Regression to the Mean Girls：『Dancing with the Stats』— 学部生が挑んだ生存時間解析

This is Regression to the Mean Girls, where we chat with brilliant data people about their silliest side quests. We're mean as in average, but actually very nice. Come sit with us. I'm Lucy D'Agostino-McGowan. And I'm Sarah Lotzbeach. And today we're super excited to have three current and former Wake Forest undergraduate statistics students with us to talk about one of their awesome projects. So we have Grayson Weevil, Wesley Krieger, and Madeline Merrill. So welcome you three. Thank you so much for having us. That was so in sync. It was. It's that good group project dynamic. Yes, exactly. So we usually kick off with what we call a media minute. It can be literally any kind of media you're consuming right now one week I literally used snacks because I couldn't think of anything I liked um we did what was mine going to be oh I am deep into because I'm prepping

for a dance competition I've been listening to a lot of like Irish dance sounding music but it's taken me down this like a Newfoundland Scottish Canadian side of Spotify um honestly delightful So shout out to the East Pointers, which apparently someone in our department knew. That's awesome. I've been watching East of Eden on Netflix. It's really good. And Florence, she's really good at acting in it. And I feel spiritually connected to her because one time when I was in London, someone stopped me in the street and told me I looked like her. And that was the best day of my life. I don't think I do. But like, I'll hold that in here forever. Because that was literally the best day of my life. And the other thing I'm watching is obviously Dancing with the Stars. Because I love it. And that's where we're here today. But yeah, those are the two things I'm consuming the most right now. And I'll follow that with I've also been watching Dancing with the Stars.

I'm on it every Tuesday. And then also I just got finished watching Big Brother. Which I used to watch this in high school. But I just got back into it. And my favorite player won the whole season. And he won, like, the People's Choice Awards. So he got the maximum amount of money that he could have won from the show. And he's just, like, this cute dad. And he was also, like, an ex-Survivor player that never won Survivor. But he won Big Brother. And I just love him so much. Rick Devins is his name. I think it's so interesting that people make an entire career out of not just, like, particular genres of reality TV shows, but just, like, reality TV. Yeah. Yeah. It's crazy. and he like left his kids for the whole summer to play this show at least he won a lot of money yeah true patrick uh lucy are you okay if patrick goes to win big brother and you'll just no i don't think i am no i didn't come back with the most money and a people's wish award well i get no i no not because if it was his dream he could you know if it were something

like that so if it was patrick's like big dream to do big brother then of course we would try to make it happen but if they were just to win the money no there's better ways to make money that don't require you to be away from me and my children for an entire summer and the show i just watched and loved um was called your friends and neighbors yeah the john ham shout out yeah john ham um it's on apple tv i love it i'm majoring in finance I'm still awake. So it was kind of funny to see that side of finance. He's an ex-hedge fund manager who turns to robbing his rich friends, and it's really entertaining. He's a great actor, and I love that show. I like that show a lot, too. My husband works in finance, so we've enjoyed watching it because he's a little bit back to home. I'm not sure how realistic it is, but it's definitely a good watch for anyone who likes those types of shows.

That's one that I watched a little because my husband was watching it, but he watched it. He's getting his MBA, so I feel like that's been his foray into these finance shows, but he's also watched Billions. So I have piggybacked and watched a lot of Billions also, which I'm sure at least finance would like you to think that's not how things are run. It feels like it could be bad for the brand. say between the two i do love your friends and neighbors but i like loved loved billions i really liked billions and i watched all of that it was yeah that was a have you watched billions i haven't no it'll have to be next on the list though yeah it's a little older i guess because it's it's done right sarah it's like it's finished so yeah that's the if anything about the right one has paul Giamatti yes he's like the DA yeah yeah and it has so the guy who plays Axelrod in Billions is like the main guy in Band of Brothers which we just finished for the first time it's a very heavy watch as you might imagine from like a like a World War II yeah um but he's like

a lieutenant winners who's like one of the main guys the whole time but it's like I keep looking at him and it's like but i've seen you do so many terrible things on that other show how can i respect you jimmy fallon has a cameo in band of brothers which really brought me out of the whole thing he had no business being there he's so good but it's just like in the middle of people retreating at the border there's jimmy fallon driving a truck and being like the germans are coming for virtually no reason i can't remember if we talked about this like here or just elsewhere so apologies to our listeners that they've already heard us say this but jimmy talent is one of those faces that like he just like he's not allowed to cameo like there's some people that are just like character actors through and through that they get to do whatever they want because they always are so good and like jimmy not that jimmy collin isn't good but he just is himself like he can't like he you just i can't see him in something and be like oh he's just you know i don't know i don't know i i i feel like yes yeah i agree i only like seeing him in the

