SIO485: It's Not Rocket Science

Transcript

This transcript was generated automatically and may contain errors — the audio is the record.

0:00:11 Thomas: Hello and welcome to Serious Inquiries Only. This is episode 485. I'm your host, Thomas Smith, and this is part two. Make sure you listen to part one. This is part two about the so -called voting irregularities, the problems with the machines. I'm not talking about 2020 and those allegations. I am talking about, of course, 2024 and Rockland, New York, Which right where you would expect an election hacker to target. No, it's where you'd least expect it. But that's why it's so dastardly. I think you already heard a lot of reasons last episode why this is not likely anything. Um, there's some great stuff in this episode as Jenessa is going to take us through these statistics, the numbers that are offered in support of this theory and oh boy, this is fun. This, this, uh, this was a fun time. I'm really excited for you to hear this. Something I didn't know I needed in this episode that really, uh just cheered me up and I'm weird. So I don't know if that'll apply to all of you, but there is an incredibly bad graph that was used that warms my heart. So fun. Got to hear it. So why don't we get on over to it? Part two with Dr. Jenessa Seymour on what happened in Rockland, County, New York. Support the show. Please ensure these episodes continue to come out. I think it's useful stuff. I think it's stuff that you don't get in a lot of other places. And we're a very small media company here. So your support really goes a long way. Thank you so much to patrons who support at patreon.com/seriouspod. You're the reason this show happens. And all my shows, really, you're the best. You like what we do, maybe you can spare a dollar or two for us. Thanks so much for the patrons who support the show. Everyone who listens. After this break, we'll get to part two here.

0:01:56 Thomas: So you'd mentioned some statistical stuff you wanted to go through oh

0:02:00 Jenessa: There's a document I just sent it to you um I think I know this isn't a visual medium but looking at this is um very storytelling this is our analysis from our um who is claiming that um is so statistically unusual and there are many red flags in this document for me if you go to um uh the third page on it you'll start to see graphs. So, okay, you've got a couple of towns here, right? And

0:02:30 Thomas: Guess uh I see some weird graphs oh Clarkstown yeah there's a lot coming at me here

0:02:35 Jenessa: So there's, I think there's like seven towns in this county. For some reason, we're only analyzing four of them.

0:02:42 Thomas: Is that like a township or something because it says 87 districts. Oh, okay. Works different out here, I think.

0:02:46 Jenessa: Yup. So God. Yeah. New York's got its own thing going on. Have 307 total in the county, in Clarkstown, which is like, you a region, basically. There were 87 of said districts. Flags instantly looking at this. So Y -axis, what you'd expect. It's the president down ballot difference. So if the president got a vote, the Republican president, but not the Republican senator, that's like plus one R. the Republican President didn't get a vote that the Republican Senator did, that is a minus one and put that together. And you do the same thing for the Democrat. They did it for Republican 2020 and Democrat 2020, Republican 2024, Democrat 2024. So that's Trump, Trump, Biden, Harris. Sense on the y -axis. If we have a big positive number, that's the president getting a vote and the senator getting less votes. Then negative number, that would be the senator getting more votes than the president. We talked about, we don't know that that was a Democratic voter or who it was, but you know, we got a number. Got a scatter plot. So each little dot is a district. So there's 87 dots on this graph. It's not inherently a problem. The X axis is just the district number in ascending

0:04:13 Jenessa: Order.

0:04:15 Thomas: I was wondering what that was. Naturally assume it was the year, the election or whatever. You'd assume this was like a change over time, but it's just the the district number.

0:04:25 Jenessa: Something. You know know what I don't expect to have an increase or decrease in down ballot difference? A sending order of districts. It's a number, but it have could been a letter. It's just a name. It's a categorical variable. You can't do this.

0:04:39 Thomas: That's so funny. That's the column they selected in Excel to auto -generate the chart. So that's what it's going to be.

0:04:50 Jenessa: They made a scatter plot instead of making what is obviously supposed to be a bar graph um so that's why we have our line of best fit here is just a horizontal line like

0:04:59 Thomas: Line of best fit? What would that be? Why would you have a line of best Fit in that?

0:05:00 Jenessa: Oh god yeah it's useless yeah why do we have horizontal lines for all of them it's

0:05:09 Thomas: Well, you see, as the district district number goes up, that tends to mean... This is so funny.

