Matt Grizzard is a communication scholar who studies how people relate to characters in entertainment media as a sign of how much they enjoy one story versus another. A guiding framework behind this work is “affective disposition theory.” What is that? Well, listen to the episode! We talk about this theory, how it helps us understand people’s reactions to what happens to characters in media, and what it means for the importance of entertainment in our everyday experience.
Transcript
Andy Luttrell (intro):
You’re listening to Opinion Science, the show about our opinions, where they come from, and how we talk about them. I’m Andy Luttrell. This month, it’s a privilege to share this conversation with Dr. Matt Grizzard. Matt is an associate professor of communication at Ohio State University where he studies entertainment media, and in particular, how our entertainment tastes are wrapped up in our sense of moral right and wrong. I just wrapped up a sabbatical year, and although my plan was to just take the opportunity to get work done at home in Columbus, OH, where I live, and not drive out to Indiana every week where I work, I realized that I’m just down the road from Ohio State, where they have a really great communication department. So I was lucky to get to spend some of my sabbatical time there. And in particular, I’m grateful that Matt Grizzard was cool with me crashing his lab meetings every week because he and I have similar science styles even though we can come at them from slightly different directions. It was a great year, so I was happy to hear Matt was up for talking with me for the podcast, and we popped over to the WOSU public radio facilities so we could do this in person and have it recorded very well without me needing to monitor all the buttons.
Anyhow, I wanted to get Matt to walk us through a big idea that guides a lot of his work—affective disposition theory. The gist is that it’s a way to understand what draws people into or out of a story by thinking carefully about who the characters are and what they’re doing. As an aside, one thing I loved about attending meetings for an entertainment media lab was how often they started with a casual debrief of the shows everyone was watching. Okay, you know what we’re doing here. Let’s jump to Matt and I getting settled in the studio…
Andy Luttrell:
Have you used the NotebookLM thing to create a podcast? Have you heard of this?
Matt Grizzard:
No, I have not.
Andy Luttrell:
Okay. It’s extremely disconcerting. So I was like, “Oh, maybe I’ll try this as a…” ‘Cause I had a little trouble understanding what do I prep to talk to you? ‘Cause I’ve been hanging out in your lab for a while and I’ve been kind of low-key keeping abreast of ADT and just what it is, ’cause it’s not a home theory for me.
So I’m like, “I don’t become an expert in this before I talk to you.” But I also don’t want to just free wheel it. And so, one of the things I did was I loaded just six of your papers into this Google AI thing, and I think its whole point is to sort of consolidate and help you find patterns. I don’t really know, I did this very quickly. But one of the things it does, and I think it’s kind of become known for is it will create a 20-minute podcast of two people talking to each other-
Matt Grizzard:
Oh, interesting.
Andy Luttrell:
… as an explainer deep dive into some concept. And so I listened on the way here to the ADT Intro podcast with these two people who don’t exist. And I was tipped off to this as a tool, again, because there’s a great show on YouTube called, what is it? Howtown, I think. And they did a thing recently where people had kept saying that these two people sound like AI bots. They’re like, “Are you guys real?” And so they had this very meta episode where they’re like, “How would we know if we’re real or not?” And one of the things they noticed, and I noticed in listening to the podcast it created is it injects a lot of that friction-
Matt Grizzard:
Oh, interesting-
Andy Luttrell:
… into the sound. And so there’s a lot of lip-smacking. There’s a lot of hemming and hawing.
Matt Grizzard:
Oh, interesting.
Andy Luttrell:
And they had a very good point. They’re very media savvy producers, and they’re like, “This is all the stuff you cut out. When you’re producing a podcast, you’re trying to cut out all of this stuff, but it’s like there is a way of authenticating, to make it feel authentic. And my biggest complaint was that the last five minutes were complete fluff. They kept trying to land the plane, talking about how much we don’t know, and like, “Oh, how cool is it that we know all this stuff?” And I was like, “There’s no new impact of this.” So yeah, so I am very confident that our conversation will be better than that one.
Matt Grizzard:
I hope so.
Andy Luttrell:
But it also was sort of, I feel a little okay about the content ’cause I loaded in the papers, so it’s not just kind of making stuff up, I don’t think. But they quoted something, and maybe this is a number you put in, which is 75% of something can be explained by character perceptions or something.
Matt Grizzard:
Yes.
Andy Luttrell:
Is that a number that feels right?
Matt Grizzard:
It does. It’s not just character perceptions though. We were able to predict, and I don’t think it was quite 75%, but we were able to predict 75%, around 50%, I suppose, I had have to go back and look at the paper exactly, of the variance associated with liking of the ending of a story.
Andy Luttrell:
Okay.
Matt Grizzard:
And that was just based on whether or not or how much on a continuum you like the character, and how much you are feeling that they’re experiencing either positive or negative emotions. So these two things connected together. If I liked the character a lot and I’m thinking that they’re feeling positive things, I really liked the ending and I can predict that pretty well. And so, it was one of the first times I’ve ever seen a perfect crossover interaction, where there’s no main effect of one of the variables, and it’s all predicted by the interaction. It was kind of cool.
Andy Luttrell:
Okay, so this is a good excuse to start from the beginning and set the ground rules. My impression is that ADT’s main contribution is that characters matter a lot, and who a character is and what they’re doing is essentially what a story is in terms of if I’m consuming a story, that experience is almost completely me tracking this person or group of people and what they’re doing, which that doesn’t feel very radical.
And I want to get your take on how much is this like, “Oh yeah, so you just defined a story,” as opposed to, “No, no, no, we needed to articulate this in this way,” ’cause before we really didn’t have a sense that this was what stories were. So maybe to backtrack a little bit about the origins of where this came from and the key insight is, and then ultimately what this whole theory is meant to do for us.
