The results are in…
As election day draws near, there is 1 question on EVERYONE’S mind… who’s gonna win? If you look online you’re sure to find a slew of predictions and poll research that all basically contradicts each other. Before we give you the results of Group Chat’s research we figured hey! why not break down what polls actually mean.
Are these things even right?
If you’re like me you probably thought Hilary had the ’16 election locked. BOY WERE WE WRONG! Not only that…guess what guys, the error in national polls were historically low in 2016. But how were all the polls off and how did America end up having an orange reality TV host as a president?
A fun little thing those in “the biz” (read: pollsters) are calling the “shy Trump voter”. Here’s how that cookie crumbles:
Polls predicted Hillary would win and they were only off by 2 points! Hillary won the popular election by 3 points in 2016 but because of the electoral college Trump wins. That’s not the whole story though. There’s 2 fun little bumps in the road that came up:
- Most polls didn’t account for social-desirability bias, or more commonly known as the “shy Trump voter”. Basically, people vote with their hearts and that’s what ends up on the ballots but fear of being ostracized or overly criticized for their political affiliations make them say they’ll vote one way but then the switch is flipped come ballot time.
- The last few days are what really counts. Trump’s victor in 2016 was in part due to the independent voters he was able to sway in the few days leading up to the election, something these polls don’t take into consideration as they typically wrap up the week prior.
How are polls documented?
Originally, polling started as mail-in or in-person. With the introduction of new technology, the practice has evolved over time and phone call polling became popular for a long time. Now, pollsters are looking to the internet with opt-in polls but with that comes some challenges. Like the inability to source large random samples, since users have to opt-in which may skew results. Or the fact there are still populations of people who are largely off-line who behave and vote differently than the majority who are online. The challenge is to find a hybrid which is randomly selected and reflects all of the population – right now that hybrid is online and phones.
How can you tell a good poll from a bad one?
Based on Pew Research
, good polls….
✅are those that are paid for and fielded by a neutral source;
✅have selected a probability-based, random sample of the public (or the population of interest, such as registered voters);
✅dial cellphones in addition to landlines;
✅make multiple attempts to reach people;
✅use live interviewers;
✅and make public both the questionnaire and a detailed methodology.
How are participants chosen?
It is supposedly completely random. Numbers are selected through random digit dialing. Here’s the thing though… have you ever met someone who was ~actually~ polled? No?
Well then, we have some GREAT news for you. One of our Kathys, Dustin Godsey gave us an inside look to being polled. Here’s our *exclusive* interview:
- How did you come about being polled? Was it volunteer or randomly selected? I’ve been polled five times during this election season, all of them random phone calls that I happened to answer on our landline (yes, we still have a landline which never gets used…except during election season when it rings non-stop). Who knows how many additional polling calls I’ve gotten that I didn’t answer.
- Was the process in-person or digital? Can you tell us more about the format of the poll? All of them were in-person (via phone), with pretty standard polling/survey formats. I was quickly screened out of two of the five I’ve answered, due to samples being full in either my demographic or education groups. Two of the remaining three were presidential preference polls – asking questions about whether the country is on the right or wrong track, party preference, my approval of Trump’s job performance and who I planned to vote for in the election. The latest poll also included questions about the local state assembly race, which led me to believe this was likely an internal poll for one of those campaigns, as I don’t usually see public polling for local races.
- Do you think the questions were posed in an un-biased and neutral way or potentially swayed in one direction or the other? All of the polls I answered were definitely scientific, non-biased polls – none of them delved into issues or seemed to attempt to sway opinion one way or another. I’ve definitely gotten calls for push polls in the past that would reword questions to attempt to either get you to answer one way or another or, likely, to measure efficacy of messaging, but the polls this year have all been pretty straightforward election preference polls.
- What are your thoughts on the sample sizes of polling to derive large generalizations from that sample size? I think polling is one of the most misunderstood subjects; everyone pays so much attention to them but really have no understanding of what they mean, or how they’re derived. Given how data driven our society has become, I actually think statistics should be a part of core math curriculum in high school. I took a ton of statistics classes and deal with a lot of market research, so I understand the limitations of sample size, but also the diminishing returns involved in expanding sample sizes beyond a certain point. I think the most important thing for people to remember is what a poll is supposed to be – it’s a snapshot of a sample of people at a specific time. Polls are never going to do anything other than provide insight or direction, using math to extrapolate out a best guess. Individual polls can swing fairly widely (and should, just based on statistics), but when you get multiple polls or the same poll repeated over time, you can identify trends. I think the average error on presidential polls over the last 50 years is something like 4 or 5 points…which can mean the wrong person can be shown to be in the lead in a close election, but it is actually pretty impressive. Imagine guessing, within 4-5 points who would win based simply on how many yard signs you see or what is being posted on Facebook! The country is so big, and we have sorted ourselves into bubbles that align with our views that it is easy to be completely blind as to what else is happening. Polls at least give us some measure of the whole.
- What region of the US are you located in? Do you think you were polled because of your geographic location? Or chosen for other factors? I live in Wisconsin, one of the top battleground states, so I definitely think there is a ton more surveying going on here than there are a lot of places. I’ve lived in a battleground or purple state for every election since I’ve been old enough to vote (Iowa, Pennsylvania, and Wisconsin) so this feels totally normal for me.
So, who’s winning this election?
Just like 2016, it’s a tight race.
Many polling aggregators are saying Biden will win, some in the talks of 7+ points, but here’s the thing… these aggregators are only as good as the original sources they’re pulling from. By now you’re probably asking, how accurate are those original sources? According to the Washington Post, not many pollsters are accurate. Again, tying back to the “Shy Trump voter”.
From the 2016 election, the 2 most accurate pollsters included IBD/TIPP Tracking and McClatchy/Marist, a close 3rd being the Trafalgar Group. IBD/TIPP recently reported that Biden’s camp hit a low last week, to bounce back slightly up to 49.2%, with Trump seeing a high at 44.1%… is this a repeat of 2016? We’ll have to wait to find out. In either case, IBD/TIPP, Marist and Trafalgar are showing a much different story than most of the other pollsters.
Although these polls could show the popular vote accurately, it all comes down to the electoral college and The Swing StatesTM . The whole world is watching these 8 states:
Group Chat’s got you covered.
Based on an aggregate of data from the top 3 pollsters we predict the winner will be…..BIDEN! But honestly, it’s anyone’s game at this point.
Need more info? Check out Project FiveThirtyEight they have a pretty fun and interactive site!
And in case you were wondering… yes, we did do our own poll.
We entrusted a third party to do a randomly selected poll following ALL of the best practices to find out who is the #1 podcast in the world. The results are in and might we say…we’re not surprised.