mechanics, or the interpretations of the dual slit experiment.
As the phenomena have become more complex, so the use of statistical methods has taken over in physics, much as elderly physicists hate it. Astronomy nowadays is largely a matter of statistical analysis.
You can argue that psychology is a hard science and the physical sciences were until recently soft sciences because experiments were relatively easy.
Wrong:
Facebook, who don't care.
Others, who care a bit, but not much.
Anyone that really cares knows to stay the hell away from FB.
This is really silly though. Everyone already knows that FB tracks all types of crap to figure out your vulnerabiolities to serve you ads and get youre "friends" to suggest stuff to you. That manipulation is surely far, far, larger than just getting you slightly up or down by feeding you news in a different order.
Possibly also required, reading the study itself...
http://www.pnas.org/content/111/24/8788.full
E.g. "On Facebook, people frequently express emotions, which are later seen by their friends via Facebook's "News Feed" product (8). Because people's friends frequently produce much more content than one person can view, the News Feed filters posts, stories, and activities undertaken by friends. News Feed is the primary manner by which people see content that friends share. Which content is shown or omitted in the News Feed is determined via a ranking algorithm that Facebook continually develops and tests in the interest of showing viewers the content they will find most relevant and engaging. One such test is reported in this study: A test of whether posts with emotional content are more engaging."
The following paragraph is also informative as to what process was followed.
The exaggerated, hyper-indignant, internet outrage bandwagon rolls on.
The ironic thing is that it is FB and other social media which have brought us this mob rule, 'grab the pitchforks and light the torches', collective complaining and whining "me too" group-think culture on such a scale.
It used to be a British national sport but now seems to be an international one.
Its not news that Facebook does all kinds of "grey" things with the data they collect. It wasn't a new revelation 2 years ago when they did the study. I understand from a scientific study perspective its certainly more ethical to have express consent. However, lets not act like they were lighting kittens on fire and timing how long it took them to stop twitching.
Much like harmjschoonhoven said - are we sure this isn't the study? To track the social media outrage? I would personally find that far more interesting as it seems the outrage is never ending.
"researchers from the University of Cornell in San Francisco"
Clearly, Kelly Fiveash, the author of the Register article, did not read the actual article and has gaps in his knowledge of American institutions of higher education.
First of all, "Cornell" is Cornell University, not "the University of Cornell". This can be easily ascertained by simply looking in the gazetteer at the back of the better American dictionaries or even by using that new-fangled Google thingie. I admit that the nomenclature of American universities is not consistent; some are "University of X", others are "X University", but the proper response to this inconsistency is to look things up and get them right instead of winging it and getting the facts wrong.
Moreover, Cornell University located in Ithaca, New York, not San Francisco. How Mr. Fiveash managed to make such a hash out of the name of a very well known American university is beyond me.
Further: Facebook's "Core Data Science Team" works out of Menlo Park, California (as the PNAS article clearly states), not San Francisco. Important note for Mr. Fiveash: "San Francisco" and "the (San Francisco) Bay area" are not synonymous. I suggest he take a gander at Google maps.
Given the Register's generally high standard of writing, these errors come as an unpleasant surprise.
"General high standard"…?! I point the honourable gentleman in the direction of Mr. Haines's pieces, most of which are utter bilge (it's not really his fault, to be fair, most of what he writes is on health topics, and he's only regurgitating the sewage spewed by the NHS. Still, he could do better - or, to be honest, SOME research, a 5-minute Google search would throw up myriad articles which prove that the NHS wouldn't know 'healthy' if it tripped over it! The more I read, the more I become utterly convinced that the UK doesn't HAVE a health service - it probably never has - all it's got is a 'keeping-folk-not-sick-by-giving-them-appallingly-bad-dietary-advice-so-Big-Pharma-can-milk-'em-dry' service, aka the sales and distribution arm of Big Pharma Worldwide, plc. Oh, and if they DON'T believe they're getting their money's worth out of you, they stick ya on statins - if ya weren't sick before, you will be after taking those for a year - soon you'll be popping so many pills you won't have much time for anything else! Oh and if your quack tells you they prevent CHD and stroke, that's marketing spiel, they don't - they cause them. Or significantly increase your risk of suffering one and/or the other, at any rate. Why…? Because it's HDL which caused an increase in overall cholesterol and statins lower overall by lowering HDL, just like those trans-fat and O6-laden spreads do. We NEED cholesterol - without it, we die. The research which 'proved' it caused CHD was conducted on herbivores which CANNOT process it. Oh yeah, they fed it to some dogs too, but there were only 2 of them and they didn't die, so they were deemed an anomaly).
That's the problem with El Reg, it doesn't really do its homework; when it comes to matters of health, Reg, I expect FAR better than just a regurgitation of the standard NHS BS.
I sent you an email on the subject. You ignored it. If this were a paper mag, this would be the point where I'd be informing your subs dept that I no longer wished to continue.
I really DO expect better in future…
The more I read, the more I become utterly convinced that the UK doesn't HAVE a health service - it probably never has
The NHS may not be perfect, but as someone who unfortunately had to use their services to *save my life* just a couple of years ago, and is still getting followup remission support, and hasn't had to go through the bullshit some 'real life' America friends had to endure, my I politely say:
FUCK YOU AND YOUR PARANOID TIN-HAT RANT
The responsibility of the Journal is to ensure that the research is conducted properly - that is what the review process entails. For PNAS to say "it was him" is an abrogation of their role. Looking at their statement, they say that the research was not performed under Cornell's research ethics standards - how then did they accept this paper? Every paper I have read using human subjects has included the specific reference to the relevant ethics committee approval. Did the authors lie about having this approval or did the journal editor/reviewers not ask for this?
PNAS cannot avoid their responsibility in this way.
... the research was also crap: for instance they were using automated emotion scoring of posts using tools (LIWC, I think) that are only considered reliable for longer texts.
It seems a shame to make such a massive ethical cock-up for such a weak result. Or maybe it's just karma.
Isn't there something in the Facebook terms that state all data is theirs and they have the right to do what they want with it? Tailoring the news report to evoke an emotion is no different to what every news outlet does every day...
The ethics may be a little grey, but that is what you get when,you give up your life for free....
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
