CODEO’s PVT estimates, the electoral computational errors, and matters arising


The Coalition of Domestic Election Observers (CODEO) held a press conference on Thursday, December 10, 2020 to present its Parallel Vote Tabulation (PVT) estimates for the 2020 presidential elections following declaration of the official results by the Electoral Commission of Ghana (EC) on Wednesday, December 9, 2020.

The EC, by law, has the mandate to declare election results, and consistent with CODEO’s protocols and international best practice, the Coalition released its estimates after the EC’s formal declaration of the presidential election results. CODEO has followed this practice since 2008 when it first introduced the PVT methodology in its comprehensive election observation.

The PVT or sample-based observation methodology was used for both the first round and run-off presidential elections in 2008 and has been deployed in all subsequent elections (i.e. 2012, 2016, and 2020) to independently verify the integrity of the official results for the presidential elections as declared by the EC.

The release of CODEO’s 2020 PVT estimates has generated some public interest and reaction, particularly following the EC’s press release on December 10, 2020 in which the Commission corrected certain errors in its previously announced official declaration. In light of this development, CODEO wishes to address questions and misconceptions about its PVT statement. 

Computational errors by the Electoral Commission

In its press release on December 10, 2020, the EC stated that “The Chairperson of the Electoral Commission inadvertently used 13,433,573 as the total valid votes cast.

The total valid votes cast is 13,119,460.” CODEO finds these computational errors on the part of the EC very unfortunate, as they create room for doubt about the integrity of the presidential election results declared by the Commission on December 9, 2020 and other disputed parliamentary results. CODEO is of the view that, given what is at stake, the EC must go beyond passive publication of a press release to hold a public event to address the underlying issues with respect to the computational errors identified, so as to allay any lingering concerns and doubts. In addition, the Coalition urges the EC to ensure that source documents signed by the political parties, which support the instrument of declaration, are consistent with the tally sheet for the collation of the national results for the presidential candidates. 

Do the EC’s computational errors affect CODEO’s Parallel Vote Tabulation (PVT) estimates?

CODEO wishes to emphasize that the post-declaration errors detected and corrected by the Electoral Commission do not in any way affect the Coalition’s PVT estimates as shared on Thursday, December 10, 2020.

The PVT is based on polling station data submitted by CODEO stationary Polling Station Observers who witnessed the counting and declaration of those results, having observed the processes of voting during the course of the day.

CODEO has absolute confidence in its PVT estimates, reaffirms the results that it has already published, and assures the public that the results declared by the EC closely match the Coalition’s PVT estimates.

The final results released by the EC after correcting the detected errors in the original declaration still fall within the confidence range of CODEO’s PVT estimates, and thus further provides CODEO the confidence to reaffirm its position on the PVT estimates (Refer to CODEO’s Press release on the PVT). CODEO wishes to use this opportunity to further explain the PVT methodology. 

PVT methodology and how it works

The PVT is an advanced and scientific election observation technique that combines well-established statistical principles and Information Communication Technology (ICT) to observe elections.

The PVT involves deploying trained accredited Observers to a nationally representative random sample of polling stations. In essence, the sampled polling stations reflect the characteristics of polling stations at the national, regional and constituency levels due to the application of stratification in the sampling process, thereby making the sample nationally representative.

On Election-Day, PVT Observers stationed at the polling stations observe the entire polling process and transmit reports about the conduct of the polls and the official vote count in real-time to a central election observation database, using the Short Message Service (SMS) platform.

The PVT technique has been deployed successfully around the world to promote electoral integrity and help defend the rights of citizens to vote and to protect such rights when exercised. The first PVT was conducted in the Philippines in 1986. Since then, the methodology has been deployed in election observation in Bulgaria, Chile, Croatia, Guyana, Albania, Bangladesh, Belarus, Indonesia, Georgia, Macedonia, Slovakia, Montenegro, Nicaragua, Panama, Peru and Ukraine. In Africa, the methodology has been deployed successfully in Madagascar, Malawi, Mozambique, Uganda, Burundi, Sierra Leone, Zambia, Ghana, Nigeria, Zimbabwe, Burkina Faso, Côte d’Ivoire, Kenya and Tunisia.