Matt Grizzard:
Yeah, so the origins of this are actually kind of cool. It was started by Dolf Zillmann and two of his students, Joanne Cantor and Jennings Bryant. And they were actually trying to explain why jokes matter, “Why do we enjoy a joke, particularly something that derides someone else?” And so they were getting the ideas about, “Oh, maybe it’s similarity, and if we’re making fun of somebody who’s like us, we don’t enjoy it so much, but if we’re making fun of somebody that’s not like us, we enjoy it more, and maybe even more so if we particularly dislike that person.”
And so, some of the early studies they did were comics showing a professor and a student at graduation, and one version was the professor putting a pie in the face of the student. And then the other version, it was the student putting the pie in the face of the professor. And they showed this to professors and students, and clearly you could predict who would like what there, but they also showed it to professionals in the city, bankers and stuff like that. And bankers also preferred the professor kind of ridiculing the student more so.
And so they were thinking it’s probably has to do with how much we like the character and not necessarily just similarity. And so they extended this from humor into drama, because they said, “Drama’s not just about ridiculing people, there’s also a lot of benefaction of others that we see, like somebody succeeding, somebody excelling.” And so they were trying to think through, “What is the best predictor of if I will like something good happening to a character or something bad happening to a character?”
And so they started to think about character liking as this big determinant of it, but then also the role of moral judgment processes in it. So I think you’re right. When you say that it’s really the characters that matter, I think if we were listening to any story and you don’t care anything at all about the characters, you don’t care what happens to them, you don’t care if it’s a good thing that happens or a bad thing that happens. You might feel a little sorry if it’s a bad thing, but if you really like that character, you might be moved to tears in a situation like that.
Andy Luttrell:
So before the Zillmann stuff, was this kind of the spark for understanding narratives in that, or was there other ways of thinking about this before then?
Matt Grizzard:
Yeah, people have been thinking about this forever. Aristotle talked about this, and it goes back to Ancient Greece and what stories are like and how they should be structured. I think the thing that Dolf Zillmann did that was really important is he brought it to the scientific community. And so, my advisor was his advisee. And so you feel like he’s my academic grandfather, and so I know all of these inside stories about everything that happened.
But basically, when they were first trying to publish this work, they were getting a lot of pushback from the journal saying, “Fine, you did a decent study. Findings are interesting, but the study of entertainment stories and narratives and things like that, that’s not really serious enough for a journal that is focused on scientific research.”
So we had theories about how stories work, and people have been thinking about this forever, like I said, but I think what Dolf Zillmann did was he was trying to bring it into the scientific realm. So, “Can I identify what are the basic building blocks of a story? And if I can, how good are those basic building blocks of predicting overall enjoyment?”
And I think that’s a cool thing about ADT, it’s a relatively limited number of variables that you’re using to try to predict something that is really varied across people. And so it really just breaks down into how much do you like the character, ranging from extreme hatred to extreme positive feelings toward the character, through indifference as well as in there? And then what’s happening to them in the story? Are bad things happening to them? Are good things happening to them? And even do you just anticipate that bad things might happen or good things might happen? That’s kind of where suspense is coming from.
Andy Luttrell:
So, to test it in the scientific realm, what did that look like in those early experiments?
Matt Grizzard:
Yeah, so in some of those early experiments, what they were doing was they were testing with children a lot. So they were doing simple stories. And so it would be, one of the famous ones is a story about two princes. And so, there’s a good prince and a bad prince, and the mother of the two princes says, “I want you to divide the kingdom up equally between you two.” And clearly, the bad prince betrays the good prince.
And at the end of the story, essentially, the good prince gets freed from his prison and has an opportunity to then meet out vengeance upon the bad prince, who is his brother. And they would manipulate things like, how does the story end? So, the buildup is all the same and we’ve got a good character and a bad character.
At the end of the story, does the good prince punish the bad prince in exactly the same way, which would be like equitable retribution? Does he punish him even more, leading to even worse situations for the bad prince than he even faced himself, so torturing him instead of just banishing him? Or does he forgive the bad prince entirely and just say, “You’re my brother, I’m going to divide the kingdom up altogether and everything”?
And so, what they found was that that over retribution ending was too much for most people. Young kids loved it. Young kids, the sterner, the jester. “Bad person gets punished, great. Yeah, heap it on.” But older kids didn’t really like that ending as much. Everybody hated the forgiveness ending. And that ending with it was where it was equal justice, that was the one that tended to be preferred. You saw more smiling and things like that from the kids when they were doing facial coding of the films that they did of the kids’ faces and stuff like that.
So that’s kind of what the scientific study is. We take a story and we try to manipulate it so that we have a very clean similar story between different features, but in one version the main character is moral and another version they’re immoral, or in one version they’re competent and the other version they’re incompetent.
And then also altering what’s happening at the end of the story. So, you couldn’t really do this with Cinderella, for example, but imagine if Cinderella was evil to all the mice at the beginning and was constantly shooing them away and scaring them, and was just a terrible person while she was doing all of this cleaning.
And then at the end of the story, instead of her getting the glass slipper on her foot, her sisters get it. So, you can’t do that with the story that everybody knows, so we write them and try to hold as much constant as possible between the different versions. So it’s akin to medicine studies, where you have a placebo and a control and different treatments and things like that.
Andy Luttrell:
So expanding it out, you end up with a handful of different versions of what is ultimately the same story in terms of its constituent players, but then ultimately you want to know, “Did you like the story?” That’s why I wonder, what is the problem that we’re most trying to solve? Is it what stories do people like when they’re done? Is that sort of the key judgment that we’re trying to predict as accurately as possible?
Matt Grizzard:
I think it depends on how you want to break it down. And so, I see ADT, affective disposition theory is kind of trying to predict three different things. And so, one of them is kind of like the evaluation of characters. So, there’s lots of things that make characters likable or dislikable to people. And what is the central version of that?
So you could think about in psychology, we have person perception research and what makes people perceived to be competent or warm and things like that. Stories tend to be an area where you focus a lot on morality and what’s approval, really, of a character’s behavior. “Do I approve of what the character’s doing? Do I disapprove of what the character’s doing?” And so that’s one version of it, where you’re trying to predict feelings toward the character.