0:05:18 Jenessa: So bad this is literally statistics 101 in that I taught caught Statistics 101 and I taught this. So literally what you needed to do was a fucking bar graph. So you would have four bars for 2020 and 2024 and Democrat and Republican. And the height of the bar would be the down ballot difference. And you would put an error bar on here for the love of Christ. They never give me error bars. Nobody does this anymore because the standard deviation is really important. You can see there is a scatter of these dots. There is a range. I want to know what that looks like. And then I could look at it in a way that makes logical sense. And instead, we have a horizontal line because the best predictor is the average because the way to minimize the distance of a point to the line is to just pick the average

0:05:26 Thomas: You actually... Yeah.

0:06:18 Jenessa: And go straight right across, it means nothing.

0:06:18 Thomas: So funny. Is really funny.

0:06:21 Jenessa: So a bar graph would be an appropriate way to display this information. Oh my God. All right. So all four of the towns, despite those not being all the towns those are going to get used. I think

0:06:21 Thomas: I love that. There's an asymptote on the district zero. It'll never reach... If you look at your numbers, it'll never reached district zero, It can always get closer and closer, but it can't quite be named District 0.

0:06:47 Jenessa: I, yeah, this is like, if you haven't taken statistics or like how to display data, like that might not be obvious, but this this is a PhD in astrophysics, like you should, yes.

0:06:57 Thomas: I feel like that skill set should know this. It's

0:07:01 Jenessa: And here's what's then, therefore, I don't know what tests you did. I can take a guess. We've got up above is the, really the only useful thing.

0:07:09 Thomas: So funny. I'm so sorry. It's so funny. I love the significant lines are District 20, 40, 60, and 80. Because you want to keep your eye on those significant ones every 20 districts.

0:07:22 Jenessa: Yeah, really helpful, right? Like,

0:07:26 Thomas: Matrix. Sorry. Okay, good. Just so good. I love this.

0:07:27 Jenessa: It's, what we have have not analyzed is the correlation or the regression line because it's obviously zero because it is a horizontal line. What we must have done, I am seeing a Z score here. Two lines up above are really the only useful thing. Appears what we did is we compared the Republican 2020 to 2024 drop -off rate. So that would be two numbers. We have averaged it from 87 data points. And we come up with a Z score and a P value. It's a little different, but it's not statistically significant. So what that says is for those 87 districts, the drop -off rate for Trump from 2020 to 2024, it's positive. He got more votes than the Republican senator each time. And it's not that different um a little different it's not a big deal um we're doing this by township I have not the faintest clue you could have done all 307 and improved the power of your statistic it

0:08:27 Thomas: Yeah, well, it's hard to fit all those on the x -axis.

0:08:34 Jenessa: Well and they wouldn't go all the way up to 307 it would repeat back to clarkstown one and then it would go ramapo one and you would have to read Now what do you do? You literally couldn't.

0:08:43 Thomas: That's right. Yeah well, you need a three -dimensional graph. This

0:08:48 Jenessa: I lost my mind when I was staring at this. OK, if you look at the Democrat one, 2020 to 2024, keep in mind the previous one is the same candidate. And now we have Biden versus Harris. So have 2020 versus 2024.

0:08:50 Thomas: Is so funny biden versus harris wait what right

0:09:05 Jenessa: If we do Republican 2020 versus Republican 2024, those are both Trump. If we do Dem 2020 versus Dem 2024, we shouldn't just assume, like, they're not the same person.

0:09:09 Thomas: Right right I see this graph is the best I'm so sorry it's so dumb it's just not the same election you know like it's just it's good point

0:09:19 Jenessa: It's truly so useless. NOT THE SAME ELECTION. Conditions, factors going on, all kinds of things. And for some reason we've restricted it to 87 out of the 307 districts and it's also saying clarkstown 2024 87 districts remember that the districts were redrawn in 2020 so I don't quite know how we're correcting for that whatever we've got z scores I'm assuming what they did is people might be familiar with a t test where you take two groups uh the averages and and take into account the standard deviation and the sample size, and see if those averages, is they're always going to be a little different, or is the difference statistically significant? Order to do that, you have to meet certain assumptions. For the t -test, it needs to be normally distributed. You need homogeneity of variance. It basically means the variability in 2020 and the variability in 2024 are the same roughly. There's is the problem is you would do that. You can do a test for homogeneity of variance and your stat software will spit this out. It will automatically pull that test and if you pass it it'll tell you and then it's appropriate to do the t -test so

0:10:23 Thomas: Not sure how you'd know that

0:10:41 Jenessa: It pulls the data and it knows to do that

0:10:43 Thomas: I guess I was just thinking like Like, with this little amount of data points, it's like, I don't know how you could conclusively say the variance is a certain thing. You know, like, you still have a small sample of, like. The universe here.

0:10:57 Jenessa: I mean it also like none of this is making sense because if you're plotting these districts what looking at the 2024 scatter plot and then they just didn't plot the 2020 ones or like I don't know what I'm looking at Yeah, but what's the line?

