Category Archives: Polling

Challengers With a Lead – Part I

By Jim Ellis

Sept. 29, 2020 — We’ve already seen eight 2020 US House challengers unseat their incumbent opponents so far — obviously all in the primaries. Is it possible that that’s a precursor to a larger number of incumbents losing their seats in the general election?

In typical election years, well over 90 percent of incumbents who seek re-election win. Therefore, eight members denied re-nomination in their own party primary elections is an unusually high number. What’s more, turning to the general election, 22 House incumbents have trailed in at least one poll since July began.

Below is the list of the first 11 incumbents, alphabetically by state, who are facing what appear to be the most competitive challengers in the country. The list includes 14 Republican incumbents and eight Democrats for a total of 22. Tomorrow, we will cover the remaining 11. All challengers have led the incumbent in at least one political poll of their race.

Rep. Don Young (R-AK-AL):

• Public Policy Polling (July 7-8)
Challenger: Alyse Galvin (I/D) margin: +2 points
2016 Presidential: Trump, 51-37%
—   Galvin ran in 2018 and saw similar polling numbers even as late as Oct. 29. Rep. Don Young (R-Ft. Yukon) would still go onto win the race 53-46 percent. Galvin is one of the many close finishers returning for a re-match this year.

Rep. David Schweikert (R-AZ-6):

• Greenberg Quinlan Rosner Research (Aug. 6-12)
Challenger: Hiral Tipirneni (D) margin: +3 points
2016 Presidential: Trump, 52-42%
—   Rep. David Schweikert (R-Fountain Hills/Scottsdale) has pleaded to 11 ethics violations relating to using his government resources for political purposes in addition to campaign finance irregularities. With the district becoming more Democratic in addition to his personal situation, Rep. Schweikert faces his toughest re-election campaign against physician Hiral Tipirneni who twice ran close races in the adjoining 8th CD.

Rep. Mike Garcia (R-CA-25):

• Normington Petts (Sept. 21-23)
Challenger: Assemblywoman Christy Smith’s margin: +3 points
2016 Presidential: Clinton, 50-44%
—   One of the most recent polls found for this House study shows Rep. Mike Garcia (R-Santa Clarita) — who won his seat in a May 12 special election by surprisingly taking a seat back from the Democrats — trailing state Assemblywoman Christy Smith (D-Newhall), the woman he defeated four months ago. With California going heavily Democratic in the presidential election, it’s certainly not out of the realm of possibility that this race again becomes a toss-up despite Rep. Garcia’s 10-point win with a high special election turnout.

Rep. Donna Shalala (D-FL-27):

• 1892 Polling (Sept. 2-6)
Challenger: Maria Elvira Salazar margin: +3 points
2016 Presidential: Clinton, 58-39%
—   This is a re-match of the 2018 open seat campaign that saw former Health & Human Services Secretary and President of the University of Miami, Donna Shalala, top former Spanish language television news reporter Maria Elvira Salazar. Shalala’s victory margin was 52-46 percent. The only public poll released so far came in early September and produced a surprising result with Salazar pulling slightly ahead.

Continue reading →

Sen. Collins of Maine Hanging On

By Jim Ellis

Maine Sen. Susan Collins (R)

Sept. 25, 2020 — Our original quick analysis may be incorrect. Soon after the announcement that Justice Ruth Bader Ginsburg had passed away, Sen. Susan Collins (R-ME) made a statement indicating that she would not support until after the election a motion to proceed that would allow the Senate to vote on confirming a Supreme Court replacement.

Responding, we believed that Sen. Collins’ decision had sealed her own doom in regard to winning her re-election campaign against state House Speaker Sara Gideon (D). We reasoned that at least a part of Maine’s conservative bloc would likely turn away from Sen. Collins, one who they do not particularly trust anyway, yet a group she needs to close the polling deficit she faces against Gideon. This was also predicated upon the belief that her move would gain little from the center-left or mostly left voter who may have supported Sen. Collins in previous elections.

