WEBVTT 00:00:00.000 --> 00:00:02.339 Welcome back. It's incredibly good to have you 00:00:02.339 --> 00:00:04.019 with us again today. Yeah, it's great to be here. 00:00:04.219 --> 00:00:07.200 Whether you are prepping for a meeting or catching 00:00:07.200 --> 00:00:09.300 up on your field or, you know, you're just one 00:00:09.300 --> 00:00:11.560 of those naturally curious people who loves pulling 00:00:11.560 --> 00:00:14.080 on a thread to see what unravels, you are in 00:00:14.080 --> 00:00:16.539 the right place. Absolutely. We are taking you 00:00:16.539 --> 00:00:18.980 on another deep dive today, and the source material 00:00:18.980 --> 00:00:21.420 we're looking at is, well, it's one of those 00:00:21.420 --> 00:00:24.320 unexpected gold mines of information. It really 00:00:24.320 --> 00:00:27.000 is, because when you first glance at it, it just 00:00:27.000 --> 00:00:29.570 looks like a standard reference document. Right. 00:00:29.730 --> 00:00:31.690 But once you start looking at the proportions 00:00:31.690 --> 00:00:35.369 and the actual timelines, a very human narrative 00:00:35.369 --> 00:00:38.350 starts to emerge from the data. Today, our mission 00:00:38.350 --> 00:00:41.310 is to unpack a Wikipedia article detailing the 00:00:41.310 --> 00:00:44.469 list of United States Senate elections in Colorado. 00:00:45.530 --> 00:00:48.850 Now, I know. It sounds dry. It really does. A 00:00:48.850 --> 00:00:51.969 massive chronological list of dates, candidate 00:00:51.969 --> 00:00:54.590 names, and voter turnout numbers spanning from 00:00:54.590 --> 00:00:57.689 the 1870s all the way to the 2020s sounds like 00:00:57.689 --> 00:00:59.810 a dusty spreadsheet. A very long spreadsheet. 00:01:00.130 --> 00:01:02.390 Exactly. Yeah. You might be wondering how an 00:01:02.390 --> 00:01:05.010 encyclopedia table translates into an engaging 00:01:05.010 --> 00:01:08.170 story. OK, let's unpack this, because hidden 00:01:08.170 --> 00:01:11.129 within these raw data points is a gripping century 00:01:11.129 --> 00:01:14.569 long record of razor thin margins, massive population 00:01:14.569 --> 00:01:17.989 booms and some truly wild political anomalies. 00:01:18.049 --> 00:01:20.150 And what I love about looking at a raw historical 00:01:20.150 --> 00:01:22.870 ledger like this is that it completely strips 00:01:22.870 --> 00:01:25.150 away the political noise. Yeah, the bias is gone. 00:01:25.290 --> 00:01:27.670 Right. We aren't here to debate the merits of 00:01:27.670 --> 00:01:30.129 the Democrats or the Republicans or the various 00:01:30.129 --> 00:01:32.569 third parties that pop up. We are just looking 00:01:32.569 --> 00:01:35.969 at the pure math of democracy and what those 00:01:35.969 --> 00:01:38.629 vote counts tell us about the electorate at that 00:01:38.629 --> 00:01:41.469 specific time. We're just data detectives. Exactly. 00:01:41.549 --> 00:01:43.989 The data acts as a neutral mirror reflecting 00:01:43.989 --> 00:01:47.069 the priorities of the voters. We're totally impartial 00:01:47.069 --> 00:01:49.769 here. We are just following the numbers. So to 00:01:49.769 --> 00:01:51.629 get started, let's establish a quick baseline 00:01:51.629 --> 00:01:53.769 with the mechanics of the Senate as laid out 00:01:53.769 --> 00:01:55.849 in the text. Good idea. We all know that every 00:01:55.849 --> 00:01:58.930 state is allotted two U .S. senators. But the 00:01:58.930 --> 00:02:01.269 key structural element that shapes this entire 00:02:01.269 --> 00:02:04.310 data set is that Senate terms are staggered. 