Kenya 2022: Lies, damn lies, and statistics

As in 2013 and 2017, disputes over the electoral process mean that the losing candidate – Raila Odinga and his Azimio coalition – immediately rejected the announcement and launched a legal battle.

It is too early to say which way the court will decide. There are three main reasons for this. One is that it is unclear what burden of proof judges will require to nullify the polls, and how heavily this court – now led by a new chief justice – will weight procedural inconsistencies as opposed to evidence of ballot fraud. Another is that a large number of petitions and affidavits have been submitted, and it is not yet clear which will stand up to scrutiny.

The third reason it is hard to second guess the court’s verdict is rather different and speaks to one of the greatest challenges facing democracy in our time: so much misleading information has been shared that it has become hard to separate the truth from the fake.

Even those of us well seasoned in analysing rigging claims have struggled to cope with the bewildering monsoon of misinformation that has rained down on social media. When you wake up every morning to a new set of messages about how the process was manipulated, it can be hard not to believe there must be something in them – even as every claim you investigate turns out to be unconvincing.

The challenge facing the court – and everyone else invested in the process – is how to thoroughly investigate such a wide range of claims in the constitutionally stipulated timeframe of just fourteen days.

The monsoon of misinformation

While the announcement of Ruto’s victory triggered a flood of fake news, misinformation had been part of the election campaign from the beginning.

Politicised keyboard warriors used a variety of strategies to try and discredit rival candidates, while Mozilla warned that platforms such as TikTok were “teeming with political disinformation”.

When it became clear that Ruto would be announced the winner, these systems went into overdrive. The claims of rigging that followed can be broken down into three overlapping categories. First, Azimio aligned commentators alleged that the results had to be false because there was no way their candidate could not win.

Mutahi Ngunyi, for example, tweeted that the turnout figures “BEATS scientific LOGIC”. One of the main characteristics of these claims has been the tendency to use ALL CAPS in place of evidence or careful reasoning.

Second, anonymous messages circulated on WhatsApp and Facebook alleged substantial discrepancies in the results, either between the turnout announced by Chebukati and the final figures, the turnout in different election races, or the results from different stages of the tallying process.

The key characteristic of these claims was that the provision of specific numbers gave them an air of scientific authority, even though it often unclear where the data had come from. Moreover, such claims hit home because significantly more votes being cast for the presidential election than for other races has long been recognised as a potential indicator of ballot box stuffing, given that Kenyans vote in all six elections simultaneously.

While the jury is still out on some of these allegations, it is now clear many were fabricated. The idea that there was a massive difference between presidential and gubernatorial turnouts across the country is contradicted by Azimio’s own petition, which claims a much smaller variation than was initially circulated on social media. Meanwhile, Charles Hornsby’s analysis, which finds that the match between different races was “better than previous elections”.

The third main type of misinformation was the set of conspiracy theories that often fly around Kenyan elections from both sides. Ahead of the elections leaders worried – not unreasonably – that the “deep state” would manipulate the polls for their rivals.

This kind of language led to an escalation in claims that the entire set of results had been fabricated – despite no evidence of discrepancies between the IEBC’s digital public portal and the hard copies given to party agents at polling stations – and even that foreign governments had rigged the election in favour of Ruto.

Despite being outlandish, the sheer number of these accusations, in the context of elections that have regularly been unfree and unfair, has exacerbated a major national divide, in which hardcore supporters of both Odinga and Ruto believe that there is a concerted attempt to unfairly deny their man victory.

Precisely because these claims are so eye catching and colourful, they have been shared many more times than the findings of the domestic Elections Observation Group, whose parallel vote tabulation based on a sample of polling stations projected that Ruto secured a greater share of the vote (50.70% as opposed to 50.49%) than was announced by the IEBC. The consequences for national unity and the legitimacy of the next government – whoever leads it – are likely to be profound.

