Tag Archives: FiveThirtyEight

Florida Polling – What to Expect

By Jim Ellis

Aug. 20, 2021 — Two pollsters released Florida ballot test data yesterday, and the combined results are a likely prelude of what we can expect from the vast multitude of survey research firms that will be testing the Sen. Marco Rubio – Rep. Val Demings general election campaign in the coming year.

Florida Sen. Marco Rubio (R)

Susquehanna Polling & Research (Aug. 4-10; 700 registered Florida voters, live interview) posted their survey result, which found Sen. Rubio topping Rep. Demings by a relatively substantial 50-39 percent clip.

This was immediately countered by a St. Pete Polls survey conducted later in the month (Aug. 16-17; 2,068 registered Florida voters, online) that sees the race already dropping into a virtual dead heat, with Sen. Rubio only holding a two-point edge, 48-46 percent.

Florida polling history suggests we will see this type of divergent pattern among pollsters probably until the next election. In Sen. Rubio’s 2016 re-election race, for example, where he defeated then-Rep. Patrick Murphy (D) with an eight-point victory spread (52-44 percent), most of the pollsters were forecasting a much closer finish.

During the period from Oct. 25 through election day 2016, 11 polls were released covering the Rubio-Murphy race according to the Real Clear Politics polling archives, and while all but one correctly predicted Sen. Rubio would win re-election, only five were within the correct final margin range. The others were forecasting a very close Rubio win of between a virtual tie and four percentage points.

Looking at the Biden-Trump 2020 Florida aggregate research studies tells a similar tale. Again, beginning with polling occurring from Oct. 25 through the election, 19 Florida presidential ballot test polls were published. Only six of the 19 correctly predicted a Trump Florida victory and all of those were close to the final margin of 3.3 percentage points. One of the pollsters who called this race almost exactly was Susquehanna Polling & Research. St. Pete Polls missed, wrongly projecting a close Biden win.

Looking at the FiveThirtyEight statistical organization’s polling firms rating chart, Susquehanna and St. Pete Polls are at parity. Susquehanna rates as the 92nd firm of the top 100, while St. Pete finishes three slots behind them at number 95. Both receive an accuracy letter grade of B+.

Continue reading →

Youngkin Within Two of McCauliffe in Virginia Governor’s Campaign

By Jim Ellis

Glenn Youngkin (R), Terry McCauliffe (D), left-to-right, Virginia gubernatorial candidates

June 15, 2021 — A just-released Virginia governor’s campaign survey yields a surprising ballot test result, especially when seeing that the data was collected before last week’s Democratic primary.

WPA Intelligence (WPAi), in their most recent poll for the Glenn Youngkin for Governor campaign (June 2-6; 506 likely Virginia voters, live interview), already finds the Republican gubernatorial nominee climbing to within two percentage points (48-46 percent) of the newly crowned Democratic nominee, former governor, Terry McAuliffe.

While certain elements point to a slight Republican survey skew, it does signal, as other research has, that the 2021 Virginia governor’s race may be tighter than the state’s most recent elections that have definitively favored the Democratic candidates.

Several analysis points need reviewing. First, WPAi is a credible Republican pollster. According to the FiveThirtyEight statistical website pollster rankings, WPA Intelligence has a very respectable 0.8 predictive rating with just a slight 0.7 percent Republican bias factor.

Second, some of the geographic segmentation returns, however, appear a bit too optimistic for the Youngkin team. While the Washington, DC DMA sector breaking 56-39 percent for McAuliffe and the Roanoke/Lynchburg DMA posting a 49-46 percent edge for Youngkin are both wholly believable, the projections for two other segmented regions raise eyebrows.

Youngkin leading 51-44 percent in the Norfolk/Portsmouth/Newport News DMA, which also includes the Virginia Beach area, isn’t particularly consistent with the way the Tidewater region has voted in the most recent elections. Additionally, the same should be said for the Richmond/Petersburg DMA where WPAi sees Youngkin pulling ahead with a 48-44 percent split.

Third, these latter numbers appear too favorable for Youngkin, and particularly so in the Richmond/Petersburg area where the McAuliffe favorability rating in this same poll is 45:40 percent favorable to unfavorable. Conversely, however, if the trend were to continue for these two places and is proven accurate, it would basically signal a return to the voters’ previous electoral pattern. The Virginia historical election matrix was much better for Republicans than the numbers found in elections since the early 2000s.

Continue reading →

Disputing the Georgia Poll

By Jim Ellis

Former DeKalb County, Georgia chief executive officer and ex-state representative, Vernon Jones

May 11, 2021 — At the end of last week, the Remington Research Group released a survey of the Georgia 2022 Republican gubernatorial primary for their client, former DeKalb County chief executive officer and ex-state representative Vernon Jones. The results were met with some derision.