The second element that ADT is trying to predict is moment to moment emotional responses. So, films use suspense constantly in what I like to call, and what the discipline and the literature has called micro-plots. So you go to a Batman movie, you know he’s not going to die at the end of the movie, but you also that there’s going to be moments throughout the movie where you feel nervous, something bad might happen to Batman.
And so stories will set up these situations where we’re kind of anticipating what’s going to happen next. It’s the damsel in distress tied to the train tracks. It’s Batman being tied up by the Joker, it’s Lex Luthor having Superman on his knees with kryptonite in his hand. It’s all of these moments where we have a light character in a really bad situation, and that produces anxiety for us. It’s akin to seeing a good friend in a bad situation. It produces these feelings of, “Oh no, I hope this character ends up okay.” And so that’s the second component. It’s the moment to moment emotionality of it.
And then finally, the third thing is the enjoyment, the ultimate response to the story. “Did like the story? Are you talking about it afterwards? Would you want to go see a sequel for this?” I think this is where industry is probably a lot more interested in trying to understand this because they’re putting a lot of money into movies. “What plot elements do I need here? How do I improve the plot elements?” And so I think understanding the enjoyment of that resolution. So we’ve anticipated something. When we see the resolution to that, how do we feel in that situation as well? So that’s kind of the third thing it’s trying to predict.
Andy Luttrell:
The industry thing makes me think of, “Well, how are they…” My stereotype is that focus groups are the way to go, where you’re really looking for people spontaneously volunteering, like, “I didn’t love this.” Is there any sense that when you hand the keys to people, they’re also talking about, “I didn’t like that guy, but good things happened to him,” or is this theory helpful at carving these stories at their joints when you’re designing them from the beginning? Or is it also people are like this truly is mapping onto how people are engaging with stories, just in their own experience?
Matt Grizzard:
I think it’s a little bit of both. So when we look at what stories exist, we see a very biased landscape. There aren’t a ton of stories that are really light characters experiencing really terrible things at the end. Most stories have light characters experiencing positive things at the end and disliked characters being thwarted at the very end.
There are still stories that do have bad things happen to light characters. So, Saving Private Ryan’s a great example of a very thought-provoking movie, and at the end, spoiler alert, it’s a little sad. Tom Hanks dies and then you have Matt Damon as an older gentleman trying to figure out, “Was it worth it? All these soldiers died to bring me home, was this kind of worth it?”
And so it’s a really kind of sad thing in the film when Tom Hanks character dies, because you’ve grown to really like him throughout, you’ve seen what a good leader he is, how thoughtful he is for his soldiers under his command. And so stories will present us with these kinds of sad things. It’s not like every story ends happily, but those things are typically kind of brought in for some bigger purpose.
He’s dying at the end, but also he dies in a way that we can find value in it, and so kind of like this meaningful search. So, I think a lot of times, story creators are using those moments to prompt something else and in the audience, get them to think more about it outside of the story world, or try to reason through why this is still kind of an okay situation.
‘Cause he could have died a lot of ways. He could have died off camera where we didn’t see anything whatsoever, he could have been tortured to death. And Matt Damon’s character could have died as well, but the story didn’t end that way. It tries to present some kind of upwards turn almost, and some way to find value in it. So, when they do deviate from that ending is positive, there’s still something positive about the ending usually, where you’re seeing some kind of virtue that’s represented that’s not just, “Yay, we won.”
Andy Luttrell:
And if the movie just ends at that turn and is just the end, goodbye, that wouldn’t be quite so satisfying. So it’s still evidence that people are really keyed in to who everyone is and what’s happening to them, even if you go like, “Oh, well this defied the thing that you wish happens. That the fact that we did that in the story is still an important piece of the puzzle.” Yeah.
Matt Grizzard:
Yeah. There are some stories though that just don’t have anything redeeming happening at the end, and they just tend not to be very well liked. They’re very polarizing. And if they are liked, it’s usually for a different reason, where it’s like, “Oh wow, that was super interesting. I’ve never seen a story happen that.”
One of my favorite ones that I talk about when I teach this to students is The Mist, which is a Stephen King novel. It got turned into a movie, and basically the dimensional portal has been open and monsters are killing everyone. At the end, a father, his son and three people are trying to escape by driving out of the mist.
And they drive, drive and drive, and they finally run out of gas. And the father’s like, “Oh my God, I’ve got a gun. I’ve got four bullets though. And so I can make it so that we don’t get ripped apart by monsters,” which is maybe a positive thing. That’s maybe the best situation it can be. And so it’s really sad and terrible scene.
But then after the father has killed everyone and he can’t kill himself, he jumps out of the car and he’s yelling at the monsters to come and kill him. And then at that point in time you realize that they had been driving away from the army the entire time. And the army and the tank comes off right behind him, and if they waited five more minutes, they would’ve been saved. Or if they had driven in the opposite direction, they would’ve been saved.
And it just ends basically with him crying in the middle of the street. And there’s very little to feel good about in this film, and it’s very hard to see any kind of positive ending of this. And so there are stories that go full dark like that, but there’s just not that many of them.
Andy Luttrell:
People can make the story they want to make, but as a social scientist, the question becomes, “How good of a choice was that?”
Matt Grizzard:
Yeah. And imagine being a movie producer and somebody comes in and they’re like, “I’ve got two stories here for you. One is the super light character, and in the end they triumph. And this other one, it’s a super light character, and everything terrible happens to them and it’s just a real downer, which one will you buy? Which one will you option for me?” They’re going to pick one of those, I think. Yeah.
Andy Luttrell:
So, the discrepancy between the movies that already exist or the stories that already exist, and then it’s self-selecting. You’re trying to maximize enjoyment, presumably, in most cases. So, it’s got me thinking about, especially with the kid studies that launched a lot of this. I’ve been thinking a lot in recent years about how stories are really how we convey moral lessons. That’s sort of like the vehicle through which a culture preserves its sense of moral value.