context he has like the late night show right that's the only context i like to see him in he did um so he did a a couple of like early 2000s rom-coms so he did one with drew barrymore called fever pitch which i don't remember like obviously there's a baseball sub scene but i don't remember i don't think he plays i don't know why um well this is great we've had and this is some callbacks to recent episodes because i think uh some the time maybe like to your friends and neighbors potentially um okay i will do this time i've been enjoying a tv show called dark matter so uh it's on apple tv it's a sci-fi ish show i watched the first season so it's based on a book um and the author blake couch he crouch he also is like involved with it so i feel like that's always kind of better when the author's involved but the first season covers the whole book and then they started doing a second season um so it's interesting because he's sort

of getting to like explore other ideas that weren't in the book i really like it it's like about um like physics there they build this like box that can move between um you know what dimensions so like you can see other realities which i feel like normally something like this i would check out kind of early because once you get into anything that feels infinite i'm like never mind i don't really i'm not so interested but i actually really like and even more than the the first season was like very good but the second season they've been just doing a lot of these bottle episodes or they just sort of like hone in on a specific character and i have been loving it i don't know that the general fan base there's a group of people that are not loving it as much but i love it and margo martindale who is a character actor we're talking about character She just showed up a second ago in, like, the last two episodes, and she's so amazing. She's otherwise known as esteemed character actress Margot Martindale because that's what they call her in BoJack Horseman, which is also a callback because we just talked about BoJack.

Yeah. So, anyways, it's great, and I love it. And, yeah, that's what I'm enjoying. There's our media minute. That was, like, pretty restrained. That was not that many minutes. Nothing crazy. We go on a little too long. And, like, our audience should know that the ones that they think are too long, we usually are cutting, like, 25 minutes from them. You're getting the, like, pie-down version of these media minutes, which do sound long even without them. Yeah. Great. Well, do you guys want to tell us maybe a little bit about the context of what you analyzed or what you did? And basically for our set, I think it was 379 course with Dr. L, we were supposed to get a data set and do survival analysis on it. That was what the whole course was about, was like survival analysis, time to event stuff. And we picked Dancing with the Stars. I was scrolling the Kaggle data set website. And this idea kind of came to me of like, I love Dancing with the Stars. Let's do that because that's reality competition based.

but also I just think it's like to me a little bit more interesting than the Survivor or whatever because I like the show Grayson like the show Wesley has come to love the show I'm sure after all the time we spent talking about it but I just thought that it would be interesting to you know do kind of analysis of that because it's like the same premise as other shows but I don't know There's more aspects, I think, to it. And, yeah, so then we did our project about it. And, yeah, we're just using different kind of variables to see, like, what matters when someone gets kicked off the show or gets eliminated. So can you take us through a little bit, like, what is the premise of the show? So, like, talk about who the players are, like, how the week proceeds. Can you talk about how the game is played? I can talk about it a little bit. Okay, so essentially you have, like, at the beginning of the season, you have 11 to 14 celebrities or just people who are semi-well-known in America, I guess, are paired with professional ballroom dancers.

And each week they, like, train for the whole week and learn. The professional ballroom dancer teaches the celebrity a dance. So there's different dances that can be done, like the cha-cha, the tango, the waltz, the foxtrot. And so each week, all the dancers are performing, like, a different dance. Like, they're not all doing the waltz or something. It's varied. And so they film the performances on Tuesday. So it's all live. You're all watching this, like, go down live. So you're going to see the mistakes and anything that happens during their performance. And then there's three judges, and each of the three judges scores the contestants based on how they think their dance is. And there's Carrie Ann, Derek, and I don't know. Who's that? Bruno. And Carrie Ann is notoriously very, like, she knows her dance stuff. And, like, if you have a wrong, if you have a misstep in the tango, like, she's going to call you out for it.

