There is other ways to be sneaky, he could go undercover or he could use a more trustworthy influencer as a proxy. He could use his audience as an inside man to trick others or make his audience think they are an inside man when he is really tricking them. Then there is always not letting them know there was a survey at all.
Tailored group member here! I loved being a part of the study. I feel like the "contrarian" information actually had the opposite effect on me throughout the study. Instead of thinking "MY opinion must be bad", I was shocked believing how could EVERYONE ELSES opinions be so bad. I feel like this lead to me taking all the information given to me with a grain of salt, and changing my answers less than I maybe would have otherwise. Either way, I love your content and I'm looking forward to more in the future, thank you for letting me participate in the study!
Before watching the video : I participated, and I have to say, I gave the same answers every time ignoring the inputs because they looked absurd to me and I thought "if Viveros already has their data, they wouldn't be asking for mine". I think I brain'd myself into failing the assignment. Sorry. Edit : Now that I've watched it, the one thing I have to note is that the "it's a lot of questions every time" effects was very strong for me. The first round, the first few tanks, I paused and asked myself "do I rank them compared to everyone else, or do I rank them compared to other tanks" for example, but half-way through I was making less effort, and the last few were just "can I be done yet ?". The farther we got into this, the less effort I put in, because the data given to me didn't feel like relevant information and it was just way too much.
30:32 im pretty sure the last survey was around the mid-season patch when mauga got buffed and the dps passive got nerfed which could explain why his final score is way higher than the original
That's what I thought, but even when I specifically only filtered for respondents who got all their surveys done before the patch came out that trend still showed. It was strange and I still don't get it.
go off girlboss and honestly, finding out about how you lie to us is a huge part of the fun of participating in studies like this. Finding the real reason for specific questions or formats is a neat lil gem
I usually try to avoid splicing together gameplay from different matches in the middle of a section bc I don't want it to be too jarring, but this one had such long sections that I had to a few times. I'm glad the transitions ended up looking smooth, I spend a lot of time trying to line up the frames for stuff like that lol
The real insult isnt getting lied to, thats part of the fun, its finding out that you managed to call hundreds of people trash by grouping them all as "low rank respondents" ;) 14:57 but who really cares knowing that you had to subject yourself to looking at .... *that* to do it. Also the line "You were helpfull to the point of actively making my life harder for a few weeks ... i hope it makes you feel better that i very much so had to suffer physically and mentally to lie to you all so extensively" in the debrief had me howling, reading that alone was worth the time spent on the study. Really, when people sign up for a study like this, from a content creator that they know likes to do social studies, you really should already go in with the expectation of being lied to to some extent, if the creator tells you outright what the data is about it might influence your answers after all. On a more serious note, i am genuinely baffled at the amount of people that fell for some of these. At the absolute latest i would have expected people to realize at the influencer numbers that something here aint adding up. Not only were their numbers through the ceiling / floor by a huge margin, but just noticing that these supposed influencer numbers dont line up in the slightest with the discourse surrounding the heroes at the time - which is largely skewed by influencers, because influencing peoples opinions is literally their job title - should have raised way more eyebrows than it seemed to do. However, saying that, i am honestly impressed with the amount of Variance in Group 1 in general. People seem to put a LOT more stock in other peoples opinions over their personal experience than i thought. Also Mauga consistently having stupidly large variance across both groups was ... interesting to see for sure.
I knew from the start that it was going to be something like that old video. Despite that i did truly answer honestly looking forward to this video. It did not disappoint ;)
@@TheViveros You might be right honestly, The devil works hard but Art students need to work even harder faster and all night for a project due the next day :D. Nice having a background thing to listen to though. love your videos man keep it up!
TH-cam did not send me this video when it came out, so here I am. This is truly fascinating. While there is certainly some level of predictability (as you said, manipulating data is nothing new), it's interesting to see data on how that actually works and plays out. I wish I could participate in these studies, but as I don't play OW anymore, my input would skew the data unfairly. But perhaps there will come a day you can branch out and do one of these in either another game or more generically. One day...
The Data did look weird to me but since i wasnt really actively Playing and discussing Overwatch Online i couldn't really know for sure if the Data is Weird. The Influencer Data was the only one i straight up did not believe. The E-Mail answer did not satisfy my disbelive but i just let it go. Anyways feel free to lie to me again in the next Study
I felt bad whenever people emailed me because the best I could really do was repeat what I'd already said or I'd risk tainting the results by giving people too much info. And depending on which topic wins the next members' choice poll there might be more surveys happening soon, although those ones will actually be straightforward (which might be a problem if people assume I'm lying lmao)
i came from that seagull reaction video you mentioned when it came out. like, i didnt really watch him, i just saw it in my recommended and thought a statistics video about overwatch would be fun. one second youre teaching a blizzard shill about data analysis, and the next youre getting death threats. glad you decided to stick around despite that. you made the point that its easier for youtubers and generally politicians and the like to intentionally act in poor faith for financial incentive. this is in sharp contrast to yourself, though we'll never hear you say it yourself; it must be said regardless. you have surpassed seagull and any other content farm in every way that matters, in keeping true to yourself and acting without bias. thank you for your time and effort.
Sorry to hear about the Seagull incident, I didn't even know he reacted to your video and his audience went crazy.. I love these studies you do and a lot of your Overwatch content. This video in itself is a really good study and take on misinformation, especially the final "ramble". This topic is very relevant to the people around me in my life and I enjoyed listening to you speak on it. Keep up the good work :)
I imagine some of the weirdness in the data, as it could not be possibly accounted for, is people playing the game throughout the survey, and having their opinions be either more or less malleable as they started to re-examine their previously held opinion. If you think a hero is good, and you get told they're bad, you'd probably at least start paying more attention to that hero's performance in-game; furthermore, if you thought a hero was bad, but were told they were good, you may start noticing when they do better in game. However, this could also have the opposite effect, where a player feels very strong that X hero's ability is bullshit/OP or underpowered/useless, they may double down when they are shown a differing opinion, given that they are directly seeing information in their games that "prove them right" in their mind. For example, while I didn't take the survey (wish I could have!), I am a Reinhardt main, and consider myself extremely knowledgeable on his ins and outs, and would be extremely hard pressed to change my opinion on a hero I have so much experience with, but, on the other hand, I have little personal experience with lots of DPS heros, so I would likely be more willing to change my opinion of heros I don't feel as though I personally know very well.