00:02:04.310 --> 00:02:07.150 They serve six -year terms, but those terms are 00:02:07.150 --> 00:02:09.490 broken up into classes. And because Senate terms 00:02:09.490 --> 00:02:12.590 are staggered into those classes, this data set 00:02:12.590 --> 00:02:15.409 isn't just a single timeline. It is effectively 00:02:15.409 --> 00:02:18.530 two parallel tracks of Colorado history running 00:02:18.530 --> 00:02:21.030 side by side. That's a great way to put it. When 00:02:21.030 --> 00:02:23.150 Colorado was officially admitted to the union 00:02:23.150 --> 00:02:26.849 back in 1876, the state was assigned a class 00:02:26.849 --> 00:02:29.469 two seat and a class three seat. Right, right. 00:02:29.629 --> 00:02:31.909 The purpose of this system, obviously, is to 00:02:31.909 --> 00:02:34.090 ensure that only a third of the entire U .S. 00:02:34.129 --> 00:02:36.770 Senate is up for election every two years. It 00:02:36.770 --> 00:02:39.710 prevents total turnover. So when we look at this 00:02:39.710 --> 00:02:41.770 list, we're really tracking the evolution of 00:02:41.770 --> 00:02:43.750 class two and class three completely independently. 00:02:44.090 --> 00:02:46.909 But the way those early seats were filled. It 00:02:46.909 --> 00:02:49.530 looks completely foreign compared to modern elections. 00:02:49.830 --> 00:02:52.150 There is a massive pivot point in our source 00:02:52.150 --> 00:02:55.270 material right at the year 1913. Oh, yeah. The 00:02:55.270 --> 00:02:58.150 17th Amendment. Before 1913, you don't see massive 00:02:58.150 --> 00:03:00.710 popular vote totals in the tables because the 00:03:00.710 --> 00:03:02.870 state legislatures were the ones actually choosing 00:03:02.870 --> 00:03:06.409 the senators behind closed doors. Then 17th Amendment 00:03:06.409 --> 00:03:09.210 passes. And everything changes. Practically overnight, 00:03:09.490 --> 00:03:13.310 the system shifts to direct election by the voters. 00:03:13.819 --> 00:03:15.860 What's fascinating here is how fundamentally 00:03:15.860 --> 00:03:19.280 that constitutional amendment changes the very 00:03:19.280 --> 00:03:22.180 nature of the data we are reading. Totally. Prior 00:03:22.180 --> 00:03:24.979 to 1913, a list of senators is essentially a 00:03:24.979 --> 00:03:27.620 record of political consensus among state level 00:03:27.620 --> 00:03:30.500 power brokers. You're looking at backroom deals 00:03:30.500 --> 00:03:33.400 and legislative majority established. But the 00:03:33.400 --> 00:03:36.520 moment you cross that 1913 threshold, the entire 00:03:36.520 --> 00:03:39.900 nature of the ledger transforms. These tables 00:03:39.900 --> 00:03:42.539 and percentages become a direct reflection of 00:03:42.539 --> 00:03:45.080 the public will. It completely democratizes the 00:03:45.080 --> 00:03:47.750 spreadsheet. Just imagine being a citizen right 00:03:47.750 --> 00:03:49.909 at that transition point. You're living in Colorado 00:03:49.909 --> 00:03:52.889 and suddenly you have direct, tangible power 00:03:52.889 --> 00:03:55.270 over a massive federal choice. That's a huge 00:03:55.270 --> 00:03:57.729 shift. And that shift in power leads directly 00:03:57.729 --> 00:04:00.610 into the actual voter data. And man, did the 00:04:00.610 --> 00:04:02.810 voters of Colorado make some of these early direct 00:04:02.810 --> 00:04:06.169 elections incredibly close. The razor -thin margins 00:04:06.169 --> 00:04:08.150 are where the statistics really start to show 00:04:08.150 --> 00:04:10.770 the tension of the era. When you transition to 00:04:10.770 --> 00:04:12.849 a popular vote, you suddenly introduce the reality 00:04:12.849 --> 00:04:15.650 that a tiny fraction of the population can literally 00:04:15.650 --> 00:04:18.689 dictate the direction of the entire state. I 00:04:18.689 --> 00:04:20.790 really want to look closely at the tightest races 00:04:20.790 --> 00:04:23.569 we found in this list, because focusing on the 00:04:23.569 --> 00:04:26.829 raw numbers grounds the history. Let's pull up 00:04:26.829 --> 00:04:30.399 the Class III special election in 1932. Oh, the 00:04:30.399 --> 00:04:32.899 