Can domestic observers serve as impartial arbiters?: evidence from Zambia’s 2021 elections

ABSTRACT
 With international election observation subject to increasing criticism, this article evaluates how effectively domestic observers can play the role of impartial arbiters relative to their international counterparts. It reviews academic arguments about the strengths and weaknesses of domestic and international observers, with a focus on 1) their methodologies, resources, and reporting practices; and 2) their credibility. It presents a case study of Zambia’s 2021 elections, detailing the major observation missions and their activities. It then compares the media coverage and popular perceptions of domestic and international observers, showing that many Zambian citizens still have reservations about the partiality of domestic election observation initiatives. The article concludes by theorizing the conditions in which domestic election observation can 1) provide impartial evaluations of election conduct; and 2) be perceived as trustworthy sources of these evaluations.


Introduction
In recent years, domestic election observation initiatives (DEOIs) have become increasingly prominent. Since the first major domestic election observation mission was staged by The National Citizens' Movement for Free Elections in the Philippines in 1982, 1 DEOIs have proliferated to over half the countries in the world. 2 Arguments regarding the relative merits of domestic and international observers were central to 1990s academic commentary on election observation. 3 One proposal from this literature was that domestic and international observers should work together more closely, so as to take full advantage of their relative strengths. More recently, the idea that DEOIs should eventually assume full responsibility for the elections in their own countries has become more common. For example, David Caroll, Director of the Democracy Program at The Carter Center, has argued that "the future is domestic observationit has to be, and everyone knows that". 4 A range of international governmental and non-governmental organizations now facilitate domestic election observers. 5 As the practice of domestic observation has expanded, criticism of international election observation missions (IEOMs) has been mounting. Since election observation became more widespread in the period following the Cold War, international observers have been subject to scrutiny, mostly relating to accusations of bias and incompetence. These criticisms have accelerated in recent years, due in part to the emergence of new concerns such as the role of "shadow observers" in legitimising authoritarian regimes and problems in addressing the digital components of some electoral systems. 6 In the African context, the perception that international observers' verdicts were "overturned" by court judgements in Kenya (2017) and Malawi (2019) has been particularly damaging. 7 In the aftermath of these events, Khabele Matlosa argues that "international election observation is now in crisis". 8 This makes it worthwhile to consider whether domestic initiatives can already perform the major observation activities that are the focus of international missions.
This evaluation is not a simple undertaking. Domestic election initiatives vary hugely, 9 and they are not all equally well-established, competent, or impartial. Their ability to be effective also depends on the context in which they operate, with some countries providing a more conducive environment than others. 10 Election observers also have several functions. They can deter fraud, 11 provide technical recommendations to electoral management bodies, 12 and provide public assessments of election quality. This article focuses on this final role by considering whether domestic election observers can be effective impartial arbiters of the elections in their countries. This high-profile function is critical both in creating confidence in well-performing electoral systems and in mobilizing protest after flawed elections. 13 Using a case study of Zambia's 2021 general elections, this article will evaluate the performance of DEOIs as impartial arbiters of election quality relative to that of IEOMs. Although Zambia had electoral transitions in 1991 and 2011, there was much speculation in advance of the 2021 elections that they might be manipulated. 14 Zambia, where coalitions of domestic civil society have been active in observation since the 1991 elections, therefore makes an illustrative example of how DEOIs perform in situations where there are prominent concerns about electoral quality. The article will firstly examine DEOIs' methodologies, resources, and reporting practices, to determine if they can make accurate pronouncements about election quality. Secondly, it will examine whether DEOIs are perceived to be credible, with a particular focus on the media and public opinion. Even if domestic observers are producing accurate evaluations of election quality, these findings will be of limited impact if they are not circulated by the media or if citizens believe the DEOIs to be biased or incompetent. Therefore, media and popular perceptions of DEOIs will be established, alongside an evaluation of the media coverage different types of observers receive.
The article begins by defining DEOIs. This section will note that most observation missions have both domestic and international elements but argues that the difference between domestic and international observers is generally well understood within the election observation community. Next, it reviews the existing literature on domestic and international election observers with a specific focus on observation activities and credibility. The article will then introduce the Zambian case study and the research methods used in data collection. Next, the findings will be presented, showing that some Zambian DEOIs were able to perform the technical work of election observation at a high standard, but they struggled to create trust among the media and public. The article concludes by theorizing the conditions in which domestic election observation can 1) provide impartial evaluations of election conduct; and 2) be perceived as trustworthy sources of these evaluations.

