Survey research in the United States is one of the most common methods used by companies and organizations in both the private and public sector to measure public opinion and political attitudes. The news cycle is oftentimes packed with polls and statistics on rather trivial matters such as God’s approval rate as perceived by Americans or the percentage of Americans who prefer dogs rather than cats. Particularly enough, the public’s focus on polling and numbers shifts dramatically during the months prior to a presidential election.
The main purpose of political polling is to understand the voting behavior of the nation as a whole by asking only a smaller representative sample.
In order to do that, those in charge of conducting research polls (pollsters) need to draw a random sample and modify it to closely match the characteristics of the entire population. The most commonly used method for achieving a random sample is through random digit dialing (RDD). With this method, pollsters have access to a database of all telephone numbers in the country; however it is not only about randomly calling numbers but also choosing random respondents within the household. It is common in some cases, that pollsters will request to speak to the member of the house with the most recent birthday, only with the purpose of randomizing age categories.
In the last decade, the rise of online surveys has opened up new and accessible platforms for research design and sampling. This new online approach recruits survey participants from an inclusive and representative pool such as phone or mail and brings them to a panel where they participate in multiple surveys over time, this is often referred to as probability-based online panels. Nevertheless online surveys oftentimes fail to include for non-internet users and non-english speakers, making accurate representation more challenging.
However, finding a representative sample is not the only issue that accurate polling faces. It is not only about asking people who they are going to vote for next election, but the way and the order in which questions are framed plays a key role in the polling results. Priming1 and framing2 have a particularly strong effect on those who don’t strongly identify with a certain party and haven’t given much thought to the campaign, in other words a swing voter. It is more likely for a flawed survey to sway the opinion of a swing voter, rather than a core voter who has already made up their mind on which party they are voting for regardless of the candidate; therefore misrepresenting the public’s opinion but rather influencing it.
Looking ahead to the 2020 presidential election, one can’t help but question polling credibility given the perception that in the 2016 election polling was inaccurate and misleading. For example, in 2016 college grads favored Clinton while high school grads broke for Trump; in reality state pollsters failed to adjust or weight their sample by education to reflect that college graduates are more likely to take surveys than adults with less higher education.
On the issue of misrepresentation, those states that later turned out to be key to Trump’s victory had few, if any, polls in the last few days before the election. In hindsight evidence suggests that in the last weeks leading to the election, more swing voters chose to vote for Trump than Clinton, but given that polling had stopped they weren’t able to detect such trends.
But what are polls saying about the 2020 election?
Currently the United States is simultaneously facing a pandemic, an economic recession and growing protests about police violence and racial justice, surely enough anger and fear are widespread. Polling conducted by the Pew Research Center suggests that the percentage of people who are satisfied with the current state of the nation dropped from 31% in April to 12% by June 30th.
In the run-up to the presidential election, Republican incumbent Trump trails behind Democrat candidate Joe Biden with a 10% point deficit. According to the Pew Research center, 54% of registered voters said that if the election were to be held today, they would support Biden or are leaning towards voting for him, while 44% support Trump or lean towards supporting him.
There are multiple factors that influence the numbers national polls are throwing; how voters perceive Trump’s presidency compared to how Biden’s potential presidency would look like, candidate’s personal traits and characteristics and the confidence each candidate engenders in handling most major issues. The latter being one of the most prominent ones given the current state of the nation; 31% of voters are very or somewhat confident that president Trump can handle major issues and unify the country and 68% are not too or not confident at all, whereas 45% of the voters say they are at least somewhat confident in Biden while 55% are not too or not confident at all.
When it comes to demographics, only 8% of white voters say they would vote for Trump (53%) than Biden (45%) whereas black voters favor Biden over Trump by an overwhelming margin of 89% to 7% and hispanic voters favor Biden as well with 66% compared to 32% who back Trump.
Age categories have become a key factor as well. Voters between the ages of 18 and 29 favor Biden over Trump by 68% but out of those, 84% expressed that their vote for Biden is mostly engendered by opposition to Trump while 16% say it is more of a vote for Biden. Nevertheless, Biden also has a wide advantage among voters between the ages of 30 to 49 with 60% and a much lower rate with voters among the ages of 50 and older with a 47%.
During the following months, we should expect to see growing and ever-changing numbers. As has been the case throughout his presidency, Trump has managed to create strong partisan loyalty and intense opposition, however the latest events of 2020 might alter his path towards reelection. And although polls attract the public’s attention with the promise of predicting the outcome of elections, their most important function is to aid journalists and citizens to break down how the presidential campaigns are being perceived by the public. In other words, what issues are important to the nation and how much support there is for particular policy changes. Results and numbers should be digested with a level of discretion and having in mind that no number is truly set in stone.
1Priming: phenomenon whereby exposure to one stimulus influences a response to a subsequent stimulus, without conscious guidance or intention.
2 Framing: idea that meaning is shaped by how issues are presented.
For more information and updated numbers visit: