Skip to main content
Search form
  • 00:08

    [Researching Food & Gender in Children's TVUsing Content & Text Analysis]

  • 00:14

    BETTINA CORNWELL: I'm Bettina Cornwellfrom the University of Oregon, and I'm here with Eric Setten.He's a PhD student in his fourth year in our program,and we're going to tell you a little bit about his journeyas he started his research program.So about our program--we have PhD applications that come in January or so,and then February, March is the evaluation,

  • 00:37

    BETTINA CORNWELL [continued]: and students know a bit after thatif they're coming into the PhD program.That means that that summer before the PhD programofficially starts, they have the opportunity to work on researchor learn about research.And so we try to get them that jump start evenbefore their program by saying, can youthink about, can you imagine what your research

  • 00:57

    BETTINA CORNWELL [continued]: interests are, what topics we could start to investigate?And then you have to understand that in the PhD program,they develop their research toolkit.So they develop inferential statistics skills,or they develop skills with modeling--all kinds of things that they probablydon't have when they arrive.So the question is, what can you do to get started?

  • 01:20

    BETTINA CORNWELL [continued]: So Eric began with interest in food,and I also do research in food with children.And so he said, well, I'm really interested in videos.And I said, OK, let's learn more about whatchildren's videos have to say.I'll let Eric introduce himself and that whole idea.

  • 01:39

    ERIC SETTEN: Yeah.And it's not just because I was a broney.I actually have two daughters.And so we would watch a lot of cartoons together.And one of them being My Little Pony or BarbieLife in the Dreamhouse-- things like that.And I noticed that the types of foodthat they showed in these cartoons

  • 01:59

    ERIC SETTEN [continued]: were very different from the types of foodthat they showed in the cartoons Iremembered growing up that were targeted towards me as a boy.I watched Ninja Turtles.There was a lot of pizza.So I wanted to see if there was anything to it,and I wanted to quantify that.But I wasn't 100% sure how, and Iwas discussing that with Bettina,

  • 02:20

    ERIC SETTEN [continued]: and she told me to refer to the literature.So I looked at the literature.I looked at different content analyses that were out there,and I started down that path.And I got some really good rich data from that,and I got a pretty good feel for it.But unfortunately, I couldn't expand the data set to the size

  • 02:43

    ERIC SETTEN [continued]: that I wanted.It took me about an hour or two to code a 22-minute televisionprogram.I wanted to be able to code more, because I noted--I noticed that there was a wide variety in the amount of foodthat was mentioned in each of the programs.Even within, for instance, My Little Pony, one episode

  • 03:05

    ERIC SETTEN [continued]: would center around getting cupcakes to so-and-so.And obviously, there would be a lot more food in that episode.And then other episodes would have zero food in them at all.And so I wanted to be able to take whole seasons, and notjust cherry pick individual episodes,in order to make the case that this is a systemic thing,and that there really is more food-- more sweet food in girls

  • 03:30

    ERIC SETTEN [continued]: shows than in boys shows.[How did you begin to carry out the research?]

  • 03:40

    BETTINA CORNWELL: So Eric began with whatwe would imagine as a traditional qualitative contentanalysis-- so watching programming, using a stopwatch,writing down instances, developing a coding sheetthat they could then use and otherscould use systematically.And the initial study, I think, led him

  • 04:03

    BETTINA CORNWELL [continued]: to some information that allowed him to build outin a more advanced study--and I'll let him tell you about that.

  • 04:12

    ERIC SETTEN: So the initial work did show that there was--there were a lot more cupcakes in girls shows,especially the shows that were tied to toys.And I ended up focusing on shows that were tied to toys,because they were much more targeted towards either male

  • 04:33

    ERIC SETTEN [continued]: or female gender, because toy manufacturers--they don't make any apologies about it--they need to sell to a target audience, and marketing 101--segment, target.And they do that on gender.And so I started looking more at this toy-tied media.And with the girls shows, you see characters

  • 04:55

    ERIC SETTEN [continued]: with names like Pinkie Pie and Strawberry Shortcake, thingslike that.And with the boys show, there was a lot of action.And so one of the things that I noticed justby observing qualitatively that food became a big plotpoint for the girls shows.And even though you don't quite see that,

  • 05:18

    ERIC SETTEN [continued]: even in the quantitative data that we have,it was very plot relevant.The screens were often filled with imagerythat was just cupcake architecture,whereas with the boys shows, evenwhen there was food involved, oftentimes,it would be more realistically portrayed.

