I’ve been following early voting returns as aggregated by Target Early — Target Early — and some interesting facts are developing. Mostly good, but some red flags about Nevada and (to a lesser degree) North Carolina as well.
First the disclaimers. I’m not a data analyst by training. This is just my layman’s interpretation of data being aggregated by Target Early. Second, I make no claims about the accuracy of the data presented by Target Early. It’s run by Target Smart, “a Democratic political data and data services firm.” They offer a breakdown by “Modeled Party”, which is important because several states do not have Party ID for voters (eg. Virginia and Wisconsin), but I do not know how Target Early arrived at these numbers. Overall, the actual voter data appears consistent with early voting returns in specific states reported elsewhere (notably Georgia). Third, I don’t have access to data of early voting returns from earlier mid-term elections (e.g. 2014, 2010, etc.), so the comparison is limited to view trends. Fourth, please note all the caveats that Target Early also presents on their website:
- There is no perfect precedent to compare against. Early vote access has expanded dramatically since 2018, but many states have limited access again since 2020, making comparisons to either year imperfect.
- Voting habits change. Trump and ultra-MAGA Republicans' attacks on our voting system have made early voting a political statement. That has had the effect of driving down Republican early voting, while moderates and Democrats continue to embrace the practice. So, while early voting numbers are likely to look good for Democrats, Republicans are even more likely to turn out on Election Day, in person.
- The data that we are sharing with you allows us to see if someone voted, it does not allow us to see for whom that vote was cast.
- While we update TargetEarly every day, there is a lag time between when election administrators post and when that data is available on the site. That’s because we match those voting records against the entire file so we can learn more about each voter. It’s also important to note that states do not always uniformly upload information: some jurisdictions in the same state are slower than others.
With that out of the way, let me state that my analysis kind of turns conventional wisdom on its head. It’s a truism that turnout matters. That we need “unlikely” voters and first time voters to show up, because infrequent voters typically vote Democratic. Whereas frequent voters tend to vote Republican. While I have no reason to dispute that analysis as a historical matter, 2022 does not seem to fit any historical trends.
All of this data is based on early voter data 15 days out from the election:
Early voting is generally up compared to 2018 — about 19% (6,395,858 in 2018 vs. 7,650,083). That trend is true (to a greater or lesser degree) when factoring gender, race, and age.
Early voting also looks very good when factoring for Party ID. By “Modeled Party”, voting among Democratic voters is up 44% overall from 2018, and their share of early votes increased from 44.8% to 53.9%:
National Vote Totals By Modeled Party ID
Modeled Party |
2018 |
2020 |
2022 |
Democrat |
2,863,121 (44.8%) |
15,377,096 (52.2%) |
4,126,746 (53.9%) |
Republican |
2,955,244 (46.2%) |
10,917,682 (37.1%) |
2,743,385 (35.9%) |
Unaffiliated |
577,493 (9.0%) |
3,149,172 (10.7%) |
779,952 (10.2%) |
Total |
6,395,858 |
29,443,950 |
7,650,083 |
This is also true (to a lesser degree) when looking at Registered Party ID.
Registered Party ID Vote Totals
Registered Party |
2018 |
2020 |
2022 |
Democrat |
1,468,022 (23.0%) |
7,969,426 (27.1%) |
2,043,288 (26.7%) |
Republican |
1,387,874 (21.7%) |
3,707,389 (12.6%) |
1,263,915 (16.5%) |
Other |
86,843 (1.4%) |
351,238 (1.2%) |
102,928 (1.3%) |
Unaffiliated |
3,453,119 (54.0%) |
17,415,897 (59.1%) |
4,239,952 (55.4%) |
Total |
6,395,858 |
29,443,950 |
7,650,083 |
There are several caveats to this overall view. For example, the relative share of votes by gender did not move in the direction I would expect following the Supreme Court’s decision in Dobbs. While the total votes cast by women have increased, the relative share of votes by women in 2022 has actually gone down very slightly compared to 2018.
