Facebook Admits It Sent Misinformation Researchers Flawed Data (msn.com) 81
"Facebook provided a data set to a consortium of social scientists last year that had serious errors," reports the Washington Post, "affecting the findings in an unknown number of academic papers, the company acknowledged Friday."
The company used a regular monthly call on Friday with roughly three dozen researchers affiliated with Social Science One, a consortium founded in 2018 that Facebook hails as a model for collaboration with academics, to admit the error and apologize for the impact on their work. The data concerns the effect of social media on elections and democracy and includes what web addresses Facebook users click on, along with other information. The error resulted from Facebook accidentally excluding data from U.S. users who had no detectable political leanings — a group that amounted to roughly half of all of Facebook's users in the United States. Data from users in other countries was not affected...
Gary King, a Harvard professor who co-chairs Social Science One... said dozens of papers from researchers affiliated with Social Science One had relied on the data since Facebook shared the flawed set in February 2020, but he said the impact could be determined only after Facebook provided corrected data that could be reanalyzed. He said some of the errors may cause little or no problems, but others could be serious. Social Science One shared the flawed data with at least 110 researchers, King said. The group's former co-chairman, Stanford Law professor Nathaniel Persily, said of the incident: "This is a friggin' outrage and a fundamental breach of promises Facebook made to the research community. It also demonstrates why we need government regulation to force social media companies to develop secure data sharing programs with outside independent researchers."
An Italian researcher, Fabio Giglietto, discovered data anomalies last month and brought them to Facebook's attention. The company contacted researchers in recent days with news that they had failed to include roughly half of its U.S. users — a group that likely is less politically polarized than Facebook's overall user base. The New York Times first reported Facebook's error...
The anonymized data set is one of the largest in social science history, with 42 trillion numbers.
One Social Science One researcher told the New York Times this discovery "undermines trust researchers may have in Facebook...
"A lot of concern was initially voiced about whether we should trust that Facebook was giving Social Science One researchers good data. Now we know that we shouldn't have trusted Facebook so much and should have demanded more effort to show validity in the data."
Gary King, a Harvard professor who co-chairs Social Science One... said dozens of papers from researchers affiliated with Social Science One had relied on the data since Facebook shared the flawed set in February 2020, but he said the impact could be determined only after Facebook provided corrected data that could be reanalyzed. He said some of the errors may cause little or no problems, but others could be serious. Social Science One shared the flawed data with at least 110 researchers, King said. The group's former co-chairman, Stanford Law professor Nathaniel Persily, said of the incident: "This is a friggin' outrage and a fundamental breach of promises Facebook made to the research community. It also demonstrates why we need government regulation to force social media companies to develop secure data sharing programs with outside independent researchers."
An Italian researcher, Fabio Giglietto, discovered data anomalies last month and brought them to Facebook's attention. The company contacted researchers in recent days with news that they had failed to include roughly half of its U.S. users — a group that likely is less politically polarized than Facebook's overall user base. The New York Times first reported Facebook's error...
The anonymized data set is one of the largest in social science history, with 42 trillion numbers.
One Social Science One researcher told the New York Times this discovery "undermines trust researchers may have in Facebook...
"A lot of concern was initially voiced about whether we should trust that Facebook was giving Social Science One researchers good data. Now we know that we shouldn't have trusted Facebook so much and should have demanded more effort to show validity in the data."
undermines trust researchers may have in Facebook (Score:2, Flamebait)
Re: (Score:2)
You do have to trust in whoever is gathering your data in any larger study. This is like saying that climate researchers shouldn't trust data from satellites.
Now the better question is, are social sciences even sciences in the first place considering how ephemeral their more complex claims are and as a result how difficult it is to construct a study to try to disprove their more complex claims?
Re: (Score:2)
Re: (Score:2)
Its because they dont know what the scientific method really is. They just dont really grok it. They think they do though.
At the core of the method is making a prediction. Nobody can explain what prediction these "researchers" are making because they arent.
This isnt "research", at best it is "data collection" but in practice it is probably used for 100% "data dredging" which isnt science, but has its pretenders.
Re: (Score:2)
A level of trust in the data is absolutely required on the part of the researchers towards the people who have the ability to get at that data.
You are 100 percent correct. Data that you don't trust gets tossed.
Suggesting that they are not real researchers otherwise is asinine and moronic.
Considering that Facebook has had documented experience in cooperation with some really sketchy groups, like Cambridge Analytica, I would have to say that some of them are rather loosely referred to as researchers.