A just-released Maine political survey, however, suggests that such a conclusion may not be so clear. Moore Information, polling for the National Republican Senatorial Committee and the first Maine Senate data released into the public realm after Justice Ginsburg’s death (Sept. 20-22; 500 likely Maine voters, live interview), finds Sen. Collins actually gaining support to pull into a 42-42 percent tie with Gideon. This is the first survey showing Sen. Collins even or ahead since another Moore Information poll conducted in late June posted her to an eight-point lead.

The Maine seat is critical to determining the next Senate majority. Routinely, we believe it is part of a four-state firewall that Republicans must maintain to uphold their relatively slim 53-47 chamber majority. The other three being converting Alabama, and retaining Iowa and Montana. Losing any one of these four states would likely turn the focus to North Carolina where Sen. Thom Tillis (R) would be forced to score a come-from-behind win, which could then pave the way for Republicans keeping majority control with a smaller edge.

Polling has been clear for weeks that Sen. Collins is trailing. Since 2020 began, now 18 polls have been conducted of this Senate race, and Gideon has led in all but the two Moore Information surveys. Overall, the incumbent has been trailing by an average slightly over four percentage points with a median of 4.5 percent, numbers barely within the polling margin of error range.

Continue reading →

Senate: A Polling Comparison

By Jim Ellis

Sept. 24, 2020 — Today, we look at the competitive Senate races and segment the group around polling consistency. Several races routinely report point spreads between the two major party candidates that are wildly inconsistent, while others vary over a small difference often within the same polling period.

The two most extreme surveys during the month are listed for each state with the most extreme first and the closest second. You will notice that the British firm, Redfield & Wilton Strategies, is often listed as the most extreme.

We begin, alphabetically by state, with the inconsistent group. Only the two Georgia races are in the September consistent segment.

INCONSISTENT

Arizona
Number of September polls: 14
Polling range: 16 points

Redfield & Wilton Strategies
(Sept. 12-16; 855 likely Arizona voters; combination online and live interview)
• Mark Kelly (D) – 52%
• Sen. Martha McSally (R) – 35%

ABC News/Washington Post
(Sept. 15-20; 579 AZ likely voters; live interview)
• Mark Kelly (D) – 49%
• Sen. Martha McSally (R) – 48%

Iowa
Number of September polls: 2
Polling range: 8 points

Fabrizio Ward/Hart Research (for AARP)
(Aug. 30-Sept. 5; 800 likely Iowa voters; live interview)
• Sen. Joni Ernst (R) – 50%
• Theresa Greenfield (D) – 45%

Selzer & Company
(Sept. 14-17; 658 likely Iowa voters; live interview)
• Sen. Joni Ernst (R) – 45%
• Theresa Greenfield (D) – 42%

Continue reading →

Sen. Lindsey Graham’s Challenge

By Jim Ellis

Is incumbent Sen. Lindsey Graham (R-SC) in re-election trouble?

Sept. 18, 2020 — Quinnipiac University surveyed the South Carolina political situation as part of their three-state polling series, which again produces some eyebrow-raising data. The results help identify why Sen. Lindsey Graham, a Republican senator from a Republican state, finds himself languishing in a competitive contest.

The poll (Sept. 10-14; 969 likely South Carolina voters, live interview conducted by the RDD firm for Quinnipiac) tested both the presidential and Senate campaigns. President Trump leads former vice president Joe Biden 51-45 percent in a ballot test that seems to be an under-count when looking at the survey’s supporting numbers. Sen. Graham, however, falls into a tie with opponent Jaime Harrison, at 48-48 percent, in a result that the underlying responses do seem to support.

President Trump’s six-point lead appears low because he tops Biden on virtually every personal and issue question. The Trump favorability index is 51:45 percent positive to negative, but the Biden ratio is much worse at 43:50 percent. The generic Republican-Democrat number falls 52-44 percent in favor of the GOP label.

