WEBVTT 00:00:00.000 --> 00:00:02.720 Welcome back to the Deep Dive. Today, we're taking 00:00:02.720 --> 00:00:05.480 a look at a historical ledger that, well, on 00:00:05.480 --> 00:00:07.000 the surface, it looks like a standard public 00:00:07.000 --> 00:00:09.580 record. Yeah, totally. It's just a big list. 00:00:09.779 --> 00:00:11.779 Right. We're examining the Wikipedia article 00:00:11.779 --> 00:00:14.580 titled, List of United States Senate Elections 00:00:14.580 --> 00:00:17.940 in Alabama. That sounds a bit dry at first glance. 00:00:18.320 --> 00:00:21.059 It does. It does. But we aren't just going to 00:00:21.059 --> 00:00:24.280 sit here and read dates and names to you. We're 00:00:24.280 --> 00:00:27.760 going to apply a sort of... A stress test to 00:00:27.760 --> 00:00:31.640 this data. We want to see the exact moments of 00:00:31.640 --> 00:00:34.600 political landscape fundamentally fractured, 00:00:34.600 --> 00:00:38.520 solidified and evolved over a century. It's really 00:00:38.520 --> 00:00:40.700 an exercise in spotting the hidden narratives 00:00:40.700 --> 00:00:44.799 buried inside. It is a phenomenal data set for 00:00:44.799 --> 00:00:46.899 exactly that reason. Because when you look at 00:00:46.899 --> 00:00:49.159 an uninterrupted chronological list like this, 00:00:49.240 --> 00:00:51.340 you start to realize that statistical records 00:00:51.340 --> 00:00:53.399 are, well, they're essentially just codified 00:00:53.399 --> 00:00:56.979 human behavior. But to truly decode that behavior, 00:00:57.219 --> 00:01:00.100 we first need to establish the... the mechanical 00:01:00.100 --> 00:01:01.939 parameters of the source material. Right, the 00:01:01.939 --> 00:01:04.760 ground rules. Exactly. So the U .S. Constitution 00:01:04.760 --> 00:01:08.140 allots every state two senators, and they serve 00:01:08.140 --> 00:01:11.480 staggered six -year terms. The easiest way to 00:01:11.480 --> 00:01:13.799 visualize this is to think of the Senate like 00:01:13.799 --> 00:01:16.959 a continuously running engine where you only 00:01:16.959 --> 00:01:19.180 ever replace parts one at a time. Right, so you 00:01:19.180 --> 00:01:21.040 can't just swap everything out at once. Right. 00:01:21.120 --> 00:01:23.680 The system is intentionally staggered so the 00:01:23.680 --> 00:01:26.379 entire chamber cannot be swept out in a single 00:01:26.379 --> 00:01:29.689 electoral wave. So when Alabama was admitted 00:01:29.689 --> 00:01:32.689 to the union back in 1819, the state was assigned 00:01:32.689 --> 00:01:35.069 a class two seat and a class three seat. Which 00:01:35.069 --> 00:01:37.549 means their elections run on two entirely different 00:01:37.549 --> 00:01:40.510 independent internal clocks. Precisely. Okay, 00:01:40.569 --> 00:01:43.189 let's unpack this. Why should you, the listener, 00:01:43.370 --> 00:01:46.730 actually care about the internal clocks of historical 00:01:46.730 --> 00:01:49.450 Alabama Senate elections? It's a fair question. 00:01:49.670 --> 00:01:51.989 Because this isn't just trivia. It's a master 00:01:51.989 --> 00:01:55.489 class in data literacy. By tracking these numbers, 00:01:55.609 --> 00:01:58.170 you learn how to spot inflection points. You 00:01:58.170 --> 00:02:00.609 learn how to recognize statistical noise versus 00:02:00.609 --> 00:02:03.150 a genuine signal, and how a seemingly stable 00:02:03.150 --> 00:02:05.489 consensus can just erode completely overnight. 00:02:05.750 --> 00:02:07.930 It happens faster than you'd think. Oh, absolutely. 