Skip to main content
Search form
  • 00:09

    DAVID LAZER: Hi, I'm David Lazer.I am at Northeastern University, as well asthe Institute for Quantitative Social Science at Harvard.I'm a professor.I wrote a piece with many others titled Computational SocialScience almost a decade ago that lookedat these issues of potential big datato understand social systems.

  • 00:30

    DAVID LAZER [continued]: There are multiple reasons why social scientistsshould be working with big data and new technologies.The first reason is that social sciencehas to go where humans go.And so there are these new spacesthat people are going like Facebook--not so new any more, actually, and Twitter,and Instagram and the like.And if we want to understand how humans behave

  • 00:51

    DAVID LAZER [continued]: and how they interact, if we want to understand society,those are the places we have to go.So that's reason number one.Reason number two is that a lot of the things that peopleare doing in these spaces is what they have always done.They exchange information, they get advice,they create relationships and learn about each other.

  • 01:15

    DAVID LAZER [continued]: So much more is being captured about just regularday-to-day behaviors.And so if we had been looking at humans, say, 20, 25 years ago,people might go days with almost nothing being captured.Maybe the phone company would know who you called.

  • 01:35

    DAVID LAZER [continued]: If people were using emails--not too many people--a lot of people were, but a lot of peoplewere not a generation ago--the emails would be captured.But almost none of that was being analyzed.Even that little bit what feels like not such big datawas being analyzed because we didn't have the tools,we didn't have the theories, and so on.Now today everything's changed.

  • 01:57

    DAVID LAZER [continued]: Number one, so much of life is being captured in some fashion.It's hard to go more than five minutes without somethingabout your life being captured, whether it's because youtexted, you get an email.Your locations are being tracked, what you spendmoney on is being tracked.Now an element of this is extremely

  • 02:19

    DAVID LAZER [continued]: creepy and problematic.And that's all the more reason why social scientistsshould go in there.So we get to see what would have been invisible before,number one.But number two, we have to think about whatthe broader social implications of these technologies are.And so there's a strong normative reasonwhy we have to go into these spacesand think about what the broader implications for our society

  • 02:41

    DAVID LAZER [continued]: are, and what we should collectivelychoose in terms of the structure of our technologies.And I think the current crisis around quote, unquote"fake news", which is something I've been researching,is just one of many examples of well,how should systems like Facebook and Twitterthink about the quality of information that they share?We wouldn't have thought about that years ago, about,

  • 03:04

    DAVID LAZER [continued]: let's say, what mail people would have sent,conversations in the hallway, right?But so many of our interactions are mediated.Not only are they mediated, but they'retechnologically curated.Now I don't think anyone would particularlysay that the content I see from a distant cousin,

  • 03:30

    DAVID LAZER [continued]: which may be fake news or what have you,should be cut out by Google in Gmail.But whether Facebook should promotethat content is another question because they're choosing--I'm not seeing everything.I'm seeing a tiny subset of what my friends share.And so I don't know what the right answer to that is.

  • 03:51

    DAVID LAZER [continued]: But we should evaluate the order of magnitude of the issue.What are the kinds of interventionsthat would be helpful in this regard,and how might that be hurtful?That's just one example of the normative imperative.But then, just in terms of basic science,the opportunities to theorize about society at scale

  • 04:12

    DAVID LAZER [continued]: is just so different than what we might have conceivedof just a few years ago.So to be able to think about the dynamics of everyonein a group or a society, how they're interacting,and how that ripples through is justdifferent from 20th century social science.So for example, in 20th century, say of the art of technology,

  • 04:35

    DAVID LAZER [continued]: we would generally think of the analysis wedo as being variables by cases.And so there may be 300 cases, we get data from 300 people,they answer a survey, we do an analysis of the relationshipamongst the items in the survey.We would talk about things like collective action and so on,

  • 04:57

    DAVID LAZER [continued]: but what we're really talking aboutis how do the bulk of people think?In what ways does opinion lean?Now we can actually look at the ripplesthrough society, how what I do affects other people,and how that, in turn, affects more people.If we're going to have a big crowd,

