THE much-criticised opinion poll by the state-run Research and Education for Democracy in Tanzania (REDET) that gives President Jakaya Kikwete a landslide victory in this month's general election has raised questions about the reliability of polls.
The findings of the latest poll released by REDET, a University of Dar es Salaam centre, were widely rejected by political commentators and a wide section of the public, with many dismissing it as a "rogue poll." By definition, a rogue poll is a survey that is completely unreliable.
A respected professor of law at the University of Dar es Salaam, Chris Peter Maina, said the results of the REDET poll was highly questionable.
“The poll is very unrealistic. It does not reflect the actual reality on the ground,” Prof. Maina told THISDAY.
“The poll has a lot of inconsistencies. I don't know where these shortcomings come from.”
According to the REDET poll, 71.2 percent of Tanzanians would vote for Jakaya Kikwete in the presidential election.
Dr. Willibrod Slaa of Chadema landed just 12.3 percent of the vote in the poll despite his growing popularity, while Prof. Ibrahim Lipumba of the Civic United Front got 10.1 percent.
"It is highly questionable that REDET can be regarded as an institution dedicated to the objective search for truth since this institution has always been partisan," said one commentator on Wanabidii Forum, an online discussion group.
"Do you really think if REDET would have found that Kikwete was behind by 15 points they would have published the results? I hate to say it, but some of the impression I have is that REDET has morphed into academic henchmen for the status quo!"
Contrary to the REDET poll, several online opinion polls conducted by newspapers and social networking sites put Slaa way ahead of Kikwete.
An opinion poll by an online public discussion site, JamiiForums, which has a much bigger sample size (nearly 9,000 unique voters) compared with REDET's poll of less than 2,600 respondents, shows that Slaa is comfortably winning the presidential race.
By Saturday afternoon, more than 69 percent of voters said they preferred Slaa as Tanzania's next president.
Kikwete came a distant second in the online opinion poll, with just 21 percent of the votes, followed surprisingly by Hashim Rungwe of NCCR-Mageuzi with 5.4 percent, while Prof. Ibrahim Lipumba of the Civic United Front has a dismal 2.9 percent of the votes cast.
Similarly, an online opinion poll of 161 visitors on the website of state-run newspaper, Daily News, gave Slaa a solid 60 percent lead, with Kikwete lagging far behind. The online poll by the government-owned newspaper has since been disbanded.
The website of a newspaper owned by the ruling Chama Cha Mapinduzi party, Uhuru Publications, disowned an online opinion poll on its site after it showed that more than 66 percent of voters were not satisfied with the direction of the country and prefer a change of national leadership.
On the other hand, an ongoing opinion poll started last week by THISDAY's website, www.thisday.co.tz, showed by Saturday that 93 percent of 106 visitors to the site disapprove of Kikwete's overall job performance as president.
Experts say online polls are biased because the online population is not a representative cross section of all adults because not everyone in Tanzania can be represented in an online poll.
The same can be said of REDET's poll of just 2,599 respondents in a country with more than 19.6 million registered voters.
Dr. Azaveli Lwaitama, a senior lecturer at University of Dar es Salaam, told THISDAY in an interview that it was difficult to judge the reliability of such surveys without knowing the polling methodology.
"In March, an opinion poll by the same pollsters (REDET) found that 77 percent of Tanzanians would vote for Kikwete and that has now dropped to 71 percent. This should not be good news for CCM," he noted.
"Slaa, who entered the presidential race just a month ago, managed to get 12 percent of the votes, while Lipumba, who has been around all these years, garnered 10 percent."
The unreliability of opinion polls is not uncommon even in the developed world.
In the 1980s, the gurus of public opinion polling in the United States declared just weeks before the presidential election that Democrat Jimmy Carter would easily win re-election.
But when the votes were counted, Republican Ronald Reagan defeated Carter by a margin of 51 percent to 41 percent in the popular vote — a rout for a US presidential race.
In the 2008 Democratic presidential primary in the US, polls showed that Senator Hillary Clinton was soundly outperforming Barack Obama in the nomination race. Obama later went on to become president.
Experts say if conducted in an unbiased and scientific manner, pre-election surveys can mirror the outcome of the election, although it is easy for researchers to influence the findings of opinion polls.
With a hotly-contested election in the offing, results of more opinion polls will likely be released.
But the only statistics that will matter the most are those to be announced by the National Electoral Commission (NEC) after Election Day that will determine the winner of the October 31 presidential election.