National Vote Totals By Gender
Gender |
2018 |
2020 |
2022 |
Female |
3,376,686 (52.8%) |
15,845,240 (53.8%) |
4,031,687 (52.7%) |
Gender Expansive |
0 (0.0%) |
0 (0.0%) |
277 (0.0%) |
Male |
2,929,847 (45.8%) |
12,903,809 (43.8%) |
3,458,234 (45.2%) |
Unknown |
89,325 (1.4%) |
694,901 (2.4%) |
159,885 (2.1%) |
Total |
6,395,858 |
29,443,950 |
7,650,083 |
On a more positive note, the relative share of votes cast by non-white voters has increased, especially among African Americans.
National Vote Totals By Race
Race |
2018 |
2020 |
2022 |
African-American |
558,750 (8.7%) |
3,069,467 (10.4%) |
762,617 (10.0%) |
Asian |
163,231 (2.6%) |
1,164,724 (4.0%) |
234,771 (3.1%) |
Caucasian |
5,253,150 (82.1%) |
22,375,827 (76.0%) |
6,135,468 (80.2%) |
Hispanic |
360,667 (5.6%) |
2,540,146 (8.6%) |
441,840 (5.8%) |
Other/Uncoded |
60,060 (0.9%) |
293,786 (1.0%) |
75,387 (1.0%) |
Total |
6,395,858 |
29,443,950 |
7,650,083 |
One data point that initially seems disappointing is votes viewed by age bracket. When broken down by age, the increase is not uniform. Although voting among all categories went up, the relative share of votes in each age bracket varied a lot. Voters aged 18-29, for example, represent a smaller share of the total vote than they did in 2018 (even though total votes went up). In fact, the fastest growing share of the vote occurred in the 75+ bracket, from 27.4% in 2018 to 29% in 2022.
National Vote Totals By Age
Age |
2018 |
2020 |
2022 |
18-29 |
318,396 (5.0%) |
2,475,370 (8.4%) |
332,070 (4.3%) |
30-39 |
362,651 (5.7%) |
2,919,249 (9.9%) |
448,471 (5.9%) |
40-49 |
482,254 (7.5%) |
3,256,490 (11.1%) |
553,528 (7.2%) |
50-64 |
1,568,818 (24.5%) |
7,762,730 (26.4%) |
1,766,169 (23.1%) |
65-74 |
1,905,531 (29.8%) |
7,155,774 (24.3%) |
2,304,013 (30.1%) |
75+ |
1,750,144 (27.4%) |
5,785,566 (19.6%) |
2,220,063 (29.0%) |
Unknown
|
8,064 (0.1%) |
88,771 (0.3%) |
25,769 (0.3%) |
Total |
6,395,858 |
29,443,950 |
7,650,083 |
I did not expect this result, so I decided to drill down a bit, and what I found really surprised me. Looking just at that age bracket of 75+, I discovered this:
Modeled Party ID among 75+ year olds
Modeled Party |
2018 |
2020 |
2022 |
Democrat |
703,733 (40.2%) |
2,661,331 (46.0%) |
1,091,741 (49.2%) |
Republican |
893,022 (51.0%) |
2,589,398 (44.8%) |
922,611 (41.6%) |
Unaffiliated |
153,389 (8.8%) |
534,837 (9.2%) |
205,711 (9.3%) |
Total |
1,750,144 |
5,785,566 |
2,220,063 |
The Modeled Party ID among those 75+ changed dramatically from 2018 to 2022. Democratic votes among those 75+ has risen from 40.2% Democratic in 2018 to 49.2% in 2022. And Republican vote share has declined from 51% to 41.6%! That seems pretty extreme, so I looked at the results by “Registered Party”, and the trend holds true (to a lesser degree):
Registered Party ID among 75+ year olds
Registered Party |
2018 |
2020 |
2022 |
Democrat |
375,241 (21.4%) |
1,564,338 (27.0%) |
