Printer-friendly versionSend by emailPDF version

Political opinion polls are controversial in Kenya especially at election time as some of the pollsters are owned by or allied to certain politicians. But more worryingly, poll results have been mentioned in connection with in inter-communal violence

At camp Chechelesi on the outskirts of Isiolo town, a young female crouched on the bare floor, mourning her unborn baby. The woman, a refugee, lost her baby six months to term.

The only comfort her mother - also an Internally Displaced Person (IDP) - could offer, were three painkillers Panadol, which she had struggled to buy, and another two Hedex tablets donated by a neighbour in the nearby tent.

"My granddaughter has died," mourned the tearful mother. "Next it is my daughter."

This was Isiolo County, just days before Kenya's March 4th 2013 General Elections, according to a report by the Consolata Missionaries Justice and Peace Commission (CMJPC).


What started as a simple flare up in August 2010 became a full scale war that left the rest of Kenya groping for answers.

Those who sought to investigate would try to visit the four IDP camps at Shambani, Masharikwata, Game, and Kisima, hosting about 2,700 people.

"The violence was political," explained Theresa Székely, a social worker in North Eastern Kenya. "The tribes were fighting to gain control of the County resources once a new government was in place."

But an enlightened account by elders there revealed a recurrent pattern.

Steven Ali, who heads the Pastoralist Community Development Organisation (PCDO), a lobby group pushing for peaceful coexistence in the region, traced the conflict to the ballots: early opinion polls predicting which political party was likely to win the General Elections.

According to Ali, this triggered the flare ups.

"The Turkana supported the Coalition for Restoration of Democracy," argues Ali. "The Boranas wanted the Jubilee alliance to win."

The Turkana had an edge over the Boranas because opinion polls showed that Presidential aspirant, Raila Odinga of the Coalition for Restoration of Democracy (CORD), would win the elections, he claimed.

"The Boranas feared their rivals would acquire power once a government was in place and displace them from their grazing land," he said. "They invaded the Turkanas who had settled there and caught them by surprise because they had guns."

And so conflict fanned out due to what political analysts say was the misconception by both parties that the earlier opinion poll was the actual electoral poll.

The impact of predictions by pollsters of the General Elections as a critical factor dividing Kenyans is no anomaly.

In the 1990s during retired President Daniel Moi's reign, the Kalenjins who have lived in Rift Valley region for generations with their Kikuyu neighbours would evict the later from their homeland.

The Kalenjin fear was that the Kikuyus would split the regional vote in favor of the opposition candidate, Kenneth Matiba in 1992, and Mwai Kibaki in 1997.

Retired President Moi of Kenya African National Union (KANU) won the 1992 polls with 36.3 per cent of the votes cast, while Matiba of the Forum for the Restoration of Democracy – Asili (FORD Asili) came second with 26.0 per cent.

The rest of the votes were shared among the other six Presidential aspirants.

In 1997, Moi won again with 40.12 per cent followed outgoing President Kibaki of Democratic Party (DP) with 31.09 of the votes cast. The rest of the votes were shared among the other 13 Presidential aspirants.

In both of these cases, opinion polls held secretly by the intelligence would show the incumbent, President Moi, was losing support in areas outside his political strongholds.

Violence and displacement of minority tribes would be reported in regions where he enjoyed popular support – including the Rift Valley - just before Kenyans went to the ballot.

But in 2007, the pattern was different. The post-election violence with a body count of about 1,000 Kenyans, would be linked to repeated opinion poll reports showing that Presidential candidate Raila Odinga would win against incumbent, Mwai Kibaki.

The later was controversially declared the winner by the now defunct Electoral Commission of Kenya (ECK). It was a price ECK paid by being phased out of Kenyan politics for alleged manipulation of poll results, including those of the Moi era.

The Independent Electoral and Boundaries Commission (IEBC) was formed and took charge of the country's electoral process.

"Unfortunately what the opinion polls predict is never the case when the final Presidential results are announced," argued Sofia Abdi Noor, former vice chairperson of Kenya Women Parliamentary Group (KEWOPA).

Why the waning confidence in Kenya's opinion pollsters when it comes to politics?


Insiders working for research firms revealed that the whole process – and its vast impact - is simply a business. Institutions were also described by insiders as likely to be subjective to the private clients commissioning polls, including politicians, whose names were never disclosed.

The field officers hired to do surveys answered to the wishes of the recruiting agency and may not generate data that is representative, investigations reveal.

Testimonies shared by some field agents indicate they are poorly paid. Others claimed they were never paid their dues by the recruiting pollster.

