ELECTION NEWS: Election Standards

InformAction by InformAction
Be the first to comment!




August 25, 2017

The disputed Kenyan election is now being fought legally, with the Supreme Court judgment due Friday, September 1st.[1] Public attention is on what the case reveals about the standard of the election: namely, were the elections administered properly, and are the results believable?

The Independent Electoral and Boundaries Commission (IEBC) will be defending its conduct of the elections, and the supporting documentation for presidential results delivered verbally on August 11th.  The question of an electoral standard will have to be addressed, even though there is no one identifiable, documented global standard for an election.  International observation is in fact of limited value in this regard, particularly in Africa, as standards of process tend to be devalued rather than elevated by international observers through what is described as a comparative approach.  International observers use the term ‘generally free and fair’ to legitimize standards that would not be acceptable in their own countries. They are also frequently constrained by diplomatic and strategic considerations. However, there are nonetheless recognized benchmarks and fundamentals accepted as essential to a democratic election, like a single certified Register of Voters and publicly verifiable results.

Every country can be seen to be in the process of setting domestic standards around these fundamentals. Kenya, with its particular history of disputed and violent election cycles, has worked towards setting explicit standards of process to prevent a repeat of the deadly mistakes of the past. These standards are spelled out in the 2010 constitution, which demands and prescribes elections that are transparent, verifiable, accurate and accountable. 

The question is whether these standards have been met or disregarded.

Section 1 of this update gives an overview of the key areas of concern in the August 8th election, where KYSY has identified significant errors and inconsistencies in process and results tallying.  Section 2 provides the latest observations on data errors identified during continued scrutiny by KYSY of results forms 34A and B.

Section 1:  Key areas of concern in the Election

Registered Voters

It is a fundamental standard to have one certified Register of Voters. But, in all Kenyan elections, there have been shifting numbers of registered voters. This time, the number of registered voters changed after certification of what was supposed to be the sole, final Register of Voters on June 27th.  An increase of 25,638 names in the Register remains unacknowledged and unexplained by the election board.

This change potentially impacted the number of people allowed to vote at certain polling stations and may have impacted the final results. For example, there was a 5.98% increase in the number of registered voters in Nyandarua County - a stronghold of the incumbent party – made sometime between certification and the declaration of presidential results.

Rejected Votes

The lack of consistency in the number of rejected votes recorded on results forms from the polling stations (Form 34A) and at constituency level (Forms 34B) compared to numbers on the IEBC’s results website suggests different, unverifiable sources of the results count.

Missing Data

There is still a lack of essential data on the IEBC website.  It is impossible to know the source of the presidential results until the website provides the full set of results at polling station and constituency level.  The documentation site continues to lack particular Forms 34A and 34B at the time of this release (25th August 2017).

It is also of concern that the IEBC Chair, Wafula Chebukati, announced results based only on valid votes.  This is a good example of a standard that is an accepted benchmark of a democratic election. A tally of all votes is necessary and expected – including rejected, objected and disputed – to account for any inconsistencies or errors in the count before a final declaration.

Mismatched Results Data

There is plenty of evidence – even from a cursory glance at the website – of data in Forms 34B (constituency results) not reflecting data in Forms 34A (polling station results). How could this happen in a system relying on simultaneous transmission? It suggests that results were changed at some point in the process, either between the polling station and constituency results tallying; or, after the transmission of constituency level results to the National Tallying Centre.

The results website also does not consistently reflect what is recorded in Forms 34B.

Errors on Results Forms

Forms 34A and 34B are replete with internal errors, including addition mistakes, missing information and misleading information.

Section 2: Update on Errors and Inconsistencies

KYSY continues to scrutinize and analyze results forms, comparing data on (1) Forms 34B  (2) the number in the IEBC Results Site and (3) the IEBC’s declaration of presidential results.

These are the main observations.

Missing totals

Some Forms 34B are missing grand totals of valid and rejected votes. Through a manual calculation of the data present on the forms, KYSY was able to confirm the number of valid votes in all but one constituency – Kaloleni - where the form is not uploaded onto the IEBC website. For rejected votes, KYSY was able to confirm totals for rejected votes in 279 constituencies, assuming a total of 0 rejected votes for forms in which there is no data recorded in the polling station entries for rejected votes.

Registered Voters

The IEBC certified the total number of registered voters on June 27, 2017. At the time of certification, the total number of registered voters was 19,611,423. The numbers of registered voters per county was then announced on the night of August 11, when IEBC Chairman Wafula Chebukati declared the presidential results. An aggregation of this total amounted to 19,637,061, representing an increase of 25,638 voters.  The largest increase was in Nyandarua County, where the number of registered voters grew by 5.98 percent.  The numbers announced by Chebukati differed in at least 43 counties.[2] 

Although the numerical increase of 0.13 percent is small, the significance is that it demonstrates the lack of one, final Register of Voters. 

Presidential Results

There continue to be inconsistencies related to the total number of valid and rejected votes.  A definitive total is important for determination of the winner, and critical for public trust in the results. Yet, inconsistencies are still pronounced between the different data sources  - meaning the IEBC website, Forms 34B, and the official declaration of results.

How is it possible to know what the final results were based on, when they were declared by the IEBC on August 11th?

According to the 2013 Supreme Court ruling, the winner of the presidential race must garner 50% + 1 of the total number of valid votes alone. The more valid votes there are, the more votes the winner needs to reach the threshold of victory.