0:11:09 Thomas: It has. It has a line. Wait. But it has it has a label for 2020 is what I'm saying.

0:11:17 Jenessa: Line has a label, but the little circles that are meant, and a solid line for 2024.

0:11:20 Thomas: There's dash. Wait. Sorry. There's dashes for 2020. there's dashes for 2020, solid line for 2024 for both red and blue. The drop -off oh average okay yeah you can't tell which what you can tell which which is averaged together and which isn't is I was saying like you don't know this

0:11:41 Jenessa: There's only red and blue dots. So does that represent what that, particular precinct's drop off rate in 2024?

0:11:48 Thomas: Is nothing I'm sorry this

0:11:50 Jenessa: Or in 2020? 20. Like if you're going to display them on the same graph, I needed four color. It's nothing.

0:11:50 Thomas: Is just try again redo your graph draft that's

0:11:57 Jenessa: You can also put a scatter on top of a bar graph if you wanted to show me not just the range, but where individuals were so I can see outliers. Like that's fine. This is nothing. This is absolutely nothing. So you want to do a Z test instead of a T test, you also need to meet certain assumptions. For example, knowing the population variance. If you have all the districts, then in theory, you have not just a sample variance, but a population variance, but that is like the population variants of every district for that election. And I don't know that I feel comfortable saying, well, then that's the variance because the whole point is it's different in the next year. So I don't love it, but like, let's say it's fine. What he is saying here is for these 87 districts, presuming this data was entered correctly and then a graph was generated that makes no sense, but that the data entry was correct. What we are seeing is Biden's drop -off rate is different from Harris's drop off rate. Okay.

0:11:57 Thomas: This is yeah okay

0:13:03 Jenessa: And there we go. I just needed to get it it off my chest that

0:13:07 Thomas: Cool this is people should look at this it's so funny it the more you look at it the funnier it gets like it's just anyone who does anything with stats and I'm not like I'm really rusty but at least in terms of doing some rudimentary graphs this is so funny god it's så funny what

0:13:22 Jenessa: Look at it and start asking yourself questions and you'll be like oh yeah yeah this this is wild um gonna keep seeing things like this happen with this case and with other cases that smart legislation is bringing so for example Example in that one where we were talking about New City primary and like, oh, there was bribery and like we're going to bring a case. All right. They're like, Oh, my God, statistically significant. We found what we think are stolen ballots. Ballots so we a single day of early voting so like way less voters than normal in a single precinct there was a bag of ballots that got mislabeled allegedly and

0:13:25 Thomas: Does anything mean okay

0:14:12 Jenessa: That misl labeled bag happened to have zero votes for a candidate who would go on to get in the entire election 29 % of the vote. We're going to go compare that to the 18 bags that got audited because there's always a 3 % audit and see if this is different. So first of all, they say we calculated the average and everything without that bag, the mislabeled bag. Your null hypothesis is that that bag's not weird. You can't take it out.

0:14:23 Thomas: So but right okay yeah yeah yeah that would skew

0:14:54 Jenessa: So, of course you don't take the outlier out and then go, look, it's an outlier. Like, you have to include it in the average and see if it's weird, even when it's included in the data.

0:14:55 Thomas: Things a little bit it yeah yeah

0:15:10 Jenessa: Whatever. There's 18 bags. That's a pretty small sample size. Degrees of freedom is 17 there. There.

0:15:18 Thomas: It's like just to maybe just to illustrate that I know it's complicated without people being able to see stuff but let's say you've got some number of coin flips and you think one subgroup of them for so let's Say just for fun it's Like actual coins that people flipped and you're Like well there's a bag of a hundred of them yeah there's A bag of 100 coins and let's Say that that you have the result or something like you can see the result on on the coin just to make it like more equal to like a vote. And then you're like, well, this one bag seems really weird. Like there's like 75 heads and 25 tails. That just, that just seems very weird. Let's calculate what like the overall average should be, but let's not incorporate that bag to the overall average. You'd be like, wow, I think it's a little more clear what, what you're saying where it's like, no, you need to include all that because presumably over time, it's going to even out. It's just, you have an outlier. So that outlier has to be part of the evening it out. That right?

0:15:33 Jenessa: One coin I think is, yep. Right. Precisely. Exactly. My null hypothesis is that it is supposed to be part of the set. So I keep as part ofthe set. If you excluded, excluded. What if it was two that you think are weird?