00:02:08.229 --> 00:02:10.949 And the timing for this specific deep dive couldn't 00:02:10.949 --> 00:02:14.150 be better. With today's date being March 4, 2026, 00:02:14.469 --> 00:02:16.969 we are sitting in the exact year of the next 00:02:16.969 --> 00:02:19.530 scheduled Class 2 election. We are. The historical 00:02:19.530 --> 00:02:21.830 ledger we are looking at leads right up to this 00:02:21.830 --> 00:02:25.020 very moment. But... To understand the weight 00:02:25.020 --> 00:02:27.340 of a modern election, we actually have to start 00:02:27.340 --> 00:02:30.199 at a major structural pivot detailed in our source, 00:02:30.360 --> 00:02:32.800 which is the passage of the 17th Amendment in 00:02:32.800 --> 00:02:35.800 1913. Yeah, that's really the crucial starting 00:02:35.800 --> 00:02:38.479 line for our analysis today. Because prior to 00:02:38.479 --> 00:02:41.340 1913, senators were chosen by the state legislature. 00:02:41.639 --> 00:02:43.860 Right. So the debt from that era just represents 00:02:43.860 --> 00:02:46.319 the decisions of a few dozen political insiders. 00:02:46.539 --> 00:02:48.919 But the passage of the 17th Amendment allowed 00:02:48.919 --> 00:02:51.259 for the direct election of U .S. senators by 00:02:51.259 --> 00:02:53.400 the general voting public. Which changes everything. 00:02:53.449 --> 00:02:56.469 Completely. Suddenly, the data set we're analyzing 00:02:56.469 --> 00:02:59.370 shifts from being a record of backroom consensus 00:02:59.370 --> 00:03:02.629 to a massive statewide metric of public sentiment. 00:03:02.849 --> 00:03:05.849 It introduces hundreds of thousands of new variables. 00:03:06.009 --> 00:03:08.150 The voters themselves. Exactly. And what makes 00:03:08.150 --> 00:03:10.050 the early years of that direct voting era so 00:03:10.050 --> 00:03:12.849 fascinating is just how mathematically definitive 00:03:12.849 --> 00:03:16.009 the results are. When we look at the class two 00:03:16.009 --> 00:03:19.629 election results starting in 1918, the dominance 00:03:19.629 --> 00:03:21.849 of the Democratic Party isn't just a heavy lean. 00:03:22.030 --> 00:03:24.210 No, it's not a lean at all. It is a mathematical 00:03:24.210 --> 00:03:26.669 absolute. I mean, look at Democrat John H. Bankhead 00:03:26.669 --> 00:03:29.430 in the 1918 race. The record shows he received 00:03:29.430 --> 00:03:32.490 over 54 ,000 votes and his percentage of the 00:03:32.490 --> 00:03:37.479 total vote. Exactly. 100 .00%. Yeah. There is 00:03:37.479 --> 00:03:39.919 no runner up. There are no scattered write -ins, 00:03:39.979 --> 00:03:42.319 making even a fraction of a percentage point 00:03:42.319 --> 00:03:45.560 dent. As someone used to modern elections that 00:03:45.560 --> 00:03:47.800 are decided by a few points, seeing a literal 00:03:47.800 --> 00:03:50.219 100 % next to a candidate's name in a secret 00:03:50.219 --> 00:03:53.409 ballot election is just stunning. Well, yes and 00:03:53.409 --> 00:03:55.469 no. I mean, it is a visual wall of control on 00:03:55.469 --> 00:03:57.469 paper, certainly. But if we dig into the mechanics 00:03:57.469 --> 00:03:59.610 of what a 100 percent victory actually implies, 00:03:59.849 --> 00:04:01.830 it tells a slightly different story. Because 00:04:01.830 --> 00:04:04.590 it is statistically nearly impossible to get 00:04:04.590 --> 00:04:06.569 100 percent of a massive population to agree 00:04:06.569 --> 00:04:08.530 on anything, let alone a political candidate. 00:04:08.750 --> 00:04:10.909 What that 100 percent actually represents is 00:04:10.909 --> 00:04:13.530 a complete lack of opposition structure. It isn't 00:04:13.530 --> 00:04:16.069 necessarily unanimous, enthusiastic consent from 00:04:16.069 --> 00:04:18.629 every living adult in the state. It's the mathematical 00:04:18.629 --> 00:04:21.709 result of an era where the primary election affects. 