In this past episode, she was acting a little wild and crazy and quoting definitions from about what a lift was and how you can't have a lift and all this stuff. So now there's like drama. But anyway Karian is like notoriously the tough scorer But anyways then one person one couple or pair gets eliminated per week I mean, the first week, sometimes there's, like, two eliminations, but they just keep getting eliminated. But as you go on in the season, obviously, the contestants get better. And, yeah, it's really interesting because some people have dance background. Like, for example, they were an actress. like this year harry shum that's his name yeah harry shum he was like the dancer on glee but now he's an actress in gray or an actor in gray's anatomy so we had like these kind of questions when we were doing our project like does your um occupation like what you were before you

came on the show help you help you like in your and staying in the show longer so yeah that's that's premise anything to add that's awesome i i saw the carry-on versus lift thing there was like a whole my like it was like an instagram carousel of like instances from the last few years of like carry-on versus dances with lifts and i was like if only we had known this you could have had well i guess i don't did you have dance level data or dancer level data it was dancer level so it wasn't like you know whatever because also there's certain like and I don't know that I just said to include this like each week is also a different theme so there's like sometimes it's an artist theme like there's like Prince Night where everyone is signed by Prince or there's like dedication that happens every year someone like brings in a family member or whatever and I think sometimes certain weeks can play to certain strengths for certain dancers like you know it's like it was like Super Bowl night this week it's like oh maybe if you have like an

athletic background that kind of vibe of dancer doing maybe fits more with that so certain times there's like certain dance styles but also the weeks will play more to like certain people might be better at a certain week but that's not something like that be included because yeah it was more dancer level so it was just what they are in their background also and I get this more later but I was thinking about this this morning on strawberry to work about the data points we had And I think something that would be an interesting add-on would be relationship slash history slash, like, I don't know, connection to Disney Plus, Hulu, ABC. Because I've seen some series online that people who, for example, one of the people on this season is one of the fan of bananas. And the banana bowl is streaming exclusively on Disney Plus. And he makes it through every week. And there's ads for his, you know, product during the show. there are certain like people who have a connection to the disney brand who are on dancing with stars and certain people don't and i'm seeing discourse online that like if you're not on the bachelor on

abc or on whatever show on hulu or on somebody on disney plus like you have maybe not as good of a chance as making it as someone who does have those connections and i don't know if that's true but that's something i was thinking about could be an interesting point because i mean i haven't seen Never mind. Even like this past week, the two people who had the possibility of being eliminated were Guillermo and Whitney. The people love Guillermo. He's not very good, and he gets really low scores, but he just has the enthusiasm and the energy. And then Tyler Cameron was also at risk for being eliminated. And Tyler Cameron was on The Bachelorette, which is the ABC exclusive. And Guillermo went home, and Tyler Cameron stayed. And I think that was a little bit of a shock to the viewers. Like, I think people, it was time for Guillermo to go home because Whitney is five months pregnant and needs to be off the dance floor. But I think Tyler probably, I don't know, I thought it was a little surprising.

I thought Guillermo might have stayed over Tyler. But to your point, maybe his connection with the ABC franchise is pushing him. So I should have asked this with the rules, but do the judges decide who leaves or is it like American Idol? Like, do the people vote? oh it's both i like it's a combined judges score and total votes but our project only included judges score because voting changed during the time that we were looking at it because now you text back in the day you had to either call or go on the website so the vote of like each episode basically they're like this is the record number of votes we've ever got on a dancing with the stars episode blah blah blah because more people can text now than you know maybe would pick up the own 10 years ago so but yeah it's the combined the ad like that's why it's such a usa like craze i didn't so many people watch it because the viewers are involved like every week yeah it feels like the golden era of american idol which we didn't watch once because my mom thought that

simon cowell was mean there's like a handful of things i was not allowed to watch because which my parents are pretty chill but it was like simon cowell was mean and i couldn't watch the rugrats because Angelica was mean um and but I do remember that being like that one you would have had to call in for sure but it was like call in to save Kelly Clarkson that kind of thing and I think it's funny that like I don't know how many other shows still do that today like basically give the audience the ability to like maybe influence yeah it reminds me of the uh the Love Island craze this summer I know a lot of people were obsessed with that and really liked that the audience could give input and determine who stayed. I feel like we're in an era of reality TV and America being united about it. Like Love Island, Dancing with the Stars. I don't know what else there is. Traitors. Traitors. People love Traitors. Can you tell us a little bit?

