Welcome back to ‘Polls In One Place’, a (not always) weekly newsletter compiled by Tom Smith, head of political research, and yours truly. The idea is to gather up the previous week’s most interesting and/or important survey research and bring it to your attention.
With regard to survey research in general, we are big believers that the margin of error always applies. The margin of error on local, state and national polls is 3%-4% or more.
There is no such thing as a 1-point or 2-point or 3-point lead. Each "lead” is a statistical tie. So: If Jack “leads” Jill 45%-to-43%, Jack does not “lead” Jill by 2 percentage points. Jack and Jill are statistically tied.
The margin of error “problem” was much in evidence in 2020. Joe Biden defeated Donald Trump by 4 percentage points in the national popular vote. Here are the national polls from the final seven days of the campaign, courtesy of Real Clear Politics:
Looking at the Biden “average” number, it was accurate to within a tenth of a percentage point (Biden received 51.3% of the national popular vote). The Trump number was ~3 percentage points low. But the spreads on the majority of the “final” polls of 2020 were way off. The last Economist/YouGov poll, for example, was uber-misleading (Biden +10) and it’s one of the better “news media” polls.
That’s the margin of error at work. If you don’t take it under consideration every time, you can end up believing things that simply aren’t true.
Aside from the “faux specificity” problem (Jack has a 2-point “lead” over Jill), there’s another (related) problem in survey research known as the “outlier.” The “outlier” is a poll that produces data that are obviously inaccurate. A recent example of this appeared in the pages of The New York Times.
Vice President Kamala Harris leads former President Donald J. Trump in three crucial battleground states, according to new surveys by The New York Times and Siena College, the latest indication of a dramatic reversal in standing for Democrats after President Biden’s departure from the presidential race remade it.
Ms. Harris is ahead of Mr. Trump by four percentage points in Wisconsin, Pennsylvania and Michigan, 50 percent to 46 percent among likely voters in each state. The surveys were conducted from Aug. 5 to 9.
The polls, some of the first high-quality surveys in those states since Mr. Biden announced he would no longer run for re-election, come after nearly a year of surveys that showed either a tied contest or a slight lead for Mr. Trump over Mr. Biden.
The polls are “some of the first high-quality surveys in those states since Mr. Biden announced he would no longer run for re-election.”
Not in every case. Dig a little deeper into the Pennsylvania survey and you find that Sen. Bob Casey (D) leads his challenger, David McCormick (R), by 14 percentage points.
There is not a single person who knows anything about Pennsylvania politics who thinks that Bob Casey leads David McCormick by 14 percentage points. Not one.
In 2022, Dr. Mahmet Oz (R), as dreadful a general election candidate for US Senate as one can imagine, got 46% of the statewide vote after every last vote had been counted. The notion that David McCormick is running 9 points behind Dr. Oz’s 2022 performance is preposterous.
But there it is…..Outlier presented as fact. The lesson here is that even “high-quality” surveys can be outliers. It’s the nature of the survey research beast.
P.S.: Two of the New York Times/Siena College statewide polls (Michigan and Wisconsin) align with other survey research being done in the state. All agree: The presidential race is close, the Senate races are close, the Democrats might have a slight edge.
Thus ends the Sunday, now Monday, sermon. On to the week’s polls we go:
1. “More” Americans trust Kamala Harris to handle the US economy than Donald Trump, according to a new poll that marks a sharp change in voter sentiment following President Joe Biden’s withdrawal from the White House race. The survey, conducted for the Financial Times and the University of Michigan Ross School of Business, is the first monthly poll to show the Democratic presidential candidate leading Trump on the economy since it began tracking voter sentiment on the issue nearly a year ago. Although 41 per cent of Americans still trust the former president more on economic issues — unchanged from the two previous monthly polls — the survey found 42 per cent of voters believe Harris would be better at handling the economy. That is a 7 percentage point increase compared to Biden’s numbers last month. “The fact that voters were more positive on Harris than on Biden . . . says as much about how badly Biden was doing as it does about how well Harris is doing,” said Erik Gordon, a professor at the university. (Source: ft.com. See preamble on “faux specificity”)
2. On The Other Hand:
The CNBC All-America Economic Survey finds former President Donald Trump holding a commanding lead among voters on key economic issues, but ahead of Vice President Kamala Harris by just a 2-point advantage in the head-to-head race for the White House.
On the economy, by 2-to-1, Americans think they will be better off financially under Trump. That’s largely driven by the 79% of Republicans who believe their economic fortunes will improve if Trump takes the White House. Just 48% of Democrats believe they will be better off if Harris wins, compared with 42% who say it makes no difference. As for independents, 31% think they will be better financially if Trump wins, 10% if Harris wins and 54% don’t think it matters.
The results offer an opportunity for both sides to convince independents they have better economic policies. Independents break for Harris 40-36 in the presidential ballot race, with 20% undecided. Her lead among independents is 4 points lower than Biden’s in the July NBC survey.(Source: cnbc.com)
Keep reading with a 7-day free trial
Subscribe to Political News Items to keep reading this post and get 7 days of free access to the full post archives.