00:04:21.800 --> 00:04:24.100 served as the general election, and the opposing 00:04:24.100 --> 00:04:26.100 party didn't even field a candidate. That makes 00:04:26.100 --> 00:04:28.240 a lot of sense. But the sheer repetition of that 00:04:28.240 --> 00:04:31.519 100 % figure is still staggering to look at in 00:04:31.519 --> 00:04:33.639 the table. Because this wasn't a one -time anomaly. 00:04:33.899 --> 00:04:36.639 Not at all. John H. Bankhead II repeats the feat. 00:04:37.000 --> 00:04:39.959 He wins the class two seat with exactly 100 % 00:04:39.959 --> 00:04:43.240 of the vote in 1942. Then John Sparkman steps 00:04:43.240 --> 00:04:46.420 in and wins a 1946 special election with 100%. 00:04:46.420 --> 00:04:49.100 And over on the Class III side, the internal 00:04:49.100 --> 00:04:51.100 clock is ticking differently, but the math looks 00:04:51.100 --> 00:04:54.620 identical. Jay Lister Hill takes a 1938 special 00:04:54.620 --> 00:04:58.060 election with 100%. He runs again in 1956 and 00:04:58.060 --> 00:05:00.759 pulls in over 330 ,000 votes, again capturing 00:05:00.759 --> 00:05:05.600 100 .00%. It is decades of absolute... unquestioned 00:05:05.600 --> 00:05:08.300 numerical dominance. It's a remarkable run of 00:05:08.300 --> 00:05:11.019 institutional stability. But even inside that 00:05:11.019 --> 00:05:13.100 fortress of 100 percent victories, if you look 00:05:13.100 --> 00:05:15.060 closely enough, you can spot the statistical 00:05:15.060 --> 00:05:17.920 whispers of the facts in the armor. Right. There 00:05:17.920 --> 00:05:20.079 is a highly illustrative data point from the 00:05:20.079 --> 00:05:23.120 1930 class two election. We mentioned John H. 00:05:23.180 --> 00:05:25.819 Bank had the second in that year, but his opponent 00:05:25.819 --> 00:05:28.730 is what makes the race interesting. He ran against 00:05:28.730 --> 00:05:31.430 James Thomas Heflin. Right. Now, Heflin was a 00:05:31.430 --> 00:05:33.970 known quantity. He had actually won this exact 00:05:33.970 --> 00:05:37.350 seat in 1920 and 1924 as a Democrat, pulling 00:05:37.350 --> 00:05:41.050 in overwhelming majorities. But in 1930, Heflin 00:05:41.050 --> 00:05:43.750 runs as an independent. And he doesn't just run. 00:05:43.910 --> 00:05:46.129 He pulls a massive portion of the electorate 00:05:46.129 --> 00:05:48.990 with him. The ledger shows Heflin capturing over 00:05:48.990 --> 00:05:51.870 100 ,000 votes as an independent, which translates 00:05:51.870 --> 00:05:54.129 to just over 40 percent of the total vote. That's 00:05:54.129 --> 00:05:57.120 a huge chunk. Right. So Bankhead the second still 00:05:57.120 --> 00:05:59.720 takes the seat with about 60 percent. But seeing 00:05:59.720 --> 00:06:01.779 an independent candidate carve out a 40 percent 00:06:01.779 --> 00:06:05.160 block in 1930 completely fractures that visual 00:06:05.160 --> 00:06:07.500 wall of unanimous Democratic victories we were 00:06:07.500 --> 00:06:09.199 just talking about. Exactly. And that is why 00:06:09.199 --> 00:06:11.120 you always look for the localized disruptions 00:06:11.120 --> 00:06:13.740 and long term data sets. That 1930 election is 00:06:13.740 --> 00:06:15.579 an early stress fracture. It proves that the 00:06:15.579 --> 00:06:17.620 voter base was not a monolith. Even back then? 00:06:17.759 --> 00:06:20.759 Yes. Even in an era where the official opposition 00:06:20.759 --> 00:06:23.519 party was functionally nonexistent at the ballot 00:06:23.519 --> 00:06:26.060 box, a 40 percent defection to an independent 00:06:26.060 --> 00:06:28.759 candidate reveals internal ideological divisions. 