A few acknowledged that surveys could be done through emails, or texts. Some even stated that questionnaires could be filled out by themselves to save on the meager pay received.

It is an allegation few opinion pollsters would publicly own up to, while most of the firms were hesitant to disclose how much field agents were paid.

Victor Rateng', a project manager with IPSOS Synovate, says his firm pays a minimum wage of Ksh. 700 (about US$ 9) per day, which does not include other allowances.

For a task that would require a field agent to interview as many people as possible if assigned face to face interviews, or those that engage civilians on the streets, many claim such pay is pinball.

However, Rateng' says his firm insists on household interviews which generate results that are proportionate to size, albeit for the same pay.

"We make sure that we have sent the agents to the fields at least three times for pilot testing of the questions before the actual exercise," he says. "It is an expensive process but we are working to review the payment upwards."

The media and civil society have linked such financial constraints to the emergence of rogue opinion poll firms in Kenya.

For instance, a research firm desperate to make quick money from the 2013 polls is said to have instructed its officers to collect data from public parks in Nairobi city, but claimed the details represented the whole country in the final report.

Another pollster is said to have sampled areas where its political client was popular, only briefly focusing on the rival candidate's political turf.

"I have reason to suspect the accuracy of political opinion polls because some of the pollsters are owned by Kenyan businesses," argues Mutegi Njau, a senior editor with Royal Media Services (RMS) in Kenya.

According to him, most of the polling firms rarely have strong in-house funding to do their research, hence the reliance on outsourcing for clients.

It is this weak financial base that inclines them to believe they would win government tenders once their preferred candidate takes power, he says.

Such doubts have raised concerns about the process of opinion polling, including by the Media Council of Kenya (MCK).

The umbrella body for media practitioners, the MCK demands transparency on details like: who commissioned the survey? Who paid for it? Why was it done? Who carried out the project? How much money was involved in the entire exercise? How big was the population sampled?

Faced with such questions, all pollsters claim that their research is scientific, transparent, accurate and representative.

But a sneak preview at how some of them conduct the process raises doubts.

For instance, the standard for publishing opinion polls is not a process that can be legislated in Kenya. Pollsters can withhold research indefinitely.

This is unlike countries such as Italy where pollsters are expected to publish the reports within 15 days, Peru 15 days, Bulgaria 14 days, Czech Republic seven days, Russia five days, and France within 48 hours before an election.

"The standard for publication of opinion polls is something new in Kenya and is still under discussion," says Rateng'.


But Angela Ambitho, the founder of Infotrack Research and Consulting Ltd., is convinced that validation of data alone is enough to reach a decision on whether the reports are suitable for public release or not.

According to Ambitho, validation is the final standard in opinion polling ie: a team from within the sector hired to validate information from the field.

"What would keep a competing firm from using its own agenda to inform the survey?" questioned a former presidential candidate James Ole Kiyiapi.

Competing firms could use the details given by another firm to inform its own research. And pollsters acknowledge that the research is shared with another firm hired to do the validation.

These concerns sustain the suspicion of Kenyans that some pollsters are paid by political clients to sway public opinion.

They also led to the introduction of the publication of the Electoral Opinion Polls Act 2012 in Parliament.

According to the Bill's author, Bonny Khalwale, the legislation demands that pollsters reveal information on the methodology used in the research, including the regions sampled and the source of funding.

"The Bill was strongly opposed by a section of Parliament and so it could not legislate the political process ahead of the March 2013 poll,' says Khalwale. "We hope to make it our key focus in the new Parliament to ensure transparency in future polls."

Meanwhile, Ambitho acknowledges that the paying clients could include individuals with political interests. But information about clients would not be released to the public by the research firm, she says.

"We are not allowed to divulge information about how much clients pay us because there is a confidentiality clause in every contract that we sign," she says.

"Professional ethics do not allow the researcher to divulge the name of the client because it goes against the code of practice."

For now it is still unclear whether opinion polls could be linked to the violence that unfurls each time there is a general election in Kenya, even as those like Ambitho claim otherwise.

But those like Njau are certain that the reports do sway public opinion. The IDPs know better.

* David Njagi is a member of Forum for African Investigative Reporters, a pan-African organisation of investigative journalists.


* Please do not take Pambazuka for granted! Become a Friend of Pambazuka and make a donation NOW to help keep Pambazuka FREE and INDEPENDENT!

* Please send comments to editor[at]pambazuka[dot]org or comment online at Pambazuka News.