"Ethics board"? More like "bored of adhering to ethics"! (Yes, I know that misleading survey participants is not actually unethical, but c'mon, let me have this pun)
I don't know anything about Overwatch (I found you through a friend sending me your Spike Chunsoft vid) and I was absolutely sure it'd turn out that you snuck a fake character into the list. Imagine being able to conjure up strong, deeply held feelings about a character class that everyone just assumed they had forgotten about because everyone ELSE has an opinion
the first thing I though was: huh this sounds more like a "how propaganda affects people" using ow study rather than "how gullible ow players are" so I'm glad i was kinda correct on that in the end
To quote my partner: "Why would I trust a liar?". Which, fair. We can't know whether or not this data is accurate without an unbiased peer review, but that would require more resources that would make this pilot study harder to actually pull off. Studies based on misleading or being disingenuous with participants get harder the more people that are involved. I really wanted to take part in this study, but I haven't played OW2 since the PvE missions launched, so I felt my data would be arguably useless/irrelevant. Very fascinating results, to be sure.
its incredible that i find one of the more mindful and intelligent creators in a video that i expected to be about lying to people for shits and giggles
I knew I should’ve kept yapping about all the things that didn’t add up in the study I just didn’t want to give Viv a big chore of a feedback response. 😭
as someone who hasn't played since overwatch 1, i'm wondering if maybe the reason maoga's change in ratings for the tailored group was so volatile because he's a newer hero? like maybe since people don't know what to think of him yet, their opinions on him are much more easily influenced. still doesn't explain the rest of that data tho lol
I'm glad you performed this survey. Cautionary tales on the internet have been difficult for me to appreciate before, and it will serve us all better in the future to have a reserve of salt to dip into from time to time when absorbing the content we do.
Instinctively wanted to reach out and kiss my laptop screen by the very end of the video because, mmh, God, it's all such good shit. This is the kind of stuff I stay on this godforsaken platform for. Thank you, you gorgeous statistician for creating this. And my condolences, for what they're worth, for that Seagull shiteshow.
I was part of the tailored group and I must say I was fooled. I remember that I at first thought the creator numbers seemed weird but didn't think more about it after the first time they were added. Overall amazing video and find this kind of content so much more rewarding than the average TH-cam video. Overall was a very enjoyable experience from the start of getting gaslighted in the survey to the end of the video and I truly don't understand how you had the energy to go through all that information.
This must've been a fun video for you to make. How nobody would've caught on to all the numbers being fake despite both including fake content creator numbers as well as including your numbers literally just being +2 and -2 is genuinely disconcerting, especially considering each set of surveys would contain each set of supposed averages listed for the participant's manipulation. Regret missing this survey, would've been fun to deduce the purpose of it as it was going!
I was a part of the tailored group! I personally was pretty consistent, for one major reason: my own games (form roughly P1-D5) held massive weight, and the general community held very little weight. I also didn't play a shit ton because I was starting a semester in school, so some of my responses were essentially copy paste from the last response given that my only new info was the info you supplied.
started with your rant about inescapbles and then just decided to go though your video essays and its offcial im subsrcibing, never watched a day of overwatch in my life but I guess I have to start now
I wonder if this test would have had more success in influencing peoples opinions if they weren't randomized, but instead tailored around pre-existing discourse. Also I just realized you didn't do a control group, where you just told them the truth(though this would have been impossible for the influencers segment). And it's entirely possible that most of this data is just within natural change in opinion over time, either doubling down or backing off of their previous opinions.
I was in the generic group of the survey and was immediately taken aback by the influencer responses. Like, ashe was one of the lowest and brig the highest? Influencer or not, if that was in any set of data I would be incredibly suspicious, but I had already pretty much decided to not be influenced by the data, so all it really did was get me excited for the reveal. And a great reveal it was!
i wonder how the data would look for the tailored group if instead of taking the absolute value of the difference (ignoring positive/negative) it was instead oriented based on the direction they were being polarized. so if the "answers" of experts (Viveros, Influencers, High Level Players) were above their original, keep the number the same, but if the answers were below, multiply by -1. Idk if it would be more uniform or even more chaotic.
I know this is mostly for fun, but in order to do experiments with scientific rigor, you need a control group. So instead of trying to do two experiments at once without control groups, you probably should have just done one experiment with control group and manipulated group(make the control group one that you give no data to and just repeat the same survey over time, and have the truth group be a separate "manipulated group" so you can contrast the control group and two manipulated groups). Without a control group that you aren't influencing, you haven't properly isolated the extent of your influence on people (whether telling the truth or lies), which leaves falsifying your hypothesis mostly up to conjecture. You can't tell how much natural shifts in opinion occurred regardless of your influence. That way you can tell definitively what difference lying made and isolate the exact difference isolating one variable (e.g. does lying vs. telling the truth make a difference in your influence of their opinion) did, instead of not having another frame of reference to contrast with, beyond people's responses from arbitrary falsified data (and just making educated guesses and assumptions to inform whether your hypothesis was actually correct or not). It would help a little in being able to see how much the changes in opinion was due to the lies and how much the changes were due to natural shifts in opinion. Ideally, in good experiments, you can definitively falsify your hypothesis because your hypothesis and data collected were both specific enough (and ideally quantitatively enough). Sometimes this will require follow up experiments just because you haven't gotten every question you have answered, but in science, quality and ensuring your foundation is built solidly is more important than quantity or speed. It would also give you more of an ability to evaluate the influence your choices in lies had, since your choices in lies were mostly completely arbitrary in scale (how many points you made a difference with etc.) Not to be obnoxious, but following the scientific method more closely would be more ideal to "do science". Sorry for being long-winded / maybe repetitive, typing on my phone.
