Household water (in)security in low-income settlements of Kampala

(1)

International Development Studies (MSc)

Graduate School of Social Sciences

Household Water (In)Security in

low-income settlements of

Kampala

Alex Trowell

alexander@trowell.com

Student Number: 11173300

Word Count: 21,822

Supervisor: Dr Michelle Kooy

Associate Professor of Water Governance, UNESCO IHE-Delft m.kooy@un-ihe.org

Second Reader: Dr Michaela Hordijk M.A.Hordijk@uva.nl

(2)

Abstract

Current measurements of water insecurity fail to capture the differences between households, and the perspectives and experiences of access to water at the household level. This paper demonstrates the need for a household-level experience-based water insecurity index by identifying dimensions of water (in)security from the perspectives of households in low-income settlements of Kampala. The paper identifies how these dimensions for physical water access, water quality, and water affect offer more accurate and useful ways to assess inequalities in water access. Data was collected through semi-structured interviews with households, water-point caretakers, health workers, community leaders, chairpersons, and other key informants across five study sites in Kawempe Division of Kampala. This was paired with a household water insecurity survey, using the World Health Organization’s random sample walk method. The cross-cultural survey developed by the project ‘Household Water Insecurity Experiences Scale’ (HWISE), focuses on socio-demography, the development of a water insecurity access scale, food insecurity, and perceived stress. It is found that households in Kawempe Division define water (in)security using the following dimensions: affordability, reliability, ease of access, usage and practices, physical and biochemical quality, water-borne illnesses, hygiene and sanitation, citizenship/community engagement, and political disarray. This paper finds that between and within these dimensions are unique household experiences of water (in)security and concludes that current global measurements fail to successfully capture these distinct inequalities.

Key Words:

Water (in)security; water supply; Kampala; low-income settlements; household; comparative assessment

(3)

Personal motivations

Much of my Father’s side of the family grew up in Kampala, with the initial family members arriving in 1929, to later be joined by more – perhaps rather ironic given this was during the colonial period and this thesis forms part of my Masters in International Development Studies. My Great Grandmother also went on to establish an Art School at Makerere University, later to be named in her memory as ‘The Margaret Trowell School of Industrial and Fine Arts’ in recognition of her due care to preserve African culture and heritage in the school through often arduous times. Carrying out fieldwork in Kampala was therefore a fascinating experience, made more so by my research being supervised by a project based within the same department as my Grandmother’s art school still operating today.

(4)

Acknowledgements

There are many people I would like to thank for making this research possible.

Firstly, enormous thanks go to Michelle Kooy, who has gone above and beyond her duty as my academic supervisor. Her meticulous advice and feedback throughout the process of producing this thesis have been greatly appreciated and her expert knowledge of water access and governance has been invaluable. I would also like to thank her for connecting me with key people who are also working on the T-Group project in Kampala.

I want to thank Makerere University and those within T-Group, particularly Robinah Kulabako and Jennifer Isoke for hosting and allowing me to conduct my research within their T-Group study sites. They provided essential local knowledge, helped coordinate many of my fieldwork logistics and were always on hand to resolve any issues I had. They also connected me with field guides in each of my study sites, who without them my fieldwork would not have been possible.

Additionally I would like to express my gratitude to the ‘Household Water Insecurity Experiences Scale’ (HWISE) project, especially the project lead Dr Sera Young for providing me the opportunity to be the Site Principal Investigator in Kampala and allowing me to use the results from the household survey in my thesis. I also extend my thanks to the study coordinator Shalean Collins for her

continuous quick-response support throughout.

Special thanks go to all ten of the interviewers in Kampala who conducted the household surveys. It was a pleasure to train them and supervise their activities and I am grateful for their

professionalism.

I would also like to thank Michaela Hordijk for taking the time to read and appraise my thesis as a second reader.

Lastly my thanks go to the residents of Katoogo, St Francis, Kafunda, Kawaala, and Makerere Zone C. Their hospitality and willingness to participate in my research was hugely appreciated and I sincerely hope they see a positive change in their water security soon. I wish them a happy and healthy life and hope we get to meet again.

(5)

Acronyms and Abbreviations

GWP - Global Water Partnership

HH – Household

HWIAS – Household Water Insecurity Access Scale

HWISE - Household Water Insecurity Experiences Scale

IHE - IHE Delft Institute for Water Education

KCCA - Kampala Capital City Authority

KF- Kafunda

KT – Katoogo

KW – Kawaala

MDG – Millennium Development Goal

MK – Makerere Zone C

MoLHU - Ministry of Lands, Housing and Urban areas

NSDFU - The National Slum Dwellers Federation of Uganda

NWSC - National Water and Sewerage Corporation

PI- Principal Investigator

SDG – Sustainable Development Goal

SF – St Francis

TM - Transition Management

T-Group (Umbrella Organisation)

UGX – Ugandan Shilling

UN-Habitat - United Nations Human Settlements Programme

UNDP – United Nations Development Programme

WHO – World Health Organisation

(8)

Household Water (In)Security in

low-income settlements of

Kampala

(9)

1 1. Introduction

1.1 Global issues of access to water

Over the years there have been various shifts in the ways we measure and set targets for global access to safe water, which give rise to new debates. What are meant by the terms ‘access’ and ‘safe’? How are global targets set, and what sources of water should be included in these estimates?