Schuyler and Walker race. Yes. You have Carl 00:04:32.899 --> 00:04:35.040 C. Schuyler, the Republican candidate, and Walter 00:04:35.040 --> 00:04:37.240 Walker, the Democratic candidate. When the dust 00:04:37.240 --> 00:04:41.439 settles, Schuyler beats Walker by exactly 1 ,065 00:04:41.439 --> 00:04:45.439 votes out of over 414 ,000 total votes cast. 00:04:45.759 --> 00:04:47.740 That is a margin of about one quarter of one 00:04:47.740 --> 00:04:50.120 percent. A quarter of a percent. A U .S. Senate 00:04:50.120 --> 00:04:53.100 seat was decided by barely 1 ,000 votes. Think 00:04:53.100 --> 00:04:55.480 about the logistics of that. That is the size 00:04:55.480 --> 00:04:57.480 of a single neighborhood deciding the federal 00:04:57.480 --> 00:04:59.939 representation for the entire state. It really 00:04:59.939 --> 00:05:02.920 is. A single rainy day in one specific county 00:05:02.920 --> 00:05:05.439 or a localized issue with a polling place somewhere 00:05:05.439 --> 00:05:08.420 could have easily flipped the U .S. Senate. It 00:05:08.420 --> 00:05:11.579 forces you to reevaluate how we view historical 00:05:11.579 --> 00:05:14.509 mandates. When we look back, we tend to think 00:05:14.509 --> 00:05:16.810 of election winners as having a broad consensus 00:05:16.810 --> 00:05:19.509 of the people backing them. Right. Like the whole 00:05:19.509 --> 00:05:21.810 state was on board. Exactly. But this data proves 00:05:21.810 --> 00:05:24.649 that often there is no broad consensus. There's 00:05:24.649 --> 00:05:27.430 just a fractured electorate where one side managed 00:05:27.430 --> 00:05:30.389 to find a tiny handful of extra supporters. Yeah. 00:05:30.490 --> 00:05:33.790 And that 1932 race wasn't a one -off anomaly. 00:05:34.250 --> 00:05:37.649 The text shows several of these absolute squeakers 00:05:37.649 --> 00:05:39.629 throughout the century. You see it again a decade 00:05:39.629 --> 00:05:42.980 later. You do. Take the 1942 Class II race. Democrat 00:05:42.980 --> 00:05:45.639 Edwin C. Johnson goes up against Republican Ralph 00:05:45.639 --> 00:05:48.540 Lawrence Carr. Johnson edges him out with roughly 00:05:48.540 --> 00:05:52.459 50 .2 % of the vote compared to Carr's 49 .2%. 00:05:52.459 --> 00:05:54.560 Wow. It was a difference of just a few thousand 00:05:54.560 --> 00:05:57.620 votes. Or you can skip ahead to the 1972 Class 00:05:57.620 --> 00:06:00.139 II upset. Oh, that's a good one. The Democrat 00:06:00.139 --> 00:06:02.800 Floyd Haskell narrowly defeats the incumbent 00:06:02.800 --> 00:06:06.199 Republican. Gordon Allott. Haskell pulls in about 00:06:06.199 --> 00:06:11.199 49 .4 % to Allott's 48 .3%. Barely 1%. Exactly. 00:06:11.439 --> 00:06:13.819 Haskell unseated a sitting senator, a guy who 00:06:13.819 --> 00:06:16.199 had been there for three terms, by a margin of 00:06:16.199 --> 00:06:18.620 just over 1%. When you lay those numbers out, 00:06:18.699 --> 00:06:21.660 it really dismantles the idea that incumbency 00:06:21.660 --> 00:06:25.420 makes a politician bulletproof. A 1 % margin 00:06:25.420 --> 00:06:28.439 against a three -term incumbent means the challenger 00:06:28.439 --> 00:06:30.220 essentially chipped away at the foundation until 00:06:30.220 --> 00:06:32.399 the whole building collapsed. It raises an important 00:06:32.399 --> 00:06:35.470 question, though. Why does this granular level 00:06:35.470 --> 00:06:39.350 of data from 1932 or 1972 matter to us today? 00:06:39.529 --> 00:06:42.350 In our modern world, we're constantly dealing 00:06:42.350 --> 00:06:45.550 with information overload and political fatigue. 