Real Clear Politics is an aggregation of many of the election polls, whether present or past. They do offer a link to each poll that they show the results from, but you have to dig a little deeper to determine:
- who was asking the questions
- who was answering the questions
- what questions were being asked
Also, it’s important to note that the candidates are not necessarily listed in order of voter approval.
For example, take this image that often shows Elizabeth Warren listed under Joe Biden, when Bernie Sanders clearly is outperforming.
The first poll listed, the Emerson Poll, when you click on the link says:
A new Emerson Poll finds a very competitive primary shaping up for Super Tuesday 2020 in Colorado with Sen. Bernie Sanders at 26%, former VP Joe Biden at 25% and Sen. Elizabeth Warren at 20%. Sen. Kamala Harris was the only other candidate to break double digits at 13%. Colorado Senator Michael Bennet is polling at 1%. The data was collected August 16-19, n=403, +/-4.8%.
But Real Clear Politics have Bernie Sanders as third on the list in the Colorado poll. If you click to read the whole poll you will see that Bernie Sanders is either tied or beating Trump and Biden.
You can also see that the demographics of the poll, listed at the very end are:
The Colorado Emerson College poll was conducted August 16-19, 2019 under the Supervision of Emerson College Polling Director and Assistant Professor Spencer Kimball. The sample consisted of registered voters, n=1,000, with a Credibility Interval (CI) similar to a poll’s margin of error (MOE) of +/- 3 percentage points. The data was weighted by a 2016 voter model on race, age, gender, party affiliation, education and US congressional district. It is important to remember that subsets based on gender, age, party breakdown, ethnicity and region carry with them higher margins of error, as the sample size is reduced. Data was collected using both an Interactive Voice Response (IVR) system of landlines only (n=656) and an online panel provided by Amazon MTurk and Dynata (n=344).
In other words, most of the people polled were on a landline responding to a computer. The other 344 were through an online panel provided by a Jeff Bezos company, Amazon. Just imagine what the results would be like if they were a realistic sampling of the voting public.
The second poll listed from Gravis Marketing, described their methodology and respondents as:
Gravis Marketing, a nonpartisan research firm, conducted a random survey of 926 registered voters in Nevada (382 Democratic caucus voters). The poll was conducted from August 14th through the 16th and has a margin of error of 3.2% (5.0% for the Democratic caucus question). The survey was conducted using interactive voice responses and an online panel of cell phone users. The results are weighted by the voting demographics. The poll was paid for by GMI.
Notice how they don’t break down how many were landlines and how many were cell phone users of an online panel.
The third poll listed is Fox News and they describe their methodology and respondents as:
Interviews were conducted August 11-13, 2019 among a random national sample of 1,013 registered voters (RV). Landline (222) and cellphone (791) telephone numbers were randomly selected for inclusion in the survey using a probability proportionate to size method, which means phone numbers for each state are proportional to the number of voters in each state.
Results based on the full sample have a margin of sampling error of ± 3 percentage points.
The Fox News Poll is conducted under the joint direction of Beacon Research (D) (formerly known as Anderson Robbins Research) and Shaw & Company Research (R).
Note: In 2019, the firm Anderson Robbins Research changed its name to Beacon Research; the Fox News bipartisan polling team remains unchanged.
PERSONAL NOTE: I could not find a website for Shaw & Company Research.
The Fox News poll that is linked devotes a great deal of time and questions as to the job approval rating of Donald Trump and how people feel about guns and mass shootings. Fox News is one of the more thorough in the description of their audience and their questioning.
To get to how Donald Trump compares to other candidates, one must search. One must also pay attention to how the questions are worded.
This doesn’t have to do with the Democratic Candidates directly but I thought it was an interesting result for one of the questions in their poll:
The fourth poll, from SurveyUSA has a broader sampling of participants–over five times as many people as the previous polls. Their methodology and respondents are as follows:
In a hypothetical nationwide general election today — which approximates the outcome of the USA popular vote but not necessarily the outcome of the Electoral College — Democrats Joe Biden and Bernie Sanders defeat by 8 points Incumbent Republican President Donald Trump, according to research conducted by SurveyUSA. Trump runs effectively even against Democrats Elizabeth Warren, Kamala Harris and Pete Buttigieg.
At this hour, in interviews with 5,459 registered voters nationwide, it’s:
* Former Vice President Biden 50%, President Trump 42%.
* Vermont Senator Bernie Sanders 50%, Trump 42%.
* Massachusetts Senator Elizabeth Warren 46%, Trump 44% — too close to call.
* California Senator Kamala Harris 45%, Trump 44% — too close to call.
* Trump 44%, South Bend IN Mayor Pete Buttigieg 42% — too close to call.
About: SurveyUSA interviewed 7,000 adults nationwide 08/01/19 through 08/05/19. Of the adults, 5,459 are registered to vote and were asked the hypothetical head-to-head pairings you see here. The research was conducted online.
The fifth poll shown was SurveyUSA in North Carolina. There is certain criteria, explained in detail in the linked survey, to be included. Some of that reads:
A total of 2,600 NC adults were interviewed online 08/01/19 through 08/05/19. Of the adults, 2,113 are registered to vote in NC. Of the registered, 534 passed all 5 SurveyUSA screening criteria and were counted as a “likely” NC Democratic primary voter and asked, “Which Democrat would you vote today?” All NC registered voters were asked the hypothetical, general-election Head-to-Head match-up questions. All registered voters, as indicated, were asked the questions about healthcare, patriotism and racism.
During the field period for this survey, 31 individuals were murdered in 2 separate mass shootings. SurveyUSA’s best estimate is that approximately 72% of the interviews for this survey were completed before news broke midday 08/03/19 of the 1st mass shooting in El Paso TX and approximately 78% of the interviews for this survey were completed before Americans woke up Sunday 08/04/19 to news of the 2nd mass shooting in Dayton OH. In 2020, North Carolina’s presidential primary is 7 months from today, 03/03/20. 110 convention delegates will be awarded proportionately, so the relative strength of those candidates finishing in 2nd, 3rd or 4th place is significant.
You can see that each poll is reaching different people and that Biden does best when the respondents are older people answering their landlines. As of 2017 only 47.5% of American households had landlines. I’m sure it’s far fewer today.
I hope this information was helpful to you in understanding that even polls can be biased based not only on who is asking the questions, how the questions are asked, and who the respondents are.
If something makes you question it, dig deeper.