Skip to main content
SAGE
Search form
  • 00:09

    GARY KING: Thanks for coming, or thanks for having meas the case may be.I'm going to talk about what I think of as a political scienceinnovation that we all put together.It's sort of a new constitution, or a new kind of constitutionthat solves a problem in data sharing between academiaand private industry.I'm going to explain it, but let me put it in context.

  • 00:31

    GARY KING [continued]: And I'm happy to be interrupted, or you should feel freeto yell out obscenities, or whatever it is you like to do.So the big picture is that the social sciencesare about understanding and solvingthe greatest problems that affect human society.That's what we do.I designed that for my dad who used to call me up and say, OK,Gary, tell me one more time, what is this political science

  • 00:52

    GARY KING [continued]: thing?So social scientists at one pointhad all the data in the world because we eithercreated it or we just acquired it from governments,or we sometimes bought it from companies.It was all in the university, that's where it was.Nowadays most of the data that exists in the world

  • 01:14

    GARY KING [continued]: is outside of academia.We sometimes get it.We get it by scraping the web or acquiring it in various ways.When I was in graduate school in American politics,more than half the dissertations were using the ANES.Now every graduate student creates their own data.

  • 01:34

    GARY KING [continued]: And that's terrific because we have more datathan ever before, and we've made faster progress than everbefore.But we actually have a smaller fractionof the data that exists in the world than ever before.So although it's spectacular and wonderful progress,it is wonderful to see all this progress,it is also an existential risk to us

  • 01:57

    GARY KING [continued]: that almost all the data in the worldactually exists and is tied up inside private companies.If you want to know or learn about the behaviorof individuals, they have more datathan we do by orders and orders and orders of magnitude.And so we have to find a way to work with them.We have to find a way to understand what it is they do.

  • 02:18

    GARY KING [continued]: We have to find a way to explain to them what we do.They don't necessarily care who we are or what we do.We have to change that.We have to make sure that they do care,and we have to care what they do.We sort of don't have a choice.Oh.Not only that, but they are our research subjects

  • 02:39

    GARY KING [continued]: and they control the access to our research subjectsand the data.Those folks that study mice in the university, the micedon't have little committees that decidewhether you can study the mice.So we have no choice but to engage with the world.OK.

  • 02:59

    GARY KING [continued]: So how is it that they should help us?Why is it that they would help us?So firms are legally required to maximize shareholder value.That's what they do.So they shouldn't help you.Except they are filled with real people,they are filled with our former students,they're filled with lots of other people.

  • 03:20

    GARY KING [continued]: They want to do good, just like you and I.From the perspective of shareholder valuethere's a long term and a short termand doing the kinds of things that we wantwould actually help.And everyone understands that.And so it's not completely obviousand it's not necessarily obvious when you call up companiesand say, can I have some data?

  • 03:40

    GARY KING [continued]: That they will give it to you.I've had graduate students who they allhave to go fan out to the world and acquire data and bring itback.And there's always the occasional sortof naive student who will call up a company and say, hey,do you have any research data?And the answer is no.No, we don't have any, and they hang up.And meanwhile, their HR system and their finance system

  • 04:02

    GARY KING [continued]: and their air conditioning system and their ground systemand every other system that they installhas a little spigot, which if you unscrew it,it comes spewing data.It wasn't designed as research data.They don't think of it as research data,but that's what it is, of course.So we're not allowed to speak our language

  • 04:23

    GARY KING [continued]: and expect other people to understand us.We have to understand them.So we have to understand their incentives.What is the incentives of folks in these companiesto work with us?Somehow we convinced Sage.Thank you very much.So the big tech companies have fairly intrusive dataon quite a lot of people.

  • 04:44

    GARY KING [continued]: Maybe they have responsibility.They actually think they have responsibility to help society.When companies get big enough and they become monopolies,then they actually make a bigger effort.So AT&T created Bell Labs and Microsoft created MicrosoftResearch and there's IBM Research and the Xerox PARCand there's actually a lot of others.

  • 05:04

    GARY KING [continued]: And these things work with the companies,but sometimes do things completely unrelated.And they make enormous progress that help us as welland help society.So there's the raw materials to craft somethingthat maybe everybody would want.So we have to understand them.They have to understand us.