So you had dancer level data and it sounds like you had multiple seasons. But what was your like, how did you define the outcome? It was the time until what happened? I can talk about that a little bit. So we had our dancer level data and our goal was to figure out the time to event for each dancer. And that event was that we chose was their elimination from the competition. And the time that we chose, we actually went back and forth on this a little bit, was the percent of the season that they completed and the reason that we did this was because the earlier seasons were a lot shorter than the later seasons because there are less contestants and kind of less popularity as the show um you know increased episode length and then furthermore we ended up having to censor um the data with the finalists that were eliminated in the final week because there were multiple contestants that ended up making it to the end of the show. So we ended up with about 79% of the contestants being eliminated before the final week of the

show. That's really cool. And so you talked a little bit at like some hypotheses, I guess you had, or some theories of what factors you thought might be associated with people going home sooner rather than later. Can you talk more about maybe which ones did you have data for and which ones do you wish you had had data for? Yeah, I can talk a little bit about what covariates like we thought would shape people's survival on the show or time to elimination. So we first looked at celebrity sex. So male or female. I don't really know. I have like some personal thoughts. um I think males tend to make it a bit further because they have to do less to like show improvement I don't know I I feel like the woman is kind of the one like dominating the the dance if you watch the show you know what I mean um so I thought that maybe males last a little longer than females um but I don't know I don't know what you guys think and

then another thing we looked at was celebrity industry which I talked a little bit about so we had data on what industry the the contestants were from so we kind of grouped it into levels one was actor or actress another was athlete another was tv personality and then like singer or rappers and then we had an other category but our thoughts were kind of that people who have industry and like the actor or actress realm are already used to being filmed being on tv rehearsing um sort of doing things like on the fly which when you're performing the dance you're on the fly kind of um so we thought that they would be more successful in the show and stay longer um and then like tv personalities are already kind of used to being in front of a camera and performing that like performance aspect so we thought they may last longer whereas athletes maybe they're i don't know you could kind of argue either way but maybe they have sort of more of an athletic build and physique and maybe that doesn't go like pair well

with some of the more like dainty dances so they might be eliminated sooner um but yeah if you guys can talk about there's we had some other like covariates we looked at as well i was just gonna bounce a little bit and say i even having seen a little bit i do see what i think you mean between like whether the pro is a woman or a man yeah they feel like not at all to undermine like the like the men I guess celebrities on the show but I feel like if you're strong enough you can support your woman pro and doing the crazy flips and like all the stuff like that where it's like the overall spectacle of it is fantastic and I think the same is true the other way but in the other pairing it's the celebrity it's like giada de la rentis who showed up off the food network and suddenly has to be able to flip in someone's arms and i do feel like i can see where you're coming from with that yeah um and if you like have never flipped anybody before and then you come and you do this

crazy spectacle like but you're you're you're considered like a good male partner and maybe would make it further yeah and then i guess just some of the other covariates we looked at were age and then also the first week score to see how long contestants lasted. Age, our findings kind of at least align with what I maybe hypothesized that younger contestants tended to do better than older contestants just, you know, kind of because of agility, strength, the reasons that you would expect. And the first week score I thought was particularly interesting because with such kind of a long show and time to improve. It was interesting kind of getting to grips with whether contestants actually improved a lot or whether it was mostly those that were pretty good at the beginning that, you know, made it further in the show. For the first week's score, that was actually our most significant predictor out of any of the covariates.

So the way that we phrased it was that for each additional point, an average score in the first week, the hazard ratio was 0.616 times as much. So there was a pretty considerable difference between those who scored poorly at the beginning than those who didn't. But we also found that as it goes on further in the competition, your week one score matters a little bit less and less as you go on because you get to a point where you're later on in the competition, maybe you know four or five of the couples had two sixes and a seven in the first week versus you know the people who got fours or lowers for the first week are gone at this point because they aren as good of dancers So they haven made it through And that not always the case because there can be like audience favorites who are getting held up by votes. Maybe they didn't do well the first week, but they're still there later because they, you know, people love them and are voting for them. But like for the most part, as you get later on, especially like the last few weeks, everyone's a really good dancer and did well in week one. So then it matters less what their justice score

words but for the first like six ish weeks seven weeks like yeah it had a big impact which is really interesting i wonder too how much of that is informed by like kind of what you're saying with people's background yeah like i wonder if you were to look at the like the week one scores by those same kind of career tracks or something like that would you see what you were kind of suspecting like these people came in and were like already performing pretty well yeah really neat i really wish that we would have had data on their popularity among the viewers so some type of voting data I think would have been really interesting or if there's like some way to capture their improvement from like week one to week two I think if they have a really big improvement from week one to week two they're more likely to kind of keep improving like yeah I don't know trying to think of somebody this season like Tatiana I think her first dance she got scored super low it was a

