Pirates, Pub quizzes and Park­life: How to Study Group Process­es with Games

Daniel C. Richard­son:
All right. Hel­lo. I’m going to assume that’s all work­ing and you haven’t all dis­ap­peared into the ether. So thanks very much for your patience and stick­ing with these vary­ing 10 sessions.

Daniel C. Richard­son:
So yeah, today I’m talk­ing about pirates, pub quizzes and park­life, and Jo asked me to talk about how to study group process­es with games. And when she first asked me I real­ly did­n’t under­stand what she was talk­ing about, and then I real­ized look­ing at these oth­er talks that, like sev­er­al of these speak­ers, I began as a devel­op­men­tal psy­chol­o­gist. And if you do that, you know that you have to make an exper­i­ment that’s fun and engag­ing, or your par­tic­i­pant bursts into tears and screams at you.

Daniel C. Richard­son:
So I’ve always had that mind­set, that an exper­i­ment is some­thing that has to engage the per­son doing it. And I think often what we do with exper­i­ments is we think of them as a task giv­en to an employ­ee and it’s their job just to go through it. But from the par­tic­i­pan­t’s per­spec­tive, it’s always some­thing like a game. They want to win. They are wor­ried about whether or not they seem good. They want to strategize.

Daniel C. Richard­son:
So I think always think­ing about every exper­i­ment as a type of game is a much more use­ful and valid way to approach it. But I might return to some peo­ple that dis­agree with that per­spec­tive right at the end.

Daniel C. Richard­son:
Let me tell you about what we’ve been doing, using games for group process­es. So what I look at in my lab very broad­ly is col­lec­tive expe­ri­ence or social con­text. So what is the val­ue of being around oth­er peo­ple in the way that we haven’t for a long, long year? What does that do to us emo­tion­al­ly? How do we coor­di­nate our behav­ior? How do we coor­di­nate our phys­i­ol­o­gy? We look at heart­beats with peo­ple watch­ing musi­cals. And how does it change our social affil­i­a­tion? How are we con­nect­ed to these oth­er peo­ple? And how might that change our deci­sion-mak­ing? Most deci­sions that we make when we use our lan­guage or deci­sion-mak­ing process­es, it’s with oth­er peo­ple, yet most­ly when we study them, it’s in a tiny lab­o­ra­to­ry cubi­cle that excludes oth­er people.

Daniel C. Richard­son:
So I got inter­est­ed in these sort of effects of social con­texts and ways that we could study them. And with my grad­u­ate stu­dent at the time, Jori­na Von Zim­mer­man, the Well­come Trust helped us devel­op this tool. It’s actu­al­ly Caul­dron devel­oped. I think they did this before they did Goril­la. It’s a sys­tem we call The Hive, where you have your lit­tle device in front of you and you sign in, and then you get a dot on your device and you move that dot, and at the front of the room there’s a big screen with your dot on it.

Daniel C. Richard­son:
And if you do this in an audi­to­ri­um and there’s your dot, and there’s a dot of 50 oth­er peo­ple. And this means we can ask peo­ple ques­tions and they can respond as indi­vid­u­als, and they can look up and see what the group is doing, and we can look at that group dynam­ic, at that inter­play between what I think and what I think in front of oth­er peo­ple and how I coor­di­nate with them.

Daniel C. Richard­son:
So we’ve used it in var­i­ous dif­fer­ent ways. It’s just a lit­tle exper­i­ment, look­ing at the wis­dom of the crowds, where we get peo­ple to say, how much does this crea­ture weigh? That’s a liger. And they drag their lit­tle dots to give their answer. The star there is the aver­age answer. Right now they can’t see each oth­er’s respons­es. I’ve drawn them back in. And what you find, just as Gal­ton observed a long time ago, is you get a dis­tri­b­u­tion of answers, but the aver­age of those answers is real­ly haunt­ing­ly cor­rect. He called it the wis­dom of the crowds.

Daniel C. Richard­son:
And we showed that if peo­ple can see each oth­er’s dots, if you turn those on, so they give their respons­es know­ing what oth­er peo­ple think, what hap­pens is that wis­dom of the crowd fil­ters away. Peo­ple can see what oth­er peo­ple think and they move to be close togeth­er, and grad­u­al­ly they drag them­selves far­ther and far­ther away from the true answer.

