Moment of truth awaits conflicting pollsters

It is a moment of truth for pollsters as in a matter of hours, the polls will open for the electorate to prove them right or wrong.

With pollsters across the world coming under intense scrutiny especially after giving inaccurate opinions in last year’s US elections and the Brexit poll in the UK, the pressure in Kenya is intense.

In the US, pollsters had predicted a resounding win for former Secretary of State Hillary Clinton only for her to lose to Donald Trump. In the UK, pollsters said the country would vote to remain in the European Union. The opposite happened.

Stark differences

Experts say Kenyan political dynamics are tricky for pollsters to give accurate predictions.

“Kenyans don’t vote on issues but largely by tribe. Getting a perfect sample to use in an opinion poll is a tall order,” Dr Joshua Kivuva of the University of Nairobi (UON) School of Economics said.

So far, four different polls have given contradictory predictions over whom between President Uhuru Kenyatta and Opposition chief Raila Odinga will win Tuesday’s hotly contested elections. Two pollsters however say there will be a run off.

ALSO READ:

LIVE: Jubilee and NASA rallies

Infotrak has predicted Raila will win by 49 per cent of the vote against Kenyatta’s 48 per cent. Ipsos Synovate, on the other hand, says Kenyatta will get 47 per cent of the vote against Raila’s 44 per cent.

Another poll by the African Electoral Observers Group says the President will be re-elected by 51 per cent of the vote over Raila’s 39 per cent. The National Super Alliance (NASA) has also hired the services of US polling firm Zogby which predicted a win by 47.4 per cent win by Raila against Kenyatta’s 46.7 per cent.

The difference, according to Ipsos Synovate lead researcher Tom Wolf, is because pollsters used a different sample of the population.

“The polls are not wrong. There are a number of factors that can create a difference. For example, we went to all counties while other pollsters did not,” he said. “Like in the survey we did at the beginning of July, our sample had 10 per cent of Muslims and Infotrak in its sample had four per cent Muslims. This can create some difference.”

Politicians have been engaging in a war of words for the last few weeks over who between Infotrak and Ipsos Synovate are giving the correct predictions. Three weeks ago, Jubilee sparked a war with Infotrak over an unreleased poll which they claimed was going to place Kenyatta and Raila at par in terms of popularity.

National Assembly Majority Leader Aden Duale termed the yet to be released polls as paid for by NASA and meant to “lay ground for the rejection of the results of the election.”

ALSO READ:

LIVE: Jubilee and NASA rallies

This war of words over opinion polling that is set to come to a grand finale on Tuesday has upped the pressure on pollsters whose credibility is on the line. And there is a good reason.

During the nominations, the predictions by a number of pollsters were proven wrong by voters. Infotrak had projected that Kisumu Governor Jack Ranguma would be re-elected by 58.7 per cent of votes against Senator Anyang’ Nyong’o’s 8.9 per cent. It is Nyong’o who won by 164,110 votes against Ranguma’s 96,480 votes.

The pollster had also placed Kiambu Governor William Kabogo ahead of his opponent and Kabete MP Ferdinand Waititu by 48 per cent against 31.8 per cent. Another pollster, Tifa, had placed Kabogo at 46 per cent and Waititu at 33 per cent but during the primaries Waititu beat Kabogo resoundingly.

In Nandi, the same pollster predicted former Cabinet minister Henry Kosgey would win the Jubilee primaries by 36 per cent of the vote followed by former Agriculture CS Felix Koskei at 21 per cent. Incumbent Governor Cleophas Lagat came a distant third with 18 per cent. It is however Senator Stephen Sang who was not even considered by the pollster who emerged victorious.

But there are times pollsters have given correct predictions too. In July 2010, Ipsos Synovate (then known as Synovate) predicted the Yes camp for the referendum on the adoption of the new constitution would win by 58 per cent while the opposing side would garner 22 per cent. The result was a victory for the “Yes” campaign, with 69 per cent in favour and 31 per cent against.

Approval ratings

Five years before, Ipsos Synovate (then known as Steadman International) predicted a 42 per cent win for those against the initial referendum for a new constitution against 32 per cent in favour. The referendum, which gave birth to the ODM party, was rejected by 57 per cent of the vote while 43 per cent voted in favour

ALSO READ:

IEBC faces more hurdles in court two days before poll

Its then Research Director Maggie Ireri, now with Tifa, insists whatever they have predicted this time round will be proven right.

“Even if the percentages don’t tally, the predictions on who will win at the county level will be correct. The only county which we have reservations is Nairobi, where Mike Sonko and Evans Kidero are at par,” she told Sunday Standard.

Then she gave a disclaimer: “In order to get the correct result, the most important thing is that the people interviewed actually reflect on the voter register.”

Jubilee, which has been given the most positive approval ratings by pollsters, is however reserved on its attitude towards opinion polls, insisting the real poll will be through the ballot.

“It depends on turnout,” its vice chair David Murathe told Sunday Standard.

Nasa on the other hand has been rallying its voters to turn out in large numbers in order to disapprove the pollsters. In 2013 polls consistently showed that Raila was ahead of Kenyatta only for the former Prime Minister to lose.

“Jubilee has wished us ill but we will still stand together to send them home,” NASA deputy presidential candidate Kalonzo Musyoka said during their final rally in Mombasa on Thursday.

ALSO READ:

IEBC faces more hurdles in court two days before poll

BREAKING NEWS Flying squad raid NASA’S tallying center [VIDEO]

NASA withdraws claims against KDF after assurance that military will not take part in elections