super I don't remember what type of dance it was but it was a very quick pace and a quick pace dance is typically pretty hard to pull off like successfully for your first week but then the second week she improved so much and she kept I think improving she's like my underdog I really want her to make it until i don't know if she's gonna make it to the finale but i want her to make it like far she's like my underdog i mean obviously there are people that are performing better but again that i think that they've had like more dance experience in their background that are forcing them to have higher scores but she's like my underdog who was scoring like for the first week and now she's doing so well yeah well and it's also something that speaking of dance experience that we didn't really have our data set because like we had their past experiences like actor singer whatever but it doesn't really equate for a dance experience because like someone like Jenna Dewan who's on this season she's like an actor but she was in

like the Step It Up movies and was a background dancer and was all these other things versus someone like Julia Files who's on the season who's also an actor so they're in the same category but she didn't really do like dancing in her movie she's not known as a dancer so even though that there's these industries I think they could have been like more niche and better like qualifier of seeing actual dance experience because like performing is one thing like you're saying Grayson like being able to be in front of a camera just like deal with like being in a show I think it is definitely an advantage but then there's the added layer of like actual dance experience being a trained dancer you know sometimes I just want more of an underdog to like win the show so does that happen how often does that happen i guess i would say like i don't know last season i feel like robert erwin i mean i feel like alex and robert oh yeah you had a huge dance experience

versus someone like when she did and she got voted off partially because people found her annoying from her show and like disliked her for personal reasons but also i think people thought it was unfair that she had dance experiences sometimes that is like in your not in your advantage because people are like that's not fair you shouldn't be on it but like well like it's like a moral dilemma if you will if you really think about it this could be a philosophy discussion too i think it's like they should be able to compete yeah i think i think that's like a maybe a little bit of a universal take okay well maybe not but i think a lot of people are on the same page about people like jenna duan who have had extensive like dance experience and not wanting them i i think that people won't vote her to the end because they want somebody who's more of an underdog to make it to the finale but just i i and i i've never actually watched an episode of dancing with the stars i'm like the i think i'm the only one on this i guess i'm gonna be i'm going to sound like I don't really know. You neither?

I haven't either, no. No, he's just a good team player. Oh, good. Okay, so we can meet together. So that's going to be a dumb question because you probably already explained it, but I didn't retain it because I haven't watched before. But it's a combination of the judges scores and, like, people's choice. Is that right? So, like, the people, I guess I could kind of see that it helps or in the very end, is it just the people that pick? it's still a combination okay so then i it's ambiguous as to how it's weighted which is what i want to know like how much are they weighting the judges scores and their improvement from week to week based on the judges perception they versus how much people are voting for them because like for example connor wood is somebody from this season who everyone collectively i feel Like, he's being, he's progressing in the competition because of his, because of the votes he's getting.

But each week, he's kind of been pretty stagnant with his judges' scores. So, for a person like him, I think his votes are taking him further. But for other people, I think maybe they're, just how well they're performing based on their judges' scores are taking him further. Okay, now this is, could this be an effort to think about whether this could be an analysis? Because we don't have any information on people's choice besides just, like, what you feel in the zeitgeist of the time. But you do have all the judges' scores, right? So you could, in theory, have – if we used all the judges' scores across, like, all seasons of Dancing with the Stars, you could look at the correlation or some kind of metric for how those, like, if you just treated each week as, like, independent, how well they predicted whether or not somebody stayed. And then you would assume, I mean, this kind of reminds me of what last time, you know, we're just like splitting up the variation, but you would just assume all the rest of the variability was due to the people. And you could see, it would be interesting because I feel like you could see two things.

Like maybe you could see if there are changes over time. So whether it's like been constant that the judges' scores predict the next week's outcome. Like, yeah, like maybe it's constant, how much of a variation it explains, but also maybe it's not. And they've sort of adjusted over time. I Googled most surprising Dancing with the Stars winners. The AI overview lists Bobby Bones, who I had to Google, is a real personality. Yeah, that was in the US. He's in the US. That was in the US. He's in the US. He's in the US. He's in the US. He's in the US. Yeah. Yeah. It says, widely considered the most shocking and controversial winner in Dancing with the Stars history. He consistently received low scores from the judges, often landing near the bottom of the board, but wrote a massive radio and fan voting case to beat high-scoring competitors. No. Yeah. I don't listen to the live-art radio. Oh, I don't listen to the radio. I feel like you came up. I only know Delilah from the radio.

