The topline numbers definitely look good...from 30,000 feet.
I had wanted to post regular update about early election numbers, but sometimes life gets in the way. So here is my update five days out from Election day.
Simon Rosenberg has been touting early voting results as a sign that Democrats are more enthusiastic, and there is no sign of a “Red Wave.” And for the most part, I agree with him. Early voting results just don’t square with any kind huge wave, especially not like what 538 and other poll aggregators are suggesting. Republicans may still take the House and Senate, but it definitely won’t be a huge wave.
The first data point is simply looking at national voting turnout to date. As you can see above, the “Modeled Party ID” at TargetEarly shows 50.4% of voters are Democrats, compared to 39.4% Republican. And early turnout has increased (about 5.5%) compared to 2018 , which was itself a crazy high turnout for a mid-term election. And we know how great the 2018 results were. Looking at those numbers, turnout looks very good for Democrats!
So what’s the problem? Well, as pointed out by a commenter in my previous post, the same conditions could be said about voting in Virginia in 2021. Here are the numbers:
| Modeled Party |
2017 |
2019 |
2020 |
2021 |
| Democrat |
92,283 (48.6%) |
65,842 (44.9%) |
1,335,313 (47.4%) |
648,524 (54.7%) |
| Republican |
80,029 (42.1%) |
64,094 (43.7%) |
1,089,694 (38.7%) |
355,038 (29.9%) |
| Unaffiliated |
17,579 (9.3%) |
16,748 (11.4%) |
391,298 (13.9%) |
182,867 (15.4%) |
| Total |
189,891 |
146,684 |
2,816,305 |
1,186,429 |
That is early voting results for Virginia, compared to previous elections going back to 2017. See how turnout was up overall? And Modeled Party ID for Democrats was WAY up? But even with those great numbers, Republicans won a majority of seats in the General Assembly, and won all three statewide elections (Attorney General, Lieutenant Governor, and Governor). Granted, Youngkin barely won the election...but he won.
So we have a previous example from just last year in which early voter turnout for Democrats was way up, but it was not enough for Democrats to win. Republicans votes on Election Day overcame the early vote advantage for Democrats. Increased early Democratic voter turnout — by itself — is not necessarily enough.
And we have reasons to believe that Republicans may be intentionally waiting to vote until Election Day. The Right Wing Noise Machine has been very loudly claiming that early voting is corrupt and rigged. I haven’t seen any polls of Republicans asking them how many are choosing to vote on Election Day to “make sure” their vote will be counted, but it’s not a stretch to believe that a lot of Republican voters are choosing to wait. We don’t know exactly how many, but it’s something to keep in mind when analyzing early vote turnout.
However, my main concern is when you look at 2022 early voting results by any kind of demographic breakdown. In a nutshell, early voters are typically older, whiter, more rural, and increasingly often, men. There are notable exceptions at the state level, but looking at the national numbers, the trend is pretty clear.