Daniel C. Richard­son:
So we can give peo­ple these lit­tle sort of quizzes, and we can look at that group dynam­ic of how peo­ple change their response, see­ing what oth­er peo­ple believe. And what we are par­tic­u­lar­ly inter­est­ed in my lab is not just how are we affect­ed by oth­er peo­ple, but how am I affect­ed by my peo­ple? And of course we know from social psy­chol­o­gy, there are mul­ti­ple lay­ers of these social iden­ti­ties that we con­nect to. Oth­er peo­ple like the same music or fash­ion or sup­port the same foot­ball club. And we’re all as indi­vid­u­als, a mix­ture of all these social identities.

Daniel C. Richard­son:
And we’ve been using these games to try and pick these things apart. And togeth­er with my col­leagues, this is their work actu­al­ly, they found that many of the behav­iors that we think of as pure­ly auto­mat­ic are actu­al­ly struc­tured by social iden­ti­ties. So the text­book thing is that yawn­ing is a con­ta­gious, auto­mat­ic thing, but they showed if they showed their stu­dents these pic­tures, but said these are also stu­dents from the Uni­ver­si­ty of St. Andrews, or these are stu­dents from the Uni­ver­si­ty of Glas­gow up the road, if they’re my peo­ple I’m more like­ly to have that yawn be con­ta­gious. I’m more like­ly to start yawn­ing if those are my people.

Daniel C. Richard­son:
So we’ve been try­ing to pick this apart and quan­ti­fy it. And again, we give these peo­ple these lit­tle choic­es where they drag the dot to one of these paint­ings here, and we tell them that you are either red or that you’re blue. So it’s a min­i­mal group par­a­digm to put peo­ple in a group with oth­er peo­ple who have the same artis­tic pref­er­ences as themselves.

Daniel C. Richard­son:
And then they drag their dots around. We say, “Just move to a green blob.” We give them these arbi­trary deci­sions. And of course they start copy­ing each oth­er. They move as a red team, or they move as a blue team. Even though that’s a free choice, they tend to be nudged and to clus­ter along with peo­ple in their social group.

Daniel C. Richard­son:
And so what I’ve just shown you, that’s a real exper­i­ment done at the Sci­ence Muse­um. We can also, we’ve spent the past six months putting all that infra­struc­ture of The Hive, which was designed for a room with a large screen, putting that online. So now peo­ple can play the same sort of games, but vir­tu­al­ly, online. We can use them for real group inter­ac­tions, and I’ll show you some in a minute, or we can sort of cheat a lit­tle bit and make peo­ple think they’re hav­ing a real group inter­ac­tion, but in fact they’re watch­ing pre­re­cord­ed slides.

Daniel C. Richard­son:
What I’m about to show you is all done with video. This is a video of me, that’s me on the right. That’s also me there, you can see three Daniels. That’s too many Daniels. But I’m imag­in­ing being a par­tic­i­pant in this study where we trick peo­ple into think­ing they’re engaged in a video inter­ac­tion. So it sort of looks like this.

Daniel C. Richard­son:
[crosstalk 00:05:50] for [inaudi­ble 00:05:51] more people.

Daniel C. Richard­son:
So we have delib­er­ate­ly made the videos a bit glitchy because as we’ve seen, Zoom isn’t always glitchy.

Daniel C. Richard­son:
[crosstalk 00:05:57] and now 12. Okay, great, we’re all here. Okay. [crosstalk 00:06:01].

Daniel C. Richard­son:
So this is all pre-recorded.

Daniel C. Richard­son:
[crosstalk 00:06:02] experiment.

Daniel C. Richard­son:
So these are all pre-record­ed videos of me, but I’m being the exper­i­menter, talk­ing to them, and they think they are one per­son sur­round­ed by oth­er peo­ple. So all those peo­ple are pre­re­cord­ed videos, apart from that lit­tle thumb­nail of me that’s a live view of me watch­ing it, lis­ten­ing to me the exper­i­menter. This is more con­fus­ing than I had in mind. But you do lit­tle bits of inter­ac­tion to trick peo­ple like this.

Daniel C. Richard­son:
Okay. So let me just check every­thing is work­ing. If you can see and hear me, could you give me a wave or a thumbs up? Okay, great. If you can see your­self and you can see every­one else, can you give me a wave and a thumbs up? Per­fect, all right, it’s working.

Daniel C. Richard­son:
So you just have a lit­tle bit of con­tin­gency built in. You set up the expec­ta­tion, say, “I’m turn­ing off your micro­phone,” so it’s not weird that no one’s talk­ing, and you can pret­ty much con­vince peo­ple. About 70% of peo­ple thought, yeah, this is some sort of weird Zoom hybrid game I’m play­ing with oth­er peo­ple. And then you can look at the social effects of who are those peo­ple? So here we just put them in red or blue teams, and then they had a pub quiz togeth­er. So we asked them ques­tions like this, and you can play along.

