Collected reactions to Twaweza / Sauti za Wananchi opinion poll findings
Two weeks ago, my colleagues and I at Twaweza launched our latest political opinion poll for Tanzania – including the perhaps surprising headline finding that at the start of the campaign period, CCM presidential candidate, John Magufuli had a strong lead over the Chadema / Ukawa candidate, Edward Lowassa. In case anyone hasn’t seen the poll findings (where have you been?), support for Magufuli was found to be 65%, while support for Lowassa was 25%, and Magufuli had a lead across all groups – urban, rural, male, female, all ages, all education levels:
I must also remind readers that this data comes from a nationally representative sample of 1,848 respondents from all regions of mainland Tanzania. It is not a prediction of the election results: a lot can happen between the time the data was collected and election day at the end of the month. The only poll that really counts in the one that takes place on October 25th. However, my main purpose here is simply to gather together a variety of media reactions to this poll and others published around the same time (e.g. Ipsos, TADIP).
1. Newspaper front pages, 23/9/15 On the day after the poll launch, it dominated the front pages. Some papers just report the poll’s findings and/or Ukawa’s reaction to it. Others are already rubbishing it directly, including unsurprisingly Tanzania Daima which strongly supports Lowassa / Chadema / Ukawa, but also the usually more neutral Guardian and Nipashe:
“Magufuli leads presidency poll”
“Study creates surprise, raises tough questions”
“Magufuli ahead of Lowassa”
“Twaweza prompts a debate”
“Goal by handball”
“Twaweza study lifts Magufuli”
2. Newspaper front pages, 24/9/15 – On the second day, the story has a little less prominence, and focusses mainly on reactions to the poll:
“Twaweza research raises 14 questions”
“Twaweza research still causing confusion”
“CCM: We’re not surprised”
“ACT-Wazalendo assess the research”
“I’m not giving up”
“Is it research or fraud …”
“Lowassa the most popular, he wins people’s opinion poll, conducted in anger at Twaweza’s report”
“Twaweza are liars – academic”
“Twaweza research is the talking point in every corner”
3. Newspaper front pages, 25/9/15 – On day 3, coverage focussed mostly on a new poll – by Ipsos / Synovate – with findings that were similar to Twaweza’s (Magufuli on 62%). Most of the papers gave attention to Ipsos, while Mtanzania (and Nipashe, to a lesser extent) reported on a separate poll conducted by Ukawa, which gave Lowassa 76%:
“Magufuli leads in another study” / “Ukawa ask 18 questions of Twaweza
“Another poll gives CCM a comfortable lead”
“One study gives Lowassa 76%, another Magufuli 62%”
“Magufuli leads in Synovate study”
“Ukawa surprised, Magufuli leads again”
“Magufuli still dominates”
“Ukawa: Lowassa leads, a new study gives him 76%”
“New research by Ipsos (Synovate): Magufuli streets ahead of Lowassa”
“Magufuli again, in new research, by Synovate (Ipsos)”
4. Newspaper front pages 26/9/15 and later – Another new poll, this time by a Tanzanian organisation called TADIP. Several papers report on the TADIP poll, but Raia Tanzania reveals that TADIP has close links to Chadema:
“Magufuli: The research underestimates my victory”
“New study by TADIP boosts Lowassa”
“Twaweza respond to Ukawa claims”
“Lowassa shines in new study”
“New research: It’s Lowassa”
“New research: Lowassa ahead”
“Research secret is out: the Institute that backed Lowassa is linked to Chadema”
5. Collected cartoons – We kept the cartoonists busy as well, with several suggesting that Twaweza was providing support (legitimately or otherwise) to CCM and Magufuli. Others covered polling more generally:
The Citizen, 23/9/15
The Guardian, 23/9/15
Tanzania Daima, 23/9/15 – “A big lift”
The Citizen, 24/9/15
The Citizen, 25/9/15
The (Tz) Guardian, 27/9/15
Mwananchi, 1/10/15 – “All Kipanya family members are required to declare who they support” / “And if you don’t like the results, will we be safe?”
Mwananchi, 5/10/15 – How did you get so many votes? / I lied to people in town, said I would end power rationing
6. The Twitter-verse – A curated collection of Tweets on the topic, mostly from the day of the launch itself: Sauti Politics Reactions
But perhaps most interesting of all are the imaginative (and in some cases highly fictional) reactions and responses. They illustrate the nature of politics in Tanzania at the moment very nicely. Here are some collected “highlights”:
Published on Jamii Forums days before Twaweza launched the findings, claiming that Twaweza was going to publish a fake poll. Several similarly worded posts were put on JF in the days before the poll launch.
“Twawezeshwa (We’ve been sponsored), throw your research away / Lowassa 71%, Magufuli 23%
“Twaweza, look at your figures again, Lowassa is president” – from Lindi, 23/9/15
A presumably fake WhatsApp conversation between CCM figures deciding to plant a fake opinion poll
Are you serious, Twaweza, or is it out of 200%?
“Ni Sisi / M”
Within hours of the poll launch, a fake Twaweza facebook page had been created, with over 5,000 likes, and looking very similar to the real Twaweza page. The main difference – the charts had been photoshopped to reverse the main findings.
From the forged Twaweza facebook page, two charts showing reversed findings.
Using the forged Twaweza account on facebook, with Twaweza apparently admitting that the poll was faked.
A clearly fake poll report, not even getting the organisation’s name right – should be “Research and Education for Democracy in Tanzania”, who have not anyway done a poll this year
“After Twaweza’s ridiculous research, people are doing their own polls, finding Lowassa ahead on 90%”
“Results of Kariakoo opinion poll: Lowassa 98%, Magufuli 2%. This research is open and more truthful than Twaweza’s”
Dear “We have been sponsored”, We now want stats on 1) ivory exports and remaining elephants, 2) Giraffes put on planes, 3) Tanzania’s drug dealers, 4) rate of increase in corruption, 5) wealth of leaders and their children. From Loafer son of Fool.
Some people will never believe something that they don’t want to hear.
Understanding of opinion polling, survey methodologies, and especially sampling techniques, is low. It is not widely recognised that a random, representative sample of 2,000 people is far more reliable than a clearly biased sample of 200,000.
Has Tanzania really reached a point when we assume that anyone can be bought?
Nevertheless, for perhaps the first time in the campaign, people were really discussing evidence and data, with a passion. Some of this was unreasonable and probably illegal, but in another way the rough-and-tumble of arguments about data was actually refreshing and healthy. It was messy, yes, but that’s how democracy works.