Re: (Score:2)
You should read “reflections on trusting trust”. It is a wellknown essay showing you cant rely on computers to not have abitrary trojan horses, even if you say read/wrote the source code or even low level instructions. To avoid it, you would need to build every tool you need for your reseach from the ground up. This is what I would call a medium level of trust. There are lower levels of trust, such as things to do with your own memory or language. There are higher levels e.g. Newtons quote about
Re: (Score:3)
If you have to "trust" somebody or something, you're not a real researcher, you're something else
So, all research must be done from first principles? You can't trust the people who made your mass spectrometer, you have to design and build your own from the ground up? Of _course_ you have to trust somebody or something in research, we're far past the point of research consisting of turning over a rock and making a note of what you see underneath with your own eyes.
It is, of course, always the case that your conclusions will be limited by the quality of the data. In cases like getting data from Facebook,
Re: undermines trust researchers may have in Faceb (Score:2)
Re: (Score:2)
If you have to "trust" somebody or something, you're not a real researcher, you're something else, and if the data you present is based on any kind of "trust" then absolutely no one would consider it meaningful, in any kind of real scientific field. The "social sciences" as you mentioned are not real sciences in the first place.
Okay me hearty - the basis of the research is you gather your data, analyze it against some hypothesis or even just report it. Then analysis can be done. After that before it becomes widely accepted, it needs to be verified. no one needs to disbelieve anything.
In fact, if you read the summary, you would see that the Italian researcher looked through the data, and saw right off that about half the US population of facebook who were not politicized, was not included in the data they got.
He trusted that
Re: (Score:3)
If you have to "trust" somebody or something, you're not a real researcher, you're something else, and if the data you present is based on any kind of "trust" then absolutely no one would consider it meaningful, in any kind of real scientific field. The "social sciences" as you mentioned are not real sciences in the first place.
This is a quixotic dream. In reality, any researcher than doesn't trust any data source or any past research is a researcher that will never accomplish anything due to wasting all of their time recreating past work and verifying the minutia of all data sources. The practical researcher considers past knowledge and current data sources and makes reasoned judgments on trust and acceptance. Blind distrust is not only impractical but also as misleading as blind trust.
Re: (Score:2)
You do have to trust in whoever is gathering your data in any larger study.
I find that a complete lack of trust in facebook by researchers would be about the best thing to happen to social media.
As a very reluctant Facebook user, I've placed a whacky profile, but along with the present day attempts to mine people's security questions, I've been getting a barrage of "Suggested for you" pages. Far left and far right, often at the same time. For a few days, I just clacked on the don't show (name of group) for everything, But then I figured out they were trying to profile me. So n
Re: (Score:2)
I wasn't advocating for trusting facebook. I was merely pointing out that there's simply no way in modern science to gather necessary information first hand in overwhelming majority of studies being done. That sort of low hanging fruit was picked in almost all fields long ago.
Full disclosure: I do not use facebook (to the point of having actual real life penalties, because there are groups I got left out of due to me being not on facebook) I take great care to block as much of facebook tracking both on my p
Re: (Score:2)
I wasn't advocating for trusting facebook. I was merely pointing out that there's simply no way in modern science to gather necessary information first hand in overwhelming majority of studies being done.
I was saying that researchers distrusting facebook is a good step in the right direction for ending the monetization of people. If it is untrustable, full of bad data, it is forever useless. I would love for an application that randomly poisons your search and browsing history to make monetization impossible to come out.
Re: (Score:2)
I think we're talking past each other. My take is "don't trust facebook, but trust the data unless you have better data".
It's an argument for pragmatism.
They DON'T trust satellite data (Score:2)
> This is like saying that climate researchers shouldn't trust data from satellites.
Bad example, because they DON'T trust satellite data.
What they do is look at satellite data vs measurements from near the ground. For example one of the most important climate research facilities, which has been measuring climate since the 1950s, is Mauna Loa Observatory, on the north side of Mauna Loa Volcano.
Then then apply mathematical models they've developed to map what they think the correlation might be between act
Re: (Score:2)
There's a tremendous amount of data that you cannot gather from other sources. Also, let's remind ourselves that data from the observatory you just mentioned is just as much of a "data provided by others that shouldn't be trusted" as satellite data.
I.e. even if we fully accept your dissent as correct, my point still stands.
Re: (Score:2)
You're right; you chose a horrible example is all.
For climate change stuff, the satellite data is is no way trusted. It's compared to ground sensor data measuring CO2 ON A VOLCANO, then they correct the data until it fits the model. What's trusted is the model. The data is really superfluous.
Re: (Score:2)
"Facebook accidentally excluding data" (Score:5, Insightful)
Re: (Score:2)
Cheat and steal. If caught, lie. Nothing new here.
Re: (Score:3)
Re: (Score:1)
You mean "a person at Facebook". As in: Not Zuckerberg or whoever sent out false data in the first place.
