Review data on: TargetSmart
By Jim Ellis — Monday, Nov. 4, 2024
Early Voting Wrap Up
Early Gender Numbers: Analysts Missing the Point — News reports are covering the early vote totals and report after report states that women are voting in greater numbers than men. While the raw numbers indicate such, the historical trends suggest something slightly different.
When comparing the gender participation segments from the previous presidential year of 2020 to this year, women are up just two-tenths of one percent while men, compared to their previous performance in a presidential year, are up one-half of a percentage point. This data comes from the Target Early/TargetSmart organization and is based upon more than 55 million early votes cast (at this writing) for the 2024 election.
Compared to the 2022 midterm, women are up slightly more than half a point, while men are down by that same amount. Perhaps the more significant finding is the increase in rural voting that appears evident around the country and the decrease detected in urban voter participation.
Closings: Early Voting Ends in All Seven Battlegrounds — All seven battleground states have concluded their early voting periods, and generally the results appear to be better for Republicans than past early voting years. The move to get their voters to the polls early looks to have produced results for the GOP.
The two biggest swings come in North Carolina and Nevada where, for the first time in history, more Republicans have voted early than Democrats. The Arizona vote count looks good for the GOP, too. The Harris campaign likely is pleased with what they see coming from Michigan and especially Wisconsin. Republican early turnout in the Badger State appears particularly low. Georgia and Pennsylvania show mixed patterns, meaning we are likely again headed for very close finishes in both of those states.
The early vote data is collected and reported by the TargetEarly/TargetSmart organization.
President
Iowa: Des Moines Register Poll Sees Lead Change — The Des Moines Register/Mediacom Poll is always viewed as the gold standard of Iowa prediction models. Their pre-election survey (Oct. 28-31; 808 likely Iowa voters; live interview) finds Vice President Kamala Harris all of a sudden leading Donald Trump, 47-44 percent, largely because of a huge break toward the former among women, particularly those who are aged 65 and older.
Yet, simultaneously, Emerson College released their Iowa survey and finds a much different result, one that is more in line with previous data. The Emerson survey (Nov. 1-2; 800 likely Iowa voters; multiple sampling techniques) sees former President Donald Trump holding a nine point lead, 54-45 percent. This is also closer to Iowa voting history. In 2016, Trump carried the state 51-42%, and 53-45% in 2020.
An Iowa win for Harris would turn the electoral map topsy turvy and potentially send the final result on a completely different course. This situation bears watching as we close in on election day. Iowa has six electoral votes.
Pennsylvania: Swinging Towards Trump — A total of six current polls from five different Keystone State pollsters all find former President Donald Trump leading Vice President Kamala Harris in Pennsylvania ballot test results. The range is relatively wide.
The Echelon Insights survey (Oct. 27-30; 600 likely Pennsylvania voters; live interview & text) posts Trump to a 51-46 percent count on the multi-candidate ballot. The poll swinging the most from that result comes from Fox News (Oct. 24-28; 1,057 likely Pennsylvania voters; live interview & online) and suggests the two candidates are tied with 48 percent apiece.
Viewed together, the two are a good example of how different pollsters can arrive at significantly different results even when testing the same race during the same relative time frame. With so many polls showing many different outcomes regarding the various campaigns, it is now time to sit back and let the actual votes determine the final result.
Senate
Arizona: Closer Than Expected — Earlier this week, we covered a new poll regarding the Arizona Senate race that posted former news anchor Kari Lake (R) running slightly ahead of Rep. Ruben Gallego (D-Phoenix) in the open Senate race. The poll, which forecast Lake holding a one-point edge, came from the Data Orbital firm over the Oct. 26-28 period.
Now, we see another survey arriving at the same conclusion. AtlasIntel released their latest Arizona poll (Oct. 30-31; 1,005 likely Arizona voters) and the conclusion was identical to that of the previous Data Orbital survey, that being Lake holding a small one-point lead.
Two other pollsters were also in the field during the similar period. Another AtlasIntel survey was conducted during the Oct. 25-29 period, and they saw the opposite result: Rep. Gallego leading by the same one-point margin. A fourth poll was also introduced (Noble Predictive Insights; Oct. 28-30; 775 likely Arizona voters) and this ballot test returned Rep. Gallego to a four-point edge.
While all the polls before October gave Gallego substantial leads, there is now significant data showing a closing of the race. Though Rep. Gallego is still the likely winner, this race may now join a couple of the Arizona House races in political overtime.
Michigan: Mirror Images — Two surveys were released of the Michigan electorate that ended on the same day and produced the same 49-47 percent ballot test conclusion, except each found a different candidate reaching the 49 percent figure. Mitchell Research and Communications, polling for the Michigan News Source (Oct. 28-29; sample size undisclosed) sees former Congressman Mike Rogers (R) with the slight lead, while AtlasIntel (Oct. 25-29; 983 likely Michigan voters) arrived at the opposite conclusion. The latter finds Rep. Elissa Slotkin (D-Lansing) clinging to the two-point edge.
This is an excellent example of two professional pollsters testing the same electorate at the same time and arriving at an opposite conclusion. This makes us remember that polls give us a broad snapshot within a stated margin of error at the time a survey is conducted and reminds us that surveys are meant to show trends over a long period as opposed to who might be leading a campaign at a given time.
Virginia: Outlier Poll Finds Hung Cao Down 1 — In another strange polling situation, one even the conducting pollsters admit is probably an outlier, Virginia US Senate candidate Hung Cao (R) has pulled into just one point of Sen. Tim Kaine (D) on the Chism Strategies survey (10/28-30; 520 VA likely voters; live interview & text; part of a three state series), 46-45%. No other data shows such a close result.
Two other late October polls (Oct. 25-29), conducted by Roanoke College and the Cygnal polling firm, see double digit leads (11 and 10 points, respectively), which is consistent with projections that Sen. Kaine will easily win re-election. As a point of reference, the Chism poll also sees a close Virginia presidential race. Their ballot test finds Harris and Trump effectively in a tie. This, too is refuted by other pollsters.
House
ME-2: A Huge Polling Range — Continuing the theme of divergent polling, we see another example in northern Maine where Rep. Jared Golden (D-Lewiston) is in a tough re-election battle with Republican state representative and former NASCAR driver Austin Theriault.
Survey USA, conducting a statewide poll (Oct. 24-29; 1,079 likely Maine voters; 484 likely ME-2 voters; online), finds Rep. Golden posting a 53-41 percent lead over Theriault even while Trump leads Harris, 49-44 percent in the same district, a CD that the former has carried in both of his previous elections.
In another survey taken during a slightly earlier period, Axis Research, polling for the National Republican Congressional Committee (Oct. 17-20; 411 likely ME-2 voters; live interview & text), actually sees Theriault leading 47-45 percent. Therefore, we see a 14-point polling span between these two studies. With so many surveys being released at the end of the cycle, it is rather routine to see data results producing wildly comparative swings, and the final days of the 2024 election is apparently no exception to such a trend.