Despite poor coronavirus management numbers for the president nationally, this South Carolina survey returns a 49:48 percent approval number on his handling of the issue. Furthermore, the respondents, in a 50-46 percent break, believe President Trump would do a better job handling coronavirus in the future than Biden. Not a particularly strong performance in this issue area, but better for the President than in almost any other place.

Trump also scores better in his handling of the economy (55-40 percent), the military (54-42 percent), and “keeping your family safe” (52-43 percent). Biden is favored, and only barely, 48-46 percent, on just one issue: racial equality.

Most importantly, the issue matrix sets up perfectly for Trump. The top two issues, according to these respondents, are the ones upon which the president is basing his campaign, law and order (23 percent) and bringing back the economy (22 percent). The Biden key issues rate rather poorly: coronavirus (12 percent), racial equality (12 percent), and healthcare (10 percent).

All of these underlying numbers suggest the Trump ballot test margin should be stronger than six points, which could be a signal that there is a “shy Trump voter factor” even in what is typically a safe Republican state. The “shy Trump voter” is the phrase now used to describe the individual who only secretly favors the president.

Continue reading →

Reading North Carolina

By Jim Ellis

Sept. 17, 2020 — CNN released new North Carolina poll results earlier this week, and we again see a familiar pattern unfolding. There has been a Republican under-poll in the southern states detected in the past few elections, and the North Carolina pattern appears to form relatively consistently upon studying its most competitive statewide races in 2014, ’16, and what may be happening in 2020. There were no statewide Tar Heel State contests in 2018.

The CNN poll (conducted through the SSRS statistical firm; Sept. 9-13; 787 likely North Carolina voters; live interview through landline and mobile phones) found former vice president Joe Biden leading President Trump, 49-46 percent; Democratic US Senate nominee Cal Cunningham edging incumbent Republican Thom Tillis, 47-46 percent; and Gov. Roy Cooper (D) easily outdistancing Republican Lt. Gov. Dan Forest, 53-44 percent.

How do these mid-September races compare with other campaigns at this same interval, and what does that tell us for autumn?

First, the CNN poll is one of seven polls conducted in North Carolina during the month of September, and its three-point margin for Biden is the Democrat’s second-best showing within this group. The only better Biden performance came from the Fox News poll at the beginning of September (Aug. 29-Sept. 1; 722 likely North Carolina voters, live interview), which posted him to a four-point, 50-46 percent, advantage.

Among the five other surveys, Biden is ahead in two, President Trump in two, and one has the pair tied at 47 percent apiece (Survey USA for WRAL-TV; Sept. 10-13; 596 likely North Carolina voters). From the eight polls conducted from Aug. 29-Sept. 13, Biden’s edge is just 0.7 percent, meaning the two candidates average to a statistical tie.

Recent political history suggests that this type of an average spread sets up well for President Trump, and possibly Sen. Tillis. It appears that Gov. Cooper’s margin is beyond the statistically relevant late-term Republican swing.

In September of 2016, a total of 14 publicly released polls were conducted during that month. Within this group, Hillary Clinton led in 10 of the surveys with an average spread of 2.4 percentage points. Trump was ahead in just three polls with an average margin of 2.0 percent. Two polls found the candidates tied. Therefore, Clinton’s overall September edge was an average 1.1 percent.

Continue reading →

Why Arizona is So Pivotal

By Jim Ellis

Does Arizona hold the key in a Trump-Biden election?

Sept. 16, 2020 — For several reasons, the Grand Canyon State of Arizona is possibly the most important state on the political map to determine the ultimate presidential election outcome.

Primarily, Arizona is one of five core states that President Trump must win to form a foundation for a favorable remaining state coalition map. The other four, geographically from west to east, are Texas, Georgia, Florida, and North Carolina. Should Joe Biden break through in any one of these five states, he will likely win the national election.