Target 7c of the Millennium Development Goals (MDGs) looked to “halve, by 2015, the proportion of the population without sustainable access to safe drinking water and basic sanitation” (UN, 2000). This target was met 5 years early in 2010. However, it has been argued that the MDGs correlated improved access with the development of infrastructure and technology, rather than criteria for reliability, continuity, quality and service affordability (Kayaga et. al, 2009). This meant target 7c concealed inequalities within access to safe water between countries. This issue was discussed as early as 2006 in the UNDP Human Development Report, which cautioned that “the MDGs should be seen as a floor not a ceiling” (p.55) and places the intersections between power, poverty and inequality at the forefront of the global water crisis (UNDP, 2016).

A 2012 UNICEF-WHO Joint Monitoring Programme (JMP) report highlighted that whilst “89 per cent of the world’s population, were using an improved water source in 2010... it is only 61 per cent in sub-Saharan Africa” (UNICEF-WHO, 2012, p.4). This also equated to an increase of over 2 billion people gaining access to an improved water source between 1990 and 2000. Yet of that 2 billion, almost half were inhabitants of India and China (ibid).

As well as concealing the inequalities within access to safe water between countries, the MDGs also concealed inequalities within countries. Target 7c correlated ‘safe’ water as an ‘improved’ water source, yet as highlighted by the JMP report, not all of these improved sources will be adequately maintained and therefore not always safe. Interpreting ‘safe’ water explicitly in terms of physical access to an improved water source also fails to capture the actual quality of the water coming from the ‘improved’ source, as well as failing to address the root causes of water poverty.

(10)

2

Such critiques of the MDGs propagated revised goals, now found in the Sustainable Development Goals (SDGs) enacted in 2015, which look to go beyond physical access to identify specific dimensions of ‘safe’ water and capture issues with reliability and safety of access. However, even with revised goals in place, managing global water and sanitation issues is no easy task. It is estimated that by 2050 over two-thirds of the world’s population will be living in urban areas, with almost 40% of urban expansion in slums (UNDP, 2015).

1.2 The water situation in Kampala

In Kampala, there have been various initiatives to improve provision of safe and sustainable water supply in low-income settlements. There are pro-poor units looking to improve service delivery and also experiments with pay as you go technologies which enable cost-recovery by the utility with physical access by residents. Yet these efforts have still proved insufficient and residents still rely on combinations of water sources due to lack of continuity, reliability, or affordability of water supplied by the National Water and Sewerage Corporation (NWSC). Shallow sub-surface groundwater sources and natural springs have historically been used to supplement, or sole supply, other sources. However these sources are not encouraged by the Kampala Capital City Authority (KCCA) who manage the provision of water in Kampala, as the high water table in some areas (<1.5m) is prone to contamination (Kulabako, 2005).

Currently an international research project in Kampala is investigating potential sources of groundwater for the urban poor as an improved alternative water source. These groundwater sources include boreholes, protected springs, and wells, to supplement the currently inadequate piped water supply provided through standpipes. The depths of these boreholes vary depending on the depth of the groundwater level, though generally in the study sites deep groundwater aquifers occur at depths of more than 30m, whilst “groundwater levels within the regolith are usually shallow, at depths usually less than 8m and reducing to less than 5 m’’ in some areas (Kulabako, 2005, p.6). The IHE Delft Institute for Water Education (IHE-Delft) is currently leading a project under the group organisation name, T-Group, which looks to experiment with practical transition groundwater management strategies for the urban poor in Sub Saharan Africa, including a project in Kampala.

T-Group aims to improve groundwater management for the urban poor and has conducted surveys on current levels of access, and household practices in terms of water supply, sources and

(11)

3

treatment. T-Group’s work has stayed at the community level in determining the most water insecure areas. However issues with access to safe water are not homogenous across communities and so the project does not identify differences in access within settlements themselves. There is therefore the question of who will benefit from the project in terms of improved management of water resources.

Another ongoing project which looks to investigate important macro-level linkages between water security and its impacts is the ‘Household Water Insecurity Experiences Scale’ (HWISE), led by Dr Sera Young of Northwestern University. It aims to develop a ‘’cross-culturally validated scale that measures household water insecurity” and explore “consequences for economic productivity, nutrition, disease, and psychosocial wellbeing” [HWISE Manual]. However in Kampala, the question still remains: how will T-Group groundwater interventions affect levels of water security and for whom? Answering these questions requires a better understanding of which households are most water insecure.

(12)

4 2. Theoretical Framework

Introduction

The following chapter introduces the concept of water security and unearths the benefits and drawbacks of different conceptualisations of the term by drawing on comparisons between narrow, broad, integrative, and epistemological framings. The effectiveness of examining water security through a sustainability lens is then addressed, notably assessing the progression from the MDGs to the SDGs in more detail. Measurements of water security at different scales are then reviewed; looking at some of the key methodologies and what dimensions of water security they capture. The chapter then builds on the knowledge gap by outlining why it is important to take into account the perspectives and experiences of households in defining and measuring issues surrounding access to safe water. The chapter concludes by reflecting on the suitability of different definitions and measurement frameworks to successfully capture the experiences of water security at the household level.

2.1 Framings and conceptualisations of water(in)security

Central to the term ‘water security’ is the way it is conceptualised. Discursive politics gives rise to various conceptualisations of the term, propagating complex and often conflicting definitions (Gerlak and Mukhtarov, 2015). Many academics situate debates surrounding the provision of water in terms of scarcity, which is interlinked with issues of availability. In this sense availability can encompass spatial variations and provide measurements relating to challenges of supply and demand (Feitelson, 2002). However, such a framework fails to capture the multidimensional nature of water and instead operates around a cause and effect mechanism. Yet with access, availability, human security, sustainability, human needs and ecological dimensions to consider, to name just a few, the challenge therefore lies in formulating a definition that strikes a balance between the benefits derived from both narrow and broad conceptualisations of water security. Issues of prescriptive or descriptive ways of thinking must also be considered. There is a tendency for the international donor community to adopt prescriptive-oriented governance of water, that whilst has the advantage of being transferable and comparable, lacks recognition of local values and human needs (Gerlak and Mukhtarov, 2015).