The drivers of distrust

Misinformation thrives, of course, when key institutions cannot be trusted and when it is repeated by respected figures. In the context of the Kenyan elections, many citizens went into the campaign with limited trust in Chebukati because he had presided over the 2017 elections that was nullified by the Supreme Court.

Their trust was further eroded – some might say exploded – when, just as he was about to read out the result, four “rebel” commissioners gave a rival “presser” saying they could not stand behind the results.

It did not matter that when the four commissioners subsequently provided details of their concerns it transpired that they rested, in part, on a mathematical misunderstanding so basic that it called into question both their capacity and their motivations. The sight of the IEBC once again at war with itself was enough to give credence to the claims the Commission had been politically captured and had fabricated the entire process.

The ability of the public and the court to separate the wheat from the chaff has also been complicated by the fact that many respected figures who might previously have served as neutral arbiters committed themselves to particular candidates. Boniface Mwangi, the photojournalist civil society activist who rose to fame after capturing striking images of post-election violence in 2007/8 backed Azimio.

Following the polls he tweeted “We have won this election, numbers don’t lie” at a point when no independent tally of the votes had been completed.

John Githongo, the country’s highly respected former anti-corruption tsar, also backed Odinga’s campaign, and provided an affidavit in support of the Azimio petition.

Githongo’s evidence states that a young man confessed to him that he was part of “a 56-member team of hackers employed to manipulate the forms 34A submitted in the IEBC server.”

Few would question Githongo’s veracity, but his understanding of electoral technology and the credibility of the story he had been told were called into question when it was revealed that some of the “screengrabs” annexed to his affidavit had themselves been falsified.

 

Fact checking these kinds of claims is one way to challenge misinformation, but the tidal wave of fake news can also disempower these efforts.

Along with others providing a critical commentary on rigging allegations, such Charles Hornsby and Economist journalist Adrian Blomfield, I was targeted by a coordinated disinformation campaign designed to create the impression that my analysis should be discounted because it was being paid for by Ruto’s United Democratic Alliance (UDA). The misinformation was so badly designed it was clear it was fake to anyone who looked at it carefully, but as ever in highly emotive contexts, not everyone had the time or inclination to do so.

Focussing on the real issues

Given how sensitive the election outcome is, and the extent to which any ambiguity can be manipulated, it is important to end with an explicit disclaimer: nothing in this article should be taken to imply that the election was free and fair, that the petitions brought against the result do not have merit, or that election results – and claims of digital manipulation – should not be forensically scrutinised.

Indeed, one of the most important consequences of the monsoon of misinformation is that it has distracted from the many limitations of the elections. I have spent the last ten years seeking to expose and prevent electoral manipulation, and as I have documented at greater length elsewhere, the 2022 polls had a number of weaknesses that need to be discussed and dealt with in future elections.

While the IEBC made major improvements in terms of making almost all polling station results forms publicly available on a digital platform, the inconsistent use of the manual register means that the election was not administered the same way in all polling stations.

Problems with the ballot papers led to the governorship elections being cancelled in Mombasa and Kakamega, which is likely to have reduced turnout in areas where Odinga won the presidential election. There is also some legal ambiguity over what Chebukati should have done when the four rebel commissioners refused to be part of the declaration of the results.

All these issues – and more besides – deserve careful scrutiny, and the court may yet decide they are enough to merit the nullification of Ruto’s victory, especially given that the margin of victory was so small.

This process would have been far more straightforward, however, as would the task of fostering national reconciliation around a common understanding of what has actually happened, if it was not taking place against the backdrop of a whirlwind of fake news. Sadly, whatever the court decides is unlikely to convince both sides. If the court confirms Ruto’s victory, many Azimio supporters will continue to believe the claims of rigging.

If it orders a “fresh” election, many Ruto supporters will feel that the “establishment” refused to allow their man to win despite a credible process.

Either way, the outcome is likely to be greater distrust in not only the IEBC but in the possibility of free and fair elections in Kenya, storing up further problems for the next poll, whenever it comes.