Here is voter turnout, broken down by age:
National Early Voter Turnout by Age: 2018-2022
| Age |
2018 |
2020 |
2022 |
| 18-29 |
1,885,137 (6.8%) |
7,854,801 (10.9%) |
1,607,580 (5.5%) |
| 30-39 |
2,353,127 (8.5%) |
8,597,791 (11.9%) |
2,147,031 (7.3%) |
| 40-49 |
3,020,666 (10.8%) |
9,400,654 (13.1%) |
2,770,026 (9.4%) |
| 50-64 |
7,967,711 (28.6%) |
19,968,201 (27.7%) |
7,893,376 (26.9%) |
| 65-74 |
7,105,349 (25.5%) |
14,993,991 (20.8%) |
8,146,153 (27.7%) |
| 75+ |
5,491,438 (19.7%) |
10,976,185 (15.2%) |
6,732,384 (22.9%) |
| Unknown |
20,373 (0.1%) |
209,342 (0.3%) |
84,018 (0.3%) |
| Total |
27,843,801 |
72,000,965 |
29,380,568
|
|
Here is voter turnout, broken down by race:
National Early Voting Turnout by Race: 2018-2020
| Race |
2018 |
2020 |
2022 |
| African-American |
2,629,766 (9.4%) |
7,075,158 (9.8%) |
2,611,473 (8.9%) |
| Asian |
806,930 (2.9%) |
3,281,971 (4.6%) |
1,021,573 (3.5%) |
| Caucasian |
21,853,337 (78.5%) |
53,931,285 (74.9%) |
23,230,348 (79.1%) |
| Hispanic |
2,293,590 (8.2%) |
6,918,723 (9.6%) |
2,234,480 (7.6%) |
| Other/Uncoded |
260,178 (0.9%) |
793,828 (1.1%) |
282,694 (1.0%) |
| Total |
27,843,801 |
72,000,965 |
29,380,568 |
|
And by gender:
National Early Voter Turnout By Gender: 2017-2021
| Gender |
2018 |
2020 |
2022 |
| Female |
14,840,227 (53.3%) |
38,361,536 (53.3%) |
15,381,381 (52.4%) |
| Gender Expansive |
0 (0.0%) |
0 (0.0%) |
787 (0.0%) |
| Male |
12,611,453 (45.3%) |
31,536,899 (43.8%) |
13,335,367 (45.4%) |
| Unknown |
392,121 (1.4%) |
2,102,530 (2.9%) |
663,033 (2.3%) |
| Total |
27,843,801 |
72,000,965 |
29,380,568 |
|
| Urbanicity |
2018 |
2020 |
2022 |
| Rural |
7,679,170 (27.6%) |
18,687,144 (26.0%) |
8,259,308 (28.1%) |
| Suburban |
11,832,627 (42.5%) |
29,960,472 (41.6%) |
12,465,420 (42.4%) |
| Urban |
8,331,267 (29.9%) |
23,351,277 (32.4%) |
8,655,651 (29.5%) |
| Unknown |
737 (0.0%) |
2,072 (0.0%) |
189 (0.0%) |
| Total |
27,843,801 |
72,000,965 |
29,380,568 |
For every category, the trend is in one direction. That’s not a typical recipe for election success for Democrats. And yet…
For all of those demographic categories (age, race, gender, urbanicity), the ratio of Democrats in those categories is up (at least at the national level). Often dramatically so.
Early Voter Modeled Party ID Among Voters 75+
| Modeled Party |
2018 |
2020 |
2022 |
| Democrat |
2,210,854 (40.3%) |
4,733,682 (43.1%) |
3,117,647 (46.3%) |
| Republican |
2,864,793 (52.2%) |
5,281,626 (48.1%) |
3,052,927 (45.3%) |
| Unaffiliated |
415,791 (7.6%) |
960,877 (8.8%) |
561,810 (8.3%) |
| Total |
5,491,438 |
10,976,185 |
6,732,384 |
Early Voter Modeled Party ID Among Caucasian Voters
| Modeled Party |
2018 |
2020 |
2022 |
| Democrat |
8,081,447 (37.0%) |
21,660,014 (40.2%) |
10,071,816 (43.4%) |
| Republican |
11,790,623 (54.0%) |
26,069,751 (48.3%) |
10,583,138 (45.6%) |
| Unaffiliated |
1,981,267 (9.1%) |
6,201,520 (11.5%) |
2,575,394 (11.1%) |
| Total |
21,853,337 |
53,931,285 |
23,230,348 |
Early Voter Modeled Party ID Among Male Voters
| Modeled Party |
2018 |
2020 |
2022 |
| Democrat |
4,989,100 (39.6%) |
13,660,581 (43.3%) |
5,860,461 (43.9%) |
| Republican |
6,572,429 (52.1%) |
14,547,733 (46.1%) |
6,047,558 (45.3%) |
| Unaffiliated |
1,049,924 (8.3%) |
3,328,585 (10.6%) |
1,427,348 (10.7%) |
| Total |
12,611,453 |
31,536,899 |
13,335,367 |
Early Voter Modeled Party ID Among Rural Voters
| Modeled Party |
2018 |
2020 |
2022 |
| Democrat |
2,222,579 (28.9%) |
5,741,108 (30.7%) |
2,840,576 (34.4%) |
| Republican |
4,769,053 (62.1%) |
10,847,331 (58.0%) |
4,503,616 (54.5%) |
| Unaffiliated |
687,538 (9.0%) |
2,098,705 (11.2%) |
915,116 (11.1%) |
| Total |
7,679,170 |
18,687,144 |
8,259,308 |
In every case, the percentage of Democratic voters increased — often dramatically so.