00:06:29.220 --> 00:06:31.660 It is a signal cutting through the statistical 00:06:31.660 --> 00:06:34.439 noise, proving that while the D next to the winner's 00:06:34.439 --> 00:06:36.319 name remained constant, the coalition underneath 00:06:36.319 --> 00:06:38.800 it was capable of fracturing. Here's where it 00:06:38.800 --> 00:06:41.860 gets really interesting. If that 1930 independent 00:06:41.860 --> 00:06:45.339 run was a stress fracture, the 1962 class three 00:06:45.339 --> 00:06:47.740 election is a localized earthquake. Oh, absolutely. 00:06:48.220 --> 00:06:51.379 Let's go back to Jay Lister Hill. In 1956, he 00:06:51.379 --> 00:06:54.620 secures 100 percent of the vote. Over 330 ,000 00:06:54.620 --> 00:06:58.019 votes, zero opposition. It is the absolute height 00:06:58.019 --> 00:07:00.779 of comfortable incumbency. But fast forward exactly 00:07:00.779 --> 00:07:04.819 one term, six years later to 1962. Hill is running 00:07:04.819 --> 00:07:06.879 for reelection, but this time he faces a Republican 00:07:06.879 --> 00:07:09.699 challenger named James D. Martin. And the completely 00:07:09.699 --> 00:07:12.139 flat, predictable, numerical landscape we've 00:07:12.139 --> 00:07:14.720 seen for decades suddenly violently shifts. It 00:07:14.720 --> 00:07:16.920 really does. The contrast is one of the most 00:07:16.920 --> 00:07:19.420 severe in the entire historical record. From 00:07:19.420 --> 00:07:23.040 an unopposed... 100 % in 1956, Jay Lister Hill 00:07:23.040 --> 00:07:26.519 barely survives the 1962 contest. Out of nearly 00:07:26.519 --> 00:07:30.360 400 ,000 total votes cast, Hill edges out Martin 00:07:30.360 --> 00:07:33.959 by a margin of less than 7 ,000 votes. Wow. He 00:07:33.959 --> 00:07:37.899 finishes with roughly 50 .8 % while Martin captures 00:07:37.899 --> 00:07:41.839 49 .14%. What's fascinating here is the concept 00:07:41.839 --> 00:07:44.759 of a structural break. A structural break. Yes. 00:07:44.879 --> 00:07:47.759 A shift of just under two percentage points is 00:07:47.759 --> 00:07:50.439 all that stood between a continuation of century 00:07:50.439 --> 00:07:52.860 -long single -party control and a monumental 00:07:52.860 --> 00:07:55.660 upset. It fundamentally changes the baseline 00:07:55.660 --> 00:07:57.660 of every election that follows it. It's just 00:07:57.660 --> 00:07:59.699 jarring to see how fast that consensus collapses 00:07:59.699 --> 00:08:02.120 on the page. Six years prior, there was no opposition. 00:08:02.339 --> 00:08:04.579 Suddenly, every single vote is critical. Every 00:08:04.579 --> 00:08:07.439 single one. And once that seal is broken in 1962... 00:08:07.769 --> 00:08:10.089 You can track a steady, undeniable climb in the 00:08:10.089 --> 00:08:12.230 Republican percentages. They don't take control 00:08:12.230 --> 00:08:14.430 immediately, but the baseline shifts upward. 00:08:14.629 --> 00:08:17.769 Look at the 1966 Class II race. Republican John 00:08:17.769 --> 00:08:20.250 Grenier captures roughly 39 % of the vote against 00:08:20.250 --> 00:08:23.310 incumbent Democrat John Sparkman. That is a massive 00:08:23.310 --> 00:08:26.610 elevation from the functionally 0 % opposition 00:08:26.610 --> 00:08:29.589 of the 1940s and 50s. It's like watching a regression 00:08:29.589 --> 00:08:32.269 to a completely new mean. The pressure builds 00:08:32.269 --> 00:08:34.990 over the 70s, with Republican vote shares climbing 00:08:34.990 --> 00:08:37.919 into the 30s and 40s, and then finally, The pressure 00:08:37.919 --> 00:08:41.320 releases. Yes. The 1980 class three election 00:08:41.320 --> 00:08:43.899 represents the realization of that mathematical 00:08:43.899 --> 00:08:47.039 buildup. Republican Jeremiah Denton wins the 00:08:47.039 --> 00:08:50.779 seat. And