600,964 (27.1%) |
Republican |
427,431 (24.4%) |
1,059,135 (18.3%) |
445,994 (20.1%) |
Other |
19,201 (1.1%) |
56,896 (1.0%) |
24,264 (1.1%) |
Unaffiliated |
928,271 (53.0%) |
3,105,197 (53.7%) |
1,148,841 (51.7%) |
Total |
1,750,144 |
5,785,566 |
2,220,063 |
It turns out that this trend holds true of most age brackets. For 65-74 year olds:
Modeled Party ID among 65-74 year olds
Modeled Party |
2018 |
2020 |
2022 |
Democrat |
825,114 (43.3%) |
3,579,541 (50.0%) |
1,242,635 (53.9%) |
Republican |
907,792 (47.6%) |
2,861,799 (40.0%) |
831,537 (36.1%) |
Unaffiliated |
172,625 (9.1%) |
714,434 (10.0%) |
229,841 (10.0%) |
Total |
1,905,531 |
7,155,774 |
2,304,013 |
And 50-64 year olds:
Modeled Party ID among 50-64 year olds
Modeled Party |
2018 |
2020 |
2022 |
Democrat |
691,871 (44.1%) |
3,935,025 (50.7%) |
929,630 (52.6%) |
Republican |
740,824 (47.2%) |
3,010,693 (38.8%) |
653,607 (37.0%) |
Unaffiliated |
136,123 (8.7%) |
817,012 (10.5%) |
182,932 (10.4%) |
Total |
1,568,818 |
7,762,730 |
1,766,169 |
And 40-49 year olds:
Modeled Party ID among 40-49 year olds
Modeled Party |
2018 |
2020 |
2022 |
Democrat |
242,523 (50.3%) |
1,828,034 (56.1%) |
328,470 (59.3%) |
Republican |
196,283 (40.7%) |
1,039,327 (31.9%) |
162,126 (29.3%) |
Unaffiliated |
43,448 (9.0%) |
389,129 (11.9%) |
62,932 (11.4%) |
Total |
482,254 |
3,256,490 |
553,528 |
These numbers seem too good to be believed. So I looked voting data another way — so-called Super Voters. The folks who vote in every election, no matter what. And the trend was even MORE pronounced. Among Super Voters of all ages:
Super Voters — National
Modeled Party |
2018 |
2020 |
2022 |
Democrat |
1,258,849 (43.1%) |
5,606,501 (52.2%) |
1,951,622 (54.4%) |
Republican |
1,451,242 (49.7%) |
4,348,998 (40.5%) |
1,377,171 (38.4%) |
Unaffiliated |
208,437 (7.1%) |
793,524 (7.4%) |
258,043 (7.2%) |
Total |
2,918,528 |
10,749,023 |
3,586,836 |
I don’t have any solid explanation for this, but the popular theories could easily be true — Senior Citizens voting to protect Social Security. Women (and Men) who have lived their entire lives with Abortion rights as the law of the land deciding that they will vote to protect those rights for their children and grandchildren. People who lived their entire lives believing that the US was the greatest democracy in the world and reacting in disgust again those who tried to overthrow our government on January 6. It’s hard to know for sure, but the trend is undeniable.
VOTING AT THE STATE LEVEL IS NOT UNIFORM
All this to say that voting patterns look very good for Democrats — at the national level. But unfortunately, there is some bad news when we look at state level data.
NEVADA
The most concerning is Nevada. First thing to note is that the reports of early voting overall has declined dramatically, and I don’t know enough to say why, as I don’t think Nevada has substantially restricted early voting.