Daniel C. Richard­son:
Now return to The Hive again, get­ting into the mid­dle. What is the pop­u­la­tion of chick­ens in the world? How many chick­ens are there? The cor­rect answer, sur­pris­ing­ly is almost 24 billion.

Daniel C. Richard­son:
So there you go, and of course we can just quan­ti­fy the degree to which peo­ple gave their answer going along with their peo­ple, and we intro­duce lit­tle things like that buffer­ing screen when we weren’t real­ly buffer­ing, but every video thing does buffer so we put in some sort of fake tech fail­ures to make it more plausible.

Daniel C. Richard­son:
So as I say, that was a pre­re­cord­ed video, but you can also have real inter­ac­tions between peo­ple, because we’ve tak­en all of that Hive infra­struc­ture and put it into Goril­la. So peo­ple can join from Pro­lif­ic and they just get thrown in a room with six oth­er peo­ple, also from the Pro­lif­ic pool. And we were engaged in an exper­i­ment with the sci­ence gallery Ben­galu­ru, hav­ing a exhib­it about the his­to­ry and art, lots of things around the idea of con­ta­gion. So they came to talk to us when we did the con­ta­gion of behav­ior and how it spreads.

Daniel C. Richard­son:
And we ran over 300 peo­ple from India and across the world engaged in this par­tic­u­lar thing. It was very dif­fi­cult, because of course the pan­dem­ic hit India even far­ther, then cyclones hit India. It was a very chal­leng­ing thing. But we got enor­mous num­ber of data, which was a lot of fun to collect.

Daniel C. Richard­son:
And we set this up, and our ques­tion here was, okay, we know there’s an effect of being on the red or the blue team, but is that a grad­ual dial? Is there a grad­ual effect of being more or less close to these peo­ple? So we var­ied lev­els of social iden­ti­ty. I’ll tell you how in a sec­ond. And then we asked them to make var­i­ous deci­sions and they played var­i­ous games like this Park­life game, which I’ll end on telling you in about five minutes.

Daniel C. Richard­son:
So how do we turn the dial on this lev­el of per­cep­tu­al iden­ti­ty? While they’re play­ing this game of pirates, we had a whole nar­ra­tive of me doing a pirate voice, which I’m not doing for you right now. And we asked them things like, “Okay, here’s the bal­le­ri­na who washed ashore, and what direc­tion’s she rotat­ing with?” And we sug­gest­ed that the rota­tion you see tells us some­thing about your per­son­al­i­ty. We said, “Some peo­ple think …” This actu­al­ly does­n’t at all in terms of the sci­ence of it, but we sug­gest­ed that maybe this gives us an indi­ca­tion into your personality.

Daniel C. Richard­son:
And peo­ple gave their respons­es. “I see that rota­tion or that rota­tion.” And then we told them what the oth­er peo­ple play­ing thought, but this was faked infor­ma­tion, this was fake feed­back. And of course we debriefed all of this after­wards. So in one case, we said, “Okay, you saw anti­clock­wise rota­tion, and all the oth­er peo­ple who saw that they’re all blue.” On the oth­er hand, we said, “All but one were blue, and then there’s one red.” And we can grad­u­al­ly turn the dial down on how many peo­ple on your team saw the same rota­tions you, there­fore are like you.

Daniel C. Richard­son:
So we vary this lev­el of group cohe­sion. And then we played var­i­ous games. We looked at how that social influ­ence spread. So in one case we said, “Oh, your ship is sink­ing. Do you want to stay on the ship and risk it or do you want to flee to an island?” And we had dots mov­ing in par­tic­u­lar ways. See­ing if peo­ple fol­lowed their team, and if that depend­ed on who they thought was in that team.

Daniel C. Richard­son:
And what we found, this is brand new data, that as we grad­u­al­ly increased that group lev­el, we increase the prob­a­bil­i­ty that peo­ple fol­low their team­mates. There was a more or less lin­ear increase as they felt more and more of these oth­er blue dots, yeah, they’re like me. They see the same rotation.

Daniel C. Richard­son:
So we had it on a series of deci­sion-mak­ing tasks. And we also got them to play this game, Park­life. So what is Park­life? This is a whole project, and I’m just show­ing you the last win­dow of it. On Park­life it’s like one of these sort of farm sim­u­la­tors, these lit­tle games where you tap a but­ton, you grow resources, and then you sort of grow some­thing. And in this game, what you grow is a park. Used to be a park in Lon­don, but we changed it so that it’s a park on a desert island in this case. The idea is that you tap, tap, tap, tap, tap, and every time you tap as a team, it grows a park feature.