00:06:46.569 --> 00:06:49.230 it is incredibly easy to feel like individual 00:06:49.230 --> 00:06:51.810 actions are just a drop in the ocean very true 00:06:51.810 --> 00:06:54.550 but these historical ledgers prove mathematically 00:06:54.550 --> 00:06:57.170 that a handful of engaged citizens genuinely 00:06:57.170 --> 00:07:00.110 tip the scales of history the data shows that 00:07:00.110 --> 00:07:02.009 the margin of error in a functioning democracy 00:07:02.009 --> 00:07:04.779 is often shockingly small It makes you wonder 00:07:04.779 --> 00:07:07.180 how many people stayed home in 1932 thinking 00:07:07.180 --> 00:07:09.360 their vote wouldn't change anything, completely 00:07:09.360 --> 00:07:11.579 unaware that a thousand extra people could have 00:07:11.579 --> 00:07:13.279 changed the course of the election. Precisely. 00:07:13.300 --> 00:07:15.279 And speaking of data that challenges our assumptions, 00:07:15.540 --> 00:07:17.620 here's where it gets really interesting. Let's 00:07:17.620 --> 00:07:19.439 talk about the anomaly I found in the Class 3 00:07:19.439 --> 00:07:21.120 data, because it might be the most fascinating 00:07:21.120 --> 00:07:23.399 narrative in the entire article. The Ben Nighthorse 00:07:23.399 --> 00:07:26.459 Campbell era in the 1990s. Yes. Let's look at 00:07:26.459 --> 00:07:29.779 the timeline here. In 1992, Ben Nighthorse Campbell 00:07:29.779 --> 00:07:32.040 runs for the Class 3 Senate seat representing 00:07:32.040 --> 00:07:34.800 the Democratic Party. He wins a solid victory, 00:07:35.000 --> 00:07:38.019 capturing nearly 52 percent of the vote and beating 00:07:38.019 --> 00:07:41.519 the Republican challenger, Terry Considine. Campbell 00:07:41.519 --> 00:07:44.019 goes to Washington as a Democratic senator. That 00:07:44.019 --> 00:07:46.819 is a standard entry for the ledger, a solid single 00:07:46.819 --> 00:07:49.560 digit margin victory for a major party. Standard, 00:07:49.579 --> 00:07:52.720 right. But then follow that class three seat 00:07:52.720 --> 00:07:56.040 to the very next election cycle in 1998. Ben 00:07:56.040 --> 00:07:58.079 Nighthorse Campbell is up for reelection and 00:07:58.079 --> 00:08:00.319 he wins again. Right. But he doesn't win as a 00:08:00.319 --> 00:08:03.160 Democrat. He wins as a Republican. He runs against 00:08:03.160 --> 00:08:06.300 Democrat Dottie Lamb. And Campbell absolutely 00:08:06.300 --> 00:08:08.339 crushes it under a completely different party 00:08:08.339 --> 00:08:10.600 banner. If you pause and just look at what that 00:08:10.600 --> 00:08:12.839 implies about the voting public, it is a remarkable 00:08:12.839 --> 00:08:15.500 piece of data. We don't need the outside historical 00:08:15.500 --> 00:08:18.459 context or the political backstory of why he 00:08:18.459 --> 00:08:20.639 switched to see the story here. No, the numbers 00:08:20.639 --> 00:08:23.120 say it all. The raw numbers tell us everything 00:08:23.120 --> 00:08:25.279 we need to know about candidate charisma versus 00:08:25.279 --> 00:08:28.920 party infrastructure. Oh, wait. If Campbell won 00:08:28.920 --> 00:08:31.680 as a Democrat in 92 and then dominated as a Republican 00:08:31.680 --> 00:08:34.899 in 98, does the data show his total voter base 00:08:34.899 --> 00:08:37.620 growing? Or did he just swap one half of the 00:08:37.620 --> 00:08:39.779 state's voters for the other? That's the fascinating 00:08:39.779 --> 00:08:42.139 part. When he ran as a Democrat, he won just 00:08:42.139 --> 00:08:45.320 over 800 ,000 votes. When he ran as a Republican 00:08:45.320 --> 00:08:48.539 six years later, he won nearly 830 ,000 votes, 00:08:48.679 --> 00:08:52.019 beating Lamb 62 .5 percent to 35 percent. So 00:08:52.019 --> 00:08:54.659 it went up. Yes. His total vote count actually 00:08:54.659 --> 00:08:56.860 increased and his margin of victory exploded. 