0:16:22 Thomas: What I was trying to think of. It makes it even worse because let's say, roughly speaking, our it's going to be 50 heads and 50 tails, 50 -50 essentially. Once you've taken out like, oh, this one is like, just for fun, let's stay 100 heads and zero tails to make it even, then that's going make the rest of it look like, Oh, my God, normally it should be like a majority tails. Like that's what we should have is majority tails results. But you take this one. Now it's even more abnormal because it's just a hundred heads. That right? Like you've got, like, let's say the rest of the population is a thousand flips. And so now for 900 of them, you've taken out a hundredheads. Like that is going to, odds are there is going to be like more tails in that remaining group.

0:16:55 Jenessa: 100%. Right. Exactly. Yes. That's exactly the right way to understand this. So first of all, flat out admitting they've already done this wrong. Going to see an image that is clearly a crop of the output of statistical software. I bet you I know which software it was. It looks extremely familiar. And we're going to see the N18, that's our number of bags excluding the one that was important, our standard deviation, our average, our Standard Error of the Mean, and nothing else. The text will tell us the T value and the P value that they got, supposedly. It would have been nice to see the whole thing. Um, but I just tested, I grabbed a, I went into a spreadsheet and threw some example data and I was able to get myself that number with some stuff that kind of made some sense. Um. Say the average number of votes for this candidate in those 18 bags was was 5 .5.

0:18:07 Thomas: That's not very many.

0:18:09 Jenessa: So our one mislabeled bag had zero.

0:18:12 Thomas: God. Yeah. Come on. Like nothing. And that's after skewing it, like way we talked about. Yeah.

0:18:17 Jenessa: After taking the zero out and only averaging the other ones.

0:18:21 Thomas: So therefore, you've like upped how much is going to average on be in the other ones, likely, just

0:18:26 Jenessa: We're going to do what is maybe the more complicated part to understand, but definitely the worst Oh test, they're going to take the data and do a one sample t -test.

0:18:51 Jenessa: What could be, for example, let's say you knew across the whole country, every single person who took SAT this year, the average score was a 12 .6. or maybe you're a test designer and you say, I want the average to be 1260, something like that. I pull a sample, I get their average, and I can try to test rather than two groups against each other, test my one group against the null hypothesis, the predicted value, the expected value, just a single number. That number, it's not data. It's just something I, is this this different. So I could say my sample, yeah, I'm sure they didn't exactly average 1260, but how far are they from it really?

0:19:41 Thomas: Like how weird is your sample kind of thing?

0:19:43 Jenessa: Exactly, exactly. So what they did here is they said,

0:19:48 Thomas: Already know their samples. You already know. I got a bag of 100 tails and zero or whatever. Ever. 100 heads and zero tails. Let's see how weird that is. Oh my god, this just in. It's quite weird.

0:20:00 Jenessa: So it's worse than that. We're going to take the average 5 .5, which is the average of of the 18 audited bags and test if it's different from zero. Yes, it is. You know, they got votes. They didn't get zero votes.

0:20:17 Thomas: So stupid.

0:20:18 Jenessa: Yeah, this not testing whether the zero bag is weird. What this tells me is I got an average of 5 .5 votes and I'm very confident the true population population average is not zero votes.

0:20:35 Thomas: Yeah, me too. I'm also confident of that.

0:20:37 Jenessa: In fact, based on the type of data set I have, I'm a hundred percent confident because I know how voting works.

0:20:44 Thomas: I mean, unless they slip into another dimension mid -analysis, I think we know that there's more than zero votes.

0:20:52 Jenessa: There's no negatives here and we're pretty close to zero. So

0:20:56 Thomas: Unless this precinct over here voted negative 50. Yeah.

0:21:01 Jenessa: So we got kind of a tail end of our distribution is cut off a little bit. So yes, in fact, that average of 5 .5, very confident, the true population average not zero that does not does not tell me that any one data point of zero is weird and literally that group that account is posting like my god look what happened

0:21:13 Thomas: Holy shit. Anything? No, it really doesn't. Yeah, that's so stupid because I know and I love stats, except I have a terrible memory for what the, oh, the T and the P value and the N value. I just, I can't, I don't, forget the minute you say it, I forget it. But I do know there's got to be some sort of test that is for if I've got my thousand coins and I'm going to divide them up just randomly in bags of, I don't know, 50 or something, there's got to be some test for like, oh, how weird is this one bag? Like relative to the other ones. And furthermore, like what are the odds of getting a bag like that? Like if I take bags of 50 coins, the odds are getting zero heads. It's going to be pretty low and there's, and, never zero obviously, but like there is some sort of statistical that tells you that, right? Like roughly speaking, how often you could expect to get a bag that's zero heads or something like like that?

