Design Imperatives for E-voting as a Sociotechnical System

There are a number of past and ongoing research efforts on the development of e-voting systems. These works largely focus on requirements, technical specification and implementation technologies to support different aspects of the elections from registration and verification through balloting to counting and result. A major shortcoming of these studies is their sole focus on technical aspect of e-voting solution wit/lOut significant attention paid to human and environment factors that arguably determine the successful adoption of such e-voting solutions. This paper addresses this design gap in three steps. First, it provides a conceptualization of e-voting system as a socio-technical system. Second, it elaborates a set of principles to guide a socioteclmical design for e-voting. Third, it provides concrete implications of these principles. The paper concludes on the pragmatics of this approach to e-voting adoption particularly in environment such as Nigeria.

E-voting constitutes a very important aspect of ICT­enabled democratic governance [1]. E-Voting solutions generally aim at increasing participation, improving the outcomes elections by addressing challenges associated with traditional voting practices. The notion of e-voting in this paper refers to the use of technology to support one or more of the major phases of the electoral process – from registration stage in the pre-voting phase to voting/balloting and verification to counting or tallying after voting [2], [3]. Although, the term is often associated with the use of electronic channels like the Electronic Voting machines or the internet for casting votes, its use for ICT-enabled voter enrollment or registration is gaining popularity particularly in the developing world. Generally, there are mixed sentiments regarding e-voting adoption, particularly in the aspects of vote casting. For example a number of countries like Netherlands, Germany and Ireland moved away from the use of e-voting for balloting after initial adoption [1]. E-Voting using Electronic Voting Machines (EVM) have also remained controversial in the United States. While e-voting is generally perceived as generally risky [4], there is some consensus that the benefits of e-voting outweigh the actual risks. In this light, developing countries have shown significant interest towards the adoption of e-voting. Given that elections in developing countries attract significant controversies and are fraught with several challenges, the idea of trying any credible alternative is a plausible strategy. While not completely oblivious of the risks involved in e­voting, research contributions on e-voting in Africa such as [5] and [6] have largely focused on technological design of e­voting systems. However, designs espoused in these works largely ignore: the social context in which the e-voting systems will enacted, peculiar needs of different users (e.g. voters) and the organizational context of the Electoral Authority. Direct experiences of the authors based on adoption and implementation of e-voting solution in Nigeria show that these shortcomings significantly impact the effectiveness of the solution and could potentially compromise the outcome of the elections. A way to address this problem is to adopt a Socio­technical System (STS) design framework for e-Voting systems. Socio-technical systems approaches advocates a human-centric analysis which considers the impact of the technical or computing sub-system on people and how technology can be designed more effectively for people [7]. The goal of an STS system is to collectively optimize the technical, social and environment subsystems [8]. We consider in this paper the design of e-Voting systems as a socio-technical system. Our goal is to augment existing knowledge about requirements and design of e-Voting systems by the specific design principles that must underpin any e-voting solution when considered more holistically as a socio-technical system. For illustration, we highlight implications of these principles for e-Voting solution in the Nigerian context.

The notion of e-Voting as an STS makes its inherent complexity more apparent. For instance, the need to operate e­voting system across different boundaries such as geographical, cultural and temporal boundaries, characteristic of complex systems [8] are not discussed in technical designs of e-voting. These principles make them explicit, for instance see principle #2 in Table 1 above. Omission in this respect creates potential source of failure for e-voting system. Congruent with the principles elaborated here, Chevallier et al [15] identified three core success factors for the Geneva’s e­voting project – the role of politics, organization of trials involving voters and other stakeholders and having an interdisciplinary team comprising Legal experts, security CERN, sociologist and political scientists. However, along the clear benefits for the adoption of a STS approach to e-voting design, we must point out some inherent challenges in operationalizing such approach, particularly in the developing country context. For instance, the classical STS approach requires iteration between design implementation and operation to enable continuous improvement and shaping of solution. The practicality of this in terms of participation of stakeholders and resource implication may be limited. Thus, there will be the need to carefully identify principles that could be supported in different context and under the prevailing temporal constraints for the overall election process. View More