(13)

5

In light of this, there have been growing efforts to coordinate framings of water security in terms of ontological and epistemological perspectives which account for states of well-being and emotional dimensions. Notably, Cook and Bakker (2012) posit a framing of water security that advocates for a broad and integrative conceptualisation that captures the multidimensional nature of water. They argue that adopting a broad approach ensures that attention is given to issues of both quantity and quality, whilst an integrative approach helps “mediate the trade-offs between different stakeholders, scales, and uses of water” (Cook and Bakker, 2012, p.98). However, when it comes to operationalising water security to make it suitable for policy making, Cook and Bakker acknowledge the drawbacks of a broad approach which fails to capture “spatial and social variations of water security” (ibid. p.99). This leads them to suggest that water security should be framed using an integrative and broad approach but then narrowed down when looking to operationalise it.

2.2 Water Access and the SDGs

Using a sustainability lens allows researchers to frame water security in line with Cook and Bakker’s argument by starting with an integrative and broad approach and later narrowing it down to focus on more specific elements.

However, even within a sustainability frame the definition of water security presents conflicting definitions.

“Water security, at any level from the household to the global, means that every person has access to enough safe water at affordable cost to lead a clean, healthy and productive life, while ensuring that the natural environment is protected and enhanced” (Global Water Partnership, 2000, p.12).

The Global Water Partnership’s (GWP) definition puts forth a broad conceptualisation that operates around access, supply, protection of ecosystems, risk management, and governance (Cook and Bakker, 2012). This arguably diminishes the dimension of basic needs in favour of setting baseline targets that can be used to compare water security across different levels and locations. This seems to be a common theme of the sustainability paradigm which struggles to strike an effective balance between top-down systematic governance and bottom-up participatory inclusion.

Previous framings of water security using a sustainability lens tend to place focus on water ‘access’. However, the term ‘access’ gives rise to ambiguity; what does ‘access’ mean; how do we measure it; and at what scale? Kayaga et. al (2009) argue that the terms ‘access’ and ‘usage’ are often used

(14)

6

interchangeably, causing confusion over how safe and reliable water is defined. Zawahri et. al (2011) find major disparities between looking at water supply solely as a contingent of water access. They assert that improved living standards combined with increased access generates increased demand, which in turn creates more wastewater flows which if not properly managed lead to a decline in water quality and distribution. Nganyanyuka et. al (2014) build on this notion and argue that factors outside of the water sector heavily impact upon access to water. They present findings to suggest that official water access statistics fail to capture these external factors and diminish the role of informal water supply activities (ibid). This argument is in line with Obeng-Odoom (2012) who contends that a broad interpretation of access to water by the donor community fails to account for the social consequences tied to issues of quality, reliability, cost, and equitability of distribution.

The Millennium Development Goals (MDGs) looked to provide empirical targets capable of capturing a global water security framework. However, the MDGs correlated improved access with the development of infrastructure and technology, rather than criteria for reliability, continuity, quality and service affordability (Kayaga et. al, 2009). A study by Onda et. al, 2012, argues that MDG target 7c grossly “underestimates the proportion of the population using unsafe water” (p.891) and provide evidence to suggest that 1.8 billion people, as opposed to the 780 million as estimated by the Joint Monitoring Programme, “did not have access to safe water in 2010” (p.892). As such, low-income communities and their informal strategies to access safe water are undermined in favour of quantifiable goals (Nganyanyuka, et. al, 2014). These critiques of the MDGs propagated revised goals which go beyond physical access to identify specific dimensions of safe water, now found in the Sustainable Development Goals (SDGs).

Target 6.1, Goal 6 of the 2015 Sustainable Development Goals (SDGs) –

“By 2030, achieve universal and equitable access to safe and affordable drinking water for all”

(United Nations, SDGs, 2015).

The SDGs look to provision a coordinated effort to “end poverty, protect the planet and ensure that all people enjoy peace and prosperity” (UNDP ‘SDGs’, 2015). However, such top-down mechanisms have been accused of provisioning a fragmented and disproportionally oriented approach to development, overshadowing less-easily quantifiable issues and undermining the rights of vulnerable peoples in the name of ‘sustainability’ (Muiderman, K., Hollander, S., 2013). An appropriate framing of water security must therefore acknowledge the long-term targets set through a broad sustainability lens but then be narrowed down to account for local values and human needs.

(15)

7

2.3 Measurement at different scales

The multidimensional nature of water security demands a means of measurement, or an indicator, that reflects causality and its linkages to other social, economic, political and environmental issues. However, to date, water security largely stands as a theoretical qualitative concept and lacks a quantitative measurement framework which can be used to steer policies and set water security thresholds (GWP, 2012). There have been many attempts to frame and measure water security using a human development, or humanitarian lens, such as Gerlak and Wilder (2012)1 and Stevenson et al (2012)2; however their definitions still lack recognition of basic human needs.

