Skip to main content
Search form
  • 00:00

    [MUSIC PLAYING][Gabe Ignatow Discusses TextMining in Sociology]

  • 00:10

    GABE IGNATOW: My name is Gabe Ignatow,and I'm a sociologist at the University of North Texas.[Gabe Ignatow, Associate Professorof Sociology, Director of Graduate Studies,University of North Texas] I've written a few bookson text mining for SAGE.And I've been interested in text mining for a few years.Text mining is a form of data mining.And in the social sciences, it's usedto collect information and analyze informationin textual form from communities or groups

  • 00:31

    GABE IGNATOW [continued]: where it naturally occurs.Both text mining and data mining are, to some degree, misnomers.It sounds like they involve acquiringdata, collecting data, and sort of mining it that way.But what they're about is having data and then mining itfor insights, mining it for information or some kindof useful knowledge.In the social sciences, we do text mining

  • 00:53

    GABE IGNATOW [continued]: on all sorts of different forms of data.It can be ethnographic interview transcripts.It can be ethnographic field notes, newspaper articles,newspaper reader comments on newspaper articles,social media sites, social media platforms, Twitter, Facebook,and so forth.So anything in the form of textual data,

  • 01:14

    GABE IGNATOW [continued]: generally, it takes a form of a large data file, lots of text,but it doesn't have to be that way.There are different ways to do text miningwith smaller batches of data.Text mining has a long history in the social sciences.It used to be referred to as text analysis.There's a history of content analysis.

  • 01:34

    GABE IGNATOW [continued]: There are more humanistic, interpretive formsof text-based social science.And it's all sort of coming togetherin our era as text mining.It's an umbrella term now and covers different thingsin different disciplines.So we wrote these books to try to providea guide book or a series of guidebooks for users,for students, advanced undergrads, graduate students,

  • 01:56

    GABE IGNATOW [continued]: and professional researchers to make it easierfor them to navigate the world of text mining, allthe software options, the data acquisitionoptions, the different theories and waysof synchronizing theory and method and data to producesome kind of valuable knowledge.So I became interested in text mining as an undergrad,actually, before the word text mining or the phrase text

  • 02:18

    GABE IGNATOW [continued]: mining was invented.My undergraduate advisor assigned me a book on politicsby George Lakoff who is a well-known cognitive linguist.And the idea behind that book was that our language,when we discuss politics, no matter who we are,it reflects certain underlying cognitive structures.And those cognitive structures differ systematicallyacross social groups and across different cultures.

  • 02:40

    GABE IGNATOW [continued]: So in the United States, Lakoff analyzedkind of conservative and liberal political cultures.And I was assigned to come up with a method of analyzing textsystematically that would allow us to determine sortof the political ideology of the person whoconstructed that text.And I was not able to do that as an undergrad.That was too much to ask, but it gotme thinking about these topics.

  • 03:02

    GABE IGNATOW [continued]: And when I went to graduate school in sociology,I continued to have a kind of side interest in text miningand text analytics.I kept learning new methods from all over the world.And it was all a little bit disconnected.And while I was in graduate school, I wrote three--in graduate school and then over the next nine years--three solo published, solo authored publications that

  • 03:23

    GABE IGNATOW [continued]: used text mining methodologies.The first one came out in '03.It was a five-year project, and it was an absolute nightmareto write it.It shouldn't have been that difficult.And I knew, even back then, it shouldn'tbe this hard to do social science research basedon texts.That was a study of-- a really kindof quirky study of the jargon of high-technology industry.

  • 03:44

    GABE IGNATOW [continued]: And I'd found a book of jargon, and Ithought it was very interesting the kinds of metaphorsand language people were using in high tech.And I did a comparative study of jargon, industrial jargonsacross different occupations.And no one really cares about jargon,but social scientists are interested in methodology.And that was basically a methods paper.A few years later, I did another study

  • 04:06

    GABE IGNATOW [continued]: that analyzed transcripts of shipyard unions from the 1970sfrom Scotland.So I found this kind of natural language transcriptionsof these guys debating their kind of political strategiesand tactics with regard to the British government.And I did a text analysis of thatdriven by some basic theoretical questions and the sociology

  • 04:28

    GABE IGNATOW [continued]: of culture, basic cultural analysis type questions.And I tied it to two different text analysis methods involvingboth metaphor analysis and analysisof narrative sequences.It was another kind of cool or quirky methods paper.I did a third one like this, also a comparative study,of online support groups for overeaters or people

  • 04:51

    GABE IGNATOW [continued]: with problems with overeating.And in this case, there was a religious, sortof more conservative group.And then there was a more secular group.And I scraped data from the web.This was published in '09.It was mostly done in '06 and '07 and used some differentkinds of software, different kind of analysis,to say something about how members of the two differentcultures, while facing the same sort of set of challenges

  • 05:14

    GABE IGNATOW [continued]: and trying to make changes in their lives and their eatinghabits, use culture in different ways and socially bonded witheach other in different ways.And it had something to do with the role of culturein social bonding.So these were three quirky studies.I did them on my own as side projectsthrough graduate school and afterwards.And then I took a break from text miningand did some other things, worked in other areas,

  • 05:35

    GABE IGNATOW [continued]: and decided to get back into it six or seven years agoand began to work with some collaborators,including Rada Mihalcea, who's nowat-- she's my coauthor with SAGE.She's now at the University of Michigan.And Nick Evangelopoulos, who's with me at North Texas,he's a statistician.We started to co-chair dissertation committeesand work with graduate students on text mining projects.