45:05 Would have loved to have seen the reaction when he realizes. After he reads the bronze data, he's like "weird but understandable" and then by the high rank he's like "what the hell is this bullshit" 58:00 also damn shots fired, pop off
In case people here are feeling pessimistic after watching this video, a brief reminder that a good way to prevent the effect of internet lies and other propaganda is to make friends with people that both have opposing viewpoints AND are willing to debate people WHILE ACTUALLY LISTENING TO THEIR RESPONSES. I put so much emphasis on those points because I'm not telling queer people to make friends with homophobes, that's just stupid. Unless you're able to actually win them over, in which case, go for it! (Just don't get manipulated, please. Some people aren't worth the effort.) Of course, doing this for every topic you might find on the internet is exhausting. Obviously. If it wasn't, people would just fact check every white lie and we wouldn't have this problem. But we're only human, after all. Just try to make friends with lots of different types of people to keep yourself out of echo chambers. And as long as you do that for the things that matter, who really cares if you think Sombra is a support?
what is the p value and r-squared for this? is there a statistically significant change in respondents' behaviour after seeing "bronze take" or are their responses within margin of error of their original ones despite changing slightly?
also, did you have a video going over your rebuttal to seagull's video reacting to your survivorship bias in overwatch 2 video? i think that would be a good one to make if you ever had to make an in-between big projects shorter filler video for whatever reason
When I gor the the influencer section of the survey I follow Flatz, Emongg, KarQ and a lot more. Noone had the takes that the survey implied I since I am gullible Was flabergasted that these guys responded that way. But mybanswers did not shift that much in general. Also one thing that I think happened alongside this study was that the game got patched. So that might have been also a factor. Specially with the Mauga result at the end.
i have no way to prove this but i had a suspicion that the numbers were completely fake after like the 4th round also sorry for using up like 9/10 reminder emails.. i had a busy week good study good video i hope to be a part of the next one :D was fun getting lied to (P.S. i think i had at least bottom 10% in terms of change in my responses is there any way to fact check that)
It's funny because I was literally just rewatching HBomb's James Somerton video in the bg yesterday while I was doing some writing. Personally, I never felt very deceived by him because the only video I had watched from him was his Killing Stalking video, and that was because I had heard the name make a few rounds on twitter and I wanted to find out what it was about and if it was good and his was literally the highest watched video on the platform about it. I didn't particularly gel with his content based on that because it just felt like some rubbish takes by a guy who didn't like or want to engage with BL content and was forcing himself to talk about it for views just because of how purely dismissive he was towards women in his video. Like, he kept saying "romance is dead this is what women read now" and I remember thinking "clearly you've never talked to a woman before in your life because women have written bodice rippers since basically before Dracula, it's just gayer now." I don't have much else to say atm about this video itself, I just wanted to say it was interesting you mentioned Somerton as someone you didn't fact check when I had just rewatched HBomb's video yesterday. Kinda interesting video, but I kind of wish you had given linked and sourced examples towards your examples of other behavioural studies like the A&W price study because a cursory search I made wielded no results and it'd be interesting to see these studies compared to yours by the numbers. Like, how much the price of their products shifted based on what other people they respected were saying. Would the highly influential group be more or less influential with factors being price vs viability? Does money affect how much sway people are willing to shift their opinions? Also, I think it would be interesting if we had access to those studies to compare how your study measured up in terms of method and extremeness of altered bias. Like, do smaller shifts in bias work more than extreme shifts like the 3pts you did for 'influencer response'. Obviously this would've put a lot more work into this video before you put it out, but I think it helps solidify your argument than just a vague "people can be influenced" because obviously, a lot of people can agree on this point, based on sort of vague data.
So the A&W price thing was an example of what could be done with this info, not an actual study. There wasn’t a study conducted by A&W publicly (at least not as far as I know) but that bit was providing examples of how this pilot study’s findings might affect stuff like how we advertise things if the findings were found to be consistent in a more robust study.
@TheViveros ah, I see. I still wish then that there was some sort of other study given so we could compare your result and experiment with. I'm sure it would be an interesting read.
damn we bitches a gulible bunch About my experience with the survey (without necessarily revealing my votes), I realized all the way from the 2nd survey that it would be about how influencable we would be as it's the only topic I can think of where specifying low ranked responses would be relevant, nonetheless I still tried to vote under the expectation that all we want is to know my opinion on these characters, without intentionally voting the same numbers just to "win" against the study. However I think a thing that wasn't calculated into the survey is the playtime of the contestants, because by the time the surveys were rolling in I haven't had played Overwatch for MONTHS, and I haven't even played as or against Mauga, reworked Sombra or reworked Pharah, or with the new hitboxes once, so my initial votes might already have been unreliable and already influenced by external factors. I realized there was something fishy when the influencers survey rolled out, both because the votes were unrealistically extreme and that those extremes didn't fit at all with what I knew about the opinion of those influencers, like saying Doomfist was around a 10 when I've heard so many of those influencers say he sucks, gut feeling had me believe those influencers were giving intentionally bad answers, but then I realized that what was actually happening is that you were giving wrong data. However I did not think of fact checking, or asking or talking about it online because of you asking not to talk about my responses (that my mind extended to the whole survey) and the fact I barely even talk online anyways, also because I thought I was already working with the additional (obvious) info that this survey was a test on influenceability, and didn't want to change other people's responses on the survey because of it. However I did point out the absurdly extreme values for those influencers....... as a reply to, at that time, most recent and completely unrelated community post... I was not in the Discord server and I would absolutely not post or even DM anything to you on Twitter, so replying to a community note it is, with the bonus that not many people would read it so I would not risk influencing anybody (it did work as not even you read it). At the end most of my answers were consistent as I would just see that day's responses and confidently say, nah you are all wrong, or in some cases just have that day's response already close to mine, with the most drastic changes happening after I actually played the game to see how thing were. Anyways I hope the drag queens and Vtubers video wins also, I was on the generic group
You may have mentioned this in the video and I just missed it idk, but did the participants reported ranks have an influence on how people responded? I was in the generic group and I’m like gold 5 or smthn around there and so when I found the influencers results weird and the lower ranked ones to make sense I was like damn I guess I’m just bad lol makes sense that I can’t relate to these good people, and I found myself (at least consciously) more swayed by the low and mid ranked responses bc I was like woah those are my people that is where I self identify, was there any sort of notable correlation with that?