It's only Delilah. I feel like I knew him because, like, a lot of times when I go to, like, the listening room and stuff, like, I don't know, it would come up. It was, anyways, fascinating. Yeah, that was a big upset. Also, apparently season one, Kelly Monaco, that was a big upset. I don't know who that is either. It was an upset, like, based on her. Dancing with the. it just says that she won despite early clumsy routines and heavy criticism but sounds like she's someone Grayson could have rooted for um the people like maybe because it just says her early performances were like rough so maybe she's then that's something with like week one average score doesn't matter as much because like she's doing poorly week one but since she still makes it and wins in the end you know but I was just gonna say the Bobby Bones season was really controversial also because and maybe this proves my like ABC connection theory because the second place guy was someone who had been on Disney Channel and was still like associated with the

Disney brand and he I mean I was not like as tapped into that season as I am now like I watched some of the episodes but I wasn't watching all of them and he was a great dancer and had like related performances and a lot of people think he was robbed of winning the show and bobby bones won but he had this connection to disney too so maybe that didn't save him though i guess because bobby bones won but just something to think about also have you looked at that whether like disney or abc connection like a variable for that it would probably you probably have to manually That'd have to be a manual thing. Yeah, when did Disney buy ABC? Disney? Disney, like in the 90s. Because, like, Disney, ABC, yeah. Yeah, they've had ABC for a while. Because they acquired, they then acquired ESPN. That's why Super Bowl night was called Super Bowl night because the Super Bowl from ESPN is here. So that's why they had Super Bowl night this year. That's why they had, like, one of the judges from Monday Night Football was the guest judge

because ABC Disney Plus is Monday Night Football. and they had like the trophy and stuff in there from the actual Super Bowl because the Super Bowl is in ESPN now, which is a Disney product. So that's like another connection. Oh, you know our history about it. I feel like I sound like crazy. I feel like I sound like I'm like, no, no. The mouse is behind it all. No, this is what I'm basically saying. Okay, I think this is going to be my review or two comment, but I want that analysis first. We would take some manual involvement, but I think you could do it. I believe in you. You seem to already have it all up there on the dome, so you just got to get it into the CSE file. You got to just do it back. Exactly. Fun fact is that the Super Bowl is on Valentine's Day this year. That is a fun fact. Yep. All the couples out there, they're going to be in some fights. No, I'm kidding. I thought that they were going to announce the halftime show on the episode last week,

the Super Bowl night. Yeah, but they didn't. it's disappointing so did i do you have predictions streets are saying miley cyrus is the prediction that's what i'm hearing they're saying miley cyrus okay but i i you know you don't want to quote me on that because i don't be right but that's what i'm hearing from the streets i like it i like it this is great i is this something i'm sure there's prediction marcus There is. They call me some money. Oh, yeah. Betting off Madeline's tip. Yeah, betting markets point to Viley Cyrus and Jay-Z as the leading contenders. Jay-Z right now, according to Polymart, has got a higher probability. He's at 68. Okay, here's what I want to know. Which prediction markets let you bet on this? I'm sure there's a real opportunity for it to arb. This is going to be our chance, Sarah. We're going to arbitrage. Well, I was just looking at the Dancing with the Stars. And there's over like $6 million on Kalshi on Dancing with the Stars for the season.

Yeah. I get ads for Kalshi on TikTok. It's like download Kalshi and predict who's going to blah, blah, blah for Dancing with the Stars. Does Polymart also have so here a hot tip You can arbitrage these if you just you know use one against the other They can both be you know you can like get some free money Let's see. Calci does have. So Calci is saying that Miley is up. It's 50% Miley to 40% Jay-Z. What is Polymart? I think has the opposite. Okay, okay, Polymart is 78% Miley. Yeah, there's a real opportunity here. Maybe I'm just stuck on Dolly, but it could be, like, a Dolly tribute or something like that. Yeah, that would be nice. I don't know. Like, Miley could do a Dolly tribute. I feel like that would be. Yeah. Yeah. I really want, like, a country music artist because I feel like it's been ages since we've had somebody. But I don't know.