At this point, Arizona appears to be the most precarious of the Trump core states. While the President’s numbers are improving here, the September polls find him trailing Biden in all six publicly released surveys from a range of one to nine points among likely voters, with a mean average of Biden plus-4, and a median of Biden plus-3.5.

The five states are so critical to President Trump, or any Republican national candidate, because, as a unit, they yield a relatively easy remaining victory map. Carrying the five southern sector domains and assuming no leakage in Ohio or Iowa, and even while not winning Nevada or New Hampshire, the GOP nominee then claims the presidency with a victory in any one of the key Great Lakes states: Michigan, Minnesota, Pennsylvania, or Wisconsin. For a Democrat to win under this scenario, he or she would be forced to sweep the aforementioned quartet.

President Trump won the 2016 Electoral College vote 306-232, which means he can relinquish a net 36 electoral votes in 2020 and still win the national election. Under the scenario of him taking either Wisconsin or Minnesota, along with keeping Arizona, he would defeat Biden with exactly 270 electoral votes. This model also assumes he wins the 2nd Congressional District from both Nebraska and Maine, the two states that split their electoral votes. He won both in 2016. Under this scenario, Michigan and Pennsylvania would go to Biden.

Arizona, now potentially teetering toward the Democrats, is critical to the president’s prospects because Trump cannot afford to trade it for one of the western Great Lakes States, either Minnesota or Wisconsin. Such a loss would force the president to win two of the four Great Lakes, but only one could be Minnesota or Wisconsin since those two states have 10 electoral votes and Arizona has 11.

Therefore, simply put, losing Arizona because of its 11th electoral vote would mean that Trump would be forced to carry either Michigan or Pennsylvania in addition to one of the other three remaining Great Lakes States. A further scenario involving Trump losing Arizona and replacing it with both of the 10-electoral vote states (MN and WI) could result in the election ending in a 269-269 tie. This would force a tiebreaker to be decided in the US House of Representatives.

Let’s look at the chances of Trump winning Arizona by comparing his current standing to where he was at this point in 2016. Looking at the Real Clear Politics polling archives, we find that 19 Arizona polls were conducted during the entire 2016 election cycle. In 2020, just since the July 4th holiday break, 25 surveys were publicly reported in the Grand Canyon State.

Four years ago, at the end of August through mid-September, two individual polls came from Gravis Marketing (Aug. 25-27, 2016) and NBC News/Marist College (Sept. 6-8, 2016). These surveys yielded Trump four and one-point leads, respectively. Shortly thereafter, the trend began to turn Hillary Clinton’s way. The OH Predictive Insights survey (Sept. 28-30, 2016) found the two candidates tied at 42 percent apiece, while Emerson College (Oct. 2-4, 2016) and the Arizona Republic newspaper poll (Oct. 10-15, 2016) detected consecutive leads for Clinton of two and five points.

Therefore, Arizona did not turn toward Trump for good until the Monmouth University survey in late October (Oct. 21-24, 2016), which put him just one point ahead. Going into the election from that point, and remembering the 2016 election was on Nov. 8, Trump led in the final four polls from a two to five-point margin. He would eventually win the actual vote count by 3.6 percentage points, 48.7 – 45.1 percent, meaning a raw vote margin of 91,234 votes of more than 2.573 million ballots cast.

Because of Arizona’s fast population growth, the state has changed in four years. During that interval, the overall population expanded five percentage points to 7,278,717 individuals according to the Census Bureau’s July 2019 estimate, the latest available.

Minorities, specifically Hispanics and blacks, account for approximately 69 percent of the population gain, thus at least partially explaining Biden’s improved prospects in the state. Overall, Hispanics represent 31.7 percent of the overall Arizona population, and blacks 5.2 percent, as compared to the non-Hispanic white position receding to 54.1 percent.

The population changes suggest that this already tight political state will likely become even closer as we head for Nov. 3.

