Judgment under uncertainty: a probabilistic evaluation framework for decision-making about sanitation systems in low-income countries
Sanitation, as described by the World Health Organization (WHO), refers to the provisions of facilities and services for safe disposal of human urine and faeces. On the outset of the 21st century, 40% of the world’s inhabitants do not have access to sanitation facilities and still rely on a bucket, a bush or a backstreet for excretion. Technically, all options that are required to deal with the global sanitation problem seem to have been already developed. However, the challenge remains in selection and implementation of technologies in a way that the desirable outcomes would be resulted. Some decision-making support tools have been developed so far to address this problem by assisting the decision-makers in selecting the appropriate technologies. While decision-making is about considering the likelihood of uncertain events, in most of the existing evaluation approaches the complex task of predicting and evaluating probabilities is reduced to simple judgmental operations. Forinstance, evaluation of sanitation options is often performed based on predicting the outcomes that best represent a sanitation system, with no or little regard to the factors that limit the predictive accuracy.
This thesis adopts a new evaluation approach by taking into account the real world examples from executed sanitation facilities and develops a probabilistic evaluation framework in which sanitation options are assessed based on the probabilities that specific outcomes occur in practice. Absolute judgments are replaced by probable assessments, as this approach tries to keep its distance from making the uncertain certain. Although there may be a hidden consensus that quantification of occurrence probabilities for various outcomes of sanitation options is not always possible, some quantification methods are developed and presented in this thesis for all the assessment criteria. Moreover, this thesis does not only focus on making the decisions, but also tries to channel the decisions in a way that the negative outcomes of sanitation facilities would be reduced through the measures that could be taken to improve the performance of sanitation options.
By applying the probabilistic evaluation approach for decision-making about sanitation facilities in low-income unplanned slum settlement of Nyalenda in Kisumu, Kenya (based on limited available data about this region in literature) it is indicated that while a sanitation option may be known for fulfilling a certain task by definition, through a probabilistic evaluation it may be revealed that the local conditions are not likely to allow the expected outcome to occur in practice and as a result this option would have no priority among other options. The necessity for monitoring and post-evaluation of implemented sanitation projects in order to have sufficient feedback for improvement of future decisions is also highlighted.