Defining domestic election observation initiatives
Following Max Grömping, this article defines DEOIs as "civil society groups that scrutinise elections in their own countries", 15 which are "non-state, non-profit, non-partisan and non-media". 16 As outlined in the Declaration of Global Principles for Nonpartisan Election Observation and Monitoring by Citizen Organizations, their core functions are witnessing and reporting on electoral developments, timely analysis of findings, offering appropriate recommendations, and advocating for improvements in legal frameworks for elections. 17 Generally, DEOIs are either organizations specifically created for these purposes or coalitions of pre-existing organizations. 18 A closer inspection of observation activities suggests that a neat compartmentalization into categories of domestic and international is not so straightforward. Organizations that meet the above definition of DEOIs frequently have strong international linkages. They tend to be funded by international donors, which means that their activities, at least in part, are tailored to external processes. Additionally, international experts often act as advisers to DEOIs to assist with capacity building, an area in which the National Democratic Institute (NDI) has been particularly active. 19 In a similar vein, some IEOMs draw on domestic civil society and local journalists to inform their findings. They also hire host country staff to perform roles that include providing expert advice on politics or legal frameworks and local assistance and interpretation for observation teams. 20 Nonetheless, our interviews with members of the election observation community (see below) suggest that there is agreement on which organizations are domestic and which are international. Within this consensus, the origins of the organizations' funding are not considered and the presence of some specialist staff, who do not match the organizations' overall designation, is overlooked. The key determinants are whether management of the organizations and the actual work of observation is done by citizens of the country holding the election.

Insights from the literature
When compared to international observers, DEOIs have been given relatively little attention in academic writing. Nonetheless, it is possible to identify several recurrent arguments regarding the ability of domestic observers to act as impartial arbiters relative to their international counterparts. These arguments, outlined below, centre on their ability to produce accurate and impartial findings, their ability to have these findings publicized, and popular perceptions of their credibility.

Producing accurate and impartial findings
There are several arguments in the literature that suggest that domestic observers can produce better quality findings than their international counterparts. It has often been argued that DEOIs are able to field far larger numbers of observers than IEOMs. 21 Their better numerical and geographical coverage still gives them an advantage over IEOMs in generating empirically robust data, and allows them to perform additional tasks, such as parallel vote tabulation (PVT). 22 DEOIs also have a significant advantage when it comes to observing longer-term electoral processes as they are present for the entire electoral cycle. 23 This also allows them to make a range of additional contributions to democracy promotion, including observing local elections, lobbying for electoral reform, and engaging in voter education. 24 In contrast, two long-standing criticisms of international election observers are that they are not present for long enough in the countries which are hosting elections, and that their methodologies are overly focused on election day. 25 Although these differences are often asserted, the academic literature rarely provides empirical data to support the claims.
Domestic observers have additional advantages in terms of their greater understanding of local context and languages. International observers have been criticized for failing to adapt to the settings in which they operate, or for not fully understanding local political dynamics. 26 IEOMs can also become dependent on translators, or even fail to interact meaningfully with certain population groups, in the linguistically diverse countries in which they operate. In contrast, as Patrick Merloe describes, domestic missions are "conducted in local languages and with knowledge of a country's political culture". 27 On the other hand, international observers are often credited with greater training and more advanced methodologies than their domestic counterparts. For example, in 1997, Neil Nevitte and Santiago Canton argued that international groups "have a large pool of experienced observers", as opposed to domestic groups that "lack access to a reservoir of citizens with direct experience in election monitoring". 28 Other problems that have affected citizen observer groups in the past include a lack of organizational capacity and an inability to properly vet recruits. 29 They have also, at times, been constrained by poor leadership and, particularly in the case of coalitions, internal friction. 30 However, the quality of many DEOIs has improved, particularly when they have had the experience of observing multiple elections in their countries. Since 2009, there has also been a Global Network of Domestic Election Monitors (GNDEM), which has facilitated peer-to-peer learning and promoted best practices. Increased funding for domestic civil society groups and training and support from international organizations have also proved beneficial.
Despite these improvements, there are still risks that domestic observers are either partisan or become entangled in partisan affairs. In some cases, DEOIs can either be created specifically to back a certain party or their key personnel can have close relationships with politicians. 31 Even when domestic observers would prefer to be impartial, they can still be intimidated by authoritarian governments. 32 Notably, domestic observers do not simply leave the country after the election, and their families can also be targeted by vindictive authoritarian regimes. 33 All these factors may constrain DEOIs' ability to collect and publish accurate findings. However, their statements and reports are often clearer and less equivocal than those of international observers. 34 There is also the possibility that domestic observation initiatives can hold pro-opposition biases. Many key personnel in DEOIs are drawn from civil society organizations, and, as a result, they are often already involved in disputes with incumbent regimes, particularly in cases where governments have been attempting to constrain civic spaces.