0:21:42 Jenessa: It's very low. So it's use certain probability calculations to calculate if we know the average and we know the standard deviation. Easier to think of for coin flips because we know there are two options and each is 50%. So what are the odds of getting blah, blah, blah? Way that would be really common is to just say, well, average is that. The standard deviation is this, so that's a measure of how typical it is to be a certain distance from the average. You know, the average of five standard deviation as one, four isn't weird.

0:22:24 Thomas: I think Oh yeah.

0:22:46 Jenessa: Five's Yeah. Yeah, four isn't weird, three's a little weird, two is very weird.

0:22:49 Thomas: Wait, what would that be in the coin? I'm always so bad at remembering. What would that be in a coin flip example? Is that just like, yeah, go ahead.

0:22:55 Jenessa: So coin flips are categorical data, so you can't do this.

0:22:59 Thomas: Okay. Yeah. Yeah, gotcha.

0:23:00 Jenessa: So it depends on whether we have categorical or interval or ratio data, the type of statistics that are appropriate. We would commonly do, include all the data points and calculate what is three standard deviations away. And if I know my data is normally distributed, so it's appropriate to make these assumptions, and then I show this data point, one, maybe two, depending on the size of my sample, are, oh, wow, those are three standard deviations away, I might be justified in removing them from my data set. This is truly an outlier. If you've got a lot of people on the edge of two standard deviations, like, well, that's just data, man. Like, it's a little weird. It's not that weird. The standard deviation, we Can calculate the percent odds of it but

0:23:50 Thomas: Okay, I that's what I mean by that. It's got to be like, and again, I recognize that apparently the standard deviation thing doesn't work with the coin flips and that glad. Cause I was like, what the F would that be? I couldn't think of what that would be. Um, if, yeah, like in my example, we're we've going to, we got a thousand results and let's just say it's 50, 50 overall for the sake of argument. But if we jumble them up and put them in little baggies of 50, you're the average over everything will be 50 ,50, but like, you'RE going to have some of those bags that are are like, wow, that's really skewed toward heads. How weird. And it's like, I'm pretty sure there's a statistical measure of like, what are the odds roughly that you get a bag that's zero heads, 10 heads, 20, you know, that kind of thing. Isn't there? That's okay. But let

0:24:30 Jenessa: Yeah I'm forgetting the exact formula because I don't do it very often but

0:24:34 Thomas: It suffice to say, it's something like that, that you would want to know about. Well, if you were even looking at this at all, cause it seems like a waste of time, but like, you want to Know like, Wow, this one bag of ballots, what's the likelihood that that we would expect it to be this number of votes based on our overall average of how many votes this person got, right?

0:24:52 Jenessa: And instead we have used definitely not the right statistic we've looked at a single day of early voting in a single polling place and tried to compare it to a sample of 18 other polling places on a single day of early voting for a candidate who only got 29 % of the vote to begin with. So we expect a really small number to begin With.

0:25:11 Thomas: And the lower sample size of anything, just in general, the more you're going to get wild numbers, right.

0:25:22 Jenessa: Yeah. So it reduces your statistical power, which reduces your certainty. See something that gut feels odd, we're less confident in making the statement about exactly how odd it is. So when you take into account standard deviation, the sample size, the mean, you draw your distribution out, you can start to say, well, with that mean and standard deviation it's about this odd. But then I look at that sample size, and I'm going to dial back my confidence quite a bit. Mathematically, that's already built in to the statistics. It will naturally dial your confidence back.

0:26:00 Thomas: Because when you're dealing with humans and small numbers, like let's say, so we got this third party candidate or whoever that's, that's you know, you could imagine something weird where it's like, well, a bunch of people who are very excited to vote for them all went on day one and then day two. I guess this is kind of the opposite of what's happening, but like you can imagine stuff like that where it'S like in the stats, if you were to just do the analysis, you'd be like, whoa, it is crazy outlier that this person who only got 25 % of the votes had so many votes early on. That's wild. But like, we're also dealing with humans. So the explanation could just be, well a church went over at the same time and you know like Like, you Like, it could be something like that. It still doesn't mean, like, even if you got a a wild wild result in terms of your expected kind of odds on that stuff, we're dealing with small numbers and human behavior. So it's like, it Could Be Totally Explainable.

0:26:52 Jenessa: And why you don't want your astrophysicist doing this stuff, because these data would require extreme sociological or political causes for their explanation and if you know anything about psychology and politics and voting behavior and sociology you would know that this is happening and it's exactly the same thing the weirder off the candidate is the more we expect exactly the behaviors you're describing the more we have certain religious communities, certain isolated communities, we expect, especially when you narrow down and pick out a few individual precincts or just one county, of course it looks weird.