And so, there’s sort of then an open question to me here. If kids can quickly intuit that, “Oh, the good thing is supposed to happen to the good person,” how much wiggle room is there for teaching new lessons, as opposed to kids are just kind of like, “I get it, bad things are supposed to happen to bad people, good things are supposed to happen to good people. That’s what makes things right and just”?
Matt Grizzard:
Yeah. And in a way too, it’s really hard with kids because there’s a lot of research showing that they don’t take the right moral lesson from it, or they take a really kind of bizarre moral lesson from it. So, one example I’m thinking of, there was a study that was done, and it was found basically that it was a Clifford the Red Dog story, and one of the dogs is missing a leg or something like that.
And so they asked kids afterwards, they were like, “What lesson did you learn from this?” And it wasn’t like, “You should include everyone,” or, “We shouldn’t signal people out based on things like that.” It was, “Be nice to one legged dogs or three like a dogs.”
And so, you can use these stories to teach moral lessons, but I think with kids it just becomes hard because you never know what they’re going to pick up on it. I don’t know if you remember, but I used to watch a lot of He-Man, Masters of the Universe when I was a kid, and GI Joe as well. And they always ended with a very clear moral lesson. And so He-Man always looked into the camera and he was like, “In today’s story, so-and-so did this, and that was a bad thing because of this. And then we did this because of that.”
And so, if you have something like that to hammer home, maybe it helps a little bit, but with kids, I think it’s hard to teach more complex moral lessons. I think a lot of time what they’re going to get from a lot of stories is, “The good guy wins and so that’s what makes him good. The bad guy loses, and so that’s what makes him bad.”
And then even worse, I think sometimes these stories have such, it’s not this way anymore, but when I was a kid growing up, it was lots of, “Heroes are violent and that’s how they solve their problems.” And so you could see an extrapolation of that, “That violence as a way to solve problems and heroes do it and that makes it okay.” And so I think there was lots of little things that could creep in that would be kind of problematic.
Andy Luttrell:
Yeah, I’ve been surprised at how quickly my daughter keyed into good guys, bad guys. And when I taught her the word villain, she was like, “This is a very useful word.”
Matt Grizzard:
That’s great.
Andy Luttrell:
She’s been impressing a lot of adults by talking about who the villain is. And most of the time it’s, “This guy’s not very nice.” We watch a movie and she goes, “He’s not very nice.” And then she’ll also be like, “Does that movie have a bad guy in it? ‘Cause I don’t like bad guys.” And I go, “Well, that’s a story. If it’s a story, there’s going to be a bad guy.”
Matt Grizzard:
Yeah, there’s usually somebody.
Andy Luttrell:
And it sort of is sort of relevant here because you go, “If there’s not, then what do we have?” There’s some Daniel Tiger in our household, and that does a job, but telling a story is not maybe one of the ones.
Matt Grizzard:
Yeah. Yeah, I know what you mean. It’s funny you bring up your daughter. We did a study where we were trying to see to what extent we can activate these perceptions of good person, bad person using tropes that are used in narratives all the time. So in Harry Potter, Severus Snape seems like a bad guy because of how he’s dressed, his stern look on his face, all of these things. And then you have other stories where you have characters that look like a good guy, but they tend to be a bad guy, so The Boys and Homelander is a great example of this.
So we were doing a study to see if just the appearance of characters consistent with heroic versus villainous kind of schema would activate perceptions of relevant character types. And so what we found was, yeah, they do. A heroic looking character, earth tones, kind look on their face, things like that is perceived to be more moral than a character who wears all black and has stern looking eyebrow features and things like this.
So when we were creating these stimuli, I was showing them to my daughter who was probably four or five at the time, I’m guessing, just to see, “Am I on the right track here?” And so I was like, “What do you think of this guy? Is this a nice guy or mean guy?” And I was like, “Mean guy.” And I was like, “Okay, what about this guy? Is this a nice guy or a mean guy?” And so she was really helpful in pilot testing some of these initial stimuli.
Andy Luttrell:
So it sort of brings up the work that you have done in this area. And my impression is that at least some of it is driven by what I might call stereotypes in the process. So that my notion of who’s good and bad, I’m not always learning directly from what I’m seeing, but I’m sort of guessing, based on these more superficial features.
And so, you mentioned that we get this information in all sorts of ways. What are the kinds of ways in which people are getting, filling in that calculus as they watch a movie or listen to a story or something to sort out, “Who’s good, who’s bad, and is what’s happening to them something I can stand by or not?”
Matt Grizzard:
So, visual appearance is a big one, and so this is where people really lean into it, especially casting directors and things like that. I can’t imagine ever John Malkovich being cast as this lovable school teacher.
Andy Luttrell:
Family man.
Matt Grizzard:
Family man. He’s more likely to be cast as a villain. So visual appearance is a big one, that’s a big driver of it. There’s also a protagonist biased I think, where stories are presented to us from a particular angle from the very beginning. And so, even though we don’t necessarily might consciously realize, “This is who I should be rooting for,” the way the story is introducing the character makes us kind of tipped off to like, “Oh, this is somebody who’s good or positive and this is who I should be rooting for.”
So the repeated patterns that we observe over and over and over in stories, I mean Campbell talked about this as the monomyth and the hero’s journey. Kurt Vonnegut did work on this where he talked about the shape of stories and how stories follow specific shapes of starting bad, but moving towards good, or starting good and then moving towards bad. So I think the repeated patterns we see help us figure out who we should be rooting for. I think narrative creators are well aware of those as well, even if they don’t know what ADT is, for example, they know these types of things. So those are the big ones.
What we found though in a lot of our work is that those types of cues are easily overridden. So, if we write a story and we try to point out that this really good, as soon as you see that character do something that you might disagree with a lot, you’re very willing to update your opinion of that character in the story. So if you find this character and, “Oh, he seems like a nice wholesome guy and he has all the same interests that I do,” if all of a sudden you find out that he’s running a dog fighting ring, that might be a bridge too far and now all of a sudden you’re willing to update your opinions about him. So we do engage in this kind of updating process.