00:08:57.279 --> 00:08:59.840 This suggests that a massive chunk of the Colorado 00:08:59.840 --> 00:09:02.120 electorate was voting for the person, not the 00:09:02.120 --> 00:09:05.100 party. That completely upends the conventional 00:09:05.100 --> 00:09:08.360 wisdom that voters just blindly check the box 00:09:08.360 --> 00:09:11.480 for their preferred party. To win 52 percent 00:09:11.480 --> 00:09:14.080 as a Democrat and then 62 percent as a Republican 00:09:14.080 --> 00:09:17.259 means that hundreds of thousands of voters looked 00:09:17.259 --> 00:09:19.519 at his name, saw a different letter next to it 00:09:19.519 --> 00:09:22.120 and said, yeah, I'm still voting for him. Exactly. 00:09:22.179 --> 00:09:24.919 It shows a unique personal relationship between 00:09:24.919 --> 00:09:27.899 a specific candidate and the voters that totally 00:09:27.899 --> 00:09:30.360 transcends the party line. It also highlights 00:09:30.360 --> 00:09:33.639 how fluid political coalitions can be. Voters 00:09:33.639 --> 00:09:36.679 aren't monoliths. And we see that fluidity even 00:09:36.679 --> 00:09:39.460 more clearly when we look beyond the two major 00:09:39.460 --> 00:09:42.220 parties. The third parties. Yes. When margins 00:09:42.220 --> 00:09:44.600 are as tight as that 1932 race we discussed, 00:09:44.899 --> 00:09:47.139 third party candidates stop being statistical 00:09:47.139 --> 00:09:49.399 footnotes and start acting as kingmakers. The 00:09:49.399 --> 00:09:52.059 third parties in this text are incredible barometers 00:09:52.059 --> 00:09:54.100 for what was going on in the country. Let's go 00:09:54.100 --> 00:09:56.500 all the way back to the 1914 class three race. 00:09:56.659 --> 00:09:58.840 This is right after the 17th Amendment. So the 00:09:58.840 --> 00:10:01.340 voters are newly empowered. It was a close race 00:10:01.340 --> 00:10:04.019 between the Democrat. Charles S. Thomas and the 00:10:04.019 --> 00:10:07.100 Republican Hubert work. Thomas won with about 00:10:07.100 --> 00:10:09.559 40 percent of the vote. But look at the others 00:10:09.559 --> 00:10:12.200 column. A progressive candidate named Benjamin 00:10:12.200 --> 00:10:15.120 Griffith took nearly 11 percent of the vote. 00:10:15.259 --> 00:10:18.419 And a socialist candidate named J .C. Griffiths 00:10:18.419 --> 00:10:21.879 took over 5 .5 percent. Two third party candidates 00:10:21.879 --> 00:10:24.799 pulling over 16 percent of the total electorate. 00:10:24.820 --> 00:10:27.200 Right. Benjamin Griffith and J .C. Griffiths. 00:10:27.480 --> 00:10:29.419 Practically the same last name, running for two 00:10:29.419 --> 00:10:31.700 completely different alternative parties, and 00:10:31.700 --> 00:10:34.320 together siphoning off a massive chunk of the 00:10:34.320 --> 00:10:36.419 voting base from the establishment candidates. 00:10:36.639 --> 00:10:38.879 It's incredible. And you see a similar disruption 00:10:38.879 --> 00:10:41.500 a decade later in 1924. This Farmer Labor Party 00:10:41.500 --> 00:10:43.700 suddenly appears in the data. In the regular 00:10:43.700 --> 00:10:45.840 Class 2 election, their candidate, Morton Alexander, 00:10:46.279 --> 00:10:48.899 pulls about 5 % of the vote. In the special Class 00:10:48.899 --> 00:10:51.519 3 election that same year, their candidate, Charles 00:10:51.519 --> 00:10:54.480 T. Phillip, pulls over 5 .5%. What is crucial 00:10:54.480 --> 00:10:56.820 to understand about these alternative party spikes 00:10:56.820 --> 00:10:59.639 is that they represent shifting cultural and 00:10:59.639 --> 00:11:02.899 economic priorities. A spreadsheet can't tell 00:11:02.899 --> 00:11:05.039 you the emotional state of the voters, but it 00:11:05.039 --> 00:11:07.279 can show you where they are directing their frustration. 00:11:07.500 --> 00:11:08.980 Yeah, where the pressure is building. Exactly. 