again, look at the margin. It is incredibly 00:08:50.779 --> 00:08:55.490 tight. Denton secures the win with 50 .15 % against 00:08:55.490 --> 00:09:00.090 Democrat Jim Folsom Jr., who pulls 47 .05%. It's 00:09:00.090 --> 00:09:02.570 like tectonic plates. They slowly move for decades 00:09:02.570 --> 00:09:04.490 until an earthquake finally happens. That's a 00:09:04.490 --> 00:09:06.710 great analogy. After generations of looking at 00:09:06.710 --> 00:09:08.789 a ledger completely dominated by one column, 00:09:08.950 --> 00:09:10.909 the appearance of the Republican label next to 00:09:10.909 --> 00:09:13.370 the winner's name in 1980 is the definitive proof 00:09:13.370 --> 00:09:15.409 that the political realignment of the state had 00:09:15.409 --> 00:09:17.559 reached a tipping point. But just when you think 00:09:17.559 --> 00:09:19.559 you have a handle on this new competitive landscape, 00:09:19.919 --> 00:09:21.740 the data set throws out the most fascinating 00:09:21.740 --> 00:09:23.860 anomaly in the entire ledger. The Shelby era. 00:09:24.139 --> 00:09:27.000 Yes. I want to introduce you, the listener, to 00:09:27.000 --> 00:09:28.779 the statistical progression of Richard Shelby 00:09:28.779 --> 00:09:31.340 in the class three seat. It's a story told across 00:09:31.340 --> 00:09:33.820 three consecutive elections, and it completely 00:09:33.820 --> 00:09:36.460 defies conventional assumptions about how voters 00:09:36.460 --> 00:09:38.200 behave. Let's trace the numbers. Let's do it. 00:09:38.259 --> 00:09:42.240 In 1986, Richard Shelby runs as a Democrat. He 00:09:42.240 --> 00:09:44.399 challenges the Republican incumbent, Jeremiah 00:09:44.399 --> 00:09:47.610 Denton. The very same man who flipped the seat 00:09:47.610 --> 00:09:51.059 six years prior. It is a brutal, razor -thin 00:09:51.059 --> 00:09:54.059 contest. Shelby narrowly unseats Denton by roughly 00:09:54.059 --> 00:09:57.440 half a percentage point, winning 50 .28 % to 00:09:57.440 --> 00:10:01.320 49 .72%. Which is a very standard, highly contested 00:10:01.320 --> 00:10:03.639 correction back to the previous party. Thus far, 00:10:03.740 --> 00:10:05.759 the data behaves exactly as we would expect in 00:10:05.759 --> 00:10:08.379 a highly polarized transitional electorate. Right. 00:10:08.419 --> 00:10:10.820 So we move forward one term to 1992. Shelby runs 00:10:10.820 --> 00:10:12.879 for reelection as a Democrat. And this time, 00:10:12.919 --> 00:10:15.480 he expands his base significantly. He wins comfortably, 00:10:15.639 --> 00:10:18.860 securing roughly 64 .82 % of the vote against 00:10:18.919 --> 00:10:20.740 to a Republican challenger who only manages around 00:10:20.740 --> 00:10:24.200 33%, the seat looks solidly blue again. But then 00:10:24.200 --> 00:10:26.990 we get 1998. Shelby is up for his third term. 00:10:27.049 --> 00:10:29.629 He wins the election. The margin is incredibly 00:10:29.629 --> 00:10:32.090 consistent with his last run. He captures 63 00:10:32.090 --> 00:10:35.009 .24 percent of the total vote. The electorate 00:10:35.009 --> 00:10:36.870 is giving him nearly the exact same overwhelming 00:10:36.870 --> 00:10:39.750 level of support. But there is a glaring difference 00:10:39.750 --> 00:10:42.629 in the record. In the 1998 column, Richard Shelby 00:10:42.629 --> 00:10:45.210 is listed as a Republican. Exactly. If we connect 00:10:45.210 --> 00:10:48.429 this to the bigger picture, it forces us to reevaluate 00:10:48.429 --> 00:10:52.190 what an electoral mandate actually is. Think 00:10:52.190 --> 00:10:54.490 about the cognitive dissonance of that data point. 