Public attention
As Grömping argues, public attention is critical for observer groups who are hoping to influence public opinion. 35 One way of seeking this is by soliciting mass media coverage, something that can be challenging for DEOIs. Grömping's study of over 1,000 domestic observation and advocacy groups shows that "71% of groups receive not a single mention in the domestic press, and 72% remain unnoticed in international news". 36 Although the relative levels of coverage that IEOMs and DEOIs receive is rarely quantified in the academic literature, Sarah Bush and Lauren Prather's study of the 2014 Tunisian elections suggests that the media references a wide range of observer groups, rather than focusing on a select few. 37

Credibility
Whether DEOIs are publicly perceived as credible is also a crucial issue, because if they are to play a role of impartial arbiter it is important that their findings are taken seriously. Some statistical insights into the credibility of observers are provided by Nic Cheeseman, Gabrirelle Lynch, and Justin Willis. 38 They asked citizens in Ghana, Kenya, and Uganda if they agreed that certain types of observers should be present during elections. Although they found generally high support for the presence of observers, the percentage of respondents who thought that "domestic observers" should be present was still around 25 points higher than that of "American observers" or "European Union observers" in all three countries, with "African Union observers" landing in the middle. This led them to conclude that perceptions of observation groups "cannot be separated from [citizens'] perceptions of the states from which they originate". 39 As Sharon Lean has argued, there are additional conditions that affect how domestic initiatives are perceived. In particular, she points to domestic political polarization as a dynamic that can undermine their credibility, concluding that "international democracy assistance providers should reassess the desirability of targeting domestic civil society in some contexts". 40 It is also worth considering Nicholas Kerr's finding, from a study in Nigeria, that IEOMs were more likely than DEOIs to create confidence in national-level results, but that domestic observers were more likely to create confidence in local-level vote counts. 41 This suggests that citizens may perceive domestic and international observers as playing separate but complimentary roles.
It should not be assumed that all citizens have good knowledge or strong opinions about observers of any kind. As Bush and Prather's case study of Tunisia suggests, popular evaluations of the quality of observation groups may not match more technical, expert evaluations. 42 Grömping also notes that "It is […] far from certain that people notice election watch groups' messaging at all". 43 Indeed, it is possible that many citizens may not even know that their country's elections are being observed, let alone by whom.

Zambia's 2021 elections
Zambian voters went to the polls on 12 August 2021, to vote in Presidential, National Assembly, and a variety of local government elections. At the top of the ticket, the main contenders were incumbent President Edgar Lungu of the Patriotic Front (PF) and Hakainde Hichilema of the opposition United Party for National Development (UPND). Amid a cost-of-living crisis and reports of rising corruption, there was pre-election speculation that Lungu might resort to some form of electoral malpractice to remain in power. 44 Ultimately, he was unable to secure victory, with Hichilema winning 59.38 percent of the vote, compared to Lungu's 38.33 percent. 45 As the results came in, there was some uncertainty as to whether Lungu was prepared to step down, particularly after he described the elections as "not free and fair". 46 Following diplomatic interventions from domestic and international actors, he did eventually concede defeat. President Hichilema was inaugurated on 24 August 2021.

"Observers" and "monitors" in Zambia
In Zambia, every person formally observing the election must be accredited by the Electoral Commission of Zambia (ECZ). It uses the terms "monitors" and "observers" to distinguish between domestic and international actors who engage in formal observation. Its website states: "The distinction between Observers and Monitors is that Observers are drawn from the international community while Monitors are drawn from the local civil society". 47 The descriptions of their roles offered by ECZ are similar. 48 However, there is an important difference, as "An observer may not bring to the attention of the Presiding, Returning Officer or a member of the Commission any discrepancies that may arise during the conduct of an election while a monitor may do so". 49