  • 05:17

    DAVID LAZER [continued]: is that because this goes through many people?Is it because of social media?Is it because we all saw the same news report?We could start to parse these kinds of very big questions,and it enables a completely different science that'snot focused so much on the individual that's disconnectedfrom everyone else, but to look at society

  • 05:39

    DAVID LAZER [continued]: as a connected whole.There are several major challengesthat social scientists face when theywrestle with big data and new technologies.The first major challenge is that most of themhaven't been trained with the computational skillsthat they need.So that is a huge challenge.And so the reality is that right now, I'dsay around 90% of the computational social science

  • 06:01

    DAVID LAZER [continued]: that is taking place is being done by computer scientists.And that's not a slam on computer scientists.It's a slam on social scientists, thatis, social scientists need to up their game in termsof their computational skills.And so I would say social scientistsneed to learn how to code.

  • 06:21

    DAVID LAZER [continued]: And they don't need to learn to code in the same waythat computer scientists do.They need to understand computationin a different kind of way.But that should be just par for course,if you go to grad school, that you shouldbe learning how to deal with large scale data,how to deal with messy data, how to think about cleaning up datain a very different way than, say, going almost cell by cell

  • 06:42

    DAVID LAZER [continued]: or coming up with some simple rules in a column in termsof cleaning up data or dealing with missing data.And so there's a whole big array of computational skills,and what those skills are is justgoing to evolve a lot over the coming decades.Right now I'd say some of the big thingsthat people are doing is around networks, and certain kinds

  • 07:05

    DAVID LAZER [continued]: of things around text.But I'd say we're going to see more and more around video,image, and audio, and that this all interweaveswith social media as well.And so it's like not just even a single modality of data,but that we have video, image, and text interwoven together.

  • 07:25

    DAVID LAZER [continued]: I think that we need to shift towards moreholistic computational methods, and that trainingneeds to be both in terms of general coding skills,number one, and number two, a continual renewalof those methodological skills, which ahas alwaysbeen true to some extent, but there's

  • 07:46

    DAVID LAZER [continued]: a much more rapid evolution right now than there was, say,in the 1990s.If we're thinking of emerging social scientists today, peoplewho are in grad school or junior faculty,I think that there are greater opportunities in termsof learning computational skills, someof them on their campuses, some of them workshops,

  • 08:07

    DAVID LAZER [continued]: some of them summer schools, than there were just, say,five years ago.And so there are more on ramps to computational social sciencethan there was five years ago.There'll be far more on ramps five years from now.But if you are in grad school now,you have to take the choices you have.There's summer school at Duke.

  • 08:30

    DAVID LAZER [continued]: There's a great new textbook by [INAUDIBLE] Organicon social research.There are all these emerging toolsfor learning about thinking about research designand computational skill that just weren't therejust five years ago, say.And so most of the time they will notbe available to grad students locally.

  • 08:52

    DAVID LAZER [continued]: But there are opportunities out there to learn this,and so I'd say go out and seize the day.Go out and take these summer programs.Take some of these online courses.Learn how to code.The entry point is actually not so expensive.So I think that's possible.

  • 09:13

    DAVID LAZER [continued]: It's within reach of any grad student.It's just not easily within reach.

Video Info

Publisher: SAGE Publications Ltd

Publication Year: 2019

Video Type:Interview

Methods: Computational social science, Data ethics, Data privacy, Data analysis skills

Keywords: computer training; cyberethics; data analysis; Facebook; internet data collection; Skills development; Social impact of the internet; Social media; Social research; social science; Social scientists; Societies; technology; Twitter ... Show More

Segment Info

Segment Num.: 1

Persons Discussed:

Events Discussed:



David Lazer, PhD, Professor at Northeastern University, discusses the need for social scientists to embrace big data and the opportunities it presents to study human society as a whole.

Looks like you do not have access to this content.

David Lazer Discusses the Development of Computational Social Science

David Lazer, PhD, Professor at Northeastern University, discusses the need for social scientists to embrace big data and the opportunities it presents to study human society as a whole.

Copy and paste the following HTML into your website