Headlines

Fact checking claims of rigging in Kenya’s 2022 presidential election + all observation reports

Nic Cheeseman21 mins
Facebook Twitter Email

Note: this is a copy of @fromagehomme’s Twitter thread here. For official fact checking organisations see @AfricaCheck_KE, @BBCFactCheck, @pigafirimbi and @debunkdotmedia. For observer reports and other election documents, scroll to the bottom of this page.

[Thread starts 8:50am Kenya time, 12 August 2022] It is far too early to congratulate @IEBCKenya (the Kenyan electoral commission) as the official election count has not even begun But one striking thing about the claims of rigging now coming out is that no evidence has been provided to support them Let’s take a look …

  1. This tweet is emblematic of the kind of lazy accusations some are making, based on no rigour and no evidence at all. Doesn’t even pretend to be a reasoned argument or to have carefully looked at the results and forms. (Also beware anyone who write in ALL CAPS)

2. Other accusations like this claim in the Star play to well known tropes of how elections are rigged. Here again, though, we have no evidence to back up the claims. I (@fromagehomme) covered Limuru, Kiambu, and Murang’a and didn’t see anything to back this up …

The specific claim in this piece is that “the turnout on the KIEMS kit did not tally with the number of ballot counterfoils” This is easy to show – turnout is recorded in all the forms 34A on line forms.iebc.or.ke/#/, and can be compared with data collected by party agents at polling stations. Yet no evidence is provided to show: 1. Discrepancies between forms + evidence from party agents, or 2. Estimates of how many places this occurred. It is also strange there are no pictures of other activities alleged as they are said to be widespread + public.

That said, we should investigate as much as possible and compare to other data. It is noticeable here that there is no evidence of these activities or problems in the @elogkenya report on the voting and closing of polling stations.

3. This one is important as it appears that one candidate (Odinga) got more votes than there were registered voters on the form on the IEBC official portal in, Bomani Primary School, Kilifi County (@DavidWanjiru). But @Nanjala has done great work and shown that it is correct on form 34B (constituency level form that adds up polling station 34A forms + generates the constituency level tally) + this is critical as it is what will be used to complete form 34C (final presidential tally). It therefore seems that this has not impacted the final pres results.

4. As has become commonplace, we are also seeing a lot of claims of “hacking”. These take various forms, i.e. that the register/KIEMS kits/IEBC servers/tallying process, has been hacked. It is important to take these seriously, as digital processes have vulnerabilities, and the issue of the IEBC’s servers was an important part of the controversy in 2017.

This time round, there is an accusation that UDA has been involved in an “attack on the portal”. There are two things to say here. 1) these claims are particularly challenging as it is very hard to prove or disprove them, and 2) this is especially the case in this example as, interestingly, the post claims the efforts were unsuccessful and so the results process has not actually been tampered with. There are 4 reasons to be deeply suspicious of this post however:

A) It has no author.

B) No evidence is provided.

C) There is no UN observer mission tallying the results.

D) The number of polling stations not transmitted is well below 900 (now 26)

5. Another interesting one. This time we know there was an issue, as the Google/Reuters data stream of Forms 34B suddenly showed a very high figure for Raila Odinga. This led to speculation (including from this author) that the site had been hacked. 

The error was quickly corrected. In the end, Reuters put out a statement explaining that there had been no hacking from outside the system – the error was due to “human error in data entry”. Important reminder to all of us, me included, to fully investigate and clarify asap.

6. So many more allegations of rigging it is hard to keep up. This one is important as we have had almost no evidence of problems with forms 34a, and if those forms are ok the election is much more secure. Perhaps as a result, some are now targeting 34As, but so far unpersuasively.

Take this allegation from @CaxstonePKigata. The claim is that numbers were added to Ruto’s tally in the form on the right, which were not on the original form, which is on the left. But if you look at the picture closely, it is the supposed “real” form on the left that looks the most problematic. The numbers that look most likely to have been changed by being overwritten is the 17 of the total number of votes cast in the “agent” form, where the ink is much heavier than for other numbers.

 

See also the middle “1” in the 115 on the “agent” form, which, when you blow it up, seems to have a white patch around it as if it had been digitally changed or tippexed. Very hard to say for sure, but here the IEBC form seems less problematic than what is being claimed to be the “real” form. Will keep looking.

Update: Further evidence that the party agent form shared by @CaxstonePKigata is probably fake is that the turnout on the “true” form is only 49%, and the average turnout in that area seems to be around 65%. (For the sake of clarity and completeness, the turnout on the IEBC form is much closer to the nearby average, at 71%).

This again makes the IEBC form more plausible than the “agent” one.
 
7. One way that Charles Hornsby (check out his blog here) suggests we look for rigging, and something that I always look for, is differential turnout in different races. As the elections occurred on the same day, and voters are given ballot papers for all races at the same time, significant differences in turnout between, say, the presidential race, MP and MCA are suspicious – especially if we see much higher turnout in the presidential contest.
 
This makes the claim by @paulinenjoroge interesting. She suggests that there are large differences between the presidential tally and that for women’s rep and senator in “8 counties just sampled”. It is hard to tell for sure whether this is true or false because we don’t have full figures from IEBC yet. But there are some major questions about this claim at present.
A: Figures for the other races are not available from the IEBC portal so where have they come from?
 