Research methods
Data was collected in Zambia over five months between July and November 2021, covering both pre-and post-election periods, and took place as part of a broader research project on popular perceptions and media representations of election observers. 50 The data collection had three empirical components. Although they ran concurrently, each component was designed to address specific issues which will now be outlined.
The first component focused on the observation missions. It began by identifying the organizations involved in observation of the 2021 elections through the ECZ, their own releases on the internet, and media coverage. Detail about the methodologies, resourcing, and reporting of the organizations was then obtained from their publicly available information on the internet and in the media, and, where possible, directly from discussions with the organizations themselves. During this process, semistructured qualitative interviews were conducted with the chief observers from the three largest DEOIs and high-level representatives from eight of the IEOMs. 51 This formed part of a series of wide-ranging qualitative interviews with observation practitioners, politicians, and academics, conducted by the authors along with Marja Hinfelaar, O'Brien Kaaba, and Kalonde Mutuna from our partner organization, the Southern African Institute for Policy and Research (SAIPAR). 52 Together these methods provided data for a spreadsheet that captured the following information about each organization: experience in election observation, sources of funding, composition and numbers of personnel on the mission, deployment dates, and core activities. The authors also attended the launch events of the various preliminary statements, during which they observed how the media and public engaged with the organizations' messages. Lastly, a content analysis of preliminary statements, subsequent statements, and final reports was conducted.
A second component of data collection focussed on the media. To evaluate the coverage that various election observation missions were receiving, a media tracker was created to log instances in which election observers were mentioned in reporting. It was established one month before the elections and continued until stories mentioning election observers stopped appearing, which occurred two weeks after the elections. Two research assistants from SAIPAR -Charles Simwanza and Kalonde Mutunawere tasked with reviewing five domestic newspapers and specific programmes on seven television stations and five radio stations. 53 These were selected in consultation with SAIPAR (who have experience running media trackers in Zambia) based on their popularity and reach. In each category, both public and private media were included. Rather than having to cover everything live, Mutuna and Simwanza were often able to review television and radio programmes that were posted on stations' websites and social media accounts after initial broadcast. They also used key word searches to find international media coverage of the Zambian elections online, which was reviewed to find instances in which observer missions were referenced. 54 Information gathered from this processincluding the date, the name of the outlet, the outlet's perceived political inclination, the broader topic of the report, the missions covered, and the sources citedwas collated on a spreadsheet. This enabled the authors to analyse election observation coverage across a variety of media outlets and identify several key patterns. Shortly after the elections took place, project partner O'Brien Kaaba also conducted a series of fifteen interviews with editors and journalists from a range of Zambian newspapers, radio stations, and television channels about their views on election observers and how they are covered in the Zambian media. They were purposively sampled from his established connections in the industry to cover a range of outlets and roles.
In the third component of the data collection, 163 non-elite Zambians were interviewed to assess popular knowledge of election observation, preferences for observer conduct, and which sources convey information about the activities of election observation missions. 55 Macdonald and Mutuna conducted this research in four different locations, each of which was visited before and after the elections. These case studies were selected to address both urban/rural and pro-incumbent/pro-opposition dynamics. The locations included two urban areas (Mazabuka and Kitwe) and two rural areas (Siavonga and Petauke). 56 Of these, Kitwe and rural Petauke had returned PF MPs during the 2016 elections, while Mazabuka and rural Siavonga had elected opposition candidates. In each area, the researchers aimed to interview 25 voting-age citizens in the pre-election period and 20 more in the post-election period. 57 They used a mixture of opportunistic and reference sampling, while maintaining a gender balance in responses and ensuring that respondents covered a wide range of ages and occupational categories. In the pre-election period, the questions mostly focused on historical experiences of observation as well as general preferences about who should observe and what tasks they should undertake. In the post-election period, the interviews focused more on evaluations of the observation that had occurred at the recent elections, and the ways in which information about observers had circulated.