I had originally intended to break this down further in the video to see if there were trends in terms of age, gender, or rank, but honestly just ran out of energy. The data is still there, though, so maybe once I’ve recovered from the experience of making this abomination I’ll go back and revisit it to see what’s there.
Is must have been a crazy amount of work, but I want to ask, is there something that specialy catched your attention? Like someone responding always the same
I am not completely done watching, but as a only former overwatch player, I wonder how this has affected real life topics, Trans rights being fought against regularly have false authorities (look at the PragerU video which has the Rapid Onset Gender Dysphoria woman) that claim that trans healthcare hurts children, and then people believe them.... I wrote that before the last part, I am glad to know that I am not the only one massively overanalyzing this. I wonder also how new a character is and if that affects something? If I still played, I would not have as strong of an opinion about new characters, if my opinion is lower than others, maybe they know a better way to play them, if I have a higher opinion than others, maybe they know how to counter it that I don't. Also I strongly agree that people are too trusting, I remember taking a college course and the teacher commenting that our (generally younger) group was a lot better at fact checking than her immigrant and returning to school classes. (the majority of her classes) And I myself try to figure out what information I got from hyper alt right people I used to trust. TLDR; thank you for doing something this interesting, its really fun to think about and I appreciate it a lot
1:07:14 Well if you just made some shit up, then at the very least you're a very good storyteller and got a few laughs out of me for an hour of my time :P
Oh. Oh wow. You are definitely a Sombra main. I cannot believe you'd trick random players into rank dysmorphia. "Bronze take" You will be judged at the pearly gates on Lifeweaver's chest.
My degree is in communications, but I've studied a little bit of everything. Two years of finance definitely involved some stats, but I also had to do some courses related to sociological research as well (that's where the Dunning-Kruger study from last year came from).
gonna be honest the stacked bar graphs are kind of hard to read when there are so many of them and I can't tell what's what. Maybe would have been better to split it into general roles, characters divided by role maybe?? But also you already had to do SO much work and to expect you to do more for the video to be easier to understand is asking too much.
I didn't expect so many first comments and I forgot to approve them until there were like 4 of them, but you were actually the first so you get the heart lmao
professional gaslighter, stunning gatekeeper, magnificent girlbosser
Viveros like 2 months from now: Hey everyone, I am working on a new study, and I _promise_ I won't be lying to you this time!
There's _no way_ he'd lie to us 3 times in a row, that'd be so stupid no one would believe him! :D
This man can lie straight to my face three times in a row and more and I wouldn't bat my eyelashes anyways
Next study will be about how people react to information if they already think it's an untrustworthy source
Just as my trust in Canadians was returning too.
There is other ways to be sneaky, he could go undercover or he could use a more trustworthy influencer as a proxy. He could use his audience as an inside man to trick others or make his audience think they are an inside man when he is really tricking them. Then there is always not letting them know there was a survey at all.
As a participant I just didn't care if you were gaslighting I just wanted to help with making a video
Tailored group member here! I loved being a part of the study. I feel like the "contrarian" information actually had the opposite effect on me throughout the study. Instead of thinking "MY opinion must be bad", I was shocked believing how could EVERYONE ELSES opinions be so bad. I feel like this lead to me taking all the information given to me with a grain of salt, and changing my answers less than I maybe would have otherwise. Either way, I love your content and I'm looking forward to more in the future, thank you for letting me participate in the study!
as a fellow tailored group member i must say i felt this exact same thing
having one of your data sets make no sense is such a feel
Not only that. The one that took more work as well
thats our win for the gaslighting
i am lost, i was fooled, i feel lead astray
For science
And personal entertainment
It says "gullible" in the video description 👆
@@Epicawesomeness64damn you
Before watching the video : I participated, and I have to say, I gave the same answers every time ignoring the inputs because they looked absurd to me and I thought "if Viveros already has their data, they wouldn't be asking for mine". I think I brain'd myself into failing the assignment. Sorry.
Edit : Now that I've watched it, the one thing I have to note is that the "it's a lot of questions every time" effects was very strong for me. The first round, the first few tanks, I paused and asked myself "do I rank them compared to everyone else, or do I rank them compared to other tanks" for example, but half-way through I was making less effort, and the last few were just "can I be done yet ?". The farther we got into this, the less effort I put in, because the data given to me didn't feel like relevant information and it was just way too much.
Semi-called it in DMs, you love to see it
You were definitely one of the ones who was getting close to the truth, I was sweating bullets trying to avoid giving anything away lmao
This was not a crossover I was expecting to see.
Surely we wont fall for it a third time, right? What if you did a fakeout study that was exactly what you said it was?
That'll be next year, I'm just gonna make it seem incredibly suspicious but it'll actually be a straight-up honest one.