B: If the data is from party agents, one problem is that we know from ELOG that they were not in all polling stations. Another is that so far no coalition has agreed to allow independent institutions to verify their counts.
 
 
C: Whether or not the data is correct, the framing of the message seems to be deliberately misleading – all of the counts we have suggest a very close presidential race, so claims of a “landslide” reduce credibility.
 
 
 
That said, we need to keep a close eye on the difference in turnout between the races when we get the full set of results from the IEBC for the 6 elections. We also need to continue to ask parties and candidates making charged and dangerous claims of rigging to share their data.
 
8. (Paraphrasing and adding to a series of Tweets sent from the KICC while waiting for Raila Odinga’s presser, watching the four “rebel” or “Serena” commissioners give their reasons for refusing to stand beside IEBC Chair Wafula Chebukati when he accounted that William Ruto had won the election. For context, the four commissioners gave a press conference in which they began by suggesting that one of their main issues with the election was that the final percentages for the share of the vote won by the four presidential candidate’s added up to 100.01%):
 
Serena 4 presser now on. It has started very badly for their credibility. They claim a big problem is that the results stated added to 100.01% which is “mathematically impossible” It’s not impossible it’s the product of rounding the numbers.
 
You do not usually write every percentage in full as they tend to run to a lot of numbers (eg according to Mwango Capital’s tally, William Ruto actually received 50.4898850892%). When you only present the number to two decimal places to make it easier for others to read, you round it, moving the final number up if the number that comes after it (i.e. the third number after the decimal point) is above five and not. This means that the sum of the shortened numbers (i.e. the number presented to two decimal places) often does not add up to 100% exactly. Sometimes it is a bit lower, sometimes it is a bit higher. This does not indicate rigging, it is known as a “rounding error”, and it  only comes about because the results have been presented to the public in a way that makes them easier to digest.
 
In other words, the four commissioners appear to have been misled by their poor grasp of maths into thinking that a rounding error makes the results a “mathematical impossibility”. As the full addition, with no rounding, reveals, when you sum the full numbers they actually add up to 100%.
 
Poor maths also undermines their next point. The four commissioners proceed to argue that the fact turnout was 100.01% means that there were 142,000 additional votes that have somehow been hidden. This is two misunderstandings in one. First, as we have just explained, there were no additional votes – this was just a rounding error. Second, 0.01% of 14.2 million would be 14,200, not 142,000. At this point, the presser becomes rather embarrassing as it becomes clear that the “Serena” commissioners statistics are far less convincing that the IEBC’s themselves.
 
There is also an important point about optics worth making here. The initial presser at the Serena by the four commissioners appeared to take place just as Azimio leaders walked out of Bomas, where the results were being announced. The second presser seems to be intentionally scheduled so that it takes place immediately before Raila Odinga takes to the stage at the Kenya International Conference Centre to reject the results and claim that they should be considered null and void. The sequencing is so tight and well done, that the statement of the Serena 4 is even broadcast live to the audience waiting for Raila to take the stage at KICC.
 
This creates the strong perception that the Serena 4 are being encouraged and coordinated by Azimio leaders, rather than simply operating under their own steam. In turn, this raises serious questions about the motivation behind the accusations they have raised – especially in light of the fact that some of the accusations (it is important to note that their statement included further claims that I have not yet been discussed in full in this analysis) were flawed, and no evidence was presented.
 
9. Continuing this thread, a lot has been made (including on by some on my twitter threads) about the Chair of IEBC stating a KIEMS kit turnout of 65.4%. On this basis the likes of @WehliyeMohamed have suggested there are votes missing from the count. But this is wrong … …while it would be true that the official turnout would be too low if the KIEMS kit figures had shown 65.4%, the Chair of the IEBC quickly corrected his statement, which was a verbal mistake.
 
The actual figure was lower and so there is no inconsistency with the final turnout figures. Given that this has now been pointed out numerous times on social media, with the video of the Chair providing the correction widely shared, to continue to make this claim is effectively to be sharing misinformation. This is not only unhelpful but dangerous.
 

Kenya 2022 Election Reports and Materials

ELOG (domestic observation group)

ELOG Press Statement(Final-10th Aug 2022) 

African Union/COMESA Election Observation Mission

AU-COMESA-ELECTION-OBSERVATION-MISSION-TO-KENYA-PRELIMINARY-STATEMENT

European Union Election Observation Mission

EU EOM Kenya 2022 Preliminary Statement 11 August 2022

National Democratic Institute/International Republican Institute Report

NDI-IRI Report

East African Community Election Observation Mission

EAC Election Observation Mission

Angaza Movement

PRESS RELEASE BY THE ANGAZA MOVEMENT

Kenya Nation Commission for Human Rights Press Statement

KNCHR_Press_Statement-2022_General_Election_Polling_Process_Updates

The Intergovernmental Authority on Development (IGAD) in Eastern Africa Election Observation Mission

IGAD Election Observation Mission

Nic Cheeseman is the Director of the Centre for Elections, Democracy Accountability and Representation (CEDAR) and the Professor of Democracy at the University of Birmingham. He is also the founder of Democracy in Africa and the author of How to Rig an Election.

One thought on “Fact checking claims of rigging in Kenya’s 2022 presidential election + all observation reports”

Join in the debate... let us know what you think!

Related News