It’s another one of those situations where the scientific study of it kind of has to take place with carefully created narratives, because again, when people are writing stories, they’re not engaging in the entire universe of possible actions. They’re trying to think through, “How do I keep the audience engaged with this character?” And so looking at those types of updating processes with naturally occurring stories becomes kind of difficult. You have to really kind of craft your own, because you just don’t have many where I think fairy godmother doesn’t show up or something like that in the middle.
Andy Luttrell:
So I wonder if you could talk about you’ve crafted a version of this where you can do these things. And it’s a good also maybe encapsulation of maybe there are things about ADT that we haven’t… It’s been around a long time and it’s gone through a lot of shifts, and so I’m not good at tracking whether we’ve covered the current state of it so far.
But very recently, you sort of developed this paradigm to be able to zero in on, “Okay, let’s actually carve this all at its corners and really put it to the test.” And you don’t need to walk us through all of those very complicated path diagrams, which I got real overwhelmed by. But just generally, what did you have to do to create a universe of stories that allows you to actually test some of the predictions of this theory, and then what are kind of the key things that you find when you do that?
Matt Grizzard:
Yeah, so what we did was basically we took a story kind of structure that is pretty common, where you have somebody trying to achieve something for themselves. And so in the context of our story, it was a character named Amelia, and we described her as this high school student who’s really good at computer programming but not much else, and she needs to get out of her small town by going to college and studying computer science is the goal. And so, there is a test coming up, and if she does well enough on this test, she will get a slot admission into the computer science program there and that’ll allow her to get in, even though her grades might’ve kept her out. So that’s the big broad picture of it.
And so what we did was we wrote different scenes for different components of this. So we have just an exposition at the beginning, “This is who Amelia is and this is what she’s trying to do.” Then the first scene has her interacting with another student and just having a conversation. And so 99% of that text is the same. The only things we changed was she’s talking to another student and she was either described as her best friend since kindergarten or her bully since kindergarten.
And so, after we had this interaction, then we were able to measure things, “How much did you approve of Amelia’s actions towards this student? How much do you like Amelia?” And using only those two kind of variables and our manipulation, we’re able to explain a ton of variants, right? Amelia, who’s kind to this other student, really well liked. Amelia, who’s a bully, really despised.”
And the so the next scene is Amelia has forgotten her programming notebook to study for the test, it’s the next day. She desperately goes to school to see if the custodian is still there cleaning up so she can get to her locker and get her book. She does. She gets into the school, and as she gets her book, she notices an envelope on the ground and it’s a test copy for the next day. And so she could take it with her and cheat, essentially.
And so, at that point, we basically froze the story again and we’re like, “What do you want Amelia to do in this situation? What do you think she will do, based on her previous behavior, and what do you think is the right thing to do?” There’s almost no variance in what is the right thing to do. Everybody said she should leave the envelope there.
And then, “What do you think she will do?” That’s very much predicted by past behavior. So if she was kind to the girl in the first scene, you think she’s going to leave the envelope and not cheat. If she was cruel to that girl in the first scene, you think she’s going to take the envelope and cheat. But what was kind of interesting in our findings was that Amelia, the more you liked her, the more you wanted her to cheat ’cause it’s going to kind of help her out.
So again, we had this scene where we could stop it, stop the story midway, see what’s being predicted there. Then the next scene she decides either she’s going to cheat or not cheat, and that was randomly determined. And we described it as she feels good about her choice regardless, so implying that she’s comfortable cheating and she also feels good about not cheating. And so then we were able to measure, again, liking it.
So, we actually see that if you liked Amelia before, you like her less if she cheats. And if you disliked her before, you like her a little bit more if she doesn’t cheat, if she leaves the test where she found it. So that allowed us to, again, see change in dispositions, like the liking of the character and perceived morality, based on what they’re doing in the scene, and we can see it updating between acts.
And then finally, at the end of the story, Amelia either is accepted to the school or rejected from school after she gets her letter in the mail. And so, we end up with eight different versions. We’ve got kind Amelia, to the student who cheats and who gets into school, who doesn’t cheat and doesn’t get into school.
And so by using it as this kind of time ordered experience, we know that we’ve satisfied the temporal order of causality there, that criteria of causality. And then by having these random assignments at different points, we also know we’ve ruled out a lot of non-spurious explanations and things like that. So, crafting these types of stories, it takes a long time, especially ’cause this story ended up being around 1,000 words or something like that.
And so it took people a decent amount of time to read through it and everything. And it took us a long time to get it so consistent that it was just one or two words being changed in each component to make it seem like, “Oh, it’s actually the same.” At the end, when she gets in versus she doesn’t get in, she cries in both, except in one it’s tears of joy and in the other it’s tears of devastation. ‘Cause you wouldn’t want to say she’s super happy in one and she’s super sad in the other, and there’s not the same expressiveness of it.
So it’s really hard to write these things. It takes a lot of time. But I think this is one of those areas where you need participants to really care about the character, because I thought of lots of ways to try to do this in repeated measures, designs, or within subjects manipulations and things like that, where we’re getting the same thing from the same person over and over and over again. And you can do that with small snippets, but if you were to have 15 stories that are 1,000 words each, the participant’s going to hate you by the time they get through to the end of that.
Andy Luttrell:
And so by being able to do this, you’re saying because you can start and stop the story, you’re able to get that within person change and reaction without having to show them a million different stories?
Matt Grizzard:
A million different stories, yeah, yeah.
Andy Luttrell:
So, my impression is that the impact of these little tweaks was relatively big, at least for this kind of work?
Matt Grizzard:
Oh, yeah. The [inaudible 00:33:12] I think on some of the manipulations of whether she’s a bully or not is two or something. It’s absurd, the effect sizes that we were able to generate with these really small tweaks.
Andy Luttrell:
That’s what’s crazy, right? It just goes to show how much that is the driver of things. If this stuff didn’t really matter, you’d go, “Oh, I guess people are kind of nudged to like it a little bit more,” but isn’t a, , “My enjoyment hinges on these three details that you’re messing around with.”