The poll (May 1-3; 1,040 likely Republican primary voters, interactive voice response system) finds Gov. Brian Kemp leading Jones only 39-35 percent, and an Atlanta Journal Constitution reporter responded with negative comments about the poll’s credibility and cited unnamed GOP consultants who allegedly agree with this analysis.

Greg Bluestein of the AJC said in response to the poll’s publishing, “suffice it to say we have not found one person with proven knowledge about Georgia politics who believes the poll is credible.” He further stated that, “the last two independent polls I’ve seen show Gov. Kemp leading Jones 74-17 percent and Kemp up 64-14 percent.” There is no evidence that these latter polls have been publicly released, however.

To reject the Remington poll in such a manner should not be considered a legitimate response. The Remington Research Group is a national pollster that specializes in interactive voice response systems and are certainly a well-established survey research firm.

According to the FiveThirtyEight statistical website that ranks polling firms, we find the RRG earning a B grade and placed as the 103rd entity of the 482 polling organizations in the FiveThirtyEight universe. This means the pollster lands in the top 22 percent of all tested survey research firms. Therefore, to reject Remington’s latest Georgia findings as not credible is an analysis and comment trail that should be discounted.

Gov. Kemp’s popularity among Georgia Republicans clearly suffered after the election controversy in his state. At the low point, according to the Morning Consult data intelligence firm, his personal favorability rating stood with a 62:31 percent positive to negative ratio among members of his own party, which is touching the danger zone for any office holder.

Continue reading →

A Wisconsin Senate Struggle:
Sen. Ron Johnson Trailing … Again

By Jim Ellis

Wisconsin Sen. Ron Johnson (R)

April 5, 2021 — Another tough statewide campaign appears to be brewing in the Badger State.
Thomas Nelson is the Outagamie County, Wisconsin, executive (Appleton area) and a former Wisconsin state assembly majority leader who is assessing his chances of challenging Sen. Ron Johnson (R) next year. For his part, Sen. Johnson has not indicated whether he will seek a third term. He has hinted both toward running again and keeping the pledge he made during his first campaign in 2010 to serve only two terms.

Nelson just released the results of an internal Change Research poll that posts him to a four-point 48-44 percent lead over Sen. Johnson. Looking back to the senator’s last election campaign (2016), trailing in a survey is nothing new. It was the repetitive data continually pegging him as trailing that led to the National Republican Senatorial Committee abandoning him as a lost cause until the late polls showed him rebounding with at least a chance to win.

In the end, Johnson defeated former Sen. Russ Feingold (D), 50-47 percent, in one of the most surprising results of campaign year 2016.

Wisconsin is likely the least accurately polled state during the past few elections. In the aforementioned Johnson-Feingold race, a total of 74 surveys were publicly released, and 70 of them showed Sen. Johnson trailing. In the same election year, 33 research studies were placed in the public domain for the presidential race and only the Trafalgar Group, just as the campaign was closing, correctly projected Donald Trump with an edge.

In 2020, a total of 70 presidential polls were released and 66 gave Joe Biden the lead. He did carry the state, but only by 20,682 votes translating into a 0.7 percent margin. The final 20 polls, all conducted after Oct. 1, found President Biden carrying a 7.0 percent average advantage, well beyond the polling margin of error in relation to the final result.

Therefore, with a consistent pattern of underestimating Republican strength, seeing Sen. Johnson trailing by only four points in the early stages of the 2022 campaign suggests his actual standing is likely much better.

Continue reading →

New Pollster Ratings

By Jim Ellis

March 30, 2021 — The FiveThirtyEight statistical organization released their latest political pollsters’ accuracy ranking chart toward the end of last week — a study that included 592 survey research firms.

Grades from A+ to C/D were assigned to the firms based upon four criteria:

  1. Their 2020 election cycle precision
  2. The predictive category that suggests how successful the firm will be in the future
  3. The number of polls analyzed in this most recent cyccle, and
  4. hat the FiveThirtyEight team terms as the company’s “mean-reverted bias” factor.

Four polling entities earned the top A+ rating. They are:

  • the Iowa-based Selzer & Company
  • ABC News/Washington Post
  • Siena College/New York Times
  • IDB/TIPP

Another six received A ratings:

  • Survey USA
  • Landmark Communications
  • Research & Polling
  • AtlasIntel
  • Monmouth University
  • Marist College
  • Fox News

Seven more posted A- rankings:

  • The Trafalgar Group
  • Public Opinion Strategies
  • CBS News/New York Times
  • Mason-Dixon Polling & Strategy
  • Public Policy Polling
  • Emerson College
  • Quinnipiac University

On the other end of the spectrum, 18 firms were at the bottom of the list but a dozen from this group conducted less than 10 surveys. Of those featuring more than 10 polls, the lowest ranked of the group were Nielsen Brothers Polling, the Humphrey Institute, FM3 Research, Opinion Research Associates, McLaughlin & Associates, and Brown University.