Yeah. Can I ask you all a couple more questions just to keep us on time? Yes. Yes. Get us back. Yes. I'm sorry. This is fun. But can you tell us a little bit about, like, what kind of models did you actually use to answer these questions that you have? We did do, yes, we did the Cox Proportional Hazards model. I'm, like, referencing our PowerPoint right now. They're excellent. Dancing with the Stats PowerPoint. And they danced at the beginning. They had music. I have beside myself. Yeah, there was music. There was a picture. That took me maybe an hour in Canva, by the way, getting the music uploaded and stuff. It took me like as long as some of the stuff in our studio because I was like focused on that. So we relate heavily to spending time like that. In Canva specifically, we've been making these like burn book style things for our Instagram. And I was like, these are going to be bespoke. I'm making these myself in Canva. Although I realized there I was thinking about this because I made such a big thing about like I'm making these by hand.

And like reality is if I was making a burn book by hand, I'd be cutting out little magazine things and I'm not doing that. So I guess I'm still using digital pieces, but I am placing them exactly where my little eyeballs want them. Basically a graphic designer. That's right. That's what that means. Yeah. So the cost proportional hazards model. So you modeled essentially the hazard that somebody would go home. Yes. Before the final or the hazard that, yeah, that they go home. But they get eliminated. Eliminated. Perfect. Thank you. Given like characteristics about them as a dancer. Right. Sex, age, week one, average judge score. In their profession. Cool. And so we talked a little bit about, like, which of these factors, like, I know it sounded, week one was a big, like, your week one score seemed to have, like, a pretty strong relationship there. Honestly, a hazard ratio of 0.61 is, like, wild. So, like, essentially for every point higher somebody scored in week one, the hazard that they're going to get eliminated is decreasing by 0.39 or 39%. Like, that sounds like if you were training up for Dancing with the Stars, maybe the move would be to show up as strong as possible.

you have to do that to the pre-work yeah but i know you also looked at it as like its ability to predict right like for a new player entering the game how well can we predict whether they're going to go home right yes yeah that was the like secondary objective so as the season went on it got worse and worse at predicting the area under the curve went down but like at 25 of the season you could put two people in and with like 77 accuracy the model could predict which of those two was going to go home and then at halfway through which I feel like we're about there right now maybe like next week is halfway it was like 70 percent of the time it would predict the right person and then as we got to like the final weeks like three quarters of the way 75 percent of the way through it was 63 accurate so that's obviously more like 50 50 or like who really knows but I mean a little bit better than 50 50 but it was definitely interesting to see that the model could do such a good job especially early on and I think that goes back to like week one judges four being

important and how you show up mattering and also how popular are popular you are because someone like Tucker Carlson you know he got like twos and threes his first week not like super popular among like young people also like back when he was on it he wasn't like even super well known so it's like that combination isn't going to do many favors like yeah you could probably predict that he was going to go home early that's so cool I would love to see how how well your model would have done on this season I know like I forgot like talk about these results but when looking at sex and then also industry our Kappa Meyer curves look the same when so we had two Kappa Meyer curves for like male and female and they're pretty much the survival curves look exactly the same so we can't really conclude that there's any you know your hypothesis there was not correct yeah it was wrong it was wrong and then for industry too those curves are look pretty much identical between like actor and actress curves versus the athlete curve versus the song and

rapper or singer or rapper curve so there's no relationship between their elimination and what previous industry but I feel like the data may have like I wish we could get more niche like Madeline was saying or like number of times been that you've been in front of a camera or something like that or been filmed for something well we also just don't know I mean it's very possible like intentionally balanced like the way that they're actually like at the very end I mean maybe it doesn't maybe they there's some imbalance if you look throughout the whole season because maybe they allow for more wiggle throughout but then at the end like perhaps they're not going to pick an athlete if they picked two the past two years have been an athlete they're going to like pick an actor you know they might sort of like intentionally shake it up so that you know which would have nothing to do with like who does better actually and more to do with just like how the show is run so that people stay invested I also wonder if it could be selection bias where like maybe the people who are saying yes to the show even if they are from an industry where you