The Senate Trends

By Jim Ellis

Sept. 14, 2020 — The AARP organization yesterday released a series of polls covering nine Senate races in eight states that help set benchmarks for the most recent data.

AARP contracted with four polling firms, two Republican and two Democratic, and paired each with the opposite party pollster. The Benenson Strategy Group (D) partnered with the GS Strategy Group (R) for surveying Arizona, Michigan, and North Carolina.

The Fabrizio Ward firm (R) and Hart Research Associates (D) conducted the joint Colorado, Georgia (both races), Iowa, Maine, and Montana survey research. All of the polls were live interview with large sampling universes of likely voters unless otherwise noted.

Predominantly, the ballot tests find the Democratic candidate typically leading, but with the Republican improving his or her position in comparison to the previously released polling results.

Below are the AARP results followed by the two most recent reported surveys in each state:

ARIZONA
Benenson Strategy/GS Strategy (Aug. 28-Sept. 8; 1,600 likely Arizona voters)
• Mark Kelly (D) – 48%
• Sen. Martha McSally (R) – 45%
Previous:
• Change Research (D) – (Sept. 4-6) – Kelly +6
• Redfield & Wilton Strategies (UK) – (Sept. 30-Aug. 4) – Kelly +15

COLORADO
Fabrizio Ward/Hart Research (Aug. 30-Sept. 5; 800 likely Colorado voters)
• Ex-Gov. John Hickenlooper (D) – 51%
• Sen. Cory Gardner (R) – 46%
Previous:
• Morning Consult – (Aug. 21-30) – Hickenlooper +9
• Public Policy Polling (D) – (Aug. 18-19) – Hickenlooper +9 (voters)

GEORGIA-A
Fabrizio Ward/Hart Research (Aug. 30-Sept. 5; 800 likely Georgia voters)
• Jon Ossoff (D) – 48%
• Sen. David Perdue (R) – 47%
Previous:
• Public Policy Polling (D) – (Aug. 13-14) – Even (voters)
• Garin Hart Yang Research (D) – (Aug. 10-13) – Ossoff +2

Continue reading →

A Polling Comparison

By Jim Ellis

Neck-and-neck polling in a few key battleground states between President Donald Trump and former vice president Joe Biden shows interesting parallels to the 2016 race between Trump and Hillary Clinton.

Sept. 10, 2020 — With a plethora of presidential polls being released every week providing sometimes radically diverse results, it is often difficult to draw a clear picture of where the electorate is heading.

The conventional wisdom and preponderance of polling trends suggest that Joe Biden is leading the presidential race, but that President Trump is making a comeback, and the race is beginning to show some of the same characteristics found in 2016.

Three of the key states that baffled the political pollsters four years ago were Michigan, Pennsylvania, and Wisconsin. As we will remember, President Trump was expected to lose them all but scored close upset victories in each place.

The aggregate group of 2016 pollsters missed in each of the three states and the Real Clear Politics polling archives still publicly maintains all of those survey results. Therefore, we have the historical data to draw clear parallels between then and now.

In Wisconsin, 33 polls were taken during the election cycle and only one, from the Trafalgar Group at the end of the campaign season, placed Trump in front. A total of 62 polls were conducted in Pennsylvania with only three, again including a Trafalgar poll, projecting the future president into the lead. In Michigan, 42 polls were publicly released with Trump ahead in just two.

Though it is not generally statistically significant to average polling results because the polling methodologies, and certainly sample sizes, are very different, doing so does give us a guide as to the error factor that was present in 2016, and possibly a glimpse into what might exist this year.

In Wisconsin, the average Hillary Clinton lead advancing into the general election was 6.5 percentage points. With a 0.7 percent win for Trump, the overall error factor became a whopping 7.2 percent. The Pennsylvania numbers were closer but still a significant miss. Clinton’s average lead heading into Election Day was 2.1 percent and the president won there by the same 0.7 percent that he carried Wisconsin. Therefore, the Keystone State error factor was 2.8. Michigan was a similar story. Error factor: Clinton plus-3.6 percent. Trump victory margin: just 0.3. Total Michigan error factor: 3.9.