  • 05:58

    GABE IGNATOW [continued]: So typically, a student will have a substantive researcharea, an interest of theirs.Nick will work more on the technology side or Rada.And I'll work more on the theory, research design side.And we've published a number of projects,one on human trafficking or mediacoverage of human trafficking and how human trafficking

  • 06:19

    GABE IGNATOW [continued]: journalists and their readers differin their interpretations of human traffickingand the kinds of topics they discuss when discussing, youknow, a very complex, difficult social issue.We did another project on newspaper reader commentson the Trayvon Martin controversyfrom a few years ago using topic sentiment analysis, whichwas another methodology.

  • 06:40

    GABE IGNATOW [continued]: And then we started writing these textbooks.Rada and I co-authored two text mining textbooks.The first is just called Text Mining.The second is An Introduction to Text Mining.And the idea was that no one should everhave to go through what I went through in the early zerosin terms of trying to figure out how to write a text miningproject that would be acceptable in the social sciences

  • 07:01

    GABE IGNATOW [continued]: that you could publish it, and it would make sense to people.And it would be considered a contributionto a specific literature.It was very hard in terms of research design and thinkingthrough the logic of inference of how these projects shouldbe set up.So I thought we should make things easier for peoplein those areas and also make it easierfor people to find the right kind of software,make the right decisions with regard

  • 07:21

    GABE IGNATOW [continued]: to their own research projects.So both of those textbooks with Rada Mihalceawere written to make things easier for studentsand professional researchers so they could move forwardwith their own projects and just dobetter work than we have done.I mean the more computational skills, the more softwareprogramming skills a person has coming in the better.But you don't necessarily need to have a lot of experience.

  • 07:42

    GABE IGNATOW [continued]: You can learn the different methods and software packagesonline or using YouTube or other streaming videokind of instructional videos.So it's not really about, to me, the technical skills coming in.Although, it's important to have those.A student who wants to do a text mining projector is interested in getting data from Facebook, Twitter,

  • 08:02

    GABE IGNATOW [continued]: news sites, or anything like thathas to set up the research questionand understand kind of the philosophical premises of whatthey're doing, the ethical side of what they're doing.They have to get the ethics straight and discussthe ethical dimension of what they'redoing with their institutional review boards.You have to cover that.It's not as straightforward as maybe some people would think.

  • 08:23

    GABE IGNATOW [continued]: And then set up a project with a researchquestion that you can potentiallyanswer using text mining methodologies.And I think that means not starting with the data.That's always a temptation where, OK, I'mlearning some cool software.I'm going to do web scraping.Look at how much data I can get.There are these cool Websites.And it's great to do that, and it's great to be excited,but you can waste a lot of time and go down the wrong trail

  • 08:46

    GABE IGNATOW [continued]: and have to start over.It's good to do that and, at the same time,know the theoretical literature, the theoretical backgroundof your question and what's been done in the pastand what people have concluded in terms of theoryand what's been done in terms of just empirical researchliterature.So you can make an actual contributionand not just create a shiny object,which is always fun to do.But there are lots of those.

  • 09:07

    GABE IGNATOW [continued]: So you want to get the social science rightin terms of theory and the empirical literature that'sout there so that you can make a contributionand make your decisions in an efficient way.[MUSIC PLAYING]

Video Info

Publisher: SAGE Publications Ltd

Publication Year: 2019

Video Type:Interview

Methods: Data mining

Keywords: cognitive structures; content analysis; data mining; ethical considerations; internet data collection; jargon; large-scale research; methodology; natural language processing; research design; research questions; Social media; Social network analysis; Social science research; text analysis; textbooks ... Show More

Segment Info

Segment Num.: 1

Persons Discussed:

Events Discussed:



Gabe Ignatow, associate professor of Sociology and director of Graduate Studies at the University of North Texas, discusses the use of text mining in social science research, including recent projects, resources, skills, and advice for those interested in using text mining methodologies.

Looks like you do not have access to this content.

Gabe Ignatow Discusses Text Mining in Sociology

Gabe Ignatow, associate professor of Sociology and director of Graduate Studies at the University of North Texas, discusses the use of text mining in social science research, including recent projects, resources, skills, and advice for those interested in using text mining methodologies.

Copy and paste the following HTML into your website