wouldn't think of them being in front of a camera being a performer like maybe you're getting the most camera friendly outgoing yeah I don't know like right I keep going back to Jada because I think she's the first cooking channel person I've seen so that one really stuck with me I would love to see Rachel Ray if I could like put out a wish list and we're gonna bring more chefs I would like Bobby Flay or Rachel Ray I think in a future season but that was one where I was like yes you are a personality but there's not a lot of showmanship necessarily in like a cooking show and so I think that one's really interesting so yeah is it also just that like maybe that was the fifth actor that was asked but the first four were like no way I don't want to be in front of a camera and I trip on myself and I'd be a terrible dancer and they've removed themselves from the competition yeah I agree that's kind of like Tyler Cameron like yes he was a football player but he also was on the bachelorette and so he's already been in front of a camera so maybe

felt more comfortable with it and he already has a fan base too like i feel like you you kind of have to go in the show with a pre-existing at least a little bit fan base and yeah in your package like what you otherwise you're really relying on how well you can dance on those scores it's like i'm not here to make friends or fans i'm just here to dance we're like it would be cool to look at number of followers on instagram as like a favorite i don't know i mean i feel like that's a little bit of a proxy for how popular they may be or what kind of fan base they might have or maybe like like follower growth over the season that's good yeah because like you might start out with a certain amount and then have a lot more as time goes on yes this is awesome this is so fun well do you all have any i guess like last remarks do you want to do you want to make officially your predictions for who's going to take home the mirror ball this season hmm I should have done the model before this our project was super fun because we got to you know

take what we learned in class and kind of transform it into something that was interesting to us but I think who's taking home the mirror ball it's either going to be Maura Higgins or Ezra Ezra is probably the one who I want to take it home the most because he's so good he's so fabulous and he's improving every week I think so yeah Ezra all right you heard it here first that's That's a good one. I think I have Ezra. I made a bracket with my friends. So I have like people going every week. I have my winner picked as Amber Glenn, the figure skater, because I think she's a really, she's really good. She has like dance experience, but figure skating, I think is different enough to where like people aren't going to be not voting for her because of that. And also her pro is Pasha, who is a pretty well liked pro, but he's never won the Mirable before. So I feel like people might want to like have him win it. and I like Amber Glenn a lot and I think it would be really cool to have her win especially since she's won a gold medal so you know you don't even go out for her but she was she fell on her like

shorter long program at the Olympics last year she didn't get her second gold medal I feel like she's like needs a vengeance or something so I have Amber Glenn winning but I think Ezra will be also definitely a winner for Mora because yeah she's also really good so that's probably my top My top three, but I want Amber Glenn to win. I haven't watched an episode yet, but the favorite on Cal sheet is Ezra, who they give a 55% chance. So I think I'd have to choose Ezra, but I like Jackson because I'm a big baseball fan. So maybe he can be an underdog and do well. Who's going home next? I think either Savannah Banana Jackson or Tyler Cameron. Yeah, I think so too. We'll check back with you next week. We'll see how you do. We should really pit your model against like your personal predictions. Next season. We'll do it next season. I'll check in with you. I'll schedule an email.

Amazing. Well, the last thing we do is we have you all rate yourself on a scale of one to many fetches. Sometimes we said one to five and sometimes we said one to many. Anyways, so if you were to each give a rating to your analysis, how many fetches would you, how fetch would it be? This might be before your time, so you don't have to know what it means. You just have to give it a number. You have to know mean girls. It can't be before their time. Sarah, I was teaching the other day and someone didn't know who Amy Poehler was. They had never won a part of her, so I don't know. But if you had said the mom from Mean Girls, I think they would probably have known that. They just don't know her like you know her. How fetch? I think it was five fetches. Many fetches. So fetch. Fetch. That's what I think. So fetch. Mm-hmm. Yeah, I do, too. I think it's really in accordance with the times. We're really on brand for the season being going right now.

And, yeah, it was fun. And I feel like our covariates were cool to look at. And, yeah, our results were cool. I think I like how we looked at survival prediction. Yeah, five out of five. I think Dr. Earl can't say this, but, like, if she had to, she would say ours was the most entertaining. But if you had to, like, rank all the projects, you'd probably, like, that one was probably, like, the best and most entertaining. And I know you can't say that, but, like, it's up here. Madeline will say it for you. She'll say it for you. Yeah, it's okay. It's okay, I'll just say it. It's fine. No, I was just going to agree with Madeline. I think we had the most fun, and I think that's what matters. So it's a five out of five. Well, thank you all for sharing with us. You go, Glenn Coco. We're Regression to the Mean Girls. We're mean as an average, but actually very nice. Tune in next time. Thank you.