  • 05:43

    ERIC SETTEN [continued]: [How did you gather and further analyze data?]Beyond this qualitative research,I wanted to see if this played out across the vast landscape.And through different conferencesthat I was able to go to, through guest speakers

  • 06:05

    ERIC SETTEN [continued]: that the department brought in to talk to the PhD studentsand faculty, and through also my coursework,I started developing a skill set in statistics.I was able to learn more about new research methods,and one of the research methods thatgot me really excited about this was automated text analysis,

  • 06:26

    ERIC SETTEN [continued]: because you could take a large corpus of data--unstructured text data-- and make sense of it.I started exploring this option more,and then I found out that there wereways to get transcripts offline offof streaming websites, such as Netflix, and Hulu, and Amazon.

  • 06:47

    ERIC SETTEN [continued]: And also, fan sites had pretty good transcripts available.So that allowed me to develop a corpus of transcriptsthat represented 777.7 hours worth of cartoons, whichwas far more than I ever would have been able to code manuallyusing traditional qualitative methods.And so I was then able to use the automated text

  • 07:10

    ERIC SETTEN [continued]: analysis to separate all the words that appeared in allthe documents into something like 45,000 individual terms,pared that down to about 1,000 terms related to food,and then we created dictionaries of wordsthat were important in categories that we thoughtwere important-- primary one being sweets and desserts,

  • 07:32

    ERIC SETTEN [continued]: but we also looked at general foodto act as an overall barometer of how important foodis to these types of programs.Junk food was another one we wanted to look at,and vegetables, because that's somethingthat is promoted for healthy eatingin shows that are publicly funded,and also shows that are targeting preschoolers.

  • 07:56

    ERIC SETTEN [continued]: [How did the qualitative & quantitative analyses informone another?]

  • 08:02

    BETTINA CORNWELL: Eric's work then-- hehad a large enough body of various programming,movies, serial children's shows to startto say that this was a generalizable phenomena.And then he could take the variables that he coded--for example, was this program publicly funded or was it

  • 08:23

    BETTINA CORNWELL [continued]: commercial--and ask himself, well, were there indeedmore vegetables mentioned in the publicly funded programsthan there were in the commercial programs?And that turned out to be the case.So the quantitative research revealedin a large body of programming that, indeed, there's

  • 08:45

    BETTINA CORNWELL [continued]: a difference between publicly funded programsand commercially funded programs.There are, however, values that camefrom the qualitative research that you could not findin the quantitative research.For example-- and I love the way he portrays this--and that is the food, for example,with older children and vegetables ended up being--

  • 09:08

    BETTINA CORNWELL [continued]: yuck-- vegetables or broccoli.So you have a food count in the quantitative study thatisn't anchored by the evaluative assessment of that foodthat you find in the qualitative study.So the qualitative study said, well,if there are vegetable mentions, but those vegetablementions are jokes, if you will, or characterizations

  • 09:31

    BETTINA CORNWELL [continued]: you wouldn't want.And so what you found was that the early primary schoolprograms that were publicly fundedhave the wholesome mention of a vegetable,where your commercially oriented programs hadvegetables as part of a prop or a joke in the program.The beauty of the research was combining

  • 09:52

    BETTINA CORNWELL [continued]: that early qualitative research with the quantitativebecause they help us answer different questionsabout the data.And I think that leads to a more comprehensive understanding.[What were the findings?]

