Why Opinion Polls are Wrong

Opinion polls released just a week to the general election on August 1st showed that Raila Odinga of National Super Alliance (NASA) was the most popular presidential candidate. The results by Infotrack research firm indicated Raila had a 49 percent lead. The bulk of his support came from Nyanza, Western, Nairobi and coast regions. President Kenyatta had a popularity of 48 percent, mainly from Mt Kenya region, Rift valley and the North Eastern. Two percent of the respondents said they were undecided on who to vote for.

Meanwhile, the Ipsos Synovate pollster rated Uhuru as the leading candidate with 47 percent, and Raila came second with 44 percent. Both polling companies are of the view that the elections will not have a round one winner since both candidates are rated at less than 50 percent. Opinion polls are released intermittently by research firms, to portray the popularity of various politica aspirants. Nonetheless, on many occasions, opinion polls have failed to reflect the actual elections results.

Prior to the Jubilee party primaries in April this year, opinion pollsters had rated a win for Kiambu Governor William Kabogo and Kirinyaga Governor Joseph Ndathi, but Kabogo was beaten by Ferdinand Waititu, and former Cabinet Secretary Anne Waiguru defeated Ndathi in the race for Jubilee party ticket. Days to the March 2013 elections the Infotrack research firm rated Raila Odinga of Coalition of Reforms and Democracy (CORD) as the leading candidate with 46 percent and Uhuru Kenyatta of Jubilee Coalition with 44 percent.

Both Ipsos Synovate and Consumer Insight pollsters released opinion polls showing that the elections would not have a round one winner and would go for a runoff, since the election was too close to call. However, the polls were proved wrong after Uhuru won the 2013 elections with 6,173,433 votes (50.51%). Raila came second with 5,340,546 votes (43.70%). Uhuru met the constitutional threshold of 50% + 1 by 8,100 votes. Interestingly, an Ipsos Synovate opinion poll conducted in September 2015 showed that 64 per cent Kenyans did not want Raila to run for president in 2017. Yet the same company and other pollsters are now rating Raila as the leading presidential aspirant.

The Trump Effect
Before the November 8th 2016 elections in the USA, various opinion polls rated Democrat Party’s candidate Hillary Clinton as the likely winner of the elections. Republican Party’s Donald Trump trailed Clinton in the opinion polls throughout the campaign period. The projection before the elections showed Hillary would win 48 percent of the total poll and Trump would come second at 44 percent. According to the USA Today, only four national polls rated Trump as the leading candidate.

Yet 63 national pollsters projected a Clinton win. Such polls were conducted by Reuters, Fox news, CBS news, Bloomberg Politics, Economist, NBC news among others. Although Clinton won the popular vote by 2.8 million votes, the opinion pollsters’ demographics did not consider Trump’s support base of rural white voters who tilted the vote in his favor. Trump won 63 percent of white men voted for Trump and 53 percent of white voters.

On the other hand Clinton received the vote of 94 percent of black women and 68 percent of Hispanic voters. According to PEW Research Centre, the poll outcome was determined by the ‘shy
Trumpers’ syndrome. This is a scenario where supporters of Trump did not want to publicly identify with his candidature. Unlike Clinton’s supporters that were more vocal in their support.

PEW elucidated that, “support for Trump was socially undesirable, and that his supporters were unwilling to admit their support to pollsters.” This phenomenon is driven by the ‘Bradley Effect’, which Newsweek describes as a situation where white voters want to appear politically correct by telling pollsters they are going to vote for a black candidate when, in fact, they are not willing to do so.

The theory is linked to African American Tom Bradley, the former Los Angeles mayor who ran for California’s gubernatorial race in 1982 on a Democrat’s party ticket. He was rated highly by opinion pollsters and was predicted to win, but on the elections day he lost to Republican’s George Deukmejian who was of the white race. Prof Patrick Egan of New York University
described it as situation where, “a number of white voters may give inaccurate polling responses for fear that, by stating their true preference, they will open themselves to criticism of racial motivation.” BuzzFeed’s writer, Anne Helen Petersen wrote on November 2nd 2016 an article titled ‘Meet the Ivanka Voter’.

She described silent Trump supporter were most likely to be suburban white women who did not want to be publicly linked to Donald Trump, however in the secrecy of the voting booth, they would pick Trump over Hillary. Petersen described the Ivanka voter as a woman who “has internalized the fact that voting for Trump is not something you do in public. Either because she thinks talking about politics is tacky, or because she simply doesn’t want to deal with others lecturing her about her vote.”

She further wrote that the Ivanka voter knew all about Ivanka’s clothing line and brand, and thought she would be great in the White House. Although the Ivanka voter did not agree with some of Trump’s utterances, “She does not think of herself as a racist. She describes herself as ‘socially moderate,’” Petersen wrote.

Still, there are analysts who argue that Clinton’s win was hampered by the FBI Director James Comey’s announcement on new investigation just a week to the elections. The FBI probed Clinton’s email use when she was the Secretary of State in Obama’s administration. The negative publicity had an effect on her candidature and swayed the undecided voters to vote for Trump.

Florence Gichoya is an Associate Fellow with the Nkrumah Center for African Affairs and Global Peace (AAGP), at the Africa Policy Institute. Email: flogichoya@gmail.com

Be the first to comment

Leave a Reply

Your email address will not be published.