00:11:09.419 --> 00:11:12.519 When you see a progressive and a socialist capturing 00:11:12.519 --> 00:11:16.220 double digit percentages in 1914 or a farmer 00:11:16.220 --> 00:11:18.759 labor party establishing a solid baseline across 00:11:18.759 --> 00:11:22.259 multiple races in 1924, you are seeing the voters 00:11:22.259 --> 00:11:25.059 explicitly voicing their demands. They are essentially 00:11:25.059 --> 00:11:27.980 telling the major parties, neither of you are 00:11:27.980 --> 00:11:30.500 addressing our needs. So we are going to park 00:11:30.500 --> 00:11:32.899 our votes over here until you figure it out. 00:11:32.980 --> 00:11:35.759 Exactly. The data captures the mood of the era. 00:11:35.879 --> 00:11:38.340 It shows that the electorate was deeply concerned 00:11:38.340 --> 00:11:41.039 with labor rights, agricultural policy and progressive 00:11:41.039 --> 00:11:43.960 reform during those decades, even if those third 00:11:43.960 --> 00:11:46.139 party candidates didn't actually win the seat. 00:11:47.000 --> 00:11:49.899 So what does this all mean when we zoom out and 00:11:49.899 --> 00:11:52.860 look at the entire century as a single continuous 00:11:52.860 --> 00:11:56.019 timeline? Because for me, as I was reading through 00:11:56.019 --> 00:11:58.279 this list, the most shocking thing wasn't the 00:11:58.279 --> 00:12:01.559 1 ,000 vote margins or the party switching. What 00:12:01.559 --> 00:12:04.720 was it? It was the sheer, unbelievable scale 00:12:04.720 --> 00:12:08.320 of the numbers as time marched forward. Ah, yes. 00:12:08.580 --> 00:12:10.980 The tides of turnout. The demographic footprint 00:12:10.980 --> 00:12:13.679 of the state changes entirely. Let's compare 00:12:13.679 --> 00:12:16.720 the total vote counts from those early 1900s 00:12:16.720 --> 00:12:18.960 races to modern times because it illustrates 00:12:18.960 --> 00:12:20.779 a fundamentally different logistical reality. 00:12:21.460 --> 00:12:24.720 Going back to that 1914 Class 3 election with 00:12:24.720 --> 00:12:27.240 the progressives and the socialists, if you add 00:12:27.240 --> 00:12:30.000 up every single vote cast across all the parties, 00:12:30.179 --> 00:12:33.860 you have roughly 250 ,000 total votes in the 00:12:33.860 --> 00:12:36.320 entire state of Colorado. A quarter of a million 00:12:36.320 --> 00:12:38.340 people participating in the Democratic process. 00:12:38.799 --> 00:12:40.860 Now, hold that number in your head and fast forward 00:12:40.860 --> 00:12:43.519 to the 2020 Class 2 election. You have Democrat 00:12:43.519 --> 00:12:45.340 John Hickenlooper running against Republican 00:12:45.340 --> 00:12:48.240 Cory Gardner. Hickenlooper pulls in over 1 .7 00:12:48.240 --> 00:12:51.529 million votes. Wow. Gardner gets over. 1 .4 million 00:12:51.529 --> 00:12:53.350 votes. Yeah. If you combine that with the other 00:12:53.350 --> 00:12:55.769 candidates, you were looking at over 3 .2 million 00:12:55.769 --> 00:12:58.090 combined votes cast in a single Senate race. 00:12:58.289 --> 00:13:00.570 If we connect this to the bigger picture, this 00:13:00.570 --> 00:13:02.950 Wikipedia list isn't just a record of political 00:13:02.950 --> 00:13:05.789 winners and losers. It is a mathematical timeline 00:13:05.789 --> 00:13:09.789 of Colorado's population explosion. You can literally 00:13:09.789 --> 00:13:12.990 trace the state's growth decade by decade just 00:13:12.990 --> 00:13:15.250 by tracking the rising tide of total ballots 00:13:15.250 --> 00:13:18.169 cast. Think about what that means for the candidates 00:13:18.169 --> 00:13:21.669 themselves. In 1914, a candidate only needed 00:13:21.669 --> 00:13:24.090 to reach and persuade about 100 ,000 people to 00:13:24.090 --> 00:13:26.629 win a Senate seat. You could theoretically travel 00:13:26.629 --> 00:13:29.269 by train, stop in the major towns, shake enough 00:13:29.269 --> 00:13:31.730 hands, and build a coalition. That was the playbook. 