  • 10:11

    ERIC SETTEN: And yes, there were a lot more cupcakesin the girls programs.We found that there were about five times as many mentionsof sweets.So ice cream, cake, cupcakes--anything like that-- sugar candy--we saw a lot more of that, and wewere able to quantify that, even controllingfor these other variables that couldimpact the number of words that we

  • 10:32

    ERIC SETTEN [continued]: would see in the transcript--also adding to that from the qualitative research.Part of the reason for this is because thesewere very plot-relevant.Also, character names oftentimes were sweet words--Pinkie Pie, Strawberry Shortcake.So you just get this entire ecosystem

  • 10:53

    ERIC SETTEN [continued]: that is based off of sweets.And you do see that in the quantitative data,and then the qualitative data helpsyou explain and understand why this is happening.In the boys' shows, we see a lot less food,and part of the reason for that is you get a lot more action.You get fighting robots.You get drama, suspense, violence.

  • 11:17

    ERIC SETTEN [continued]: But with the girls shows, there was something elsethat you needed in order to draw attention, I think.And by combining those two, you can say,yes, this really does happen, and you can say,this is why I think this happens.[Were there any methodological challenges?]

  • 11:38

    BETTINA CORNWELL: Another value to that qualitative research isto understand the visual and how muchbase it takes, how long it's on the screen,because your quantitative text analysis does notanalyze the visuals.So when you have cupcakes as wallpaper,and cupcakes as primary visual, and on clothing,

  • 12:01

    BETTINA CORNWELL [continued]: and the color combinations are cupcake colors,if you have something that's very different frommentions and so that's again how the different approachesyield different understanding.[Why is the research so important?]

  • 12:23

    ERIC SETTEN: Some of those takeaways from the qualitativeresearch-- that sweets are a relevant plot point,in that they make up a mandatory food group for girls,and vegetables, just in general, or ucky--are important food messages that shouldbe taken into consideration when evaluating television shows.

  • 12:43

    ERIC SETTEN [continued]: So there are third-party evaluators,such as Common Sense Media, that lookat things like sex and violence, for concerned parents thatwant to shield their children from these kinds of messages,and have them watch shows that have wholesome messages.And currently, they don't evaluate food messages.They look at alcohol and drugs, but not food messages.

  • 13:04

    ERIC SETTEN [continued]: And so it'd be really great if a third-party watchdogwould start looking at this aspect of it.And there's a lot of variability within genresas well, so even for shows that targetgirls that are commercially funded,there's variability there.And so being able to look at a rating

  • 13:25

    ERIC SETTEN [continued]: would be really helpful for parents--that this show's going to tell my kids thingsthat I don't necessarily agree with,in terms of eating habits, and this showdoes a better job of telling childrenabout the eating habits that I agree with.And I hope that this research ultimatelywill start this conversation, especially since nowadays, kids

  • 13:47

    ERIC SETTEN [continued]: are streaming everything.The old Saturday morning cartoonsfilled with commercials in between--that is no longer the case.Saturday morning cartoons don't exist.It's all moved to cable, and now everythingis migrating towards streaming with or without, oftentimes,commercials.

  • 14:06

    BETTINA CORNWELL: So I think we can say, importantly,about Eric's research that it is in contrastto a lot of the research that has proceeded that is reallyfocused on the advertising surroundingthe media and the programming.But now, with that not being an option with streaming,we have oftentimes, the messages essentiallyare captured in the programming itself.

  • 14:27

    BETTINA CORNWELL [continued]: And it's important when this programming is alsotied to toys in-store and other ways in which communicationis taken to a child that is in programming,not in advertising.


Bettina Cornwell, Head of the Department of Marketing, and Eric Setten, PhD candidate, at the University of Oregon, discuss using content and text analysis to study food and gender in children's television.

Looks like you do not have access to this content.

Researching Food & Gender in Children's TV Using Content & Text Analysis

Bettina Cornwell, Head of the Department of Marketing, and Eric Setten, PhD candidate, at the University of Oregon, discuss using content and text analysis to study food and gender in children's television.

Copy and paste the following HTML into your website