The Water Poverty Index (WPI), born out of a need to address important water poverty and human needs linkages is a “composite, inter-disciplinary tool, linking indicators of water and human welfare to indicate the degree to which water scarcity impacts on human populations” (Sullivan et. al. 2003, p.191). Formulated around five key pillars; resources, access, capacity, use, and environment, comparisons can be made between different locations and across different time frames (ibid). However, there are two key limitations of applying the WPI below the community level. Firstly, the WPI’s multidimensional nature must be questioned; much of the qualitative data does not always best translate into a quantitative index value and so it does not always accurately reflect the finer details of water security issues. In other words, “the information is in the components rather than in the final single number” (Lawrence et al, 2002). The second issue is that the WPI lacks recognition of local politics, power dynamics, and psychosocial dimensions. Feitelson (2002) draws attention to the difficulty in “distinguish[ing] between situational failures in human adaptive capacity and structural impediments” (p. 271).

For these reasons, there is the need to develop a water security index at the household level that goes beyond issues of physical access as per the SDGs and provides a more accurate reflection of household experiences of water security than the WPI. The Asian Development Bank (ADB) has formulated a Household Water Security Index composed of three sub-indexes; access to piped water supply, access to improved sanitation, and hygiene (ADB, 2013). Although the index also looks to implement policy levers to improve household water security, namely; finance, management, social and technology strategies, it still draws too heavily on MDG water supply targets and therefore fails

1

Gerlak and Wilder (2012) definition of water security: Water access: ability to use water and sanitation that is available, affordable, safe and clean.

2_{Stevenson et al (2012) definition of water security: Insufficient and uncertain access to adequate water for an} active and healthy lifestyle.

(16)

8

to capture informal water security strategies within low-income communities. Tsai et. al, 2017, put forth an 8-item Household Water Insecurity Access Scale (HWIAS) which moves beyond focusing on just water availability and quality. The scale reveals “significant intra-household gender differences in perceptions of household water insecurity”, providing further evidence that a means of measuring household water security that captures these unique household experiences is required (ibid. p.9). However, a key limitation of Tsai et. al’s HWIAS is that it has been developed by adapting previous scale item studies, rather than being constructed via in-depth qualitative methods which would foster culturally-specific items (ibid).

Acknowledging these limitations, Hadley and Wutich 2009 argue that previous household water security scales do not make the link between access inequities and their impacts, which are socially embedded practices. Instead, psychosocial dimensions in previous scales are reduced to just proxy measures, rather than being based on culture, experiences, and perceptions (ibid). Further studies by Wutich in Bolivian squatter settlements developed a cumulative experience-based water insecurity scale, which explores the relationship between water-related emotional distress and inadequate supply, poor water networks, seasonality, and gender (Wutich and Ragsdale, 2008). Although SDG 6.1 places emphasis on universal access, goals for equity, and safe and affordable water, it fails to consider issues of psychological or social stress and its focus on sustainability detracts from the local politics of water. Wutich’ scale goes beyond such previous measurements of water security and factors around cumulative experiences, as opposed to scale item analysis as per Tsai et. al’s HWIAS. This entails an integration of water distress issues, rather than them being treated as separate entities. However, as recently highlighted by Jepson et. al 2017, a major flaw of the scale is its failure to embody dimensions of emotional distress within the water security index itself, choosing to compartmentalise them separately. As a result, Jepson has developed a water insecurity scale structured around three cumulative scale pillars: water access, water quality acceptability, and water distress (ibid, p. 8).

2.4 Understanding the perspectives and experiences at the household level

In order to successfully capture the multiple dimensions of water security, it is important to understand the perspectives and experiences of access to water at the household level. As noted by Jepson et. al 2017, politics of water “are often restricted to water supply rather than the ways in which exclusion and changing social relations of water are produced” (p.48). Focusing on the

(17)

9

household level therefore allows key differences between population groups and water security to be better understood, which a city, regional, or national lens fail to capture.

However, current literature on household water security acts on the assumption that measurement at the household level is needed to help interpret linkages between access to water and other development spheres, without delving into ‘why’ or ‘how’. In the paper “Bathing without water, and other stories of everyday hygiene practices and risk perception in urban low-income areas: the case of Lilongwe, Malawi”, Rusca et. al, 2017, discuss the understandings of hygiene practices and the implications of how these understandings are framed. Similar to Jepson’s viewpoint that the politics of water must be socially deconstructed, Rusca et. al argue that hygiene practices and the ways they are prioritised are “socially constructed and are the product of embedded practices” (p.9). They use the term ‘hygiene hierarchies’ to explain how examining local socially embedded practices reshape the relationship between affordability and access of water (ibid). Understanding these embedded practices therefore requires an examination of the perspectives and daily lived experiences at the household level. These perspectives are important to understand existing practices, in terms of how to change them and also how they relate to inequalities.

2.5 Jepson’s definition

In the paper ‘Progress in household water insecurity metrics: a cross-disciplinary approach’ Jepson et. al (2017) review some of the leading debates surrounding water security and the way it is framed. From this they identify four key spheres of water security: 1) human needs and development, 2) ecological sustainability, 3) geopolitics and international relations, and 4) vulnerability and risk (p. 3). Building on Cook and Bakker’s (2012) framing of water security, as discussed in section 2.1, Jepson et. al advocate for a narrowed definition born out of an initially broad and integrative approach. Acknowledging the benefits of using a human needs lens to capture the multidimensional elements of water security at the household level, Jepson et. al (2017) define water security as:

“...the ability to access and benefit from affordable, adequate, reliable, and safe water for wellbeing and a healthy life.” (ibid. p.3)

The objective of my research in Kampala is to understand the differences in water security between population groups by examining the perspectives (ranking/weighting of conditions of physical access) and experiences (related social distress/psychological factors) of access to water at the

(18)

10

household level. Utilising Jepson et. al’s (2017) definition of water security therefore allows elements of social distress to be integrated into this research’s understanding of household water security. In doing so, it moves “beyond the physical dimensions of access or quality” and provides a better understanding of water security at the household level than previous means, including the SDGs (ibid. p. 8). Integrating the ‘affect’ dimension into the sub questions on access to water and safe water quality also helps make the link from how households define dimensions of access and quality to how they experience them, which gives different household level perspectives on what water (in)security is.