Observer groups, methodologies, resources, and reporting practices
Several Zambian civil society and religious organizations were involved in domestic monitoring. The most prominent of these was the Christian Churches Monitoring Group (CCMG), a coalition launched in 2014 with four member organizations. It worked in partnership with the National Democratic Institute (NDI), which used USAID and FCDO funds to provide technical assistance. CCMG also received funds for the pre-and post-election periods from the EU through Diakonia. 58 On election day they had over 1,600 monitors and mobile supervisors covering all constituencies in the country. With technical assistance from NDI, they also conducted a PVT exercise. Another organization with previous experience in election observation was the Foundation for Democratic Process (FODEP), which was established in 1992 as a successor to the Zambia Election Monitoring Coordinating Committee, and undertook monitoring funded by ActionAid Zambia and OSISA. 59 On election day, FODEP deployed 1,450 monitors across eight of the ten provinces. A large-scale monitoring mission was also conducted by the recently-formed Governance, Elections Advocacy Research Services (GEARS) Initiative, with funding from SIDA via ActionAid Zambia. 60 It provided 11,787 monitors, covering all constituencies in all ten provinces. Good Governance Zambia and other local organizations, which were working with the SADC Good Governance and Election Monitoring group, were reported to have accredited some 3,000 observers for election day, although there is little open-source information about their activities. 61 A range of other groups also received accreditation for smaller numbers of monitors.
Several IEOMs were also present in Zambia during 2021. The African Union (AU) deployed a short-term mission to Zambia comprised of 30 observers, four electoral experts, and a technical team led by former Sierra Leonen President Ernest Bai Koroma. The European Union (EU) deployed 75 observers on its election observation mission, with an 11-person core team that arrived six weeks prior to the elections, 32 long-term observers who followed two weeks later, and several locally-recruited shortterm observers from the Lusaka-based EU diplomatic community who covered voting and tallying. Even these larger IEOMs could not match the DEOIs' geographical coverage. EU observers were able to visit 423 of the country's 12,152 polling stations, and the AU managed to cover 250 polling stations. 62 Other international organizations which deployed missions included The Commonwealth Observation Group (COG), the Common Market for Eastern and Southern Africa (COMESA), and the Bujumbura-based International Conference on the Great Lakes Region (ICGLR), who were working with the Lusaka-based Levy Mwanawasa Regional Centre for Democracy and Good Governance. The COG maintained a smaller mission of 13 observers supported by a similar number of Commonwealth Secretariat staff who variously arrived a week or two before the elections. It drew its personnel from Commonwealth member countries, with former Tanzanian President Jakaya Kikwete as chief observer. COMESA deployed 42 short-term observers, led by Ambassador Ashraf Gamal Rashed of Egypt, and, like the AU, visited 250 polling stations. The Southern African Development Community (SADC) also convened a "Virtual Electoral Observation Mission" involving "virtual stakeholder consultations" in the ten days leading up to the elections.
Several African NGOs deployed more modest missions, such as the Johannesburgbased Electoral Institute for Sustainable Democracy in Africa (EISA), whose technical mission involved 13 election experts from across Africa, and the Zimbabwe Election Support Network (ZESN), who deployed three personnel on what they described as a "learning observation mission". 63 Other NGOs that were present included The Carter Center, whose Election Expert Mission focussed on key aspects of the electoral process, including the campaign environment, and disinformation and misinformation trends. Some African electoral management bodies also sent delegations, including electoral commissions from South Africa and Kenya, whose parliament also sent a delegation of 33 personnel.
Missions that were able to deploy for longer periods had a much greater ability to observe the electoral process. Therefore, DEOIs with a constant presence, such as CCMG and FODEP, had a clear advantage. Their long-term monitoring included on-the-ground assessments of constituency delimitation, voter registration, voter education projects, candidate nomination processes, campaigning, voting, and counting. 64 CCMG, for example, was able to deploy 330 long-term observers across the country on 8 March and issue statements throughout 2021. 65 In contrast to the longer-term domestic initiatives, shorter-term missions such as those of the African Union (which deployed its observers eight days before the elections) and COMESA (whose observers were deployed three days before the elections) focused on the final days of campaigning, the distribution of polling materials, election day voting, and the counting and collation of ballots. The Carter Center began operating in-country three weeks prior to the elections, while the EU deployed almost a month before election day.
The first post-election preliminary statement was issued on 13 August by CCMG. Although largely supportive of the electoral process, they were not afraid to take a critical stance, especially in their assessment of pre-election matters such as the issuing of national registration cards and the voter registration process. CCMG presented many of its election day findings in quantitative terms based on reports from its nationally representative sample of 1,500 polling stations. It also reflected on wider issues relating to, for example, incidents of violence and internet shutdowns. 66 Although CCMG had completed its PVT by midday on 13 August, it was legally required to hold back on releasing this data until the ECZ declared the official election resultswhich it aimed to do within 72 h of polls closing. 67 The next day (14 August) saw several press conferences from the larger international missions, including the AU, COG, and EU, as well as some brief and uncritical statements given by the COMESA and ICGLR missions. The AU was largely positive about the voting process but outlined both the unlevel playing field during campaigning and the politically-motivated violence of the pre-election period. 68 The COG statement was also fairly positive about election day, but noted allegations of the police's lack of impartiality in the application of the Public Order Act, and added its concerns about the deployment of army personnel prior to election day. 69 The EU was the last of the major international missions to hold its press conference, and its preliminary statement was the most critical. They described "a technically well-managed electoral process marred by unequal campaign conditions, restrictions on freedoms of assembly and movement, and abuse of incumbency". 70 The timing of the IEOM statements was particularly significant, as they served to undermine Lungu's claims that the elections had not been free and fair, which he made on the same day. CCMG were also able to publish detailed data that contradicted Lungu's specific claims that his party's agents were chased away from polling stations in some provinces.
The ECZ released the final election results early in the morning of 16 August. As they were in line with its PVT, 71 CCMG did not need to use their data to publicly challenge the ECZ, which they had otherwise been prepared to do. 72 Ultimately, the PVT was still published to "provide independent non-partisan verification of the accuracy of official results" after they were declared centrally by the ECZ. 73 Even after the ECZ announced the final results, it was still unclear if Lungu would be willing to accept defeat. CCMG also used their PVT to inform the heads of other missions, including Koroma and Kikwete. 74 Although this arguably went beyond the mandate of election observers, the two former presidents were then subsequently involved in the negotiations to persuade Lungu to concede, which he did on the afternoon of 16 August.
Further statements were issued later in the process by FODEP (on 18 August), GEARS (23 August), and The Carter Center (23 August). These also included much criticism, but the major drama of the election had concluded with Lungu's concession. In the months following the election, several missions, including the AU in November and the EU in February 2022, published more detailed reports. As they were released after citizens had moved on from the elections, these final reports generated less public attention than the preliminary statements. However, their level of detail makes them suitable for the ECZ and other stakeholders who are aiming to improve Zambia's electoral processes.
This review of the observation missions and their activities has provided empirical data that supports some of the claims in the literature relating to the advantages DEOIs have over their international counterparts. They provided thousands of monitors and started activities early in the electoral process, both factors that enhanced the validity of their findings, giving them a strong position from which to act as arbiters. CCMG's PVT also provided an important safeguard in case false results were announced by the ECZ. In contrast, the international observers were in Zambia for a relatively short period of time and deployed relatively small numbers of observers. A review of the statements issued within 48 h of the close of polls shows that CCMG were prepared to be positive about some aspects of the electoral process while being as critical as the sternest IEOMs about others. Their report was neither suggestive of incumbent pressure to moderate their findings, nor of a pro-opposition bias that might have prevented them from playing their role as arbiters impartially.