  • 05:25

    GARY KING [continued]: All right.So now, that means when any of usgo visit companies they are treated to our lobbyingcampaign.If I visit, you have no choice but to hear a little bitof my lobbying campaign.And I was at Facebook lobbying in part,I was there for lots of other reasons,

  • 05:45

    GARY KING [continued]: but I was also asking them for data.And it was right before the Cambridge Analytica scandal.It was the worst times lobby eventin the history of the world.In fact, I was in my hotel room packing to go homeand I got an email from my friendsat Facebook that said, hey, what do we do about this thing?

  • 06:07

    GARY KING [continued]: So I thought, OK, I'm just going home.But they called a few days later and they said, hey,could you do a study of, I don't know, the 2016 electionand maybe you could tell us, did wechange the outcome, or if we did is there something we could do?Could we change what we do to improve the situation,

  • 06:28

    GARY KING [continued]: or can we just learn from this?And I thought wow, that would be really great to do that study,and I thought about it for a minuteand I realized there's no way we can do this.There's no way we can do a study like this because Ineed two things, and you Facebookare only going to give me one.So the two things I want is academic freedomto publish anything I want.It can't be the case that I write an article, you read it

  • 06:49

    GARY KING [continued]: and you say, no, let's not publish that.Then it has no academic value whatsoever.So first is academic freedom and secondis complete access to your processes and papersand people and data and platformsand everything I need inside the company.And so the answer usually is you can onlyhave one of those two things.

  • 07:10

    GARY KING [continued]: You want to be an academic, you get academic freedom.You want to get complete access to the company,you're basically an employee.And employees don't have freedom of speech completely.They do when they're speaking about thingsorthogonal to the purposes of the company,but this was quite tied up.And so there's two things.I said I need two things.You have to give me both, but you're notgoing to give me both.

  • 07:30

    GARY KING [continued]: And the answer was, you're right,we're not going to give you both,but how do we solve this problem?And I said I just told you we can't solve that problem.And so we went back and forth a few times until I finallysaid, OK, how about this?And this actually worked.And that became this project and Social Science Onethat I'll describe to you.

  • 07:52

    GARY KING [continued]: So here's how it works.What I said was and what we're actually doing is first,we agree on the scope.So the scope of the study is the effectof social media on elections and democracy anything.About the effect of social media on elections in democracy.OK.So that's the scope.

  • 08:13

    GARY KING [continued]: But wait a minute, there's still two thingsand we can only have one.How do we solve that problem?So we solve that problem within the scope by instead of megetting two things, we create two me's.Well, that would be more efficient.But instead we have two groups.One group is a commission of distinguished academicsthat agree not to publish.

  • 08:34

    GARY KING [continued]: So they don't get that first thing,but they get complete access in the company to everythingappropriate to what they're doing,anything that they need for their mission.Now, what is it that they need for the mission?What is it that they get access to?They get access to the knowledge of any data set thatcould be useful for studying the effect of social mediaon elections and democracy.

  • 08:54

    GARY KING [continued]: They can look all around Facebook.They can ask anybody.They can get access to the code books.They can understand what's going on.They can get trained on the platforms.They can get access to the computersystems, things like that.Am I saying this right, guys?And just like employees, they get access

  • 09:15

    GARY KING [continued]: to the things that are appropriate to whatthey're doing.They can then pull out a data setand have an RFP, request for proposals,from outside academic researchers.Those outside academic researcherscan apply for access.And if they're granted access theycan publish with no constraints.

  • 09:37

    GARY KING [continued]: So I mean, no prior pre-publication approvalby Facebook.So that's the idea.That solves the problemThis commission and Social Science One we put together.Nate personally is a professor at Stanford,and I have organized this thing in orderto make the outside academics be able to do this.

  • 10:02

    GARY KING [continued]: They'll get access to the data but will also fund them.So the way that works is that is we collect moneynot from Facebook.So Facebook's money doesn't get to influencethe process inappropriately, or at all actually.

  • 10:24

    GARY KING [continued]: And so we have collected seven nonprofit foundation,soon to be eight actually.And the non-profit foundations range from the most liberalto the most conservative foundations.Their money all gets pooled and put into one pot.That pot is administered by a non-profit organization called

  • 10:44

    GARY KING [continued]: the Social Science Research Council.The decisions about who gets that moneyand simultaneously who gets the data access at Facebookis solely determined by the commission at Social ScienceOne.So I'll explain to you how it works.I would as a side point really like the money from Facebook

  • 11:07

    GARY KING [continued]: to also go to this project, but this is probablythe most highly charged partisan issue there is around.And so everybody agreed that it was a better idea to justkeep it totally separate.And having the foundations that normallydon't agree on things like this normally don't get together,

  • 11:28

    GARY KING [continued]: and in fact have never to my knowledgeever gotten together in this way that spanssuch an enormous ideological terrain.That's very helpful to us to be able to say this is actuallya truly nonpartisan effort.So no one, foundation, or viewpointhas too much influence.