Influenced by Jepson’s previous work on developing experience-based scales and classifications to assess household water security in colonias on the US-Mexico border (2014), Jepson et. al use the following definitions to structure the three pillars of water security:

Water access

“The capacity to access water for consumptive purposes, including physical access, affordability, and reliability” (Jepson et. al, 2017, p. 109)

Water quality acceptability

“The broad range of biophysical characteristics of water quality (taste, color, smell, biochemical, etc.) that influences water usage and health/well-being” (Jepson et. al, 2017, p. 109)

Water distress

“The emotional, cultural, and subjective experiences of water” (Jepson et. al, 2017, p. 109)

The focus of my research in Kampala is on household ‘experiences’ of water insecurity, whereby ‘experience’ equates to the product of dimensions of physical access plus their associated affect. In other words, the term ‘experience’ therefore captures both how households define and their social/psychological experiences of what they define. Since Jepson’s index is developed by using qualitative data in the preliminary stages, my research endeavors are similar in that I first identify common concerns/priorities (how households define water insecurity) and then common social dimensions. Household experiences of water insecurity will also determine the means of measurement, or units, which best describe these household definitions. For example, while a household may define affordability of water as an important dimension of water insecurity, their

(19)

11

unique experiences of affordability may give rise to units of measurement that go beyond simple monetary value. Adopting Jepson’s definition of water security and structuring my research around her three key pillars will allow me to explore and capture the experiences of water insecurity which are unique to the study sites in Kampala.

Theoretical Framework Summary

This chapter has introduced the conceptual framework of the research. I have presented theory to demonstrate that many elements of water security are socially embedded practices that therefore require a specific examination at the household level. A thorough literature review argues that an integrative and broad conceptualisation of water security should be narrowed down when looking to operationalise it. Positioning this research’s problem statement within international scientific debate leads one to suggest that previous means of measuring water security fail to capture the experiences and perceptions of water security at the household level. The transition from the MDGs to the SDGs has been a welcomed progression, though this chapter argues that the SDGs still fall short of successfully recognising local values and psychosocial dimensions unique to water security at the household level, and instead place too much emphasis on water access. The Water Poverty Index (WPI) has gained recent traction in the field of measuring water security, however this chapter reviews theory to suggest that the WPI, though the suitable at the community level, lacks an appreciation of politics, power dynamics, and social dimensions at the household level. Drawing on the above literature and academic debates reviewed, I assert that Jepson et. al's (2017) definition of water security is to be used to structure the proceedings of this thesis to help formulate and provide direction to data analysis.

(20)

12 3. Summary of Research Objective

Demonstrate the need for an experience-based water insecurity index at the household level which can be used to assess/measure the differences in water security between population groups. This requires demonstrating how perspectives and experiences of access to water at the household level are different than current global indicators for water access. This exercise is carried out by identifying criteria/indicators for physical water access, water quality from the perspective of households in low-income settlements of Kampala and their related water affect (socio-psychological impact).

Conducting an extensive series of semi-structured interviews with various households, water-point caretakers, health workers, community leaders, chairpersons, and other key informants across my five study sites in Kawempe division of Kampala; Katoogo, St Francis, Kafunda, Kawaala and Makerere Zone C, I identify recurring themes and group together water insecurity issues. These qualitative findings are compared and contrasted with the results from a cross-cultural study of household water insecurity in the form of a survey which focuses on the following topics: socio-demography, a water insecurity access scale (WIAS), water quality, water acceptability, water quantity, water utility, water stability/reliability, food insecurity, perceived stress, infant feeding, and socioeconomic status.

Approximately 250 surveys are conducted with households in the study sites (50 surveys in each of the five sites in Kawempe). The data is then analysed to produce descriptive statistics which provide an indication of what are the key water insecurity issues as defined by residents in the areas themselves.

Together, my qualitative and quantitative data are used to support the development of an experience-based water insecurity index. Since there are many different dimensions of water insecurity being measured in the surveys, my research focus examines the relative importance of certain indicators which are proven to be of importance in my study sites. With regards to T-Groups ongoing work, it is hoped that an index will help the Kampala team determine how to identify the most disadvantaged/most water insecure households in the project sites, and be able to measure how improved access to groundwater sources will/not affect these households, or how to adapt groundwater interventions to address the factors/criteria/experiences of water insecurity which seem to matter most to households.

(21)

13 Research questions

Main research question:

How do household experiences of water (in)security in low-income settlements of Kampala compare with global measurements?

Sub-questions

1. How do households experience physical access to water? 2. How do households experience safe water quality?

3. How do household experiences of access to safe water in Kawempe Division compare with the global household water insecurity experiences scale being prepared by the HWISE project?

(22)

14

2.6 Conceptual Scheme

A conceptual scheme helps visualise the relationship between the key concepts. Here, household water (in)security is conceptualised as an experience defined by households themselves. This is formulated as a combination of how households define and experience physical access to water and safe water quality. These two categories are positioned as sub questions 1 and 2 respectively. These categories are each composed of sub-categories which households have defined; when combined with their respective social and psychological affect these sub-categories formulate household experiences. The double-ended arrows represent that the two categories are interrelated.