Facebook is a bunch of humans too. ... Well, ... mostly humans. ;)
Re: (Score:3)
Re: (Score:2)
"An Italian researcher, Fabio Giglietto, discovered data anomalies last month and brought them to Facebook's attention." Nope, a researcher brought it to Facebook's attention.
Indeed. And then FB decided they could not cover it up and tried to spin it like they were the good guys here. See above for a moron that fell for it.
They had "tust" in Facebook? (Score:5, Funny)
Are these people mad?
Re: (Score:2)
Re: (Score:2)
In related news . . . (Score:2)
after a child snatching children everywhere are shocked and dismayed, and have lost trust in the creeps in trench coats handing out candy from vans community . . .
Zuckerberg is a lying slimy fuck.... (Score:1)
Re: (Score:2)
Re: (Score:2)
Re: (Score:1)
You don't come to that conclusion from everything you have observed up to this point?
Even if he were the nicest guy ever if you're his personal friend, what does it matter to us, who don't ever experience that? What we experience, is effectively what something is. Things are defined by our experiences with them. If all you ever see is white swans, you are allowed to declare there are no black swans for as long as you don't encounter any, directly or indirectly. Not because there might not be any outside of
Re: (Score:1)
You mean "robotic". And throw "uncanny valley" in there too somewhere
Re: (Score:2)
Uncanny Reptilian would be a great band name.
What are they complaining about? (Score:4, Insightful)
First FB gave them a bunch of data they could write a paper about, feel all warm and fuzzy about being scientists and brag about it with their buddies.
Now FB gave them a bunch of fixed data they can write basically the same paper about with a small introduction along the line "I did everything right but FB botched the data, so here's my new low effort paper".
Chances are they only have to dump the new data in their existing Jupyter notebook and be done with it.
Also now they can write a meta paper about the impact of botched data on the results.
More papers, more warm and fuzzy scientists, more bragging rights. Isn't that what they want?
Re: (Score:2)
Re: What are they complaining about? (Score:2)
There's a difference between mentioning en passant and non-committally, in a list of possibilities, that the virus might have had origined in a lab, and hammering this specific hypothesis -- turned "certainty" because it fits an all-encompassing, pre-existing ideological narrative, in what amount to a clear case of bottom-line reasoning.
The first isn't a conspiracy theory. The second is. In fact, the second remains a conspiracy theory even if the hypothesis is later found to be true, because everything in i
Re: (Score:2)
[] the virus might have had origined in a lab, and hammering this specific hypothesis -- turned "certainty" because it fits an all-encompassing, pre-existing ideological narrative, in what amount to a clear case of bottom-line reasoning.
The first isn't a conspiracy theory. The second is.
Since when does applying Occam's razor make the result a conspiracy theory? The lab leak theory needs fewer inexplicable steps to be convincing than the wet market theory.
Or is your's another Chinese sock puppet account? That would explain it all.
Re: (Score:2)
Since when does applying Occam's razor make the result a conspiracy theory?
It seems you didn't actually read my comment. Try again, slower this time.
Re: (Score:2)
Your jaundiced view of researchers makes it clear you don't know very many, if any at all. Most are serious scientists that are interested in knowledge. You, on the other hand, seem to be projecting what you would do if you had their jobs.
Re: (Score:2)
Re: (Score:3)
The two social scientists that I was forced to put up with had their own agendas. These agendas show pretty obviously in their work. As in "keep performing the statistical analysis using different transformations (ex
They already pay off their advocacy groups (Score:5, Informative)
https://twitter.com/rmac18/sta... [twitter.com]
4 of the 5 privacy groups that they consulted about their smart glasses have public statements listing Facebook as a donor. I'm sure they are beyond repute.
* Future of Privacy Forum (https://fpf.org/about/supporters/)
* National Network To End Domestic Violence (https://nnedv.org/content/technology-safety/)
* National Consumers League (https://nclnet.org/wp-content/uploads/2021/07/NCL-2020-Annual-Report-final-july-2021.pdf)
* Information Technology and Innovation Formation (https://itif.org/our-supporters)
any researcher worth their salt knows (Score:2)
Re: (Score:1)
No, that is not how normal humans behave. At least offline, face-to-face, we go a conscience, and an interest in being there for our fellow humans. What you describe is the behavior of a mental illness called psychopathy or sociopathy, aka how humans act when they can't feel empathy, e.g. because online we're just pseudonyms and hence objects to each other.