Credibility
If domestic observers are to play their role of impartial arbiters effectively, it is crucial that their findings are viewed as being credible. Otherwise, they will struggle to have an impact. This section will begin with consideration of the media before turning attention to popular perceptions.

The media
At the time of the elections, Zambia's media environment was very polarized, with public media showing a strong pro-incumbent bias and private media more likely to support the opposition. This had an impact on the coverage of election observers. For example, one senior journalist at a state-owned newspaper, explained that at their outlet: [the] ruling party is covered because the public media are considered to be a government mouthpiece.
[…] We usually only cover the opposition when there is something malicious about them. It is from this angle that election observers are covered by government media. If they say something good about the government, then we cover them. If not, then we ignore them. 75 Beyond this, the interviewees made it clear that they have no specific strategy for covering election observers, or even elections more broadly. As former ZNBC news editor Hyde Haguta commented, "It is a fragmented method. There is no actual strategy. It does not exist in Zambian media". 76 This suggests that the media organizations decided whether to cover the activities of observation missions on an ad hoc basis.
The interviewees did, nonetheless, have some strong opinions on the credibility of election observation groups. In this regard, the EU performed particularly well, with six of the journalists and editors identifying them as the best observer group, despite not specifically being asked to rank them. The reasons given for this were a combination of perceived impartiality, a willingness to criticize, and technical expertise. The AU and SADC were viewed as the least credible, as they were seen to be too close to incumbent parties and unwilling to produce firm criticisms of flawed elections. Although they were not universal, some concerns were raised about the quality and partiality of local observers. For example, former Daily Mail (Lusaka) editor Martin Nkolomba offered that "Local observers tend to have people who do it simply for money. Their intentions are not well established. We see them disappear off the scene just after elections. Their reports are not credible". 77 News reporter Fatima Mawere was even more direct in stating: External observers tend to be more credible. They have no interest in the outcome of the elections. They have no relationship with candidates. Local ones are not objective. Many are associated with the ruling party, or they just want the opposition to win. 78 The media's perceptions of election observers have the potential to impact the amount and quality of coverage that different organizations receive. Data from the media tracker will now be presented to see if that was the case.
Most of the international media's coverage on the Zambian elections did not reference election observers at all. However, the media tracker identified 45 international reports that mentioned them. Of these, it was unclear which observation missions were being discussed in 16, as non-specific phrases such as "observers" or "international observers" were used. Where the missions were specifically named, international observers were more prominent than domestic observers, who were only mentioned in two reports. In particular, the AU (14 mentions) and EU (9) were covered most, while other international missions were rarely mentioned. The coverage of all missions tended to be uncritical. These findings broadly support Susan Hyde's argument that international media typically consider IEOMs to be a credible source. 79 However, this should be qualified, since international media coverage is not always clear on the precise missions it draws upon and it may not be focused on observer judgements in cases where the election outcome is not particularly contested.
In domestic media, there was a much more even balance in the prominence of domestic (21 mentions) and international observers (28 mentions). The EU were the most cited (10 mentions), with the rest of the coverage scattered across a range of international and domestic groups (this is similar to Bush and Prather's findings). In Zambia's campaign period, pro-government media did cover some of the Lungu administration's attempts to discredit international observers. This appears to highlight the importance of media bias in determining which stories about election observers are covered. However, in the period immediately after election day, observer statements were generally conveyed uncritically by media from across the political spectrum. These findings suggest that the domestic media's ad hoc coverage of election observation in Zambia was not always driven by evaluations of the credibility of different observer types.