Physical Access to water

HOUSEHOLD WATER (IN)SECURITY

Safe Water Quality

Social & psychological affect

ADD ADD

Affordability Reliability Ease of Access Usage and Practices

Physical and biochemical quality Water-borne illnesses Hygiene and Sanitation Citizenship/ Community engagement

Political disarray Social & psychological affect

(23)

15 4. Methodology

Introduction

The study location is first introduced which provides basic demographic and geographical context relevant to the research question. The reasons behind choosing a mixed-methods research design are then explained and reflections of my own epistemological position are discussed before outlining the sampling methods used. The data collection methods are then divided into three sections: semi-structured interviews, participant observations, and household surveys. Each of these sections presents the research techniques employed and justifies why they have been chosen. Following this, the data analysis methods are outlined for both qualitative and quantitative data, and the reliability and validity of these approaches are reflected upon. Finally, the ethical aspects of the research are examined and the limitations of the research are presented.

4.1 Research Location

The study sites have been selected based on where previous and ongoing work in low income urban settlements is conducted by T-Group and where a wide scope of household water security issues can be captured. The five zones; Katoogo, St Francis, Kafunda, Kawaala, and Makerere Zone C, are found in Kawempe Division, Kampala District (Fig 1).

(24)

16

The Division which consists of 13 parishes is located in the Northern area of Kampala around 7km from the central business district and occupies approximately 7600 acres, with around 290,000 people living in informal settlements (Ministry of Lands, Housing and Urban Development, 2014). A more detailed description of the research location, including historical and geographical context can be found in section 5.

4.2 Research Design and Epistemology

A mixed methods approach was employed to document and analyse how households in low-income settlements of Kampala experience water security. Qualitative research consisted of semi-structured interviews with households and in-depth participant observations, whilst quantitative research was structured around household surveys.

It is therefore important to acknowledge the epistemological foundations of the research. As outlined by Bryman (2012), epistemology relates to the ‘’understanding of the social world through an examination of the interpretation of that world by its participants’’ (p. 380). Adopting an interpretivist position allowed the research to examine the social constructs of household water security and gain a greater understanding of the embedded practices through the eyes of households themselves. However, it is important to also reflect upon my own positioning as a researcher. I must be particularly conscious of my own role within my research and data collection and ensure that I develop a good appreciation of my own bias; not only what I observe and record but also my own direct and non-direct actions. As noted by Hammersley and Atkinson (2007), ‘’there is a conflict with the naturalistic realism built into older ethnographic approaches’’, whereby I come to ‘’construct the social world through [my own] interpretations of it’’ (p.11). Because my research is situated within a much larger ongoing project carried out by T-Group and its umbrella organisations there is a danger that I may become complacent recognising my own positioning and become too involved in the institutional setup already in place. Continuous epistemological reflection is therefore vital.

4.3 Sampling

The research wanted to capture a wide spread of household water security issues in order to help identify the recurring issues they experience and the ones that are most prominent within the

(25)

17

communities. To ensure this, purposive sampling was first used to strategically determine which zones in Kawempe would be included in the study. The chosen zones are: Katoogo, St Francis, Kafunda, Kawaala, and Makerere Zone C.

Theoretical sampling - a form of purposive sampling- was employed when looking to gather data from semi-structured household interviews. As outlined by Glaser and Strauss (1967) ‘’the purpose is to develop theory as it emerges’’, which in the case of this research allows for categories of household water security issues to develop into comparison groups (p. 364). This is continued until a point of theoretical saturation is reached. However, with regards to this research, theoretical saturation is more of a conceptual idea as there will likely be unique household water security issues, particularly those revolving around household perceptions, which the study fails to capture during the relatively short fieldwork period.

The sampling method used to conduct the HWISE household survey was the WHO random sample walk. When training the interviewers I asked them to survey every third household, unless they personally knew the members of a household, in which case they were told to move onto the neighbouring household instead then continue to survey every third household from then on. Approximately 50 surveys were conducted in each of the 5 chosen zones in Kawempe, Kampala.

4.4 Data Collection

4.4.1 Semi-structured interviews

Qualitative research predominantly consisted of face-to-face interviews with households and speaking with key informants in these areas. 18 households, 2 community leaders in the different zones of Kawempe, 12 water-point operators (otherwise known as caretakers), and 1 local health worker who regularly has to deal with water borne illnesses were interviewed.

After consulting with the Kampala T-Group team it was decided that I would utilise the field guides in place in the zones to help coordinate my selection of households to interview. I wanted to minimise community fatigue through over-researching as much as possible, since many of the areas had previously been subject to extensive research, data collection and improved water source construction projects. My study aimed to capture the water insecurity issues that the most vulnerable groups of people face in low-income settlements of Kampala, so it was decided that my research focus on just five zones: Katoogo, St Francis, Kafunda, Kawaala, and Makerere Zone C. Prior

(26)

18

to undertaking the semi-structured interviews I first took several site visits to different zones to get a feel for the research location and context surrounding water security issues. It soon became apparent that there was a high level of widespread government mistrust within the study locations and was therefore was decided that the semi-structured interviews would not be digitally recorded to limit any mistrust the respondents may have felt from my presence which would have affected the quality of data collected.

Semi-structured interviews with households were conducted at the home of the respondents, or at the water source. In total I conducted 34 semi-structured interviews. When approaching the household, the field guide would normally give a brief introduction to the research, normally in the local language, Luganda. From there I would ask who is the most knowledgeable about water acquisition within the household, though I kept an eye on the ratio of males and females interviewed to try and capture an even spread. 20 females and 14 males were interviewed (7 in Katoogo, 3 in St Franics, 12 in Kawaala, 6 in Makerere Zone C, 6 in Kafunda).

