Tag Archives: coronavirus

Beneath the Numbers

By Jim Ellis

Is Maine Sen. Susan Collins (R) in trouble?

July 30, 2020 — Two Senate polls were released earlier this week, and though the ballot tests in Maine and South Carolina didn’t show us anything particularly new, the pollsters asked their respective sampling universes some interesting ancillary questions.

Colby College, a Waterville, Maine institution housing approximately 1,800 students, released its second political poll of 2020 in conjunction with the McVey Data Science Initiative. The questions touched upon the presidential and Senate ballot tests but delved deeper into the attitudes and perspectives of their 888-person sample over the July 18-24 period.

The presidential ballot test found former vice president Joe Biden leading President Trump statewide, 50-38 percent. The Senate ballot test was closer in that former state House speaker Sara Gideon (D) leads Sen. Susan Collins (R), 44-39 percent.

In looking deeper, it is probable that the actual race standings are likely a bit closer. There appears to be a Democratic skew in that all favorability indexes for Democratic public officials were positive and those of the tested Republican officials were decidedly negative. Additionally, 52 percent of the sampling universe comes from the state’s southern congressional district, the 1st, which is decidedly more liberal than the northern 2nd District. Together, these factors provide us the indication that the results are a few points more favorable for Democrats than what we might see in actual voting.

Continue reading

Coronavirus Polling Numbers

By Jim Ellis

COVID-19 virus

June 30, 2020 — The Pew Research Center yesterday released the results of their national poll about how the public is viewing the COVID-19 response, which enables us to put the data in a political context. The polling results contain some good news for both presidential candidates and the respective major party leaders who are attempting to craft national campaign agendas in unique times.

According to the Pew methodology report on page six of their synopsis, the survey was conducted from June 4-10 via “the American Trends Panel (ATP), as created by the Pew Research Center, [which] is a nationally representative panel of randomly selected U.S. adults. Panelists participate via self-administered web surveys.” The ATP has a total of 19,718 adults of which 11,013 were sampled for this poll and 9,654 responded.

The sampling error is reported to be plus or minus 1.6 percentage points, but Asians (8.2), Blacks (5.3), and Hispanics (4.5) were well over the average. While the pollsters show all segments falling between a plus or minus 1.8 and 8.2 error factor, they still list the overall sample rate (1.6 percent) as falling below even the low number on racial segmentation.

The best news for former vice president Joe Biden is that the Trump Administration scores the lowest rating relating to whom and what the respondents trust most about coronavirus information. The administration is believed either almost all (eight percent), most (21 percent), or some of the time (29 percent) by a combined 58 percent of the respondents. In contrast, the Center for Disease Control is the most reliable cited source with a combined 88 percent rating (22 percent almost all; 42 percent most; 24 percent some of the time).

With President Trump and his team scoring low on the believability scale, the better news for his campaign is that fewer people are following the disease coverage closely. Furthermore, it is clear that large segments don’t know what to believe from news accounts of the disease’s effects.
Continue reading

Is Biden Under-Performing?

By Jim Ellis

Former vice president and current 2020 presidential candidate, Joe Biden (D)

May 28, 2020 — Currently, former vice president Joe Biden is leading in virtually every battleground state poll that has been released into the public domain; but do the survey numbers tell the whole story? Other available data, that derived from actual votes being cast, suggest there may be cracks forming in his political armor.

In terms of battleground states, a new Arizona poll was released Tuesday, in what could become the most important must-win domain for the Trump campaign. There, the Phoenix-based HighGround Public Affairs consulting firm publicized their latest Arizona statewide poll (May 18-22; 400 likely Arizona general election voters), and the ballot test projects Biden holding a tighter 47.0 – 45.3 percent edge over President Trump.

Though Biden leads, the trend is a clear improvement for Trump since the OH Predictive Insights poll (May 9-11; 600 likely Arizona voters) became public during mid-May and posted the former vice president to a 50-43 percent advantage. Redfield & Wilton Strategies followed with their survey release (May 10-14; 946 likely Arizona voters) that yielded a closer 45-41 percent Biden lead.

Tuesday’s HighGround survey also published some interesting secondary questions. In answering whether the respondents felt President Trump job performance was excellent, very good, or OK, the combined positive calculated to 49.5 percent. Among those believing the president’s job performance is poor or failing, the negative totaled to 49.0 percent.

Accordingly, 29.3 percent of the Arizona respondent sample believes President Trump is to blame for the spread of the coronavirus, while 20.3 percent pin the responsibility on China. All other responses: the people, federal government, Democrats, or other unnamed sources each registered less than 10 percentage points. A total of 25.3 percent say they either don’t know or believe no single source is to blame.

Over the Memorial Day weekend, the state of Hawaii hosted its Democratic presidential primary. The unique format is interesting in that there were two votes the participants cast: the first included choosing among all of the names who originally qualified for the Hawaii presidential nomination ballot. The second isolated only Biden and Sen. Bernie Sanders (I-VT). Just under 35,000 people cast mail votes in the stand-alone presidential primary.

Continue reading

Super June & Oregon Primary Results

By Jim Ellis

May 21, 2020 — With so many early primary states moving their elections to June due to the coronavirus shut down, no less than 24 states will hold their nominating event in the upcoming month, making this the most active primary month during the election cycle.

A dozen of the states are still observing their regular political calendar, but 12 more moved into June from earlier dates. Therefore, the following adjusted calendar has been locked into place:


JUNE 2

Connecticut (presidential only)
Idaho (from May 19; though mail voting began on the original primary day)
Indiana (from May 5)
Iowa
Maryland (from April 28)
Montana
New Mexico
Pennsylvania (from April 28)
Rhode Island (presidential only)
South Dakota
West Virginia (from May 12)


JUNE 9

Georgia (from May 19)
Nevada
North Dakota
South Carolina


JUNE 20

Louisiana (presidential only)


Continue reading

Montana Shock Poll

By Jim Ellis

(Left) Montana Sen. Steve Daines (R); Gov. Steve Bullock (D)

May 7, 2020 — Montana State University at Bozeman’s research arm just completed a survey of their state’s electorate (April 10-27; 738 Montana adults, 458 likely Montana voters) and produced a surprising tally in the Senate race.

According to MSU Bozeman, Gov. Steve Bullock (D) has jumped out to a 46-39 percent lead over first-term Sen. Steve Daines (R) even with President Trump posting a 5.6 percentage point advantage over former vice president Joe Biden (46.3 – 39.7 percent) within the same sampling group.

Though Montana is viewed as a Republican state, and it generally performs as such in most federal races, the margins usually aren’t particularly lopsided, and Democrats have done well in statewide contests up until 2016. Except for Gov. Bullock’s 50-46 percent re-election victory that year, Republicans running in the wake of Trump’s 20-point landslide win over Hillary Clinton, swept the other races.

Several notes about this poll: first, the questioning period lasted 18 days, a very long time for a likely voter sample size of 458 individuals. Typically, such surveys are conducted over a three-day period. Such an implementation interval substantially increases the error rate.

Second, though the error factor is stated as 4.6 percent, the chairman of the university’s political science department, Dr. David Parker, stated in a local Helena KTVH television news story, that the Senate race is within the margin of error and in reality too close to call. While his conclusion may well be accurate, the ballot test shows a margin between Bullock and Daines of seven percentage points, meaning that the result is well beyond the polling margin of error. Therefore, Dr. Parker’s comments suggest the methodology actually yields an error factor larger than stated, which is more consistent with the elongated sampling time feature.

Continue reading