@@TheViveros😎📸 cant wait to pull out this screenshot a year from now (and then go along with whatever BS you put on the survey)
30:32 im pretty sure the last survey was around the mid-season patch when mauga got buffed and the dps passive got nerfed which could explain why his final score is way higher than the original
That's what I thought, but even when I specifically only filtered for respondents who got all their surveys done before the patch came out that trend still showed. It was strange and I still don't get it.
go off girlboss
and honestly, finding out about how you lie to us is a huge part of the fun of participating in studies like this. Finding the real reason for specific questions or formats is a neat lil gem
*_"We've Been Tricked, We've Been Backstabbed and We've Been Quite Possibly, Bamboozled"_*
No way you hide this clean af transition in here holy
Ah shoot lmfao meant to timestamp it 14:23
ps- hope resource extraction wins!
I usually try to avoid splicing together gameplay from different matches in the middle of a section bc I don't want it to be too jarring, but this one had such long sections that I had to a few times. I'm glad the transitions ended up looking smooth, I spend a lot of time trying to line up the frames for stuff like that lol
The real insult isnt getting lied to, thats part of the fun, its finding out that you managed to call hundreds of people trash by grouping them all as "low rank respondents" ;)
14:57 but who really cares knowing that you had to subject yourself to looking at .... *that* to do it.
Also the line "You were helpfull to the point of actively making my life harder for a few weeks ... i hope it makes you feel better that i very much so had to suffer physically and mentally to lie to you all so extensively" in the debrief had me howling, reading that alone was worth the time spent on the study.
Really, when people sign up for a study like this, from a content creator that they know likes to do social studies, you really should already go in with the expectation of being lied to to some extent, if the creator tells you outright what the data is about it might influence your answers after all.
On a more serious note, i am genuinely baffled at the amount of people that fell for some of these. At the absolute latest i would have expected people to realize at the influencer numbers that something here aint adding up. Not only were their numbers through the ceiling / floor by a huge margin, but just noticing that these supposed influencer numbers dont line up in the slightest with the discourse surrounding the heroes at the time - which is largely skewed by influencers, because influencing peoples opinions is literally their job title - should have raised way more eyebrows than it seemed to do.
However, saying that, i am honestly impressed with the amount of Variance in Group 1 in general. People seem to put a LOT more stock in other peoples opinions over their personal experience than i thought.
Also Mauga consistently having stupidly large variance across both groups was ... interesting to see for sure.
I knew from the start that it was going to be something like that old video. Despite that i did truly answer honestly looking forward to this video. It did not disappoint ;)
And yes I am a demon and watch videos on 2x speed.
Being able to watch my videos at 2x speed feels like a sign to consume less caffeine lmao
@@TheViveros You might be right honestly, The devil works hard but Art students need to work even harder faster and all night for a project due the next day :D. Nice having a background thing to listen to though. love your videos man keep it up!
TH-cam did not send me this video when it came out, so here I am. This is truly fascinating. While there is certainly some level of predictability (as you said, manipulating data is nothing new), it's interesting to see data on how that actually works and plays out. I wish I could participate in these studies, but as I don't play OW anymore, my input would skew the data unfairly. But perhaps there will come a day you can branch out and do one of these in either another game or more generically. One day...
The Data did look weird to me but since i wasnt really actively Playing and discussing Overwatch Online i couldn't really know for sure if the Data is Weird. The Influencer Data was the only one i straight up did not believe. The E-Mail answer did not satisfy my disbelive but i just let it go. Anyways feel free to lie to me again in the next Study
I felt bad whenever people emailed me because the best I could really do was repeat what I'd already said or I'd risk tainting the results by giving people too much info.
And depending on which topic wins the next members' choice poll there might be more surveys happening soon, although those ones will actually be straightforward (which might be a problem if people assume I'm lying lmao)
Viveros' fake rating actually influencing peoples opinion that much is wildddddd
You would NOT believe where it says gullible!
I was in the tailored group and was immediately suspicious because you said the bronzes ranked mercy between 2 and 3. Yeah, nah aint no way lmao
Holy. More than a red flag, that was a war banner
minute 30: wow, this is a lot of interesting data im hearing about!
minute 67: i dondt,, know,, anythign,
i came from that seagull reaction video you mentioned when it came out. like, i didnt really watch him, i just saw it in my recommended and thought a statistics video about overwatch would be fun. one second youre teaching a blizzard shill about data analysis, and the next youre getting death threats. glad you decided to stick around despite that.
you made the point that its easier for youtubers and generally politicians and the like to intentionally act in poor faith for financial incentive. this is in sharp contrast to yourself, though we'll never hear you say it yourself; it must be said regardless. you have surpassed seagull and any other content farm in every way that matters, in keeping true to yourself and acting without bias. thank you for your time and effort.
I really do appreciate that
Sorry to hear about the Seagull incident, I didn't even know he reacted to your video and his audience went crazy.. I love these studies you do and a lot of your Overwatch content.
This video in itself is a really good study and take on misinformation, especially the final "ramble". This topic is very relevant to the people around me in my life and I enjoyed listening to you speak on it.
Keep up the good work :)
I imagine some of the weirdness in the data, as it could not be possibly accounted for, is people playing the game throughout the survey, and having their opinions be either more or less malleable as they started to re-examine their previously held opinion. If you think a hero is good, and you get told they're bad, you'd probably at least start paying more attention to that hero's performance in-game; furthermore, if you thought a hero was bad, but were told they were good, you may start noticing when they do better in game.
However, this could also have the opposite effect, where a player feels very strong that X hero's ability is bullshit/OP or underpowered/useless, they may double down when they are shown a differing opinion, given that they are directly seeing information in their games that "prove them right" in their mind.
For example, while I didn't take the survey (wish I could have!), I am a Reinhardt main, and consider myself extremely knowledgeable on his ins and outs, and would be extremely hard pressed to change my opinion on a hero I have so much experience with, but, on the other hand, I have little personal experience with lots of DPS heros, so I would likely be more willing to change my opinion of heros I don't feel as though I personally know very well.