Matt Grizzard:
Exactly.
Andy Luttrell:
So this is a hypothetical story. I asked you to come prepped with a real story that showcases this, and so I won’t make you have done that for nothing. So, in this very controlled situation, you can sort of see people reacting to this, and we’ve sort of given a handful of examples along the way, but is there sort of a good go-to story people recognize, where if you have an ADT hat on, you go like, “Oh, this makes sense that this is why the story works the way that it does”? It’s not the plot. I think writers or beginner writers are tricked into thinking like, “Oh, you need to have a good premise and a good plot,” but maybe if character is everything, this can help us appreciate that.
Matt Grizzard:
Yeah, so I agree. I think there’s a big emphasis on plot, and there are books on, “How do we make a likable character?” One’s called Save the Cat, and it’s about, “This is how you generate liking towards the character very quick.” But a great story that I think really adheres to ADT and shows you both the global predictions of how much you’re going to like the story, and then also the scene to scene kind of components, is Cinderella, the 1950s Disney version of Cinderella.
It starts with Cinderella waking up in her room and all the birds love her, and she’s super kind to them, and she’s kind to all the little mice that live in the house, and she’s kind to the dog, and she’s kind of the horse, and then she’s kind to even the cat, she’s kind to. And they find a new mouse, Gus, and she makes him a little outfit for him and everything.
And so, you see this character that is just unbelievably kind, despite the terrible situation she’s in with her mother being dead and being a slave basically to her stepmother and evil stepsisters. And then the interesting thing about it is the micro-plots that happened throughout this. So it’s, “Oh, there’s going to be a ball, but you can’t go. Oh, but you can go if you get your chores done.”
And so it’s like, oh, okay, now she’s doing her chores and she’s singing and it’s lovely. And then, oh, she got her chores done, but she doesn’t have a dress to wear. Oh, but her mice friends have made her a dress, and so now we have this positive feeling. Oh, but her stepsisters tear the dress to pieces and now we’re devastated again. Then the fairy godmother comes along and it’s like, “I’m going to make you a new dress,” and we feel positive again because we like Cinderella. If we didn’t care about who Cinderella was, all of these events that are happening are not that important. And if we hated her, we would be like, “Good, I’m glad that her stepsisters tore that dress off of her.
Andy Luttrell:
“What are you doing mice? Stop it.”
Matt Grizzard:
Yeah, exactly. And then the ending, it really builds up to this, where she gets locked in the tower and she’s trying to get out, and the king’s helper is trying to get all these shoes on these women and everything. And he’s trying to leave and she finally escapes thanks to her mice friends. And then she’s getting ready to try on the shoe and the stepmother trips the guy and it shatters into 1,000 pieces, and so our heart drops again. But then she has the other shoe with her, and so it’s this repeated pattern.
And there’s this great scene at the end where Cinderella’s joy, it comes right after you see just a devastated look on her stepmother’s face. There’s like a zoom in shot to how upset her stepmother is that she’s going to get to try on this shoe and leave and she’s going to have to find a new housekeeper. And so again, it’s just how much we like Cinderella and then how much we hate her stepmother getting combined together. If you’re lukewarm towards both of these characters, it’s not a big payoff. If you really like one and really hate the other, it’s a huge payoff.
Andy Luttrell:
It’s a good example too, that this isn’t just like, “Oh, I just happen to like it when these good things happen to good people.” It’s that, “I want good things to happen to good people.” And so it’s really kind of this forward momentum as opposed to just a passive, “Oh, I see a scene and I go, ‘I like this one. I don’t like this one.'” It’s that, “I’m actually invested in an outcome that is going to satisfy my just craving for justice.”
Matt Grizzard:
Absolutely.
Andy Luttrell:
So why do we do this? Why do we care? You mentioned, I think I can understand the editors in the early days of those journals being like, “Oh, cool, you want to talk about story characters.” To you, why is this important beyond the scratching an itch of, “I wonder why people like this movie”?
Matt Grizzard:
I think what this reflects is it’s not just stories where we’re doing this. I think this is something about who we are as people. I think you see this everywhere in your life. I think this explains sports, and effective disposition theory has been explained with sports. If we think about, we’re on OSU’s campus right now and I’m a professor at OSU, I like to tell my students the only thing better than OSU winning is Michigan losing. And if you’re a big fan, those things are huge payoffs. I think we saw it this year with the Super Bowl as well, which characters or players are being portrayed as good people versus bad people. Who deserves it more, who deserves it less? So I think we see it there.
So I think stories reflect us. They’re like a mirror and they allow us to kind of test things from a psychological or communication perspective that we wouldn’t be able to test in other situations. So if I asked you for example, “Are you racist?” Everybody’s going to say, “No, of course not.” But maybe I could show that you dislike films with Black characters more than you dislike films with white characters even with the same plot elements. And so, I think it’s a way for us to identify these biases that are playing a role in our everyday lives, but they’re also driving everything that we’re thinking about with stories.
Andy Luttrell:
And you probably guessed that this is where I was going to go, is this different than just understanding people in our world? That’s sort of my social psychologist glaring at media’s communication work is, “Is that different? Just ’cause it’s on a TV screen, is it different?”
And also, to turn it back on me, is what we’re doing any different when we use a bunch of vignettes to… ‘Cause I have a paper looking at I update my impression of a person when they’ve done something bad and I thought they were good. And I’m quick to say, “Oh no, now you’re a bad person.” But that wasn’t a real person in my study, I essentially created a story about a person. Except when I look at the data, I go, “See, this is how we react to the social world around us.” And when you look at it, you go, “Well, this is how we react to stories.”
And so is there anything unique about this sort of mediated context that entertainment lives in that would make our engagement with these kinds of variables different than if this were about, “My family member is going through the same exact struggles”? Am I approaching that any differently?