Continue reading →

New Hampshire Primary Today

By Jim Ellis

Sept. 8, 2020 — Winding through the final state primaries, voters in the Granite State cast their ballots today in order to nominate candidates for US Senate, governor, and two congressional districts. After today, only three primaries remain: next Tuesday in Delaware and Rhode Island, and the Louisiana jungle primary that runs concurrently with the general election.

Sen. Jeanne Shaheen stands for a third term this year and draws only minor opposition on the Democratic ballot. On the Republican side, businessman Corky Messner, who has already loaned his campaign approximately $4 million, is favored to top retired Army General Don Bolduc.

Messner certainly has the resources to run a competitive race against Sen. Shaheen, but there is no question she is a heavy favorite in the general election. Prior to being elected to the Senate in 2008, Shaheen served three two-year terms as governor but lost her first Senate bid opposite then-US Rep. John E. Sununu (R) in 2002.

Since the turn of the century, however, New Hampshire has been one of the most volatile political states, and swingingly wildly from the top of the ticket all the way down the ballot has become a frequent occurrence. Therefore, incumbents from both parties can never be considered completely safe.

Gov. Chris Sununu (R) stands for a third two-year term – New Hampshire and neighboring Vermont are the only two states that mandate two-year gubernatorial terms – and faces only Franklin City councilwoman and radio talk show host Karen Testerman and a man named Nobody, who frequently runs for New Hampshire political office as a Republican or a Libertarian Party member.

The Democrats feature a two-way gubernatorial nomination race between state Senate Majority Leader Dan Feltes (D-Concord) and Executive Councilor Andru Volinsky. New Hampshire’s unique Executive Council is a five-member panel elected in districts and serve as gubernatorial advisors and a check on the governor’s power. The Executive Council has veto power over pardons, nominations and large state contracts. Polling suggests a close race.

Continue reading →

Senate Sleepers in Minnesota
and South Carolina

By Jim Ellis

Are incumbents Sen. Tina Smith (D-MN), left, and Sen. Lindsey Graham (R-SC) in re-election trouble?

Sept. 8, 2020 — At the beginning of the 2020 election cycle, it appeared that Sens. Tina Smith (D-MN) and Lindsey Graham (R-SC) were set for re-election. That no longer appears to be the case.

Polling is now consistently showing close races in the two states, and Sen. Smith’s recent comments about the police have seemingly moved her re-election campaign closer to the toss-up category according to a just-released Harper Polling survey.

Looking at the HP data (Aug. 30-Sept. 1; 50 likely Minnesota voters via live interview), Sen. Smith’s ballot test edge has dropped to just 43-41 percent over former US Rep. Jason Lewis (R).

In May, Harper found Sen. Smith posting a 46-35 percent advantage over Lewis. After this release, Public Policy Polling forecast a 48-39 percent split toward the end of July. Lastly, Emerson College published what appears to be precursor data to the latest Harper findings. The Boston-based university poll (Aug. 8-10; 73 likely Minnesota voters) found a three-point split between the two contenders, 48-45 percent, again in Sen. Smith’s favor.

To the southeast of liberal Minnesota, in conservative South Carolina, veteran Sen. Lindsey Graham (R) has also fallen into a competitive contest with former state Democratic Party chairman Jaime Harrison, who had raised an incredible $29 million for his smaller state race through the end of June.

The most recent poll we have seen from the Palmetto State was released at the beginning of August. Quinnipiac University (July 30-Aug. 3; 91 registered South Carolina voters) detected a 44-44 percent tie between Sen. Graham and Harrison. From mid-July through the beginning of August, six surveys were conducted, including the Quinnipiac study, and Sen. Graham’s average lead was only 2.8 percentage points.

Continue reading →