Popular opinion
It should not be assumed that domestic media and the broader public view election observers in the same way. Rather, gathering the opinions of non-elite citizens directly is key to understanding how much they know about election observers and how they might perceive the DEOIs relative to international observers.
It was evident during both pre-election and post-election interviews that many respondents, particularly but not exclusively those in rural areas, had a limited understanding of election observers. Indeed, in several cases, the respondents were completely unaware of them. For example, one elderly woman in rural Petauke explained "This [interview] is the first time I have ever heard of them!". 80 Many other respondents appeared to confuse election observers with other actors involved in elections, something that became clear when they were asked to outline their understanding of the role of observers. They would often describe work that polling station staff, party agents, or voter educators would generally perform. Even among those who had an accurate understanding of the role of election observers, few could remember the name of even one specific election observation group, either international or domestic.
When questions were asked concerning the information about election observers that respondents are exposed to through the media and other sources, it became clear that it is difficult for non-elite citizens to access details of their activities. This was particularly true in rural areas, where respondents tend to have less access to media, but it was also an issue in urban areas, despite many respondents having access to a wider range of social and traditional media. Some of the respondents did hear about the presence of observers, but even during the post-election interviews few had heard the details contained in observer statements and reports. This may be partly due to a lack of controversy surrounding the outcome of the elections, as a more contested result may have led to greater media focus on the observers. It is also possible that a more controversial outcome would have influenced more citizens to intentionally seek information about observer verdicts or would have resulted in more popular discussion about their performance.
Despite not being aware of many specifics, most respondents did display strong preferences about what election observers should be doing and how they should do it. Of particular interest to this article are informants' responses to questions about who they thought should be observing elections and why. In this part of the interviews, respondents were specifically asked about three types of observation group: domestic, observers from other African countries, and observers from outside Africa. The clearest finding was that respondents did not have good perceptions of observers from other African countries. They were often seen to be too close to incumbents or just generally corrupt. For example, one businesswoman in Mazabuka expressed her view that "African leaders want to cling to power and thus will not dissuade their colleagues", 81 while a young man in Kitwe simply stated his view that "in Africa there is too much corruption". 82 Despite these negative perceptions, many of the respondents argued that African observers should be invited anyway, as that is in the interests of transparency and that there are benefits to having multiple perspectives.
There was more variety of opinion concerning domestic monitors and other types of international observers. Respondents were very keen that, at future elections, either one, or both types of observers should be present. Indeed, only one respondent said that she did not want any observers at all to be present. A handful of the 163 respondents felt that local monitors should be solely responsible for observing future elections. The strength of the monitors was generally believed to be that their superior local knowledge gave them greater ability to detect electoral malpractice. For instance, one middle-aged man from Kitwe argued that locals "know all the corners, all the hiding places to check". 83 Although respondents often explained that they were reluctant to criticize fellow citizens, they nonetheless raised various concerns about domestic monitors taking sides, being corrupted, being influenced by ethnic ties, or having less freedom to make criticisms. For example, one young man in Siavonga argued that "if it was just the Zambians then they would be manipulated", 84 while a retired farmer in Petauke contended that "domestic observers have always shown they have political favourites". 85 Due largely to these concerns, the number of respondents who thought that non-African international groups should conduct observation alone was far higher than those who wanted domestic monitors to work alone. These respondents often took the view that international observers "appear to operate fairer, and they can speak freely", 86 and that "they are objective and show no favouritism". 87 However, the most common response, accounting for over half of interviewees, was that some combination of different types of international and domestic observers would be best. Many of the respondents advocating this seemed to envisage international observers supervising the work of domestic monitors, to prevent their concerns about the latter from being realized. For example, a young woman from Siavonga said "if you use local observers alone, elections will be rigged, thus regional and international observers are necessary to keep an eye on them". 88 Similarly, an older woman in Petauke argued "international ones can help the Zambian ones improve and keep standards high". 89 The findings presented in this section run counter to survey data from other countries, such as those of Cheeseman, Lynch, and Willis (see above). This may be partly explicable due to differences in country context and the wording of the questions that were used. It is also likely that methodological differences cause part of the variance. As discussed above, the qualitative approach that was employed in this research showed that many respondents had incomplete understandings of election observation or conflated observers' roles with those of other actors. When they became aware that this was the case, Macdonald and Mutuna were able to ensure that respondents who were confused about the topic were given a basic explanation of election observation before they were asked questions about their preferences for how it should be conducted. This clarity allowed respondents to provide better informed responses. In contrast, in a survey it can be unclear how many respondents understand election observation well enough to accurately state their preferences. Indeed, the responses of those who do not know what it is or confuse it with something else are muddled in with the responses of those with better understandings. As a result, caution should be exercised when using survey methods to investigate perceptions of election observation.
Overall, the discussion of different types of observers suggested that non-elite Zambians still have concerns about the credibility of domestic monitors that relate to political bias, potential ethnic alliances, and susceptibility to corruption. Despite the arguments in the academic literature regarding their various biases, international observersspecifically those from outside Africawere generally viewed as more independent and impartial. This section also raises concerns about how little information from either DEOIs or IEOMs actually reaches the public. This suggests that there are some limitations to the extent to which domestic observers can play the role of impartial arbiters in Zambia, as the poor circulation of their findings and a lack of trust means they may struggle to create public trust in elections or to mobilize protest in the event of significant electoral malpractice. More broadly, this finding points to political polarization, high perceptions of domestic corruption, and the salience of ethnic politics as factors that may undermine the credibility of domestic observers.