00:13:31.870 --> 00:13:34.610 But by 2020, you were trying to communicate with 00:13:34.610 --> 00:13:37.710 over 3 million voters spread across booming metropolitan 00:13:37.710 --> 00:13:41.149 areas and vast rural counties. The logistical 00:13:41.149 --> 00:13:43.250 nightmare of campaigning at that scale is hard 00:13:43.250 --> 00:13:45.940 to comprehend. The scale of the electorate entirely 00:13:45.940 --> 00:13:48.340 changes the mechanics of how a politician builds 00:13:48.340 --> 00:13:51.179 a mandate. And as that population grows into 00:13:51.179 --> 00:13:53.620 the millions, we also see fascinating patterns 00:13:53.620 --> 00:13:56.460 in the ebb and flow of power. The data reveals 00:13:56.460 --> 00:13:59.059 a constant tension between the advantage of incumbency 00:13:59.059 --> 00:14:01.519 and the volatility of the electorate. You definitely 00:14:01.519 --> 00:14:04.019 see long streaks of stability where incumbent 00:14:04.019 --> 00:14:07.220 seem locked in. Gordon Allott is a great example 00:14:07.220 --> 00:14:09.539 in the Class 2 seat. He wins his first term in 00:14:09.539 --> 00:14:13.620 1954, successfully defends it in 1960, and defends 00:14:13.620 --> 00:14:16.879 it again in 1966. He was a fixture. He secures 00:14:16.879 --> 00:14:19.379 three solid terms before that narrow defeat in 00:14:19.379 --> 00:14:22.279 1972. Or look at Michael Bennett in the Class 00:14:22.279 --> 00:14:25.059 3 seat in more recent history. He wins in 2010, 00:14:25.360 --> 00:14:29.340 defends in 2016, and defends again in 2022. The 00:14:29.340 --> 00:14:31.990 incumbency advantage is real. Voters often prefer 00:14:31.990 --> 00:14:34.570 the familiar name or at least the stability that 00:14:34.570 --> 00:14:37.250 comes with a known quantity. But the data also 00:14:37.250 --> 00:14:39.730 provides a sharp counter -narrative. Just when 00:14:39.730 --> 00:14:41.450 a party seems to have established a dynasty, 00:14:41.690 --> 00:14:43.950 the pendulum swings violently in the other direction. 00:14:44.190 --> 00:14:46.610 The recent Class 2 data is the perfect case study 00:14:46.610 --> 00:14:48.889 for that pendulum swing. Let's trace the last 00:14:48.889 --> 00:14:52.009 few cycles. In 2008, Democrat Mark Udall wins 00:14:52.009 --> 00:14:53.990 the seat against Bob Schaefer. Udall gets over 00:14:53.990 --> 00:14:57.350 1 .2 million votes. So Udall goes to Washington 00:14:57.350 --> 00:14:59.269 and becomes the incumbent heading into the 2014 00:14:59.269 --> 00:15:01.789 election. Right. But in 2014, the pendulum swings. 00:15:02.009 --> 00:15:04.649 Republican Cory Gardner unseats Udall, winning 00:15:04.649 --> 00:15:07.169 by roughly a 2 % margin. A highly competitive 00:15:07.169 --> 00:15:10.190 race that unseats an incumbent in a rapidly growing 00:15:10.190 --> 00:15:13.350 state. Highly competitive. So now Gardner is 00:15:13.350 --> 00:15:16.269 the incumbent heading into 2020. He has the advantage 00:15:16.269 --> 00:15:18.590 of the office. But what happens? He swings again. 00:15:18.809 --> 00:15:20.850 The pendulum swings right back. Democrat John 00:15:20.850 --> 00:15:23.649 Hickenlooper unseats the incumbent Cory Gardner 00:15:23.649 --> 00:15:27.470 in 2020 by a margin of about 53 .5 percent to 00:15:27.470 --> 00:15:31.370 44 percent. It is this constant shifting battleground. 00:15:31.809 --> 00:15:34.509 Just when you think a party has a lock on a specific 00:15:34.509 --> 00:15:37.250 seat, the voters decide to completely change 00:15:37.250 --> 00:15:39.450 course. It perfectly illustrates the accountability 00:15:39.450 --> 00:15:41.789 built into the staggered six year term system 00:15:41.789 --> 00:15:43.909 we discussed at the very beginning. Every six 00:15:43.909 --> 00:15:46.289 years. The incumbent has to return to that ever 00:15:46.289 --> 00:15:48.350 -changing, rapidly expanding population of 3 00:15:48.350 --> 00:15:50.690 million people and ask for their mandate again. 