The 2016 presidential election showcased the ever-growing and increasingly hostile partisan divide within The United States of America. The highly publicized battle between Donald Trump and Hillary Clinton catapulted controversial issues like immigration, healthcare and gun reform into the forefront. Aided by the recent explosion of social media, Donald Trump infiltrated an untouched population of lower- and middle-class white Americans who were feeling the brunt of an increasingly globalized environment and a democratic government that were not pertaining to their needs. However, in the weeks and months prior to the election, opinion polls historically failed to quantify Trump’s support. In the days after Trump’s presidential success, many people were left stunned by the inaccurate pictures painted in public opinion polls (Enns et al., p.42). Without reference and knowledge to their properties in question, samples sizes, methodology and potential for bias, it can be easy to draw false inferences. Critically analyzing 2016 opinion polls is imperative in understanding the collective inductive weakness of their findings and the misguidedness that they inevitably created.
In the month prior to the election, Pew Research conducted a series of opinion polls to gauge support for the two running competitors. In one survey conducted over a five-day period in late October, Pew asked how much respect Donald Trump and Hillary Clinton respectively had for America’s democratic institutions. By using a scale from ‘none at all’ to ‘a great deal’, Pew concluded that 63% of Americans agreed that Clinton had some to a great deal of respect. Trump scored lower with only 43% (Doherty et al., p.1). Pew surveyed 2,583 adults over the age of 18 with 2,120 of the participants having likely already registered to vote. Their large sample left a ±2% margin of error with a confidence level of 95% (Doherty et al., p.76). As the margin of error was so small, Pew had little risk of hastiness in their findings. The sample surveyed were from all 50 states in the USA and Pew made sure that the survey was accessible for both English and Spanish speakers. To make sure their sample was representative of the American public, Pew went to great lengths to include a wide variety of different people. By using a mathematical procedure, attributes like race, religion and educational attainment were weighted to match the “2014 Census Bureau’s American Community Survey and population density to parameters from the Decennial Census” (Doherty et al., p.76). By implementing this weighting technique, Pew was able to use their findings to conclude that they had used the principle of total evidence. In Pew’s opinion, their survey could be used to generalize to their target population of 18+ citizens in The United States. However, Pew’s method of attaining their information was not so conclusive. Pew surveyed by telephone. They reached 647 participants by landline and 1,936 participants by cell phone. When a call was answered, the pollster would ask for the youngest member of the household who was over 18 and then ask them for their opinion on how much respect Trump and Hillary had for America’s democratic institutions. On reporting their findings, Pew only used information from participants that did have an opinion. They failed to mention how many people chose not to answer in their poll so the findings are inaccurate because there is no percent to represent the non-answers. Even though Pew tried to gather a representative group of people to engage in their survey, their sample was atypical. Not only was the census data used to calculate their representative sample two years old, but their survey question was extremely vague and difficult to answer. For the average person, being asked ‘how much respect do you think Donald Trump has for America’s democratic institutions’ gives no visceral reaction. More likely, questions regarding what America’s democratic institutions are and what defines respect enter the mind. The question is not relatable to people that do not follow politics or have a grasp on democratic institutions. It can also have the unintended consequence of ostracizing people with lower levels of education because the answer is incorrect or inconsistent with their actual political beliefs. Furthermore, the important variable of youth is disproportionately present in Pew’s methodology. Results from the 2016 general census estimate the mean age of US households to be 48.6 (US Census Bureau, 2016). These statistics suggest that Pew’s sample may be favoring younger people over the older population. Pew infers that this method of finding survey participants “improves participation among young people, who are often more difficult to interview than older people because of their lifestyles” (Pew Sampling).
Even though Pew may have gathered information regarding a hard-to-reach section of the population, the results from their survey were generalized to the entire population of The United States and without factoring in people of all ages, they may have failed to represent the true proportion of nationwide beliefs. Pew’s inductive argument regarding how much respect Trump and Clinton respectively have for America’s democratic institutions can be rated a 4/10. Pew tried hard to gather information for their survey in an unbiased and inclusive way but failed to create a truly representative sample. Their survey question was too vague and could not be objectively analyzed because of the different personal interpretations of the meaning of respect and/or what America’s democratic institutions are.