Total Votes In Nevada
Modeled Party |
2018 |
2020 |
2022 |
Democrat |
41,504 (49.1%) |
145,724 (53.8%) |
16,494 (49.0%) |
Republican |
37,704 (44.6%) |
102,992 (38.1%) |
14,911 (44.3%) |
Unaffiliated |
5,337 (6.3%) |
21,895 (8.1%) |
2,233 (6.6%) |
Total |
84,545 |
270,611 |
33,638 |
Note the fairly small numbers involved here, even compared to 2018. But the relative vote share is what I am looking at mostly. Democratic share of early votes has remained flat so far. This is true when looking at older voters (remained flat). But of significant concerns is the decline among Super Voters:
Super Voters in Nevada
Modeled Party |
2018 |
2020 |
2022 |
Democrat |
16,836 (46.7%) |
52,916 (55.5%) |
5,526 (45.5%) |
Republican |
17,656 (49.0%) |
37,472 (39.3%) |
6,078 (50.0%) |
Unaffiliated |
1,521 (4.2%) |
4,978 (5.2%) |
543 (4.5%) |
Total |
36,013 |
95,366 |
12,147 |
One saving grace may be First Time Voters, where the relative share of Democrats voting for the first time has gone up dramatically. However, the total number of votes at play here is very small, so take the percentage increase with a grain of salt:
First Time Voters In Nevada
Modeled Party |
2018 |
2020 |
2022 |
Democrat |
2,197 (52.0%) |
12,986 (45.9%) |
1,566 (72.1%) |
Republican |
1,600 (37.9%) |
11,072 (39.1%) |
415 (19.1%) |
Unaffiliated |
428 (10.1%) |
4,251 (15.0%) |
191 (8.8%) |
Total |
4,225 |
28,309 |
2,172 |
NORTH CAROLINA
Another state with a troubling trend is North Carolina. First a look at overall vote totals by Modeled Party ID, in which the total number of early votes reported has declined, compared to 2018. But also the relative share by Democrats has declined as well.
Total Vote Share In North Carolina
Modeled Party |
2018 |
2020 |
2022 |
Democrat |
304,141 (52.2%) |
1,017,813 (54.2%) |
196,565 (51.9%) |
Republican |
249,842 (42.9%) |
744,458 (39.6%) |
155,314 (41.0%) |
Unaffiliated |
28,568 (4.9%) |
116,084 (6.2%) |
26,568 (7.0%) |
Total |
582,551 |
1,878,355 |
378,447 |
Thankfully, the relative vote share by Democrats among voters 75+ increased (abeit slightly).
Model ID Vote Share Among Voters 75+ In North Carolina
Modeled Party |
2018 |
2020 |
2022 |
Democrat |
46,721 (48.2%) |
158,219 (50.4%) |
43,346 (49.7%) |
Republican |
46,018 (47.4%) |
140,010 (44.6%) |
39,019 (44.7%) |
Unaffiliated |
4,277 (4.4%) |
15,439 (4.9%) |
4,865 (5.6%) |
Total |
97,016 |
313,668 |
87,230 |
Among Super Voters of all ages, there was also a slight increase of Democratic vote share:
Model Party ID Vote Share Among Super Voters in North Carolina
Modeled Party |
2018 |
2020 |
2022 |
Democrat |
134,801 (54.0%) |
368,356 (57.3%) |
96,468 (55.8%) |
Republican |
106,317 (42.6%) |
250,847 (39.0%) |
69,207 (40.1%) |
Unaffiliated |
8,608 (3.4%) |
24,171 (3.8%) |
7,111 (4.1%) |
Total |
249,726 |
643,374 |
172,786 |
PENNSYLVANIA
Now for some (very) good news. The early voting results in Pennsylvania look very good.
Total Vote Share in Pennsylvania
Modeled Party |
2018 |
2020 |
2022 |
Democrat |
25,895 (42.9%) |
566,293 (69.1%) |
357,847 (73.3%) |
Republican |
30,089 (49.8%) |
201,839 (24.6%) |
112,878 (23.1%) |
Unaffiliated |
4,439 (7.3%) |
51,366 (6.3%) |
17,410 (3.6%) |
Total |
60,423 |
819,498 |
488,135 |
That is an increase of 30.4 percentage points! WHAT?!? And the trend holds even by Registered Party ID.
