Kenyan pollster Ipsos explains why they missed the mark

Today Ipsos Synovate provided their own internal analysis (see here, pdf) of the election results vis-a-vis their poll numbers right before the March 4th election.

According to the final IEBC tally (Which Mr. Odinga is challenging in court) all the eight candidates except Mr. Kenyatta performed within the margin of error of Ipsos’ last poll before the election.

Mr. Kenyatta outperformed the last poll by 5.25%, well outside the margin of error.

How did Ipsos miss this?

Their answer on page 23 basically agrees with my observation that differential turnout, especially in the candidates’ respective strongholds, made the difference.

According to the final IEBC numbers, Mr. Kenyatta’s 20 biggest vote-basket counties averaged a turnout rate of 88%, compared to Mr. Odinga’s 84%.

In related news, tomorrow Mr. Odinga will officially file the petition that seeks to nullify Mr. Kenyatta’s election as president.

CORD may seek the nullification of the whole election or narrow their challenge to just whether Mr. Kenyatta actually crossed the 50% threshold.

Crucial figures to think about as we await to see the content of the petition tomorrow are (1) 10.6m votes were cast in the 47 governor races compared to 12.3m in the presidential race, a difference of 1.7m votes; and (2) Mr. Kenyatta crossed the 50% threshold by less than 10,000 votes.

More on this next week.

3 thoughts on “Kenyan pollster Ipsos explains why they missed the mark

  1. Mr Opalo, you are an excellent analyst and I enjoyed reading your analysis of the Kenyan political environment. You only need to dig a little deeper into the IEBC results to unmask the criminal organizational that is the IEBC. The truth is laid bare.

    -There is nothing unusual about Luo and Kalenjin turnout. It has always been 85-93% since the 2005 referendum.
    -Kikuyu turnout was apparently 95%; historically, even in the high-stakes 2007, it peaked at 82%.
    -In NEP turnout suddenly jumped to 88-93%. A region where some constituencies could barely manage 10% vote registration is suddenly doing 93% turnout?
    -Turkana, West Pokot, Samburu regions have never had such a high turnout of up to 90%. Regions that could barely manage 40-50% voter registration?
    -What stake did Maasais have in this election that 90-95% had to turnout and vote?
    -How come turnout increases everywhere in the country to between 80-96% but most of Coast remains at 60-68%?
    -As part of the implementation of the Kriegler Commission and Recommendations and to eliminate a variance in the ballots cast for the Presidential and other positions, the IEBC was issuing all voters with the all 6 ballots yet we still end up with 1.7m more votes for the Presidential race?
    -Then there is all the other school boy errors like duplicate entries in multiple constituencies consistent with an exercise that was being doctored.
    -How is it possible that the IEBC added up to 8000 votes in some constituencies after vote reconciliation during register cleaning? Throughout the voter registration exercise, IEBC would release updated stats in PDF format with its logo or an IEBC watermark. The last one that was quietly uploaded on Feb 24 could have been made by anyone.
    -Then there is the issue of the results that were announced based on Form 36 without verification. Some polling stations had 130% turnout although this is hidden in the constituency-level reconciliation included in Form 36.
    -This election was bungled at almost every stage. All geared towards giving someone a Round1 win. It is very dishonest for IPSOS to carry an analysis based on cooked results.


  2. Opalo, I like your blog but if it is true and confirmed that their is a gap of 1.7 million votes existing between presidential and other elective posts then someone was seriously cooking results at Bomas. I have heard Raila talking about this in a recent rally at Mombasa he held on Monday 18th. If he can substantiate this claim then he might stand a chance of overturning the election victory of Uhuru though he will need more evidence than just that. But this is even worse than the ECK discrepancy of 2007 between the presidential and parliamentary results.

    It is also amazing that almost two weeks after the elections, the IEBC while it has released the names of elected leaders for all the other posts, has failed to give us a mathematical breakdown of turnouts across these other posts. They act like people who are trying to hide something.

    Having had a look at the cord petition filed at the Supreme court and the responses from the respective respondents I seem to think that all the three respondents seem to be engaging in personal attacks against Raila rather than trying to address and rebut all the claims made by Raila. Note that Raila has raised pertinent issues regarding on how he thinks the elections were bungled by the IEBC the rest are rather bashing the character of Raila.


  3. Pingback: Why Raila Odinga Lost « Opalo's weblog

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.