Matt Grizzard:
I think in some ways, we’re not approaching it any differently, but I think in other ways, stories do things that real life doesn’t do, or they allow us to see things and give us the opportunity to make judgments that we wouldn’t make in real life. I think Breaking Bad is a great example of this. So whenever I tell people about ADT, usually the first thing somebody says is something like, “Oh, I like Tony Soprano though, explain that.” Or, “I like Walter White, explain that. He’s a really bad person.”
And he is at the end of the story, but when he started, he was really kind of a good person, right? High school teacher, passionate about his job, good at it as well. We see him in the classroom teaching and things like that. And he gets a devastating cancer diagnosis and he has to try to find some way to generate income to pay for his cancer treatments. And so we start with a likable character-
Andy Luttrell:
Cinderella.
Matt Grizzard:
Yeah, exactly. Exactly. Right. And he gets worse though. Where Cinderella’s doing all these positive things, he starts selling meth. And it’s weird that people liked him so much at this point. You hear about your neighbor setting up a meth lab in the garage next door, you’re not like, “Oh, I wonder why he did that. I hope that that he’s doing okay and this isn’t because he got diagnosed with cancer.” You’re just pissed that somebody has set up a meth lab in your neighborhood. So stories give us this opportunity to think more deeply about characters or be kind of okay with these kinds of settings.
So I think that’s kind of different. And I think stories do it in an interesting way. They serve as kind of like a microcosm. And so, if you look at Walter White in season one, who is the antagonist? It’s Krazy-8, who is a low-level drug dealer and Walt’s better than him. Well, when Walt kills him at the end of the season one, he’s gotten as bad as him, but the next antagonist isn’t another low-level drug dealer. It’s now all of a sudden Tuco Salamanca who’s connected at the Mexican cartels.
And then by the time Walter gets as bad as Tuco, we’ve got a new antagonist, which is the cartels itself. And then by the time he’s gotten as bad as them, we’ve got a new antagonist. It’s Gus Fring who’s a total psychopath and sociopath, and willing to kill his best friend. And so then Walt gets worse and worse and worse, but the story always is giving us this daylight between him and who he’s up against.
And so by the end, Walt has gotten so bad that he’s betrayed his colleague Jesse Pinkman and left him in the arms of the antagonist. But at that point, the antagonist has to get so bad that they’re literally a Nazi. You rely on like, “Wow, we’ve made Walt so bad, that if I want any kind of daylight between him and the character, I’ve got to go to Nazi biker gang, essentially.” And then by end of the show, Walt does die of cancer, but not before he comes back for Jesse to rescue him. And so we see this moral upturn at the end.
So I think that’s what makes stories a little bit different than real life is they operate as if it’s real life. And because of who it’s focusing on and who we’re thinking about, it gives us these opportunities to engage in slightly different moral reasoning processes. It’s not like we have a media lobe of the brain. It’s all the same thinking, I think. But the context that they’re in, and the lack of direct ramifications for us in the real world gives us a little bit more of a latitude of acceptance.
So, I think it’s a little bit looking at moral judgment processes in a cleaner way in some ways, because it allows you to cut away all of the excess, right? It’s like when Harry Harlow was doing his experiments on connection and whether or not it’s food or comfort that causes these rhesus monkeys to attach to their mothers and dominant thinking at the time, whether it’s a food thing. And then he shows that, “No, actually, it’s cuddling, it’s comfort, it’s those things matter.
And it’s not like any of this stimuli looked like real rhesus monkeys. They were little wire things, and with terry cloth on them and weird looking eyes. If you look up pictures on them, there’s no rhesus monkey in the world that thought that was a real rhesus monkey. But the utility of stripping away everything else and having just these very clean manipulations, I think is something that stories can do, that shows a lot of bridges between communication media and psychology that would allow us to answer deeper questions.
Andy Luttrell:
And yeah, it’s sort of like a safe playground to play with these ideas. I would be comfortable watching what happens on Breaking Bad unfold, but if I had to experience that by going to a neighborhood where this stuff is going down, I’d be like, “I don’t want to.” And it’s different because it’s not a thought experiment where I’m having to sort of invent everything that’s happening, but I’m being fed things to consider.
And it seems like that is also maybe what makes a story unique, is that I get to be like, “Okay, you’re giving me a safe space to explore what it would mean to do this, that, or the other, and you forced me to grapple with something that I wouldn’t have invented on my own.” And presumably, that plays out in reality.
Although, it’s like all the work on moral dilemmas is sort of the same thing, of these hypothetical situations. And it’s unclear, are those powerful because they allow for this kind of safe thinking, or are they misleading because I’m not actually invested in what happens?
Matt Grizzard:
I think what really makes moral dilemmas interesting is when we’re comparing them to one another. And so it’s like an individual moral dilemma doesn’t tell us a whole lot. It tells us that maybe people are utilitarian in this particular instance, but when we compare the trolley dilemma to the bridge dilemma, that’s where we can see, “Oh, okay, what’s going on here that’s weird because the ending is the same, five people saved or one person saved?” But in one situation, you’re pulling a switch, and the other situation, you’re throwing somebody off a bridge. So I think that’s another thing, where stories are useful to compare to one another, and not necessarily to think about them as these global things.
Andy Luttrell:
That’s a broader issue just in general about how we do things of there’s questions about when we measure people’s opinions. There’s, “Am I getting exactly what their opinion really is?” And I go, “I don’t really care. What I care about is would they have said something different when prompted differently?”
And so I don’t actually care that I’ve nailed your opinion perfectly. I care that I can detect when it’s going to shift based on some other information. And so here, it’s like, “Okay, I don’t know, would you really push this guy off a bridge? That’s not really what I care about. I care about, would you have come to a different conclusion if I told you a different story?” And that’s what you’re saying?
Matt Grizzard:
Yeah.
Andy Luttrell:
Yeah. Actually, I feel like maybe we’ve talked about this before, of what if I told you a story but I said it was real or you know it’s a story, does that change the game at all? Do we know much about that, actually?