Semi-structured household interviews were formulated around three key water insecurity themes in line with Jepson’s definition found in section 2.5 ; access, affect, and quality (See Appendix).

First, I began by exploring the access theme, asking questions to identify what they themselves defined as important access and quality issues, and how some of the key water issues faced by households which emerged from the household survey in the areas are experienced. Many of the questions included in my interview guide were open-ended in order to allow respondents to elaborate on their answers where they deemed necessary. This method was chosen because I wanted to ensure that the questions asked allowed for respondents to share their perceptions and experiences of water security, whilst still maintaining a relationship with the concepts discussed in my theoretical framework. Once the access questions had been asked I moved onto the theme of water quality, before finally delving into the affect dimension. I chose to ask quite open-ended questions with regards to water affect, as it was intended to reflect emotional, cultural, and subjective experiences of water rather than the more straightforward answers which respondents tended to give when asked about their issues relating to water access and quality. However, I soon found that much of the social and psychological issues that households face are closely tied to issues of access and quality, and so many of my findings are recorded as an integration of themes. The water affect dimension helped investigate and provide a means of measurement and comparison for less quantifiable forms of water security. Over time, the semi-structured interview guide was adapted since new water security issues and themes would emerge as more participants were interviewed.

(27)

19

4.4.2 Participant observations

Observations allowed me to develop a sound understanding of the research locations and the contexts which surround the water security issues. Four lengthy observations, generally between 3-6 hours, were carried out intermittently during the semi-structured interview period. This was done to help enlighten the interview process and provide a greater understanding of the community dynamics. The nature of my participant observations involved immersing myself in the study location for some time. Therefore although my research was grounded with the concepts derived in my theoretical framework, much of the finer details of theory was an ‘’outcome of an investigation rather than something that precedes it’’ (Bryman, 2012, p. 384).

Initial semi-structured interviews, along with findings from previous T-Group site reports found that power relations between households and landlords/caretakers played a pivotal role in the provision of water. Observations were therefore conducted at various water sources during different times of the day to try and capture some of these experiences. Many household experiences had become so regular that they were now common place. Intra and inter household dynamics and gender related practices were highlighted as interesting areas to further investigate as a result of observations. Observations helped me gain a deeper understanding of these experiences that respondents would otherwise deem as normal activities and would not share their detailed perceptions of during standard face-to-face interviews. Observations were often carried out alone, or with a field guide nearby. Since the majority of conversations between residents were in Luganda, having a field guide meant that they could point out and translate interesting observations to me.

4.4.3 Household surveys

246 household surveys were conducted in Kawempe division, equating to approximately 50 surveys in each of the five zones (Katoogo, St Francis, Kafunda, Kawaala, and Makerere Zone C). The surveys developed by the Household Water Insecurity Experiences Scale (HWISE) project, led by Dr Sera Young of Northwestern University, consist of 32 questions about water insecurity, along with other questions that interact with water issues, headed under the following key themes: sociodemography, water quality, quantity, accessibility, reliability and utility, food insecurity, perceived stress, and infant feeding (HWISE Manual, 2017). The data collected was a combination of qualitative and quantitative data. The survey looks to explore important linkages between

(28)

20

household water insecurity and its impacts, which the HWISE project expects “will have consequences for economic productivity, nutrition, disease, and psychosocial wellbeing” (ibid, p. 7).

During fieldwork, I was assigned the role of the Site Principal Investigator in Kampala for the HWISE project. As site PI my role in the data collection process was to coordinate a team of 10 interviewers to conduct the surveys in a responsible, accurate and ethical manner, and coordinate other project logistics.

The surveys in the form of questionnaires were printed out and approximately 5 surveys were handed to each interviewer at the start of each day and then collected at the end of the day and checked they had completed properly. Each interviewer would first ask if the person they are speaking to is the most knowledgeable about water acquisition within the household and if not who they could speak to. Each interviewer was selected on a number of credentials, including their English language ability; however most interviews were conducted in the local language, Luganda. The majority of questions in the HWISE survey are structured using a Likert scale to answer them; for example:

In the last 4 weeks, how frequently has your household water supply from your main water source been interrupted?

A…..Never (0 times in the last 4 weeks) B…..Rarely (1-2 times in the last 4 weeks) C…..Sometimes (3-10 times in the last 4 weeks) D…..Often (11-20 times in the last 4 weeks)

E…..Always (More than 20 times in the last 4 weeks) DK…..Don’t Know

NA…..Not applicable/ I don’t have this

Household surveys were used to complement the data already collected from semi-structured face-to-face qualitative interviews and participant observations. The results could then be compared and contrasted with each other, whereby the survey data provides support to the household definitions of water security and the interviews provide more detail into the experiences and perceptions of these definitions. Since the surveys also included questions on perceived stress, including open-ended questions, the survey data also somewhat reflected experiences and perceptions. In order to address the main research question of this study, to understand how households in low income settlements of Kampala experience water security, it is important to explore the linkages between

(29)

21

the different dimensions. The household survey therefore helps to distinguish some of these linkages as defined by household themselves and also identify the associated severities of the various water insecurity issues.

4.5 Data Analysis

The qualitative data collected through semi-structured interviews was recorded in note form at the time and then elaborated on straight after the interview whilst the findings were still fresh in the memory. Many respondents opted to answer in Luganda, meaning the field guide had to translate their answers at the time. If there was a particularly interesting or relevant issue that the respondent raised then the quote would be scribed in its entirety.