Table 1: Inconsistencies between the IEBC’s Data Sources


IEBC Results Website (National Aggregate)

IEBC Official Announcement of Presidential Results

Forms 34B


Total Valid Votes in Presidential Race





50% + 1





Total Rejected Votes in Presidential Race


Not Given



Valid Votes

The IEBC’s final declaration of presidential results, which included total valid votes per county, continues to differ from data on the IEBC’s website and from data in Forms 34B. It continues to be true that no announced county-level total of valid votes corresponds to the total valid votes recorded on the IEBC’s results website (except in the diaspora). The largest differences are outlined below.

Table 2: Differences in Valid Votes at the County Level



Announced Valid Votes

IEBC Results Website Valid Votes (Aug. 23)

Difference between and Website and Announcement (Website Minus Announced)

Forms 34B Valid Votes

Difference between Forms 34B and Announced (34B Minus Announced)
























































Announced valid votes match aggregated county totals extracted from Forms 34B in 5 counties (10.6 percent). These counties are Laikipia, Nyeri, Trans Nzoia, Isiolo and Nandi. Announced valid votes match the county-level data in the IEBC results website only in the diaspora.

At the constituency level, the number of valid votes in the IEBC’s results website matches the corresponding data in Forms 34B in 41 constituencies, including the diaspora (14.1 percent).

Rejected Votes

The number of rejected votes at the county level on the IEBC’s results website matches the number of rejected votes recorded in Forms 34B in 0 counties. Furthermore, the data from Forms 34B all show LESS rejected votes than what is recorded in the IEBC’s results website. The largest differences are shown below.

Table 3: Differences in Rejected Votes at the County Level



Rejected Votes on IEBC Results Website (Aug. 23)

Forms 34B Rejected Votes

Difference (Forms 34B Minus IEBC Website)





























*These totals are based on partially incomplete data. Certain Forms 34B are missing totals of rejected votes.

At the constituency level, the number of rejected votes on the IEBC’s results website matches the number of rejected votes recorded in Forms 34B in 3 constituencies, including the diaspora (1.0 percent). These constituencies are Suba North, Samburu North and the diaspora.


According to the IEBC Results Site, total turnout for the presidential election was 15,593,050, which represents 79.5 percent of registered voters. Since Forms 34B do not consistently show all votes cast (valid, rejected, rejected-objected-to and disputed), it is impossible to calculate turnout as recorded on Forms 34B and compare it to the website. The IEBC also did not give a breakdown of turnout during the announcement.

Regional Impact?

KYSY continues to identify errors and inconsistencies in the uploaded Forms 34A and Forms 34B. There are mathematical errors, irregular forms, and missing information on the forms. It is not yet clear what impact these have had in particular regions. For example, KYSY notes the following in Isiolo:

  • Bibi Water Supply Polling Station, Kinna, Isiolo South, Isiolo: All agent information and signatures appear to have been recorded by the same individual.
  • Daawa Primary School Polling Station, Stream 1, Garbatulla, Isiolo South, Isiolo: There are at least two different forms, each of which includes the same numbers but different other markings. The presence of more than one Form 34B taints the legitimacy of the final polling station result, because it makes it impossible to verify which form is original and true.
  • Daawa Primary School, Stream 2, Garbatulla, Isiolo South, Isiolo: There are at least two forms, each of which has a different total number of valid votes.
  • Dabasiti Primary School in Isiolo South, Isiolo: Agents’ information is recorded in what appears to be the same handwriting. 


  • There is still missing data from the IEBC website, as from 4pm 25th August 2017 – the day before the Supreme Court hears the case.
  • There continues to be major errors and inconsistencies in all the publicly available results data from the IEBC, as from 4pm 25th August 2017 – the day before the Supreme Court hears the case
  • Public trust in the election and the legitimacy of the result depends on a clear resolution of all the inconsistencies and errors documented by KYSY.
  • Public trust and regime legitimacy will depend on a clear explanation of the source of the final declaration of presidential results.
  • The Supreme Court ruling will determine Kenyan electoral standards for results counting and tallying for future elections.


[1]          The Supreme Court timeline dictates filing any challenge be filed within seven days of the declaration of presidential results, and a court ruling made fourteen days after the petition has been filed.

[2]               Since the live broadcast cut away from the announcement during the numbers announced for Embu, it is not possible for KYSY to verify what number was used for this county. KYSY analysis is based on an assumption that the certified number for Embu was used in the announcement. The number for prisons was not included. The number of registered voters in the diaspora was the same as what was certified.

3.               Missing total valid votes from prisons (was not included in announcement)

[4]               Missing data from prisons because there is no publicly available Form 34B for prisons.

[5]               This is the difference between Forms 34B and Announcement (Forms 34B Minus Announced)

[6]               Missing data from prisons because there is no publicly available Form 34B for prisons. Also, the total number of rejected votes contained in Forms 34B is missing from 18 constituencies, either because the forms contain no data for rejected votes or because KYSY has not yet completed the manual calculations of rejected votes for forms on which the totals were nor provided.

[7]               This is the difference between Forms 34B and the IEBC website.

Read 654 times

Leave a comment

Featured news article

My Right to Protest

          My Right to Protest is a booklet produced by InformAction, based on Article 37 of Kenya’s Constitution. It gives information and insight into this right, and is designed to help people exercise it, and...

Read more