Matt Grizzard:
I feel like we think we do, and we don’t. I feel like a lot of studies have been done where it’s like, “This is real or this isn’t real.” But I think it’s one of those times, where as soon as you tell somebody, “This is real,” you’re leading them towards something and do they really buy it or not? And as soon as you tell them, “This is a fictional story,” especially, they’re like, “Okay, that’s changing how I’m thinking about it,” and things like that.
And you don’t go into Breaking Bad being confused about whether it’s real or fictional. You know it’s fictional. So, I feel like we think we know a lot about that, but I don’t know that we actually do. Every study that I’ve kind of seen that has done this has tried to use some kind of prompting of participants that this is real or this is fictional. And I think that’s a problematic way of looking at it because I don’t think we know that they’re perceiving it to be real or fictional. And by just bringing their attention to it, I think we’re altering the process in such a way that it doesn’t reflect how they’re really making these judgments.
Andy Luttrell:
To the point about people are bringing some of their own understanding of the world, like the hero, villains, stereotypes, archetypes. I kind of come preloaded with some expectations. It seems like that’s also happening with different kinds of stories. I don’t have to say, “This is a fictional story,” but this is on TV, I go, “I’m going to go ahead and boot up the fake TV show program,” as opposed to something that is either based on a true story or it’s a documentary. I’d bring a different set of assumptions to it.
I feel like there’ve been times where you watch what you thought was fiction, and then you discover that it was real, and you go, “The fact that that makes me rethink everything is, to me, a little bit of evidence that I’m using a slightly different program to parse that story.” If you go like, “Wait a minute, this all actually happened. Now, I worry that I have the wrong take on what happened.” So I don’t know, maybe this is fodder for another time, maybe.
Matt Grizzard:
No, I don’t don’t know. I think that’s interesting, because this is something that I’ve struggled with and thinking through this and where does this go next? Where does ADT research go? One thing is the retrospective analysis of things. That’s something that’s prompted by the story, but when we’re in a theater and we’re watching something, a lot of times we’re taking it as real and we don’t even realize it.
And it’s usually something that brings us out and reminds us that it’s fake that makes us identify that. There’ll be something in the plot that it’s just like, “That’s absurd. This would never happen in real life,” or, “I can’t believe somebody would respond that way to it.” Or you see a character that you know from something else and all of a sudden you’re like, “Oh yeah, that’s that actor.” And so, when we’re in situ watching it, I feel like the processes are probably the same. And so there’s usually something that’s pulling us out of that, that’s making us think differently about it.
Andy Luttrell:
You sort of got me right to where I was going to wrap up, which is what are the open questions? It’s been around a while. What do we still go, “We haven’t quite cracked the code of this in terms of how all these things are playing out,” or does it feel pretty well established at this point?
Matt Grizzard:
I think feels pretty well established, but I think there are a lot of open areas for it. I think the big thing that we don’t know a lot about is everything else that influences character liking. Right now, that’s kind of treated as noise in the model. The big thing is character morality. But I think there are other things that are really kind of important.
And so, I think the character’s needs, people have different needs. Some people need to feel incredibly competent. I think as a professor, that’s a big thing that I’m always concerned about, “Does this person think I’m smart? Does that person think I’m smart?” Or, “Am I saying something really stupid right now?”
So I have huge competence needs. And so I think if I were going into a film, seeing a character having their competence needs thwarted makes maybe be attached to that character more. So I think there’s a lot of things that we’re treating as noise right now that are meaningful and that we could systematize and understand better. Things like personal needs versus character needs, competence, other evaluative components of how we think about people. I think that’s a big area that we don’t know a lot about.
I think the emotions and how they unfold over time isn’t really well understood. There is work in FMRI and using bio-physiological measures that look at responses to stories over time. But again, a lot of that is using pre-existing stories. A lot of those FMRI databases of people watching things, you’re watching stories that exist. Well, what stories get made? The ones that people like. So you’ve got a lot of reduced variance there.
So when you have that kind of situation and you have reduced variance, it gives the possibility that the big systematic factors that would vary between stories are minimized. And what you end up with is noise amplifying just weird things that are happening in this story. So I think that’s another area where we really don’t know a lot about, is the in-process responses to stories and particularly how they relate to big plot structures. So I think those are big questions that we don’t know a lot about.
Andy Luttrell:
Okay. Well, I expect you to answer them all.
Matt Grizzard:
Yeah, totally. Absolutely.
Andy Luttrell:
And in the meantime, thanks for catching us up on this theory of characters and stories.
Matt Grizzard:
All right, thanks.
Andy Luttrell:
Thanks.
Andy Luttrell (outro):
Alrighty, that’ll do it for another episode of Opinion Science. Thanks to Matt Grizzard for taking the time to talk about characters and what they mean for the stories we love. You can find a link to Matt’s website on the episode webpage, which where you can learn more about his work. Also a thanks for Matt for letting me crash his lab meetings this year. They’re doing some neat stuff that I’m sure you’ll see roll out over the next few years.
For more on this show, just fire up your favorite website on the whole internet, www.OpinionSciencePodcast.com. You can find all the past episodes, find ways to support the show both with cash and with your bright spirit. Transcripts are back, too, so make use of them as you need. Do your podcast-listening duty and rate and review the show on Apple Podcasts. Oh, and I just learned that you can leave comments on episodes on Spotify. I know that because listeners have been doing it! Thanks! Now that I know they’re there, I’ll make a point to read them more often.
Okay, one final announcement for the Opinionistas who made it this far. This summer, I’m reviving the Hot SciComm Summer tradition. In 2022 and 23, I reserved space on the feed in the summer for a bunch of interviews with great science communicators—people who use all sorts of media to help science, especially social science, break out of insular academic circles. Well, it’s happening again! The next slew of episodes on the podcast will be the SciComm Summer feed. I get that maybe that’s not what you all are here for, so just know that Opinion Science Classic will be back in September, so happy summer vacation. But, I’m really excited about the SciCommers I talked to, so I hope you’ll stick around.
Alright, that’s it for now. Thanks for listening, and I’ll see when I see ya for more Opinion Science. Buh bye…