I guess I'm sleeping an hour less tonight to watch this haha. Already enjoying the hilarious introduction to what the study is
Gonna take about 5 years now to rebuild up trust again to make another gaslighting survey lmao
Hey if Blizzard are gonna lie to Overwatch players to sell them Overwatch 2, I think it's fine when you do it for research.
"Ethics board"? More like "bored of adhering to ethics"!
(Yes, I know that misleading survey participants is not actually unethical, but c'mon, let me have this pun)
I don't know anything about Overwatch (I found you through a friend sending me your Spike Chunsoft vid) and I was absolutely sure it'd turn out that you snuck a fake character into the list. Imagine being able to conjure up strong, deeply held feelings about a character class that everyone just assumed they had forgotten about because everyone ELSE has an opinion
Man, so glad I discovered you tonight. I gotta stop watching before I finish all of your videos in a day. And only 12k followers? Ugh. Keep it up man
Luv ur vids btw, I so glad this video is finally out since I was one of the participants!
I haven’t even played overwatch and I watched this whole thing. Great video!
the first thing I though was: huh this sounds more like a "how propaganda affects people" using ow study rather than "how gullible ow players are" so I'm glad i was kinda correct on that in the end
To quote my partner: "Why would I trust a liar?". Which, fair. We can't know whether or not this data is accurate without an unbiased peer review, but that would require more resources that would make this pilot study harder to actually pull off. Studies based on misleading or being disingenuous with participants get harder the more people that are involved. I really wanted to take part in this study, but I haven't played OW2 since the PvE missions launched, so I felt my data would be arguably useless/irrelevant. Very fascinating results, to be sure.
I just have to say that you are clinically insane for doing this and I am here for future insanity
This is a genuinely a really well designed study. Good work!
Especially love your visualizations. I've gotta steal some of those graphs.
its incredible that i find one of the more mindful and intelligent creators in a video that i expected to be about lying to people for shits and giggles
love that you used the aitsf mafia office theme in the intro. rohan kumakura WISHED he could girlboss this hard
I knew I should’ve kept yapping about all the things that didn’t add up in the study I just didn’t want to give Viv a big chore of a feedback response. 😭
as someone who hasn't played since overwatch 1, i'm wondering if maybe the reason maoga's change in ratings for the tailored group was so volatile because he's a newer hero? like maybe since people don't know what to think of him yet, their opinions on him are much more easily influenced. still doesn't explain the rest of that data tho lol
You are too powerful. You know that right?
I'm glad you performed this survey. Cautionary tales on the internet have been difficult for me to appreciate before, and it will serve us all better in the future to have a reserve of salt to dip into from time to time when absorbing the content we do.
Instinctively wanted to reach out and kiss my laptop screen by the very end of the video because, mmh, God, it's all such good shit. This is the kind of stuff I stay on this godforsaken platform for.
Thank you, you gorgeous statistician for creating this. And my condolences, for what they're worth, for that Seagull shiteshow.
The way my votes changed simply because of how my most recent interaction with the character went
I was part of the tailored group and I must say I was fooled. I remember that I at first thought the creator numbers seemed weird but didn't think more about it after the first time they were added. Overall amazing video and find this kind of content so much more rewarding than the average TH-cam video. Overall was a very enjoyable experience from the start of getting gaslighted in the survey to the end of the video and I truly don't understand how you had the energy to go through all that information.
Side note: Your baby dva gameplay is great
Viveros, the truth enhancer
Ah yes, a Sombra player of all time.
_Truly_ a *liar*
*_BLACKMAILING AT ITS FINEST_*
This must've been a fun video for you to make. How nobody would've caught on to all the numbers being fake despite both including fake content creator numbers as well as including your numbers literally just being +2 and -2 is genuinely disconcerting, especially considering each set of surveys would contain each set of supposed averages listed for the participant's manipulation. Regret missing this survey, would've been fun to deduce the purpose of it as it was going!
I was a part of the tailored group! I personally was pretty consistent, for one major reason: my own games (form roughly P1-D5) held massive weight, and the general community held very little weight. I also didn't play a shit ton because I was starting a semester in school, so some of my responses were essentially copy paste from the last response given that my only new info was the info you supplied.
This shit was chill af
This is so cool I love your videos
This video has NOT helped my trust issues, but I think this might be the only time it’s been for the better.
started with your rant about inescapbles and then just decided to go though your video essays and its offcial im subsrcibing, never watched a day of overwatch in my life but I guess I have to start now
If i had a nickel for everytime viv gaslit his audience, id have two nickels! but its weird that that happened twice
I wonder if this test would have had more success in influencing peoples opinions if they weren't randomized, but instead tailored around pre-existing discourse. Also I just realized you didn't do a control group, where you just told them the truth(though this would have been impossible for the influencers segment). And it's entirely possible that most of this data is just within natural change in opinion over time, either doubling down or backing off of their previous opinions.
31:25 🔥
31:28 🔥
I was in the generic group of the survey and was immediately taken aback by the influencer responses. Like, ashe was one of the lowest and brig the highest? Influencer or not, if that was in any set of data I would be incredibly suspicious, but I had already pretty much decided to not be influenced by the data, so all it really did was get me excited for the reveal. And a great reveal it was!
mauga is dependent on healing, so as people adapted and got to experience more with the dps passive maybe they changed their view as time went on
i wonder how the data would look for the tailored group if instead of taking the absolute value of the difference (ignoring positive/negative) it was instead oriented based on the direction they were being polarized. so if the "answers" of experts (Viveros, Influencers, High Level Players) were above their original, keep the number the same, but if the answers were below, multiply by -1. Idk if it would be more uniform or even more chaotic.