  • 11:49

    GARY KING [continued]: Facebook doesn't have too much influence.And that's basically the arrangement.Now, suppose we go into Facebook and wefind a data set and we say, oh, let's have that data set.Researchers would be able to do something useful for it,and Facebook says, wait a second.If you analyze that data set we're going to look bad.

  • 12:10

    GARY KING [continued]: OK.What happens if that happens?OK.And the answer is that the commission inside Facebooknot only has access to everything,but they have one thing that employees do not have,which is complete freedom of speech about the process.So if that happens, we are allowed toand indeed obligated to report to the public

  • 12:31

    GARY KING [continued]: that Facebook is reneging on its bargain.Now, the way backward induction worksis that they agreed to that, so of course that's not happening.And I can tell you it's not happening.OK.So all right.So how does this process works?So now think of yourselves, which is your job.

  • 12:53

    GARY KING [continued]: So how do you get this money and the data access?How does it work?It's an amazing amount of data.It's more informative data than has ever been--I don't even know exactly how to finish that sentence.But it's very informative data.I'll give you examples of it.So here's how it works.

  • 13:15

    GARY KING [continued]: You write a proposal, you take your proposaland you bring it to your universityIRB, your Institutional Review Board,and you say this has to be appropriate.You have to get approval.Once you have approval you then send that proposalin to the Social Science Research Councilwhich is organizing the peer review process.

  • 13:35

    GARY KING [continued]: So the Social Science Research Councildoes a normal merit based peer reviewwith anonymous reviewers.And they go out, they write their views, they come back.Sort of like a journal.We then have layered on top of that a special ethical reviewover and above beyond the IRB.

  • 13:58

    GARY KING [continued]: IRBs were usually created mostly for medical researchand they don't always understand our social scientists.Actually the most outrageous studies I've ever done the IRBdeclared them not human subjects,including when we randomly assignedcommunities in Mexico to either receive health careor to not receive health care.

  • 14:19

    GARY KING [continued]: Receive health care meant we built hospitals,not receive health care means we didn't build hospitals.And that was not human subjects.OK.But we have another layer of ethical reviewthat's designed for the types of datathat we're going to be making available.I'll talk more about the types of data.

  • 14:39

    GARY KING [continued]: So after it goes through those levels of reviewand also the ethics of the researcher--we don't need this Cambridge Analyticaguy that was at the University of Cambridge thatcreated that scandal involved in this process.So your ethics are important as well.Then we take the recommendations from the Merit Reviewfolks and the Ethical Review folks

  • 15:01

    GARY KING [continued]: and they all go to members of the academic commission.They then make the final decisions.We make the final decisions.Why is that?That's because we have private informationfrom inside Facebook that nobody else has.So, for example, we would know that thereis a lawsuit going on about some particular subject.

  • 15:23

    GARY KING [continued]: And if we allowed researchers to pursue that subject,they would be subject to discoveryand would be in courtrooms for the rest of the year.And so we would have to eliminate that.And there's various other cases which we wouldn'tallow researchers to go into.In fact, like when you apply for research access

  • 15:44

    GARY KING [continued]: to various types of government data,you don't just say send me the data.I'd like to explore.You say, here's the subject I'm going to pursueand the quantities of interest that I have,and I will go after those.And you apply, you ask for those,and then you're granted access.And if you have a new idea, that's fine,

  • 16:05

    GARY KING [continued]: but you have to send in a change of the proposal.That second process by the way will be faster.But you have to do that.So we're going to know what it is you're analyzing,and the commission we're going to pay very close attentionto what you're proposing.You really have to do what it is you're proposingand not other things.But if you have other ideas we'll help you do that.

  • 16:26

    GARY KING [continued]: And if you propose something thatis scientifically meritorious and ethically appropriate,but there's no way you're going to get it because it's steppingon a landmine that you don't know about,we'll try to help you sort of steer around that, OK?So that's the job of the commission.The members of the commission are notallowed to publish from the data.That's their deal.