Selzer & Company has long been recognized as one of the country’s better pollsters based upon its strong record predicting the Iowa Caucus presidential results and other races most often from the Hawkeye State.

In 2020, 12 of 19 published polls projected Iowa Democrat Theresa Greenfield as leading Sen. Joni Ernst (R). Approaching Election Day, only two pollsters posted the incumbent ahead four or more percentage points: Selzer & Company and Insider Advantage. Sen. Ernst’s final victory margin was 6.6 percentage points. The IA firm came closest to the end result (Ernst, plus-6; actual, 6.6 percent). Surveying for the Des Moines Register newspaper, Selzer & Company again landed within the accuracy realm (Ernst plus-4).

The ratings also included the polling firms’ partisan bias factor. The bias swing in favor of one party or the other ranged from 4.7 points toward the Democrats (Survey Monkey) to a 3.1 sway for Republicans (MRG Research). The overwhelming majority of pollsters with the highest bias rating favored the Democrats. Of the 28 firms that registered a 2.0 or greater partisan bias factor, 26 favored Democratic candidates.

Continue reading →

Newsom Recall Election Likely

By Jim Ellis

California Gov. Gavin Newsom (D)

Feb. 15, 2021 — Proponents of the recall effort against California Gov. Gavin Newsom (D), with more than a month remaining until the March 17 deadline to collect the necessary number of valid signatures, are already close to forcing a removal election.

Five other attempts have been made to recall Gov. Newsom, but this is the first that had a serious prospect of qualifying. According to the latest poll of the state’s electorate, dissatisfaction with both the governor and the state government’s response to the COVID-19 pandemic has led to the uprising.

Under California election law, state officials can be forced to stand for a recall election if a number of registered voters equal to 12 percent of the total number of votes cast in the most recent governor’s election, in this case the 2018 contest, sign a state certified petition. The total number of qualified signatures needed for the 2021 recall must equal a minimum of 1,495,709. Yesterday, the proponents reported that they have collected 1.47 million signatures. To ensure qualification, the organizing committee members have a goal of obtaining over 1.8 million signatures.

CALIFORNIA GOV. NEWSOME RECALL EFFORT:

 • Signatures Needed • Signatures Collected
 1,495,709 1,470,000
 • Signatures Reviewed • Signatures Validated
 485,650 410,087
 • Signature Approval Rate • Total Signatures Needed
 84.4% 1,772,169

At this point, the Secretary of State’s staff has reviewed 485,650 of the submitted signatures according to the FiveThirtyEight statistical website, and 410,087 have been ruled valid. This translates into an approval rate of 84.4 percent. If this ratio were to continue, the proponents would have to submit a minimum of 1,772,169 signatures. At their most recent reported gathering rate of over 100,000 signatures per week, they should easily reach their quota.

The University of California at Berkeley’s Institute of Government Studies released their latest statewide survey (Jan. 23-29; 10,357 registered voters from online stratified random samples) and compared it with the results from their September 2020 poll. The results show a significant deterioration in the governor’s support base in the short period between the two research studies.

Continue reading →

Graham Getting Close

(Former South Carolina state Democratic Party chairman, Jaime Harrison’s negative ad on opponent Sen. Lindsey Graham (R))

By Jim Ellis

March 27, 2020 — A new statewide South Carolina political survey finds Sen. Lindsey Graham (R) falling into a surprisingly close race with former state Democratic Party chairman Jaime Harrison. According to the Brilliant Corners Research and Strategies firm (March 3-11; 804 likely South Carolina general election voters), Sen. Graham’s lead over Harrison has dwindled to just 47-44 percent.

Several points about this survey. First, the Brilliant Corners operation is a relatively new firm, one that we are only seeing for the first time, and the entity doesn’t even have a reliability rating from the FiveThirtyEight statistical operation that evaluates all polling firms. Therefore, little exists with which to compare their work. Second, this poll is not in line with other published data and, third, the survey was already two weeks old before publishing.

The other two surveys released this year both give Sen. Graham double digit polling margins. East Carolina University went into the field in early February (Jan. 31-Feb. 2; 1,756 registered South Carolina voters) and projected the senator to be holding a 51-38 percent lead over Harrison. NBC News/Marist College followed a couple of weeks later (Feb. 18-21; 2,382 registered South Carolina voters) and found a similar ballot test result, 54-37 percent, in the senator’s favor.

On the other hand, the national Democrats believe this is a sleeper race for them, and based upon Harrison’s prolific fundraising, they might be right. At year’s end, Harrison had raised an impressive $7.6 million with no candidate investment. That amount cost a fair amount to raise, however, as he was left with $4.7 million in his account to begin the election year.

This is a far cry from the gross amount obtained, but is still more than enough to already run a credible campaign effort in a smaller state with inexpensive media markets. Accounting for likely outside spending and considerably more that Harrison will raise, the South Carolina Senate campaign could become an “A” level race in terms of competitive effort.

Continue reading →