I was part of the tailored group! sorry for misbehaving x
I know this is mostly for fun, but in order to do experiments with scientific rigor, you need a control group. So instead of trying to do two experiments at once without control groups, you probably should have just done one experiment with control group and manipulated group(make the control group one that you give no data to and just repeat the same survey over time, and have the truth group be a separate "manipulated group" so you can contrast the control group and two manipulated groups).
Without a control group that you aren't influencing, you haven't properly isolated the extent of your influence on people (whether telling the truth or lies), which leaves falsifying your hypothesis mostly up to conjecture. You can't tell how much natural shifts in opinion occurred regardless of your influence.
That way you can tell definitively what difference lying made and isolate the exact difference isolating one variable (e.g. does lying vs. telling the truth make a difference in your influence of their opinion) did, instead of not having another frame of reference to contrast with, beyond people's responses from arbitrary falsified data (and just making educated guesses and assumptions to inform whether your hypothesis was actually correct or not).
It would help a little in being able to see how much the changes in opinion was due to the lies and how much the changes were due to natural shifts in opinion.
Ideally, in good experiments, you can definitively falsify your hypothesis because your hypothesis and data collected were both specific enough (and ideally quantitatively enough).
Sometimes this will require follow up experiments just because you haven't gotten every question you have answered, but in science, quality and ensuring your foundation is built solidly is more important than quantity or speed.
It would also give you more of an ability to evaluate the influence your choices in lies had, since your choices in lies were mostly completely arbitrary in scale (how many points you made a difference with etc.)
Not to be obnoxious, but following the scientific method more closely would be more ideal to "do science". Sorry for being long-winded / maybe repetitive, typing on my phone.
45:05 Would have loved to have seen the reaction when he realizes. After he reads the bronze data, he's like "weird but understandable" and then by the high rank he's like "what the hell is this bullshit"
58:00 also damn shots fired, pop off
In case people here are feeling pessimistic after watching this video, a brief reminder that a good way to prevent the effect of internet lies and other propaganda is to make friends with people that both have opposing viewpoints AND are willing to debate people WHILE ACTUALLY LISTENING TO THEIR RESPONSES. I put so much emphasis on those points because I'm not telling queer people to make friends with homophobes, that's just stupid. Unless you're able to actually win them over, in which case, go for it! (Just don't get manipulated, please. Some people aren't worth the effort.)
Of course, doing this for every topic you might find on the internet is exhausting. Obviously. If it wasn't, people would just fact check every white lie and we wouldn't have this problem. But we're only human, after all. Just try to make friends with lots of different types of people to keep yourself out of echo chambers. And as long as you do that for the things that matter, who really cares if you think Sombra is a support?
When you started describing the way you conducted the study my first thought was that you used your own data from round 1 and fucked with it a bit
It's here!
what is the p value and r-squared for this? is there a statistically significant change in respondents' behaviour after seeing "bronze take" or are their responses within margin of error of their original ones despite changing slightly?
also, did you have a video going over your rebuttal to seagull's video reacting to your survivorship bias in overwatch 2 video? i think that would be a good one to make if you ever had to make an in-between big projects shorter filler video for whatever reason
the feminine urge to get locked out of your Google account
I may never trust another human being ever again.
aww viv I trust you wholeheartedly :3
cool video, sadly im too stupid to understand all these funny numbers
the true goal here was to give me an aneurysm with those graphs ;-;
When I gor the the influencer section of the survey I follow Flatz, Emongg, KarQ and a lot more. Noone had the takes that the survey implied I since I am gullible Was flabergasted that these guys responded that way. But mybanswers did not shift that much in general. Also one thing that I think happened alongside this study was that the game got patched. So that might have been also a factor. Specially with the Mauga result at the end.
The Ai somnium files yakuza music :)
i have no way to prove this but i had a suspicion that the numbers were completely fake after like the 4th round
also sorry for using up like 9/10 reminder emails.. i had a busy week
good study good video i hope to be a part of the next one :D was fun getting lied to (P.S. i think i had at least bottom 10% in terms of change in my responses is there any way to fact check that)
It's funny because I was literally just rewatching HBomb's James Somerton video in the bg yesterday while I was doing some writing. Personally, I never felt very deceived by him because the only video I had watched from him was his Killing Stalking video, and that was because I had heard the name make a few rounds on twitter and I wanted to find out what it was about and if it was good and his was literally the highest watched video on the platform about it. I didn't particularly gel with his content based on that because it just felt like some rubbish takes by a guy who didn't like or want to engage with BL content and was forcing himself to talk about it for views just because of how purely dismissive he was towards women in his video. Like, he kept saying "romance is dead this is what women read now" and I remember thinking "clearly you've never talked to a woman before in your life because women have written bodice rippers since basically before Dracula, it's just gayer now." I don't have much else to say atm about this video itself, I just wanted to say it was interesting you mentioned Somerton as someone you didn't fact check when I had just rewatched HBomb's video yesterday.
Kinda interesting video, but I kind of wish you had given linked and sourced examples towards your examples of other behavioural studies like the A&W price study because a cursory search I made wielded no results and it'd be interesting to see these studies compared to yours by the numbers. Like, how much the price of their products shifted based on what other people they respected were saying. Would the highly influential group be more or less influential with factors being price vs viability? Does money affect how much sway people are willing to shift their opinions? Also, I think it would be interesting if we had access to those studies to compare how your study measured up in terms of method and extremeness of altered bias. Like, do smaller shifts in bias work more than extreme shifts like the 3pts you did for 'influencer response'.
Obviously this would've put a lot more work into this video before you put it out, but I think it helps solidify your argument than just a vague "people can be influenced" because obviously, a lot of people can agree on this point, based on sort of vague data.
So the A&W price thing was an example of what could be done with this info, not an actual study. There wasn’t a study conducted by A&W publicly (at least not as far as I know) but that bit was providing examples of how this pilot study’s findings might affect stuff like how we advertise things if the findings were found to be consistent in a more robust study.