And you mentioning "the truth" like it's some absolute thing, shows that really, you need to learn a lot more about that (like the foundations of science
Re: (Score:1)
Re: any researcher worth their salt knows (Score:2)
Empathy can be conditioned to varying degrees. For instance I was taught to call waiters by their name to make a human connection. Likewise I grew up with pets, so it gives me a better ability to show patty to animals based on how I was taught to treat them (also a better ability to cope with death). In every way I mention here, I can find Westerners who share this understanding. I now live in China and generally I love it but empathy is very different here, especially towards animals.
So the data FB has is bogus? (Score:5, Insightful)
Erh... isn't that basically all they have to sell?
Dear advertisers, you might want to wonder whether the stuff you get from there is worth what you're paying for...
Re: (Score:1)
It isn't. "Not worth what you are paying for it" is literally the definition of "profit margin". :)
Re: (Score:2)
Ok, let me rephrase it: If the data is bogus, it's worth literally zero.
Re: (Score:2)
This has always been true. Nobody who actually is familiar with facebook and has two brain cells to rub together thinks they actually offer targeted advertising. They assign "interests" to users based on auto-generated keywords of content that users have interacted with by commenting or by using a react emoji option. And if you delete them, they put the same ones back. Every ad exec who thinks they're getting targeted ads from facebook for their ad money is a gross incompetent.
What's the problem? (Score:2)
It sounds like the misinformation researchers got what they research.
Re: What's the problem? (Score:1)
What did they expect? (Score:5, Funny)
FB only sends quality data to paying customers.
Re: (Score:2)
I wouldn't even be sure of that.
Remember how psychopaths work:
FB will only be sending quality data if it thinks it can't get away with sending effortless crap.
Basing social sciences on the scum again? (Score:1)
I remember when social sciences had to literally rewrite all their textbooks, because one researcher found that all studies had been done with American students (I think economics students, to make matters worse), repeated them around the world, and found exactly the opposite from what had been the basis of social sciences at that time. (Those students were super-seflish sociopaths. While most of the world would put their community above themselves and help out others. Europe was somewhere in-between. Don't
Re: Basing social sciences on the scum again? (Score:2)
Social studies we're never easy but clearly even more difficult now. All humans have a tendency to diverge into there own cliches, groups, and etc. As the world has more people and our tendency doesn't change for these smaller collections, we end up with more divergence. More so it's virtually impossible to get any conclusion that can be seen as universal for any social study, even if we supposed society was static after this point (which alone is a significant reason why social studies need to be repeated
Assumptions I'm sure (Score:2)
Re: (Score:2)
When facebook's original app ate everyone's* contacts, it might have seemed like an accident... if it didn't send a copy to facebook first.
* I was not dumb enough to run a facebook app, thanks
Facebook... a big no no (Score:1)
Entitled much? (Score:1)
The group's former co-chairman, Stanford Law professor Nathaniel Persily, said of the incident: "This is a friggin' outrage and a fundamental breach of promises Facebook made to the research community. It also demonstrates why we need government regulation to force social media companies to develop secure data sharing programs with outside independent researchers."
Er ... facebook doesn't have to give you anything, do they?
Re: (Score:2)
He should try asking Apple for some metadata regarding their filtering program, see how far he gets.
Facebook Overlords (Score:1)
The first hurdle (Score:2)
the impact could be determined only after Facebook provided corrected data that could be reanalyzed.
So people who are professionals at researching misinformation could not / did not detect the misinformation in what they were using as research data.
Did it not occur to them to check it - to notice that there weren't any "don't knows" in what they had been given.
You would think that, as with any other quality assurance system, they would have done some basic QA on the samples provided.
Is this down to lack of (real world) experience, or were they simply too trusting?
Try sending 'flawed data' to the IRS (Score:2)
They'll explain everything to you in jail.
Openness is never the answer (Score:2)
Don't communicate except when required by law. There is no upside to it.
Their mistakes just get them more grieve due to a mistake, the cooperation never got them anything positive.
Alienate headline (Score:2)
we shouldn't have trusted Facebook (Score:1)
Ya don't say!
So then, where do we go from here? We can tune out facebook, or remain addicted, give them everything they want, and continue to complain. Sounds eerily similar to our politics
Yet another Facebook "apology" (Score:2)
Anybody paying attention has watched Facebook repeatedly "apologize" for their wrongful actions. The old saying "fool me once, shame on you; fool me twice, shame on me" may have played a role years ago, but at this point Facebook has wronged people way too many times. Anyone who trusts Facebook gets what they deserve.
Attempting to control the narrative? (Score:2)
It would appear that during the 2016 election cycle, Facebook data was used in its raw form and arguably to the benefit of the Trump campaign. A lot of people didn't like that. So is this a case of the left-leaning powers-that-be at Facebook trying to pass off altered data as raw data in an attempt to prevent a future right-wing campaign from repeating what was done in 2016?
Yo dawg... (Score:2)