0:27:37 Jenessa: Well, and we drop polling places non -randomly. They are dropped in the middle of town and the town could be predominated by a very specific ethnic religious social group of people I

0:27:36 Thomas: Maybe they live near that, or the people who are supporting them happen to be geographically clustered. Right, yeah. Are almost incapable of being arbitrarily distributed when it comes to this because we will cluster and start believing the same thing. It's like not possible for us to do that.

0:28:06 Jenessa: No exactly so like in the same way that I expect you the north country the very rural upstate counties to be more republican I I don't look at that and then go go, son of a bitch. New York went blue, but this county went red. We're like, yeah, man. Like, yeah. What are the odds? Extremely low if your null hypothesis is that they should be the same.

0:28:22 Thomas: What are the odds? Yeah. Well, no, you see, every county contains like, you know, five Muslims, five Christians. It's all randomly, you know.

0:28:39 Jenessa: One in five people is from China in our state.

0:28:44 Thomas: Yeah, a lot of people from India. Yeah it's just how it works. Oh

0:28:47 Jenessa: Know, we're just going to keep seeing this. I just like scrolled through the Twitter account for this organization And like, there's a whole screenshot of their handwritten, I surveyed this poll site and was a poll watcher. And like I'm going to really conveniently put the red arrow of like, look at this data point to cover up the data point that somebody making a podcast who wanted to do the math and check your math can't see the data. A point that I needed to double check. But it was like, oh, for scanner A, we had this many votes counted. And when we look at the overall counter, it started at this and it ended at this. And that perfectly adds up. For scanner C, it's the same thing. And then for scanner B, well, scanner B broke halfway through the day. And using the estimated count by random poll worker, whose whose name I wrote down, versus the laptop, which is a slightly different number, neither of which is the freaking broken scanner. Oh my God, we in this one poll site have a discrepancy of 16 votes. We got to throw out all the scanners.

0:29:58 Jenessa: It's like the scammers that didn't break were completely accurate to the vote, even though it's already recorded like tens of thousands of votes.

0:29:57 Thomas: My god. Yeah oh boy.

0:30:07 Jenessa: It's like, it's hunting for anomalies. And you look hard enough, you're going to find them. And they don't mean anything.

0:30:15 Thomas: Especially because what is your grand theory? Because we're dealing with small candidates who nobody gives a shit about, sorry, the Theory cannot be artificially, manually somebody tinkering with numbers because otherwise these votes would be either they would have gone toward the person you're trying to get to win or they would be ignored. Like those would be the two things likely they would Be. So your only theory can be that means the machines are broken and all hell breaks loose. Like it's just willy nilly. Who knows? And if that were true, the results that you would expect would be just insane. Like you can't it wouldn't be like little bits of votes here and there are one or two off of what you think.

0:30:54 Jenessa: Yeah. And another thing to know is, so in New York, the different counties pick what machines are going to use. So some counties are using, oh God, the company is E S and S I forget. It's like electronic software and, um, then there's dominion. People are probably familiar by now. Um, dann there's like one other one that's like kind of new to the scene. Um, like, make these scanners, and different counties pick different ones. I think Rockland is an ES &S county, because I surveyed all these counties, and I know what their poll sites look like.

0:31:28 Thomas: So you saw all the fraud.

0:31:28 Jenessa: Yes, yes, I've walked around. I caused it, actually. I like crack, oh, my god. Just ridiculous. The whole thing is ridiculous. Like I surveyed so many poll sites.

0:31:32 Thomas: Distributing USB what drives was it? Rudy said USB drives. No, I always forget. It was something dumber than that. Damn it. I always forget what it was. USB drives

0:31:42 Jenessa: He said they were passing out USB U drives like it was crack cocaine, which is just like obviously racially coded because it was two black women.

0:31:46 Thomas: Was drives. OK. Well, yeah, of course, when I'm counting votes, votes I always am like can you pass pardon me can you pass the crack cocaine that's normal

0:31:56 Jenessa: And come on. It's so holy shit. Yeah, no, I've surveyed poll sites and poll workers are like, who are you? What are you doing here? You can go here, but you can't go there. You can look at this, butyou cannot look at that. And like, I have federal authority to be there. That's like part of my job. So like, yeah, randos can't come in. And so it's really exactly like you're saying, if you think the machine is just unreliable, what you would see is you'd be like, damn, blue state got a new machine and all of a sudden a bunch of Republicans won or vice versa, or like counties behaved in ways you didn't expect. If you think it is a grand conspiracy where someone has hacked the machine, which I happen to know is precisely what this organization, they're not writing it in this legal document, but it is what they spread on their social media all the time, then you would have to presuppose that this has been happening for a really long time. Because across elections, it's actually normally it's quite stable, especially statewide, who's winning, who'S not. It changes slightly over time by demographics. But like, still requires even the people who are experts in this, who think that we have a security problem, and I personally disagree with them knowing about the security procedures that we have, even they admit that the hacks they propose require an inside man. It's not connected to the internet. Can't You sneak into poll sites and do this. It requires an election worker, someone who decided that they would dedicate their life to voting to not care about voting and elections and then like either be the one doing this or let you in to hack a voting machine and just never adds up to some logical behavior that is truly executable on anything more than like maybe a township level and