  • 16:47

    GARY KING [continued]: So they are basically the other halfof what would normally be your brainand doing both things at the same time.So that's how we solve that problem.So the commissioner at Social Science Onemakes that final set of decisions.And Social Science One then is funded solely

  • 17:10

    GARY KING [continued]: by the nonprofit foundations, and so they'refree of influence from either the viewpointsof the different ideological foundations or from Facebook.So let's see.This is a very controversial subject.We have been attacked in the press pretty much continuously.

  • 17:32

    GARY KING [continued]: There have been multiple petitionsfrom academics which we appreciate very much, thankyou, well meaning.My own company, I created a company calledCrimson Hexagon, you may know that,collect social media posts.It's been attacked, and by the way, utterlyand completely exonerated.You can ask me more questions about that later.I'd be happy to answer them.

  • 17:57

    GARY KING [continued]: They've threatened to sue us that is Social Science One.They will sue us.They seem to sue Facebook like all the time, every day.The latest and most recent is we get hate mail.We've even gotten hate mail from prisoners.And I thought that's really impressive.So it is a very controversial set of issues

  • 18:21

    GARY KING [continued]: and we're working on doing it in the most conceivablyappropriate way that you could possibly imagine because we'relayering in more procedures that has everexisted before in any kind of projectthat we've ever heard of.We're building first of its kind privacypreserving computer infrastructure.

  • 18:41

    GARY KING [continued]: We have computer scientists that are specialistsin a particular type of privacy technology called differentialprivacy.I don't know if you know what this is.Sort of a cool idea.I'll just tell you this idea.So as social scientists, we don't care about anybody.We only care about everybody.

  • 19:01

    GARY KING [continued]: We don't really study individuals.Sometimes politicians, but we don't reallycare in a survey who is in the fourth row.We only care about patterns for everybody.OK.The privacy experts are of coursetremendously concerned about reidentifying or violatingthe privacy of any one of those people.

  • 19:22

    GARY KING [continued]: So that's really something that we want to avoid.So how do you do both of these things?We just want you to send us all your dataand we won't look at the individuals.We'll just look at the overall thing.So differential privacy is this really cool ideain which you add a particular type of specializedrandom noise to the data.There's other ways of doing it, but in this special case.

  • 19:42

    GARY KING [continued]: So you add random noise to the data,then I can hand you the data set and wewould know that it would be impossiblefor you to ever reidentify any one individualbecause there's all this random noise in there.But still if you ran your regressions,the regression coefficients would be almost the sameas it would have been originally.So that's the promise of this technology, whichis pretty cool because we can get

  • 20:03

    GARY KING [continued]: what we need to learn new patterns about societyand for society and also individuals get protected.So we're trying to build that in.We're also changing the model of privacyand of protections from one of individual responsibility

  • 20:23

    GARY KING [continued]: to collective responsibility.So what this guy Kogan did that created the Cambridge Analyticascandal is he basically followed like pretty much all the rulesand he just violated one.He took the data and he gave it to a private company.So he violated his individual responsibility.And basically the entire system and all of us

  • 20:43

    GARY KING [continued]: because a lot of our data access gotcanceled with that one violation,we all depended upon one researcher.So there was like a single point of failure.So what we're doing is we're tryingto change to a model of collective responsibility.So the way this would work is that therewould be no single point of failure,at least to the best of our abilities.

  • 21:04

    GARY KING [continued]: So when you get data access we'renot just going to be handing you the data.We're not going to be handing anybody the data.We're going to give you data access,and you're going to be on a specialized computersystem in specialized circumstanceswhere if you type in a line of code, that line of codeis known not only by you but by others.Not the world, but it can be monitored, audited,

  • 21:27

    GARY KING [continued]: and verified.And so that way there's yet another layer of protection.So the process that we've gone through with Facebookand the foundations and the SSRC and a large commissionof academics has been very hard implementing this thing.There's a lot of issues every single time,

  • 21:50

    GARY KING [continued]: and pretty much everything that we do has never reallybeen done before.So we have a collection of legal agreements that now numberin the--it's probably not 100, but it's getting there,if you include all the NDAs and things.Every one of the issues is sort of cross cutting

  • 22:10

    GARY KING [continued]: or cross-functional.It hits legal issues, security issues,the platform inside Facebook, privacy issues,new laws like the GDPR, which has justappeared like in the middle of our projects.I mean, I realize they knew about it ahead of time.The technical issues, like differential privacy.

