2013 Kenya Exit Poll — academic study published (updated)

Professors Clark Gibson, James Long and Karen Ferree have now published an article from their 2013 Kenyan election exit poll in The Journal of East African Studies.

The Star has an analysis in Wednesday’s edition. This is the front page, but the story is not yet up online. (Update: Here is The Star story, “Uhuru didn’t get 50% in 2013–U.S. academics“.)

See my May post with the video from an original presentation at Johns Hopkins SAIS here.

“Choosing Peace Over Democracy”

“Kenya’s 2013 Elections: Choosing Peace Over Democracy” has been published in the new Journal of Democracy by James D. Long, Karuti Kanyinga, Karen E. Ferrer and Clark Gibson.  Important and worthwhile reading for anyone interested in Kenyan politics or democratic process in Africa or the developing world more generally.

This is the first of formal publications using the exit poll and other polling data that were presented by Professors Gibson and Long at Johns Hopkins’ SAIS on May 2 and widely covered in the Kenyan media.  See my post with the video:
“Fraud and Vote Patterns in Kenya’s 2013 Election: Evidence from an Exit Poll”–Gibson and Long event in Washington Thursday.

Long and Gibson were the researchers who also carried out the 2007 IRI/USAID/UCSD Kenyan exit poll that showed an opposition victory.

Voting Procedure

Economist’s Boabob Blog features UCSD work on “smart” election monitoring

More great work to fight election fraud from my former colleagues on the USAID/UCSD/IRI Kenya exit poll, Clark Gibson and James Long:

“African elections:  How to save votes”

COULD smartphones help reduce electoral fraud in Africa and in other regions? At a recent forum hosted by the Brookings Institution on the ways that wireless technologies are affecting politics in various countries, Clark Gibson, a professor at the University of California, San Diego (USCD), presented findings from experiments in Afghanistan and Uganda which suggest that they can. Local researchers were deployed to polling stations armed with digital cameras and smartphones to take photographs of the publicly posted election tallies. The research found that this alone can cut electoral fraud by up to 60%.

The experiment was first developed during the 2010 Afghan elections by James Long and Michael Callen, then UCSD graduate students, with funding from the Development Innovation Ventures section at the United States Agency for International Development. . . . The research concluded that as a result electoral rigging was cut by 25% in the polling stations in the treatment group and the theft of ballot boxes and other election materials was reduced by 60%.

Mr Gibson replicated the experiment during the Ugandan presidential election last year, using a bigger sample of 1,000 polling stations scattered all around the country.  .  .  . using a special app developed by engineers at Qualcomm, a big technology company based in San Diego, the researchers this time were able immediately to send their data back to a server at UCSD. Academics there could then check to see if the voting numbers had been falsified by looking for give away number-patterns. They found again that vote tampering and ballot-box theft were much lower among polling stations that had received warning that a photo would be taken of their tally than among those that did not.

The technology is relatively cheap—smartphones cost around $250—and allows  more locals to get involved in monitoring elections. There is a great hunger for democracy in Africa and elsewhere, says, Mr Gibson, you can tell just by looking at the queues of voters who turn out on election day. Nothing is more dispiriting than to learn that their vote has been manipulated.

Unfortunately we didn’t have funding for separate electronic verification efforts in Kenya in 2007, but this should be that much cheaper and more readily feasible in Kenya for 2012/13. Knowing what happened last time there is no excuse not to have digital image verification this time.

Friendly Fire? IRI Chairman McCain Labels Exit Polling as Pork!

Republican Senators McCain and Coburn have issued a purported list of 100 wasteful porkbarrel programs getting funding under federal stimulus legislation–one item targeted on the list is a little over $200,000 for exit polling in Africa by the University of California, San Diego. 

Is this just a political cheapshot at UCSD for publishing the results of the Kenyan exit poll from the 2007 general election and accompanying research? 

For this Kenyan exit poll, McCain’s International Republican Institute (“IRI”), for which I was Resident Director of the East Africa Office at the time, received funding from USAID, along with an extra $10,000 from Dr. Clark Gibson, chair of Political Science at UCSD.  The poll showed the challenger Raila Odinga soundly defeating the incumbent Mwai Kibaki.  When the Electoral Commission of Kenya announced that Kibaki had won amid disputes and allegations of fraud, the US Ambassador Michael Ranneberger initially called on Kenyans to accept the results and the Bush State Department initially congratulated Kibaki (later retracting), even though the Ambassador had received the preliminary exit poll results on the evening of the vote.

Dr. Gibson and his associate James Long designed the poll under a consulting agreement with IRI and Long supervised the field work of IRI’s Kenyan polling firm Strategic.  IRI maintained a six month “exclusive” on rights to publicity on the poll under the consulting agreement and refused to let UCSD or Strategic release or comment on the results.  IRI declined to comment on the poll and then began telling journalists and others in Washington that it was flawed, eventually issuing a statement on February 7, 2008 that it had determined the poll to be “invalid” after hearings that day of Senator Feingold’s Africa Foreign Relations Subcommittee in which Feingold called on Asst. Secretary of State Jendayi Frazer and the Asst. Administrator for USAID to explain why the poll had not been released as post-election violence and negotiations between the contestants continued.

After the expiration of the six month embargo, the Center for Strategic and International Studies (CSIS) sponsored the release of the poll by UCSD on July 8.  Gibson and Long presented a detailed rebuttal to the alleged concerns raised by IRI.  The UCSD team also presented at SAIS at Johns Hopkins.  In August, more than a month later, on the day before Gibson and Long were to testify on the results of the poll before the Kreigler Commission in Nairobi, appointed to review the election under the February 28 power-sharing settlement, IRI released the poll, having found that it was valid after all. 

In the meantime, IRI continues exit polling all over on the taxpayer dime–and trumpets the “earned media” it gets for this from publications like the New York Times.  But apparently National Science Foundation funding for polling done by actual social scientists at UCSD outside the auspices of International Republican Institute is pork!

As Gibson and Long pointed out in their presentation of their research to the Working Group on African Political Economy last year, the US spends hundreds of millions on democracy promotion, but we don’t even know what motivates African voters.  Of course, if we don’t really always want to know HOW they vote, I guess maybe we don’t care why either?  And for that matter, maybe we don’t want to learn more about how effective that “democracy promotion” money is?

James Long worked tirelessly under pressure to help execute the Kenyan poll for IRI under difficult circumstances, and even provided substantial free assistance on IRI’s September 2007 pre-election poll (which was quickly released, by the way).  File this under the category of “No Good Deed Goes Unpunished”.