0:33:55 Jenessa: Who and why and

0:33:54 Thomas: You would need a little Clive Owen inside every voting machine it's

0:34:01 Jenessa: Like it's just it's not happening this is all very silly

0:34:03 Thomas: The movie Inside Man that I barely remember I had to look up which actor it was but yeah you get a little Clive Owens locks himself in the voting machine and then later on does the I think that's what it was guys been been 20 years since I've seen that or no. Okay. Oh, it looks like it was released 19 years ago ish. So, uh, we're old anyway. Yeah. Silly. The point I was made a moment ago is if it were some sort of deliberate hacking, like to what end, like you could only say someone did a really bad job of hacking and changing votes that had no conceivable effect that makes any sense. And the way you're finding out is handfuls of votes for a candidate no one cares about they're

0:34:16 Jenessa: I know what you're talking about oh my god yeah it's been a while so yeah yeah like are we genuinely proposing proposing that the LaRouche party was making a comeback. I guarantee you most New Yorkers do not know what that is, who that is. There's no way. The things I saw come across social media was the same organization proposing that this happened in North Carolina. They're looking looking at the drop -off rate. Oh my God, like got this many votes and look at the general. So now we're not looking at The Senate. We're always going to pick, right?

0:34:48 Thomas: The key they're the key wow yeah you get to pick which different thing looks weird yeah

0:35:16 Jenessa: Instantly, I just checked the comments because I don't know, I hate my life and I just want to make myself miserable by seeing what non -experts think. And actually, I was very impressed because this was on our North Carolina. So it's people who know their state and yes, it's the exact same thing where I know my state and my county and what the weird sociology is.

0:35:30 Thomas: Reddit that you subreddit yeah gotcha

0:35:41 Jenessa: Everybody was like, oh, is like the most popular person on the planet. Republicans love him and are voting for him and independents and Democrats. And like, it surprises no one that you would have somebody who voted Trump and that Democrat AG. And I think there was also something going on where, like, the Republican AG candidate had some kind of something or other going on. And was also the year, I know it's a different race, but, like I think it was the governor had the Republican candidate who was a Black guy was making extremely anti -Black

0:35:40 Thomas: Yeah right yeah yeah all right

0:36:17 Jenessa: Racist comments. Comments.

0:36:18 Thomas: That guy yeah was it mark

0:36:19 Jenessa: That sounds right. Yes.

0:36:19 Thomas: Robinson or something something like that yeah

0:36:23 Jenessa: So it was already weird.

0:36:23 Thomas: You're gonna gonna have out if you get to just survey the nation if you get to look at all of these bags of 50 coins across the entire nation that were randomly let's say hypothetically distributed you can pick the handfuls of ones that wow there's zero heads in this one wow there's no tails in this one it's not impressive it doesn't it's just not right

0:36:43 Jenessa: And kept happening in that video that a lot of people flagged was the person kept saying, you don't need to be a statistician to know. My God. Everybody was like, I precisely want a statistician. I want to see the trend over time and not just that this looks funny but how unusual is it and like that know this claim that like it's always framed as like oh this is the behavior of democrats and it's like no if you vote you're a republican who voted trump and the dem that would explain this behavior it one of these combinations could explain this and a person who didn't vote Harris and then dropped off the AG one or voted for the Republican AG for God knows what reason. So many weird behaviors happening. So it's picking out little things and using it to get people... I don't think it's an accident. This was filed in December. The motion to not dismiss it was decided in March. But this whole thing started blowing up on my social media right after musk tweeted oh you know you're welcome for the election or whatever the hell he said like

0:36:54 Thomas: You do oh yeah

0:37:57 Jenessa: Think it capitalized on that to get all over every piece of social media of like yeah we do have data for this and it's like we don't that there's no plausible logic behind this the statistics are not making any sense this is an organization where I have found multiple instances of abusing the hell out of statistics and doing things that make no sense. They were pro se until five minutes ago. Just no, no, no, I hope you all can carry this forward and give up on this belief that we're going to fix this.