  • 22:32

    GARY KING [continued]: It's very complicated, but we're making actuallytremendous progress, and I'm proud of the progress.A very large number of people involved in this effortare producing.So we have that first data set that we have announced,and you can go to socialscience.one,which is our website and the name of our project.And you can see the first code book from the first data set.

  • 22:57

    GARY KING [continued]: This data set is URL sharing, so all the URLsshared on Facebook publicly with some privacy protections.Roughly speaking that's what it is.It includes a trillion numbers, that one data set.It's a million gigabytes, OK?So just sort of try to wrap your mindabout how much information there is

  • 23:18

    GARY KING [continued]: about human behavior in that data set.So it includes from 2017 to the present,we hope to be able to go back further eventually,counts of user behavior within URL by country, by region,by age, by gender, by the type of device,

  • 23:39

    GARY KING [continued]: by the week because it's time, and alsoin certain countries by ideology and by your friend's ideology.So the number of people in each one of those things.And then characteristics of the URL, the thing you clicked on,the website, some of them are real news and some of themare false news.

  • 24:01

    GARY KING [continued]: So which is which?Well, actually we don't know.That seems like a really good research project.I wish somebody would solve that problem.But Facebook sent the news items to third party fact checkersor checked with third party fact checkers,and we have it coded as to what the third party factchecker said.Now, the fact checkers needed to be checked themselves,

  • 24:23

    GARY KING [continued]: so it's just another source of data.All these variables are outcome variables.They're not necessarily explanatory variable.So you have to figure that out.But anyway, there's views and shares and the number of peoplethat read the URL, the number of people that shared the URL,the number of people that shared the URL without readingthe website, that's a thing actually,

  • 24:46

    GARY KING [continued]: or the number of likes and wows and sads and angrys,et cetera, whether it's coded as hate speech.So it's actually a tremendously informative data set.And it's just the first of quite a sequence of data sets.I'm going to tell you about some of the others,but I really want to encourage youto apply because once you have access,once we get you through this whole sequence of applications,

  • 25:09

    GARY KING [continued]: you have some other ideas, you have another data set that youmight be interested in, we can get youaccess to those much faster.You can also get $50,000 to do what you want with.That sentence is not exactly correct.So let me give you a feel for what some of the other datasets are.Probably the next two data sets that we'll

  • 25:32

    GARY KING [continued]: release is a data set on CrowdTangle, whichused to be a startup that Facebook created on mediastories and the usage of what type of viewersviewed each media story, and what type of peoplethey were, and things like that.Second is a database on political ads.

  • 25:53

    GARY KING [continued]: Imagine all the ads, declared political ads tryingto influence somebody else.If you're a political scientist, a sociologist,social scientist, I would think thiswould be incredibly relevant.It's not only ads, but it's all the versionsof an ad that were sent to different populationsto try to influence them.

  • 26:15

    GARY KING [continued]: We are running our own surveys in probably Mexico, Brazil,and the United States, and then opting peoplein to sharing their Facebook data.So we run our own survey.We ask them at the end of the survey,would it be OK if we had all your Facebook data?And they just spent all this time with you,and so we hope they will all say yes.

  • 26:37

    GARY KING [continued]: We're making arrangements with all the big academic surveyslike the ANES and the Kanin Election Studyand the CCES and the CSES and the EES and allthe LAPOP and various others, and we'retrying to make arrangements where with each onethat we will opt people in and include Facebook data.

  • 26:58

    GARY KING [continued]: I think that this will change the waythis all works quite a lot.We learned that inside Facebook they do lots of AB testson all kinds of things, mostly about their platform,but sometimes on other things.They probably have the largest collection of experimentsever conducted in the history of the world.Some of those will be relevant to the effect of social media

  • 27:20

    GARY KING [continued]: on elections and democracy.We'll be able to take those out and have RFPs for them.Similarly they've run surveys on their platform.I don't know if you've ever been on Facebook,there's a little survey pops up and theyask a question or two or three.That also is probably the largest collectionof survey data in the history of the world.And some of those will be relevant to the effect

  • 27:42

    GARY KING [continued]: of social media on elections and democracy.So that's another data set we're going after.Another one is there's a lot of people on Facebooklike me have all their posts coded as public.So you can go to my Facebook siteand you can see all my posts.They're all public, and quite a lot of people do that.