Similar to Pew’s research poll, a few weeks prior to the 2016 election, Bay News 9 conducted a popularity poll asking Florida locals, who would you vote for? Political reporter Stuart Rothenberg calls Florida, “a ‘neutral’ landscape …and because they perform at or near the national margin, they give a good indication of the partisan direction of the cycle” (p.1). Florida is one of thirteen swing states in The United States and their pre-election opinion polls can paint a reality of election outcomes. With the whole country watching Florida’s every move leading up to the voting day, Bay News 9’s poll was of great importance. Bay News 9 conducted their survey with SurveyUSA and offered six options for the ‘who would you vote for’ question including Donald Trump, Hillary Clinton, Jill Stein and still undecided. In a tight race, Clinton led by 48% with Trump trailing behind at 45%. 1,400 Florida adults were surveyed including 1,314 likely already registered voters and 1,251 very likely determined to vote voters (News 13 Florida). Contrary to Pew, SurveyUSA asked participants questions about age, race, education etc. in conjunction with their election questions. They also asked questions about the participants voting habits and found that 1% of previously voting adults would not be voting in the 2016 election because of their dislike for both Trump and Hillary (News 13 Florida). Bay News 9’s survey sample left a ±3% margin of error with a confidence level of 95%. As the margin of error was relatively small, Bay News 9 had a small risk of hastiness in their findings. Bay News 9’s sample featured a wide cross-section of the Florida community. 28% of survey participants were between the ages of 18-34 and over 50% of participants were of retirement age, closely mirroring Florida’s age demographic. They also asked questions regarding party affiliation, religion and personal gun ownership (News 13 Florida). The methodology of the survey was simple yet effective. They implemented a mixed method mode to make sure they reached people from all walks of life. 65% of the surveys were completed via cell phone or landline but SurveyUSA also included options for participants to complete the survey through a range of electronic devices including laptops, tablets and smart phones. SurveyUSA focused primarily on registered voters and then weighted the information to be accurately representative of the Florida population (News 13 Florida).
Bay News 9 and SurveyUSA focused on easy and answerable questions that left little room for interpretation nor were they restrictive in who could understand and clearly answer the questions. Their sample, although smaller than Pew’s, focused on a wide range of different groups that reflected the diversity in Florida’s communities. By focusing on a sample of a specific community, it was easier for Bay News 9 to apply the principle of total evidence and generalize that for Florida as a larger population, in October of 2016 Clinton was the public favorite. Bay News 9 and SurveyUSA’s inductive argument regarding who the people of Florida wanted to vote for can be rated as a 7/10. By using a range of methods to engage with participants, creating easily answerable questions and by not limiting their reach in regards to who they surveyed, Bay News 9 and SurveyUSA created a strong inductive case. In the weeks that followed, support for both Trump and Hillary changed with Trump winning the Florida election with a 49% margin (‘Presidential Election in Florida’, 2016). Like most opinion surveys across The United States, Bay News 9 did not predict this change in the polls. Trump’s success came as a shock to a lot of people throughout the country. Even though the outcome of the 2016 did not reflect the opinion polls, it does not necessarily make polls like Bay News 9 bad or weak. One of the main theories for this phenomenon has been coined ‘shy Trumpers’. This idea implies that prior to the election, many people felt embarrassed to show their support for Trump in opinion polls because he was the ‘socially undesirable’ candidate (Mercer et al). The ‘shy Trumpers’ theory suggests that when people were surveyed, they did not tell the truth about their political beliefs or chose not to engage in opinion polls making it hard for almost all opinion polls throughout The USA to be truly representative and accurate in regards to predicting the outcome of the election.
The 2016 presidential election showcased a very strong divide between the left and the right. However, unlike the many opinion polls suggested, Trump triumphed over a hard-done-by group of Americans and took the presidency. By analyzing the inductive arguments in some popular opinion polls, it is clear that without understanding their individual properties in question, samples sizes, methodology and potential for bias, it can be easy to assume them as strong and accurate. By uncovering their faults, it is evident how the disparity between speculation and reality occurred.
- ‘America’s Families and Living Arrangements: 2016’. United States Census Bureau, 2016, www.census.gov/data/tables/2016/demo/families/cps-2016.html. Accessed 31 May 2019.
- Doherty, Carroll, et al. As Election Nears, Voters Divided Over Democracy & ‘Respect’. Washington DC, Pew Research Center, 27 Oct. 2016.
- Enns, Peter K., Julius Lagodny, and Jonathon P. Schuldt. ‘Understanding the 2016 US Presidential Polls: The Importance of Hidden Trump Supporters’. Statistics, Politics, and Policy, vol. 8, no. 1, 2017, pp. 41-63. ProQuest, https://login.ezp.pasadena.edu/login?url=https://search-proquest-com.ezp.pasadenaedu/docview/2178972373?accountid=28371 , doi:http:/z dx.doi.org.ezp.pasadena.edu/10.1515/spp-2017-0003.
- Mercer, Andrew, et al. ‘Why 2016 Election Polls Missed Their Mark’. pewresearch.com, 9 Nov. 2016, http://www.pewresearch.org/fact-tank/2016/11/09/why-2016-election-polls-missed-their-mark/ Accessed 2 June 2019.
- News 13 Florida. ‘Exclusive Florida Decides Poll: Presidential race to come down to wire’. baynews9, 25 Oct. 2016, http://www.baynews9.com/fl/tampa/news/2016/10/25/florida_decides_polls_presidential_race#pdfs Accessed 2 June 2019.
- ‘Pew Sampling’. pewresearch.org, www.pewresearch.org/methods/u-s-survey-research/sampling/. Accessed 31 May 2019.
- Rothenberg, Stuart. ‘Swing States, Battlegrounds and the ’12 Map’. Roll Call, Jun 14, 2012. ProQuest, https://login.ezp.pasadena.edu/login?url=https://search-proquest-com.ezp.pasadena.edu/docview/1020394786?accountid=28371.
- ‘U.S. Presidential Election, Florida, 2016’. ballotpedia.org, 2016, http://ballotpedia.org/Presidential_election_in_Florida,_2016 Accessed 2 June 2019.