00:10:54.970 --> 00:10:57.629 It is profoundly rare to see a massive block 00:10:57.629 --> 00:10:59.809 of voters continue to support the exact same 00:10:59.809 --> 00:11:02.629 individual, maintaining a highly consistent margin 00:11:02.629 --> 00:11:06.169 dropping only slightly from roughly 64 % to 63%, 00:11:06.169 --> 00:11:08.769 even as the official party apparatus surrounding 00:11:08.769 --> 00:11:11.570 that candidate completely inverts. It's so unusual. 00:11:11.909 --> 00:11:14.129 It strongly implies that the voters in that specific 00:11:14.129 --> 00:11:16.110 window were not strictly voting for the party 00:11:16.110 --> 00:11:17.750 platform. They were voting for the candidate's 00:11:17.750 --> 00:11:20.450 personal brand equity. It is entirely unique 00:11:20.450 --> 00:11:22.909 in this list. It challenges the assumption that 00:11:22.909 --> 00:11:25.850 the R or the D next to a name is the absolute 00:11:25.850 --> 00:11:28.850 determining factor for the electorate. And again, 00:11:28.929 --> 00:11:30.769 we are just looking purely at the math and the 00:11:30.769 --> 00:11:32.929 labels here. Right. Impartial data. We aren't 00:11:32.929 --> 00:11:35.429 analyzing the ideologies or the platform shifts. 00:11:35.610 --> 00:11:38.350 We are simply looking at how a historical spreadsheet 00:11:38.350 --> 00:11:41.830 documents a state's voters logging onto a single 00:11:41.830 --> 00:11:44.429 individual's name, holding steady, while the 00:11:44.429 --> 00:11:46.769 entire structural label around that name flips 00:11:46.769 --> 00:11:50.370 180 degrees. It is a brilliant example of candidate 00:11:50.370 --> 00:11:53.149 brand recognition overriding partisan reflex 00:11:53.149 --> 00:11:55.580 in the data. And that party switch by Shelby 00:11:55.580 --> 00:11:58.320 in the late 90s effectively ushers in a brand 00:11:58.320 --> 00:12:01.080 new era of stabilization in the ledger. After 00:12:01.080 --> 00:12:03.600 the highly volatile, razor thin margins of the 00:12:03.600 --> 00:12:06.379 60s, 70s and 80s, the data begins to settle into 00:12:06.379 --> 00:12:08.879 a new, heavily reinforced pattern. We see this 00:12:08.879 --> 00:12:10.720 clearly when we look at the class two seat starting 00:12:10.720 --> 00:12:13.120 in 1996. Yes. This marks the beginning of Jeff 00:12:13.120 --> 00:12:15.379 Sessions tenure. Running as a Republican, Sessions 00:12:15.379 --> 00:12:17.759 wins his first term with a competitive but solid 00:12:17.759 --> 00:12:21.019 52 .45 percent. But watch how those numbers compound 00:12:21.019 --> 00:12:23.799 over time. His margins just keep widening. He 00:12:23.799 --> 00:12:27.620 hits roughly 58 % in 2002. By 2008, he is up 00:12:27.620 --> 00:12:30.879 over 63%. The momentum is entirely one -sided. 00:12:31.019 --> 00:12:34.559 And then we reach 2014. And the ledger provides 00:12:34.559 --> 00:12:36.820 this unbelievable callback to the very beginning 00:12:36.820 --> 00:12:40.340 of our analysis. Jeff Sessions runs essentially 00:12:40.340 --> 00:12:44.139 unopposed. He captures an astounding 97 .25 % 00:12:44.139 --> 00:12:46.220 of the vote, with the remaining fraction going 00:12:46.220 --> 00:12:48.100 to scattered right ends. It is the closest thing 00:12:48.100 --> 00:12:50.240 to the 100 % club we have seen in over half a 00:12:50.240 --> 00:12:52.720 century. The visual wall of dominance has returned. 00:12:52.940 --> 00:12:55.240 It has just migrated to a different column. It's 00:12:55.240 --> 00:12:57.820 a striking historical echo. But as we established 00:12:57.820 --> 00:13:00.159 with the 1962 election, data in a democratic 00:13:00.159 --> 00:13:02.620 system rarely stays static for long. No, it does 00:13:02.620 --> 00:13:05.559 not. A 97 % victory implies an utter lack of 00:13:05.559 --> 00:13:07.759 organized opposition. But the very next entry 00:13:07.759 --> 00:13:09.779 in the class two sequence proves how rapidly 00:13:09.779 --> 00:13:12.259 an unorganized opposition can mobilize when the 00:13:12.259 --> 00:13:14.779 variables change. It completely upends the table. 00:13:15.360 --> 00:13:19.220 After that 97 % unopposed victory in 2014, the 00:13:19.220 --> 00:13:21.500 source reveals a special election held in 2017 00:13:21.500 --> 00:13:24.779 for that exact same class two seat. And instantly 00:13:24.779 --> 00:13:26.740 we are ripped right out of the era of comfortable 00:13:26.740 --> 00:13:29.179 incumbency and thrown straight back into the 00:13:29.179 --> 00:13:31.600 territory of the razor thin, highly volatile 00:13:31.600 --> 00:13:34.899 margins. The numerical shift in the 2017 special 00:13:34.899 --> 00:13:38.860 election is staggering. Democrat Doug Jones flips 00:13:38.860 --> 00:13:41.659 the seat and the margins are incredibly tight. 00:13:41.779 --> 00:13:44.519 Jones edges out Republican Roy Moore by just 00:13:44.519 --> 00:13:47.759 under 22 ,000 votes out of a total pool of over 00:13:47.759 --> 00:13:51.120 1 .3 million cast. We are talking about a margin 00:13:51.120 --> 00:13:52.879 of victory of roughly one and a half percent. 00:13:53.289 --> 00:13:57.870 Doug Jones with 49 .97 % to Roy Moore's 48 .34%. 00:13:57.870 --> 00:14:00.629 Once again, we see a monumental structural shift 00:14:00.629 --> 00:14:03.289 decided by a microscopic fraction of the electorate. 00:14:03.309 --> 00:14:05.669 It's a sudden injection of intense high turnout 00:14:05.669 --> 00:14:08.090 competition into a seat that just three years 00:14:08.090 --> 00:14:10.970 prior mathematically featured zero organized 00:14:10.970 --> 00:14:13.110 resistance. Exactly. But the data also shows 00:14:13.110 --> 00:14:15.990 us that this 2017 flip was a brief, highly localized 00:14:15.990 --> 00:14:18.320 anomaly within the modern era. When you look 00:14:18.320 --> 00:14:19.899 at the next regular election for that seat in 00:14:19.899 --> 00:14:22.259 2020, the ledger shows a swift correction back 00:14:22.259 --> 00:14:24.580 to the prevailing mean. Republican Tommy Pureville 00:14:24.580 --> 00:14:28.519 defeats the incumbent, Doug Jones, with 60 .0 00:14:28.519 --> 00:14:32.799 % of the vote compared to Jones' 39 .74%. And 00:14:32.799 --> 00:14:34.419 if we bring this right up to the present day, 00:14:34.519 --> 00:14:36.860 the most recent data point in our source document 00:14:36.860 --> 00:14:40.820 is the 2022 Class 3 election. In that race, Republican 00:14:40.820 --> 00:14:44.460 Katie Britt secured the seat with a solid 66 00:14:44.460 --> 00:14:47.929 .62 % against Democrat Will Boyd. The numbers 00:14:47.929 --> 00:14:50.350 show a firm return to wide, comfortable margins. 00:14:50.690 --> 00:14:53.190 This suggests that the 2017 special election, 00:14:53.490 --> 00:14:56.350 much like the 1930 independent run or the 1962 00:14:56.350 --> 00:14:58.710 near miss, was an inflection point driven by 00:14:58.710 --> 00:15:02.190 unique localized variables. By 2020 and 2022, 00:15:02.509 --> 00:15:04.970 the broader numerical trends indicate a stabilization 00:15:04.970 --> 00:15:07.309 back toward the wider margins established in 00:15:07.309 --> 00:15:09.710 the early 2000s. The system experienced a shock 00:15:09.710 --> 00:15:11.509 and then the system corrected. So what does this 00:15:11.509 --> 00:15:13.490 all mean? We started today by looking at a Wikipedia 00:15:13.490 --> 00:15:15.730 list of United States Senate elections in Alabama. 