  • 28:02

    GARY KING [continued]: Many, many, many, many, many more times than Twitter.I don't know how many yet, but it's a lot.Most of the scholarly literature is using the Twitter API.Imagine if you had access to something orders of magnitudelarger than the Twitter API.And we just replicated all the studies we did using that.I would think that would probably change the field,

  • 28:23

    GARY KING [continued]: or at least have a big influence.So that's another one.We then would like a large random sample,like a very large random sample of Facebook newsfeeds.Yes?Sorry, say it again.

  • 28:40

    AUDIENCE: Did you say that you weregoing to try to release a data set of all the public--

  • 28:45

    GARY KING: Public posts.We're not going to release any data.We're going to give you access to the data.

  • 28:51

    AUDIENCE: So not their information about nottheir about me stuff, but all the posting posted.

  • 28:59

    GARY KING: With certain qualifications, yes.OK.So let me explain the qualifications.So you're going to apply, you're goingto say here's what I want to do with those data.And so we're going to approve that.And the question is not going to be,I want to check into one of my ex spouseor Britney Spears is doing, OK?So you're going to propose a question.That question is probably not going to be looking up somebody

  • 29:22

    GARY KING [continued]: individually, although all these data are technicallyyou can get them already, except you could never get themalready.And you'd propose what it is and then you come to the systemthat I was describing earlier and you can get access to it.You cool with that?

  • 29:43

    GARY KING [continued]: And each one of these is going to be adjusted the aswe get to the point of release.But yeah, that will be one of them.Think of it like a giant Twitter API but much more informative.OK.So then I hope that we will be able to expand beyond this.

  • 30:03

    GARY KING [continued]: I hope we will be able to expand beyond the effectof social media on elections and democracy.One of the first areas beyond this that I'd like to go foris public health and well being and things like that.I already have a group of foundationsthat seem very, very interested in the subject.And we'll have to get everybody to agree to that also

  • 30:24

    GARY KING [continued]: and getting everybody to agree the first time was a littlelike the Arab-Israeli Peace Treaty,except there's many more players so it's more complicatedand nobody's going to kill anybody so it's probablya little easier.We have some interest from other companieswho have unique sources of data thatmight be willing to share data under specific circumstances

  • 30:48

    GARY KING [continued]: to help this project, and then eventually perhapsto use the same two part model to make dataavailable from those companies.The project seems to be sort of successful so far.It has recently been extended.So the original plan was we're going do this for a year,see what it's like at the end, and after three or four

  • 31:11

    GARY KING [continued]: months we've been at this everybody'sagreed including Facebook to say,OK, now it's at least two years and we'llfigure out the form in which it is goingto take place in two years.As I said, we've expanded the number of foundationsfrom seven to eight.We've had interest in lots of othersThe only thing we're missing from this is you.

  • 31:34

    GARY KING [continued]: So we need you to apply to be part of it,and we'll give you money and then access to data.So it seems like it would be a good thing to do.And I would do it early because if you do itearly you'll get access now and you'll be able to act faster.The application process is designed to be hard for us

  • 31:55

    GARY KING [continued]: and easy for you.So it's a short application process.It is reviewed relatively fast.You should get the answers relatively soon.Using the data is different than itwould be if you were going to analyze the NES,so you're going to have to learn some new systemsand things like that.But I think the amount of information being brought

  • 32:16

    GARY KING [continued]: to bear on the people we study for the questions we study isso much larger than all the information wehad available previously that gosh, it seems worth it.

Abstract

Gary King, PhD, Weatherhead University Professor at Harvard University, discusses why and how big data should and could be shared between private companies and academia, including who has and who controls data, the challenges of academic freedom to publish and complete data access, and how Social Science One addresses these challenges—how it works, challenges it faces, privacy and collective responsibility issues, access to databases, potential for the future, and how to apply for access to databases.

Looks like you do not have access to this content.

Data Sharing Between Academia and Private Industry

Gary King, PhD, Weatherhead University Professor at Harvard University, discusses why and how big data should and could be shared between private companies and academia, including who has and who controls data, the challenges of academic freedom to publish and complete data access, and how Social Science One addresses these challenges—how it works, challenges it faces, privacy and collective responsibility issues, access to databases, potential for the future, and how to apply for access to databases.

Copy and paste the following HTML into your website