Unlike pre-election surveys (which track voter intentions) or exit polls (which track voter re-collections), the PVT is based on the actual behavior of electorates in an election. In other words, PVT Observers do not speak to voters about the choices they made at the polls to predict the outcome of an election.

Rather, after observing the entire voting and counting process, PVT Observers record the official vote count (primary data) as announced by election officials at polling stations and rapidly transmit this information to an observation center via coded text message. Consequently, the PVT data is collected directly from polling stations. 

PVT in the 2020 elections

On election day, CODEO deployed 4000 trained stationary Observers to 4000 randomly selected nationally representative sample polling stations across the 275 constituencies in the country. Out of the 4000, 1502 Observers were designated as PVT Observers and were stationed at 1502 polling stations selected for the PVT. The Observers observed the polling processes from arrival, set-up, voting, end of polls, counting and declaration of results. It should be noted that the PVT estimates of the election results were conducted for only the presidential elections, as has been the standard practice of CODEO since it started using the PVT in 2008.

Apart from the 4000 stationary Polling Station Observers, CODEO also deployed over 300 roaming Observers and 275 Constituency Collation Center Observers. 

Collation center challenges

Media reports and reports from CODEO Collation Center Observers indicated that there were some challenges with collation center proceedings. This was the case in constituencies such as Techiman South, Savelugu, Sefwi Wiawso, Jomoro, Dormaa West, Zabzugu, Assin, Upper Denkyira East, and Sene West. Some people have opined that the incidents in these constituency collation centers should have had a bearing on the PVT findings. On the contrary, as pointed out earlier, CODEO PVT estimates are based on results declared at polling stations from where observers submit reports after results have been endorsed by presiding officers and candidate agents present and announced publicly. PVT estimates are not based on collation center data. 

Unwarranted attacks on CODEO’s Acting Chair

CODEO is dismayed by the unwarranted online attacks on its Acting Chair, Sheikh Armiyawo Shaibu, who also happens to be the spokesperson for the Office of the National Chief Imam. CODEO is a coalition made up of 42 organizations, including professional bodies, trade unions, faith-based organizations and civil society organizations that have worked together to advance Ghana’s elections and democracy since the founding of the Coalition 20 years ago. Sheikh Armiyawo Shaibu has been a longstanding member of the 16-member Advisory Board of the Coalition until his appointment as Acting Chair in 2019 following the passing of CODEO’s Co-Chair Justice V.C.R.A.C. Crabbe and the retirement of its second Co-Chair Prof. Miranda Greenstreet.

The views expressed by Sheikh Armiyawo Shaibu on the PVT findings do not represent the view of the Office of the National Chief Imam or his personal view, but rather the collective decision of the CODEO Advisory Board representing the 42-member organizations of CODEO. CODEO condemns the attacks on the person of Sheikh Armiyawo Shaibu, who presented the PVT estimates in his capacity as the Acting Chair of CODEO. 


CODEO is thankful for the support and confidence the public has reposed in it since it was founded in 2000 and would like to assure the public that the PVT methodology is a credible and tested confidence-building tool for independently assessing the accuracy and integrity of official results declared by an election management body like the EC. The PVT has been a valuable component of CODEO’s observation work. Indeed, it has been used as a reference document by all parties as well as the EC since 2008. In 2012 it was tendered as part of the evidence affirming the accuracy of the 2012 results declared by the Electoral Commission. CODEO takes its observation activities seriously and undertakes them with the utmost professionalism and objectivity.

The Coalition again wishes to express its deepest condolences to the families of those of our compatriots who lost their lives during the electoral process. We also reiterate our call to the National Election Security Task Force (NESTF) to investigate swiftly all incidents of violence during and after the elections and ensure that those found culpable are duly prosecuted. Lastly, CODEO urges all aggrieved parties to resort to peaceful and lawful avenues and processes to seek redress.

Leave a Reply