0:38:42 Thomas: And another point, I just don't know if it needs to be said but like the two parties are not equal on this I know there's a lot of people who love to try to like anytime something like this pops up they love to be like look at that it's now the blue and on and I was I was kind of joking about that there there is certainly a small and there always has been, there's always been a small subset of Democrats, especially around anti -vax behavior. I know that's sorted differently now, I think, but like they're I used to argue with Democrat or I guess I'll say left anti-vaxers all the time who would mostly either vote for Democrats or maybe they don't vote because they're mad at the system or whatever. And that you're going to have your, weirdos in everything. But I had the same experience you did when it got to the top of Reddit. Now, while that does mean like a lot of people are upvoting it, maybe they're not looking at it. They're just upvoted like, oh yeah, weirdness, weirdness.

0:39:28 Thomas: The comments for the most part were like, uh, I don't think this makes sense. Like, I don'T really see it, which it was refreshing. It is nice to see a lot of people applying actual skepticism to the situation. So I'm glad that you were here to give us more detail to that. I suspected this was nothing, but that was a lot lot of fun. I, I love someone may, I, if there's a, um, would it be like if Twitter feed of the, or whatever feed these days of just people making really stupid graphs, I'm realizing in my life, I actually want that. That must exist. Maybe there's an subreddit, subreddet r slash I am bad at making graphs.

0:40:03 Jenessa: I can't remember the name of it but this does exist there's a whole thing of like it's worst offender is like bad y -axis like truncated like look at the difference and it's like one it'S percent versus four percent but they've made the bar massive people do that a lot.

0:40:05 Thomas: I bet it does. So good. Yeah. Yeah. That's, want ones like this where it was like man that's funny I haven't seen that in a while like I've seen the yeah zoom in on the axis or make it you know whatever cut it off or do the like dot dot dot and then like you skipped know a bunch in your axis but the having the precinct number be an axis was truly that was that I enjoyed that that made my day so stupid

0:40:27 Jenessa: Like That's common. Yup. Yup, uh -huh. I don't know how you look at a horizontal line and not be like, I mean, that can't be right.

0:40:44 Thomas: Well anything else we gotta talk about or look for in this one any closing thoughts oh

0:40:52 Jenessa: Definitely I definitely agree with you that, like, you as opposed to seeing people go like, oh, my God, it's true. I know, approached me knowing I am a voting rights attorney. And what I got was, hey, have you heard about this? What do you know about this, like a request to explain and not, oh my God. God it's so like I do I am really seeing a difference of people asking and being open to hearing and like yeah I do not think this is on the level of like I just don't like when the misinformation is getting out there and I think it's more tempting when people hear a whole county got zero that's just a game of telephone is really screwing this up so no more game of telephone. Fix

0:41:35 Thomas: The the theft of our country's dignity is happening right in front of us too it's like we're looking people are looking for these like Like sneaky, oh, it seems so much more like if it's this conspiracy, these numbers of, it's like, watch the news, watch The Fucking News for five seconds where we have a right -wing misinformation machine and then a completely inept or corrupt mainstream media that I think would rather have an interesting president that gives them ratings than, a country. Country yeah and and social media all these forces like there's so many forces that are ruining our country and they're doing it in a very obvious and dumb and out in the open way and then it's not these sly geniuses hacking stuff you enhance

0:41:53 Jenessa: The nefarious is, yeah. Clickbait, man. It's killing us. Hmm. Hm.

0:42:24 Thomas: Over and over all

0:42:25 Jenessa: And answer is not that the little old lady at your poll site should hand count and write down little tally marks, hundreds of thousands of votes. That That can't be the correct answer for how to do this.

0:42:40 Thomas: Well thanks so much for the breakdown of that um plugs you want to to toss out any any place you want to send people on this or

0:42:46 Jenessa: Would say right now, primaries just ended in New York, so but we're in a bit of a lull, but that general is going to sneak up on you faster than you think. So if you are not registered, if you want to change your address, anything like that, New York state, we will have our general election for a lot of local offices on Tuesday, Tuesday, November something. Your calendars. Come up sooner than you think. So start getting educated on your local politicians.

0:43:14 Thomas: All right Jenessa seymour thanks so much and forward to talking with you again you

0:43:18 Jenessa: Good to see you. To you. Something like that.

0:43:23 Thomas: Can retake it if you want good there's an asymptote on the district zero. Let's see how weird that is. Oh my god, this just in. It's quite weird. I still scrolling in this document or are you on something else?

0:43:24 Jenessa: Good talking to you! I'm on something else, but have fun if you want.

0:43:57 Thomas: Thank God. No, it's distracting me. I got to get out of here.

0:44:01 Jenessa: Needed to do was a fucking bar graph. So