00:15:16.169 --> 00:15:18.169 On the surface, it is just a chronological pile 00:15:18.169 --> 00:15:20.450 of dates, names and percentages. Just raw data. 00:15:20.669 --> 00:15:23.309 Right. But by aggressively interrogating the 00:15:23.309 --> 00:15:25.990 numbers, we've watched an electorate evolve over 00:15:25.990 --> 00:15:29.149 a full century. You can literally trace the consolidation 00:15:29.149 --> 00:15:32.049 of power. We trace the mathematical absolutism 00:15:32.049 --> 00:15:34.190 of the 100 percent club. Watch the consensus 00:15:34.190 --> 00:15:36.789 fracture by fractions of a percent in the 60s. 00:15:37.110 --> 00:15:39.250 Marveled at the brand loyalty of the Shelby era 00:15:39.250 --> 00:15:42.149 and witnessed the sudden whiplash of modern special 00:15:42.149 --> 00:15:44.789 elections. It's a historical heartbeat monitor. 00:15:45.240 --> 00:15:47.440 And this raises an important question about how 00:15:47.440 --> 00:15:50.440 we consume information. Knowledge is most valuable 00:15:50.440 --> 00:15:52.279 when you question it and strip away assumptions. 00:15:52.740 --> 00:15:55.220 By looking at this raw data without historical 00:15:55.220 --> 00:15:58.139 bias, we can identify pure behavioral patterns. 00:15:58.220 --> 00:16:01.240 Yes. We see that absolute consensus is rarely 00:16:01.240 --> 00:16:03.840 as absolute as it appears on paper and that monumental 00:16:03.840 --> 00:16:06.279 shifts in power often hinge on a shift of just 00:16:06.279 --> 00:16:08.919 one or two percentage points. Multiple perspectives, 00:16:09.279 --> 00:16:11.600 especially purely statistical ones, force us 00:16:11.600 --> 00:16:13.700 to recognize the fragility and the resilience 00:16:13.700 --> 00:16:16.049 of electoral. systems. I want to leave you with 00:16:16.049 --> 00:16:17.929 a final thought to mull over as you look ahead. 00:16:18.610 --> 00:16:21.330 Our source document specifically notes that the 00:16:21.330 --> 00:16:23.629 next scheduled U .S. Senate elections in Alabama 00:16:23.629 --> 00:16:27.029 are fast approaching. The next class two election 00:16:27.029 --> 00:16:31.029 is in 2026. That is this exact year. It is. And 00:16:31.029 --> 00:16:34.250 the next class three election is in 2028. Here 00:16:34.250 --> 00:16:36.450 is something to consider as you watch those numbers 00:16:36.450 --> 00:16:39.440 roll in. Throughout this entire century of data 00:16:39.440 --> 00:16:42.059 we just reviewed, the total number of raw votes 00:16:42.059 --> 00:16:45.120 cast has swelled massively, from tens of thousands 00:16:45.120 --> 00:16:47.879 in the early 1900s to well over a million today. 00:16:48.179 --> 00:16:50.860 But historically, as the sheer volume of voter 00:16:50.860 --> 00:16:53.340 turnout has increased, the margins of victory 00:16:53.340 --> 00:16:55.539 have generally shrunk and become more polarized. 00:16:56.139 --> 00:16:58.440 Does a more engaged, higher turnout electorate 00:16:58.440 --> 00:17:00.720 mathematically guarantee deeper, tighter polarization? 00:17:01.120 --> 00:17:03.299 It's a fascinating trend to track. It really 00:17:03.299 --> 00:17:05.619 is. Look at the turnout totals in the historical 00:17:05.619 --> 00:17:07.880 ledger versus the shrinking margins of victory. 00:17:08.140 --> 00:17:10.619 It is an inverse relationship you can track yourself. 00:17:10.900 --> 00:17:13.359 What will the data table look like a decade from 00:17:13.359 --> 00:17:15.700 now? Will we see another sudden data anomaly 00:17:15.700 --> 00:17:18.720 or the start of a new 100 -year streak? Keep 00:17:18.720 --> 00:17:20.519 your eyes on the numbers this year. The spreadsheet 00:17:20.519 --> 00:17:22.279 is still being written. Thanks for joining us 00:17:22.279 --> 00:17:22.980 on this deep dive.