Voter Turnout: Numbers Should Always Add Up, IEBC’s Don’t

In the salient math headache from the just-concluded Kenyan elections, IEBC Chair Wafula Chebukati did not help the situation much, having announced two conflicting turnout figures.  

He first announced 65.4% which he said accounts for total voter turnout as per the electronically transmitted results received via the Kenya Integrated Election Management System (KIEMS) kits. He however disclaimed that this excludes areas where the commission had allowed manual voting.

Shortly after, Mr Chebukati sought to clarify that voter turnout is in fact 64.6%. This represents 14,164,561 voters. He added that this figure would go up once those who voted manually are fully determined.

However, what the Independent Electoral and Boundaries Commission (IEBC) chair left out in both instances is declaring what the denominator to the two percentages is. Herein lies the loophole for speculations, and regrettably one for misinformation too. Some people, including the IEBC commissioners who made earlier announcements on voter turnout based their remarks on the assumption that the denominator must have been the 22,120,458 total gazetted registered voters. This is not the case.

By the 64.6% which Mr Chebukati said represents 14,164,561 voters, it would imply that there were 21,926,565 total registered voters from the tally of electronically transmitted votes.

This can be computed as: 14,164,561*100/64.6=  21,926,565 voters.

For this reason, the percentage given by the chairman is not and should not be interpreted using the gazetted 22,120,458 registered voters as the denominator. If this had been the case, the percentage should have been 64.03376 % of voter turnout.

There may be a number of reasons for this; first,  that the numbers given excluded registered voters from polling stations that purely used manual identification and were yet to transmit their data. It could also mean that some polling stations that were to identify voters electronically had not begun voting at the time the IEBC chair made these announcements. For example in Eldas constituency were voting went on the following day. 

Yet another possibility is the human error in computing the percentages. Sometimes, errors are made by calculating the average percentages of various subsets. In this case, they averaged the percentage voter turnout from polling stations (published on Forms 34A) or constituencies (published on Forms 34B) instead of the national absolute totals at the end polls.

IEBC  needs to clarify to Kenyans;

  1. What formula they used to compute the voter turnout they announced.
  2. IEBC records should also reveal the number of voters not identified electronically at all other polling stations but voted.
  3. How many polling stations identified voters manually, how many of the same manual votes were valid or rejected.

The other common assumption is that all ballots in the presidential boxes should be equal to all the voters who turned up. This may not be the case. Presidential votes cast may not give us the exact voter turnout but should be close to it.

If you calculate the sum of total valid votes cast for the presidential post as seen on the Form 34C and the rejected votes, divided by the total gazetted registered voters, the percentage of Kenyans who registered to vote and voted for the presidential ticket is  64.76697%. This is the correct way to represent this mathematically. 

This figure should ideally be close to the actual turnout percentage because it can be assumed that people who received the 6 ballots would likely have inserted them in the respective boxes, but should not necessarily be interpreted as the voter turnout.

In future, there should be a forum to present legitimate quantitative arguments in the same way lawyers can fight out the qualitative aspects of electoral disputes. Numbers should always add up. These errors have exposed a lack of statistical competency in the commission and the secretariat, posing a great danger to the credibility of the elections in Kenya, and hence a threat to the nation’s peace and stability. 

Worse than this, the IEBC’s inability or reluctance to answer these rather obvious questions adds to rampant speculation about possible subjectivity in discharging its duties. Nobody wants an incompetent referee, much less in an election as close as this one, where every action will be interrogated by either side.

To date, the IEBC has yet to publish numbers representing manually identified voters and the eventual voter turnout in percentage or otherwise. 

Add comment

Your email address will not be published.

Purity Mukami

Purity Mukami is a data and open source intelligence journalist . She has collaborated on projects that generated stories from Sudan's protesters shooting, Kenya's general elections, sexual behaviour patterns and government procurement data-sets. Purity holds a degree in Mathematics and Computer Science and a Master of Science in Applied Statistics.