@TheViveros ah, I see. I still wish then that there was some sort of other study given so we could compare your result and experiment with. I'm sure it would be an interesting read.
damn we bitches a gulible bunch
About my experience with the survey (without necessarily revealing my votes), I realized all the way from the 2nd survey that it would be about how influencable we would be as it's the only topic I can think of where specifying low ranked responses would be relevant, nonetheless I still tried to vote under the expectation that all we want is to know my opinion on these characters, without intentionally voting the same numbers just to "win" against the study. However I think a thing that wasn't calculated into the survey is the playtime of the contestants, because by the time the surveys were rolling in I haven't had played Overwatch for MONTHS, and I haven't even played as or against Mauga, reworked Sombra or reworked Pharah, or with the new hitboxes once, so my initial votes might already have been unreliable and already influenced by external factors.
I realized there was something fishy when the influencers survey rolled out, both because the votes were unrealistically extreme and that those extremes didn't fit at all with what I knew about the opinion of those influencers, like saying Doomfist was around a 10 when I've heard so many of those influencers say he sucks, gut feeling had me believe those influencers were giving intentionally bad answers, but then I realized that what was actually happening is that you were giving wrong data. However I did not think of fact checking, or asking or talking about it online because of you asking not to talk about my responses (that my mind extended to the whole survey) and the fact I barely even talk online anyways, also because I thought I was already working with the additional (obvious) info that this survey was a test on influenceability, and didn't want to change other people's responses on the survey because of it.
However I did point out the absurdly extreme values for those influencers....... as a reply to, at that time, most recent and completely unrelated community post... I was not in the Discord server and I would absolutely not post or even DM anything to you on Twitter, so replying to a community note it is, with the bonus that not many people would read it so I would not risk influencing anybody (it did work as not even you read it).
At the end most of my answers were consistent as I would just see that day's responses and confidently say, nah you are all wrong, or in some cases just have that day's response already close to mine, with the most drastic changes happening after I actually played the game to see how thing were.
Anyways I hope the drag queens and Vtubers video wins
also, I was on the generic group
Next time I'll just lie back
1:07:20 chills
You may have mentioned this in the video and I just missed it idk, but did the participants reported ranks have an influence on how people responded? I was in the generic group and I’m like gold 5 or smthn around there and so when I found the influencers results weird and the lower ranked ones to make sense I was like damn I guess I’m just bad lol makes sense that I can’t relate to these good people, and I found myself (at least consciously) more swayed by the low and mid ranked responses bc I was like woah those are my people that is where I self identify, was there any sort of notable correlation with that?
That was a long af run on sentence hope it made sense lol 😅
I had originally intended to break this down further in the video to see if there were trends in terms of age, gender, or rank, but honestly just ran out of energy. The data is still there, though, so maybe once I’ve recovered from the experience of making this abomination I’ll go back and revisit it to see what’s there.
@@TheViverosfair enough! This was super interesting as is, thanks for doing this and sharing it w us!
Is must have been a crazy amount of work, but I want to ask, is there something that specialy catched your attention? Like someone responding always the same
30:00 maybe survey fatigue? too much information too fast and they were just like "i dont even know what i think anymore, fuck it"
I am not completely done watching, but as a only former overwatch player, I wonder how this has affected real life topics, Trans rights being fought against regularly have false authorities (look at the PragerU video which has the Rapid Onset Gender Dysphoria woman) that claim that trans healthcare hurts children, and then people believe them.... I wrote that before the last part, I am glad to know that I am not the only one massively overanalyzing this.
I wonder also how new a character is and if that affects something? If I still played, I would not have as strong of an opinion about new characters, if my opinion is lower than others, maybe they know a better way to play them, if I have a higher opinion than others, maybe they know how to counter it that I don't.
Also I strongly agree that people are too trusting, I remember taking a college course and the teacher commenting that our (generally younger) group was a lot better at fact checking than her immigrant and returning to school classes. (the majority of her classes) And I myself try to figure out what information I got from hyper alt right people I used to trust.
TLDR; thank you for doing something this interesting, its really fun to think about and I appreciate it a lot
1:07:14 Well if you just made some shit up, then at the very least you're a very good storyteller and got a few laughs out of me for an hour of my time :P
I just took the survey so I can get an ego boost about not being manipulated much when this video comes out, so don't feel bad about lying to me >:)
Oh.
Oh wow.
You are definitely a Sombra main.
I cannot believe you'd trick random players into rank dysmorphia. "Bronze take"
You will be judged at the pearly gates on Lifeweaver's chest.
You monster
Love this jeep up the gaslighting 👍
great job girlbossing really proud of u
Why would the viveros's anwser be a decimal?
can i replicate a social experiment? overwatch flavor
just out of curiosity are you studying/studied something related to statistics?
He does a study about communication
My degree is in communications, but I've studied a little bit of everything. Two years of finance definitely involved some stats, but I also had to do some courses related to sociological research as well (that's where the Dunning-Kruger study from last year came from).
I feel like this might qualify as an unethical study.
Anyway when do you plan on stepping it up to proper human rights violations?
gonna be honest the stacked bar graphs are kind of hard to read when there are so many of them and I can't tell what's what. Maybe would have been better to split it into general roles, characters divided by role maybe?? But also you already had to do SO much work and to expect you to do more for the video to be easier to understand is asking too much.
Awww now I wanna play OW again for the sole reason of getting hoodwinked by small fnuuy vtuber
Hey viveross
Girl, mental health check, u good?
Nope!
@@TheViveros 😭 treat yourself after this one
Alright buddy
Aha ha…
Commenting...
First :)
I didn't expect so many first comments and I forgot to approve them until there were like 4 of them, but you were actually the first so you get the heart lmao