One last data point — Target Early does not allow users to drill down very deeply, so I can’t pull data on Modeled Party ID for Suburban White Women. I can see that White Women make up 79.1% of all Suburban Women voters (up from 77.4% in 2018). And here’s the table for Suburban Women voters:
Early Voter Modeled Party ID Among Suburban Women Voters
| Modeled Party |
2018 |
2020 |
2022 |
| Democrat |
3,154,359 (49.6%) |
8,246,859 (51.2%) |
3,612,597 (54.7%) |
| Republican |
2,637,141 (41.4%) |
5,956,510 (37.0%) |
2,274,784 (34.5%) |
| Unaffiliated |
572,712 (9.0%) |
1,908,104 (11.8%) |
713,789 (10.8%) |
| Total |
6,364,212 |
16,111,473 |
6,601,170 |
As I noted at the start, there are exceptions at the State level. On the positive side, we have some states that look better than the national data - Arizona, Ohio, Nevada, Florida, and Michigan all look at least “okay,” but Georgia (Race and Urbanicity) and Pennsylvania (Race, Gender and Urbanicity) for example, are very positive. Pennsylvania in particular looks almost too good to be true.
Early Voter Modeled Party ID in Pennsylvania
| Modeled Party |
2018 |
2020 |
2022 |
| Democrat |
54,707 (43.4%) |
1,263,620 (64.1%) |
687,884 (71.9%) |
| Republican |
61,703 (49.0%) |
559,586 (28.4%) |
233,253 (24.4%) |
| Unaffiliated |
9,532 (7.6%) |
148,223 (7.5%) |
36,041 (3.8%) |
| Total |
125,942 |
1,971,429 |
957,178 |
An increase of 28.5 percentage points for Democrats over 2018? What?!? Even compared to 2020, the numbers are a significant increase for Democrats (7.8 percentage points). Of all the state to look at, Pennsylvania seems to be the prime candidate for early votes to predict final election results, if only because the numbers are SUCH an outlier. But we’ll see.
On the flip side, the one “battleground state” that looks bad right now is North Carolina:
| Modeled Party |
2018 |
2020 |
2022 |
| Democrat |
794,088 (51.9%) |
1,863,932 (48.0%) |
699,524 (49.1%) |
| Republican |
659,894 (43.2%) |
1,770,559 (45.6%) |
622,307 (43.7%) |
| Unaffiliated |
75,203 (4.9%) |
251,535 (6.5%) |
101,676 (7.1%) |
| Total |
1,529,185 |
3,886,026 |
1,423,507 |
On every metric — including Modeled Party ID, the early vote seems to favor Republicans. That decline of 2.8 percentage points among Democrats does not bode well. If there is any silver lining, Republican voter turnout is up only .5 percentage points, and the “big” increase is among unaffiliated voters — 3 percentage points.
All things considered, I’ll take the substantial increase in Democratic voter turnout, even though the total pool of early voters is older, whiter, more rural, and more male. But keeping in mind the 2021 election results in Virginia, we can’t count on early vote turnout to guarantee election wins. GOTV on Election Day is just as important to ensure that we remain a democracy in the US.
