Page 97 - UKZN Proceedings of the Conference Report

a MuLtI-CrIterIa deCISIoN ModeL for INforMatIoN SySteMS aNd teChNoLogy ProjeCtS for KeNyaN ParaStataLS
odero, f.a.
University of KwaZulu-Natal, Durban, South Africa Orcid: 0000-0002-7068-8301
X: @ChilavertF
govender, I*
University of KwaZulu-Natal, Durban, South Africa Orcid: 0000-0002-4499-1091
*Correspondence: 214584404@stu.ukzn.ac.za
abstract
Over the past two decades, governments in the Global South, in partnership with the private sector, have intensified the roll-out of information systems and technology (ISandT) projects to provide e-services timeously and cost-effectively to citizens and businesses. The projects have mainly focused on internet backbone infrastructure, data management technology, cloud computing, and the integration of artificial intelligence in contemporary digital platforms. While it is generally agreed that Kenya has made remarkable progress in actualising the digital economy, the challenge of selecting and prioritising the most suitable digital solution from a set of alternatives continues to face decision-makers across Parastatals. In many cases, decisions regarding selection and prioritisation in ISandT projects are made without consideration of the full scope and magnitude of influence of decision factors relative to the project’s goal. The likely result of this situation is uncoordinated investments, unforeseen loss of jobs, loss of finances, sub- standard projects, and sometimes failed projects. This study aims to identify the complexities and challenges facing decision-makers involved in ISandT projects in the face of increasingly complex project decision situations and to propose a framework for scientific decision-making at all project stages. The findings of the study should inform policy aimed at assisting decision-makers in streamlining the process of implementing national, regional, and global ICT development agendas and spurring productivity and competitiveness across the Kenyan Parastatals. A mixed-method research approach will be chosen for this study. Simple random sampling and purposive sampling will be used to collect quantitative and qualitative data, respectively. The sampling unit is the Parastatals, and the sample population is the body of ICT managers in Kenyan Parastatals. Quantitative data will be collected using self- administered questionnaires, while qualitative data will be collected using in-depth interviews.
Keywords: decision factors, ICT development agenda, information systems and technology (ISandT ) projects, scientific decision-making, Parastatals
Introduction
It is generally acknowledged in the literature that decision- making is central to human social, economic, and political activities (Tindale and Winget 2019). Take for instance the choice of who is to provide software as a service to a Kenyan Parastatal, or who is to provide infrastructure as a service, or who is to provide platform as a service. In each category, there are dozens of reputable IT companies. However, only one may qualify at any given time, and the decision must be made by organisational decision-makers entrusted with the assignment. Naturally, each decision-maker will have their way of judging who is the most suitable service provider. For harmony, standard instruments for technical and financial evaluation are normally used. However, these instruments may not incorporate the views held by an individual decision-maker regarding the magnitude of influence of decision factors that inform the final choice. Moreover, the measures used therein are absolute, like a static score against an element to be evaluated. The likely result of this situation is inconsistent and unsustainable decisions (Yuan, Wu and Tu 2023; Pouresmaieli et al. 2024). The natural choice would be for decisions that are supported by a relative scale of measurement, and that are consistent (Hendrycks et al. 2021; Pant et al. 2022).
Estimating the average opinion of a group of people is de- sirable in many real-life situations, especially in selecting the most suitable solution from a range of possible alternatives. The shape of the distribution in group decision-making “corresponds to Dirichlet distribution” (Vargas 1982). The Dirichlet distribution features prominently in mathematical modelling literature (Ng et al. 2011; Minka 2000; Li 2024). Per- haps the most significant among them is in the modelling of compositional data. Scientific studies agree that the natu- ral way to study something is by comparison rather than in isolation (Saaty 2013; Miller 1956). In further support of this established principle, weighing decision variables in deci- sion modelling is often relative to the goal or a well-defined decision variable in the case of dependencies and feedback characterised by networks. Compositional and multivariate data naturally arise from group decision-making situations and follow the Dirichlet distribution (Vargas 1982). Relative measurements are common in modern scientific measuring systems such as in the Analytical Hierarchy Process (AHP)
Proceedings of the conference on Public innovation, develoPment and sustainability | 95