Vote Total by Registered Party ID in Pennsylvania
Registered Party |
2018 |
2020 |
2022 |
Democrat |
27,740 (45.9%) |
605,724 (73.9%) |
356,681 (73.1%) |
Republican |
28,038 (46.4%) |
144,457 (17.6%) |
94,582 (19.4%) |
Other |
1,164 (1.9%) |
13,396 (1.6%) |
7,498 (1.5%) |
Unaffiliated |
3,481 (5.8%) |
55,921 (6.8%) |
29,374 (6.0%) |
Total |
60,423 |
819,498 |
488,135 |
WISCONSIN
Wisconsin does not allow voters to register Party ID, so I have to rely on Modeled Party ID. But the results so far look quite good. Note in particular the decline in Republican voter ID — 15.4 percentage points!
Modeled Party ID Vote Totals in Wisconsin
Modeled Party |
2018 |
2020 |
2022 |
Democrat |
61,140 (36.1%) |
348,850 (37.9%) |
97,942 (41.6%) |
Republican |
74,637 (44.0%) |
371,451 (40.3%) |
67,456 (28.6%) |
Unaffiliated
|
33,732 (19.9%) |
200,865 (21.8%) |
70,226 (29.8%) |
Total |
169,509 |
921,166 |
235,624 |
GEORGIA
Georgia is also looking pretty good so far (note: Georgia also does not record Party ID among voters, so this is all Modeled Party ID).
Modeled Party ID In Georgia
Modeled Party |
2018 |
2020 |
2022 |
Democrat |
229,187 (43.7%) |
756,068 (48.1%) |
434,812 (52.3%) |
Republican |
268,735 (51.2%) |
702,951 (44.7%) |
334,639 (40.3%) |
Unaffiliated |
26,713 (5.1%) |
111,872 (7.1%) |
61,451 (7.4%) |
Total |
524,635 |
1,570,891 |
830,902 |
This holds true when looking at Age and Voter Frequency (Super Voters).
Modeled Party ID Among Super Voters In Georgia
Modeled Party |
2018 |
2020 |
2022 |
Democrat |
91,183 (40.1%) |
271,376 (46.3%) |
188,623 (50.3%) |
Republican |
129,834 (57.1%) |
295,316 (50.3%) |
175,881 (46.9%) |
Unaffiliated |
6,295 (2.8%) |
19,929 (3.4%) |
10,732 (2.9%) |
Total |
227,312 |
586,621 |
375,236 |
Modeled Party ID Among Voters 75+ In Georgia
Modeled Party |
2018 |
2020 |
2022 |
Democrat |
32,285 (34.1%) |
84,799 (35.1%) |
59,233 (39.6%) |
Republican |
58,454 (61.7%) |
144,031 (59.6%) |
82,941 (55.5%) |
Unaffiliated
|
3,949 (4.2%) |
12,883 (5.3%) |
7,309 (4.9%) |
Total |
94,688 |
241,713 |
149,483 |
The trend of increased share of early vote totals by Democrats hold true among other states as well — Florida, Texas, Arizona, etc. Too many for me to post about here directly.
CONCLUSION
There are still 15 more days of early voting, plus Election Day voting. So nothing is certain. But these initial results are very encouraging. Encouraging enough to tell the pessimists and doomsayers to knock it off. Although we can’t know how early voters actually voted, we CAN say that the pool of early voters is far more Democratic than we had any right to predict going into a mid-term election.
Also note these initial results MUST be considered in light of extensive polling data suggesting that Republicans have been gaining ground the last two weeks. While I think there are unprecedented circumstances that may cause polling to be significantly off this year (loss of a Constitutional right to abortion, an actual attempt to overthrow the US Government on January 6, and ongoing Election Denialism among Republicans), it would be a mistake to completely dismiss polling trends that favor Republicans. In other words, nothing is guaranteed. GET OUT THE VOTE!
But if you needed some reason to believe that there is hope, then here it is. The early voting data is very encouraging indeed. Now if we could just make sure that voters turn out in Nevada and North Carolina!