Daniel C. Richard­son:
What we did in this game was intro­duce an inequal­i­ty. So the red team have to tap more than the blue team, because they found less trea­sure in some pre­vi­ous game. So the two teams are play­ing this games, but the red team are dis­ad­van­taged. And what we’re inter­est­ed in is whether that unfair­ness, that dis­ad­van­tage, that resent­ment they might feel, would trans­late into anti­so­cial behavior.

Daniel C. Richard­son:
So you could tap, tap, tap to grow your park, or you could flick a switch and tap, and then you start to van­dal­ize the oth­er side’s park. And that’s what we’re inter­est­ed in, how that lev­el of inequal­i­ty trans­lates into anti­so­cial behav­ior. There’s a whole back­ground from the Lon­don riots that inform this.

Daniel C. Richard­son:
So there they are tap­ping Y. And indeed, we found that as your degree of con­nec­tion with the peo­ple around you increased, then the degree to which you took that inequal­i­ty as a neg­a­tive and turned that into anti­so­cial action increased. So peo­ple are very sen­si­tive, not just to, this is my team, but these are peo­ple who I’m con­nect­ed with. And that changed their response to social inequal­i­ty. It changed the way that they approach the whole game.

Daniel C. Richard­son:
So there’s many oth­er things about that project, but I just want to end so we don’t over­run a lit­tle ear­ly. I just want to end just with a few com­ments, because I think many of us might agree with the over­all idea, since you’re here on a ses­sion about gam­i­fy­ing research. But you will run into resis­tance. And I’m in the mid­dle of a huge argu­ment with [inaudi­ble 00:12:22] at the moment, who basi­cal­ly saw these Park­life games and said, “Well, that’s just a game, right? We’re inter­est­ed in riots and anti social behav­ior.” But he said, “Well, your lit­tle game, that’s noth­ing like a real riot where peo­ple could be impris­oned or they could be phys­i­cal­ly dam­aged, they might go to prison. There’s noth­ing like a real riot.”

Daniel C. Richard­son:
And we [inaudi­ble 00:12:40] this big argu­ment, which I think you will find if ever you do games riot research, we say, “Well, it’s not sup­posed to be a sim­u­la­tion of a riot. That’s not what a game is. We’re not try­ing to get every sin­gle aspect of this phe­nom­e­na and repli­cate it, like it’s in vir­tu­al real­i­ty in a lab. A game is just some ele­ments of that psy­cho­log­i­cal mech­a­nism, that con­nec­tion between frus­tra­tion, anti­so­cial behav­ior. We’re just try­ing to cap­ture that in a small game. So you should­n’t hold us to the cri­te­ria of being an unre­al­is­tic sim­u­la­tion. That’s mis­un­der­stand­ing what a game is. And it’s mis­un­der­stand­ing what an exper­i­ment is real­ly. I mean, we’ve learnt enor­mous amounts from the pris­on­er dilem­ma exper­i­ments, but no one’s actu­al­ly gone to jail after a pris­on­er’s dilem­ma. It’s all a hypo­thet­i­cal lit­tle interaction.”

Daniel C. Richard­son:
So games are not sim­u­la­tions. And that’s a real­ly impor­tant point that if you ever do this research, one review­er will throw back at you. And in the oth­er [inaudi­ble 00:13:30] they’re say­ing well all exper­i­ments are a form of play and a form of inter­ac­tion, so I don’t think there’s any­thing unusu­al about gam­i­fied research and non gam­i­fied research. It’s all games that peo­ple will engage with.

Daniel C. Richard­son:
And cer­tain­ly, this lit­er­al­ly just occurred to me this morn­ing is well, games are impor­tant. Why? Because that’s how we learn, right? I mean, that’s how kids learn. They play fight, they play house, they pre­tend things. That’s how a team will change. They will play and they will do drills. And this whole idea of a sim­u­la­tion of a real phe­nom­e­non as being a way to learn about it, that’s a human char­ac­ter­is­tic, might be a mam­malian char­ac­ter­is­tic, I don’t know enough about it, but it cer­tain­ly is a nat­ur­al thing.

Daniel C. Richard­son:
All the time play is how we learn and how we devel­op. So rather than being an unusu­al thing an exper­i­ment like a game is get­ting right at the heart of how we adapt and how we change our behavior.

Daniel C. Richard­son:
Any­way, I’m going to end there just so we have time for a dis­cus­sion. Thank you very much for your patience and your time. I have too many col­lab­o­ra­tors to thank, but there they are.