Conclusions
The 2021 Zambian elections highlight several issues that are important to consider when evaluating whether DEOIs can act as impartial arbiters. Domestic missions, with the help of international funding, were strong in providing large numbers of monitors and broad geographical coverage. CCMG were also able to provide coverage of the pre-election period, a PVT, and the publication of regular, detailed and often critical reports. This shows that organizations that have benefited from technical assistance programmes and have prior experience of observation can be very capable. The organization's success can also be attributed to strong leadership and the ability of its coalition partners to work well together towards shared objectives. From a technical standpoint, it is not easy to identify activities performed by international observers that cannot be undertaken by an experienced organization such as CCMG.
The fact that Zambia has a robust civil society more generally is also beneficial to its DEOIs. Despite civic spaces coming under attack during Lungu's administration, some sections of Zambian civil society proved resilient. 90 As a result, established civil society organizations and trained personnel were able to contribute to observation missions. This has also been the case in other African countries with strong civil societies, for example Ghana, where domestic observers have played a critical role in furthering the democratic process. 91 In contrast, if civil society is repressed or co-opted, or if there has not been continuity in the organizations tasked with election observation, DEOIs will struggle to play the role of impartial arbiter. In Tanzania, for example, civil society has been badly affected by repressive government policies, and experienced observation initiatives have been side-lined. 92 The findings also reveal that Zambian citizens, including those working in the media, still have reservations about the partiality of DEOIs, and often view certain types of IEOM to be more impartial. It should be noted that mistrust of international organizations can be higher in some countries than it is in Zambia, and this would adversely impact the credibility of IEOMs. Nonetheless, the finding show that, for DEOIs, creating trust is not straightforward in contexts where information about specific missions' activities and statements are not always well-circulated. The uncertainty that the Zambian public expressed about the credibility of DEOIs partly reflects broader societal problems such as political polarization, government corruption, and ethnic cleavages.
Drawing upon this analysis, the article now concludes by theorizing the conditions in which DEOIs are able to perform their role of impartial arbiter well. They are most likely to provide impartial evaluations of election conduct in situations when: 1) they have gained experience during previous elections; 2) they are well funded (preferably from early in the electoral cycle); 3) they have benefitted from technical assistance programmes; and, 4) they can draw upon a vibrant and independent civil society. They are most likely to be perceived as trustworthy sources of these evaluations when operating in contexts where: 1) there are low levels of political polarization; 2) perceptions of political corruption are low; and 3) ethnicity does not play a prominent role in politics.