00:15:50.909 --> 00:15:53.669 And as the back and forth between Udall, Gardner, 00:15:53.789 --> 00:15:56.289 and Hickenlooper clearly shows, that mandate 00:15:56.289 --> 00:15:59.149 is never guaranteed, the electorate is fickle, 00:15:59.169 --> 00:16:01.070 and their priorities are constantly evolving. 00:16:01.389 --> 00:16:03.610 We have covered some serious ground today. Who 00:16:03.610 --> 00:16:05.769 knew a chronological table of election results 00:16:05.769 --> 00:16:08.379 could map out so much human behavior? It is a 00:16:08.379 --> 00:16:10.820 testament to the fact that data is never just 00:16:10.820 --> 00:16:14.159 numbers. It is a codified record of choices made 00:16:14.159 --> 00:16:17.779 by millions of individuals over a century. Thank 00:16:17.779 --> 00:16:19.559 you so much for going on this statistical journey 00:16:19.559 --> 00:16:21.700 with us today. We started with what seemed like 00:16:21.700 --> 00:16:24.460 a standard reference list, but by unpacking the 00:16:24.460 --> 00:16:27.539 proportions and the context, we found incredible 00:16:27.539 --> 00:16:30.200 stories. So many stories. We looked at extreme 00:16:30.200 --> 00:16:32.639 nail biters decided by a fraction of a percent, 00:16:32.840 --> 00:16:35.559 explored the fascinating candidate first loyalty 00:16:35.559 --> 00:16:37.580 of Ben Nighthorse Campbell winning under two 00:16:37.580 --> 00:16:40.279 different party banners, and traced a massive 00:16:40.279 --> 00:16:42.919 demographic explosion from a quarter of a million 00:16:42.919 --> 00:16:45.840 voters to over three million. And before we wrap 00:16:45.840 --> 00:16:47.779 up this deep dive, I want to leave you with one 00:16:47.779 --> 00:16:50.799 final thought to mull over. As we explored this 00:16:50.799 --> 00:16:53.080 data today, we naturally focused heavily on the 00:16:53.080 --> 00:16:57.320 past, the margins of 1932, the third -party disruptions 00:16:57.320 --> 00:17:00.639 of 1914, the massive modern turnout of 2020. 00:17:00.799 --> 00:17:03.500 Right, the history. But the Wikipedia text explicitly 00:17:03.500 --> 00:17:05.740 notes something else at the very end of its tables. 00:17:06.019 --> 00:17:08.259 It states that the next scheduled elections for 00:17:08.259 --> 00:17:10.779 these Colorado seats are in 2026 for Class 2 00:17:10.779 --> 00:17:13.539 and 2028 for Class 3. The ledger is still open. 00:17:13.900 --> 00:17:16.099 That is exactly it. I want you to consider that 00:17:16.099 --> 00:17:18.920 this historical table we've been analyzing isn't 00:17:18.920 --> 00:17:21.579 a finished artifact. It is a living, breathing 00:17:21.579 --> 00:17:24.799 document. Every single time a citizen casts a 00:17:24.799 --> 00:17:27.980 vote in 2026 or 2028, they aren't just performing 00:17:27.980 --> 00:17:30.240 a civic duty in the present moment. They're literally 00:17:30.240 --> 00:17:33.059 adding a digit to a permanent historical ledger. 00:17:33.359 --> 00:17:35.859 We've spent our time today marveling at the collective 00:17:35.859 --> 00:17:38.819 decisions made by voters 100 years ago. It leaves 00:17:38.819 --> 00:17:41.500 you to ponder, what will the data generated in 00:17:41.500 --> 00:17:44.700 2026 and 2028 eventually say about our current 00:17:44.700 --> 00:17:47.619 cultural mood to a curious reader looking back 00:17:47.619 --> 00:17:50.480 at this exact same list a century from now. What 00:17:50.480 --> 00:17:52.700 a fantastic thought to end on. You are writing 00:17:52.700 --> 00:17:54.539 the next row of the spreadsheet with every ballot. 00:17:54.720 --> 00:17:56.380 Thank you all for joining us on this deep dive. 00:17:56.480 --> 00:17:58.240 Stay curious, keep pulling on those threads, 00:17:58.339 --> 00:17:59.500 and we will catch you on the next one.