The qualitative results from semi-structured interviews were then coded by hand, employing mainly open and axial coding methods. In line with Corbin and Strauss (1990) grounded theory research, open coding was used to develop conceptual themes and form categories and subcategories. Ethnographic research in this study, particularly observation work, involved trying to see through the eyes of the participants to better understand their experiences and perceptions of household water security. Therefore open coding helped to steer away from any preconceived conceptions and bias I may have had. Axial coding was then used to fine tune the subcategories that had emerged through open coding methods and provided the tools to decipher the relationships between them (ibid). The findings from participant observation work also helped coordinate the coding process by providing additional ethnographic context to the interview results.

The quantitative data collected through the household surveys was inputted into an online database where the HWISE project survey results are contained for all the different study locations around the world. All of the participants were kept anonymous, using just a unique code to help identify them. Descriptive statistics were conducted by the HWISE project on demographics, validated scale scores, and water access question responses. I then carried out further statistical analysis on certain survey questions using Excel in order to triangulate the quantitative results with the qualitative findings and compare and contrast them.

(30)

22

4.6 Methodological reflections

Reflecting on the methodological quality of research is important to establish its strengths and weaknesses. Here the quantitative aspect of the research will be first be discussed by dealing with indicators of validity and reliability with regards to the household survey conducted. The qualitative side of my research will then be analysed, looking at the methodology used for the semi-structured interviews and observations, again in terms of validity and reliability.

The results from the household surveys provide a reasonable indication of key water insecurity issues and the descriptive statistics derived from the survey data present easily interpretable results which help to identify important linkages between water insecurity dimensions and other cross-cutting affairs such as food insecurity, stress and socio-demographic variations. However, the face validity of the statistical data does not always fully capture intrinsic water security dimensions at the household level and in some areas has produced rather misleading information. There is also the issue of concurrent validity; as the surveys being conducted in Kampala had to correlate with surveys being conducted elsewhere in the world.

Moving on to aspects of reliability, the surveys were conducted by a team of ten interviewers, there is therefore the issue of inter-observer consistency, which occurs when different observers or interviewers have conflicting interpretations of the data and information they record (Bryman, 2012). Another aspect of reliability to address is stability. Widespread government mistrust, poor perceptions of previous surveys in the research locations and cultural stigma attached to revealing sensitive information about lived experiences are the principal factors which affect the quality of quantitative data collected from surveys. Many respondents were a little apprehensive to give up their time and energy answering questions and often expressed their dismay that they had seen little progress in improving their living conditions and bettering their water situation over the years. In this sense, respondents’ answers and perceptions of their living conditions may change over time, which is something I must account for when looking to compare data collected from previous years with my own data recently collected. Since my research was positioned under the organisation, T- Group, many of the respondents I interviewed and households that were included in the surveys had been previous subjects of studies conducted by T-Group in the research locations. Therefore the inter-observer consistency of the results must be reflected on, since different staff members will observe and interpret things in different ways. The internal reliability of the surveys must also be questioned. Many different survey questions were asked which aimed to measure the same

(31)

23

definition of water insecurity; therefore the results must be closely examined to see to what degree they point towards measuring the same dimension.

As the Site Principal Investigator my role was to manage the team and survey logistics. I held several training workshops to try and limit any interviewer bias when they asked and recorded their answers. This was particularly problematic given that the interviewers either lived in the community or nearby communities themselves and so they already had strong ideas about what they thought were the key water insecurity issues in the area.

Cultural stigma also seemed to be another key issue. It often appeared that respondents were reluctant to reveal their true sufferings, instead choosing to emphasise their pride in the things that are going well or give false information, particularly in the surveys. The inter-rater reliability of the data must therefore be questioned since some respondents will be more willing to answer questions and give more truthful answers than some others will, which in turn affects the outcome of results when aggregating the data collected. Also some respondents will have different opinions of rating what is deemed as water insecure (Bryman, 2012).

Qualitative data, on the other hand, primarily collected through semi-structured interviews with households and key informants arguably provides a more accurate reflection of the unique trade-offs between water insecurity issues and inter/intra household dynamics in the study areas. However, the validity of the data must be questioned. As a Westerner, some of the observations I make and the theoretical ideas I deduce from interviews may not accurately reflect the true internal validity of what is actually occurring in the study sites; instead I may be guilty of drawing linkages between variables, in part due to my preconceived ideas as a Westerner (Bryman, 2012). With regards to external validity, my methodological procedures used to structure my semi-structured interviews were unique to my study location in Kampala and caution should be taken when attempting to generalise them across different social settings (ibid). The nature of my qualitative research which rested principally upon semi-structured interviews may have been the most appropriate method for the purpose of my research study, yet its holistic framework lacks the integrity to be repeated at a later date in the same manner, and most likely will fail to produce sound external reliability (Bryman, 2012).

(32)

24

4.7 Ethical reflections

Throughout the research, ethical considerations were continuously reflected upon. The nature of my research, being a combination of ethnographic qualitative and quantitative studies, brings with it a host of issues. Before undertaking interviews I would always ensure that all participants were made well aware of where I am from, the basic concept of the research, the purpose, and what they can expect (Bryman, 2012). Due to the sensitive data being collected on household water security issues I made sure to keep each respondent and their answers confidential. As a Westerner I also had to reflect upon my own role and positioning as a researcher within the field. I had to make clear to the respondents that no money was to be exchanged for their time and be realistic about what they can expect from my studies. Sharing my research findings with participants, otherwise known as a member-check, is one approach often taken by researchers, though given the nature of data being collected it presents a moral and ethical challenge and I must first consider what effect this would have (Goldblatt et. al. 2011).

Understandably, my presence would have affected to some degree their normal routines and activities, which is something I had to acknowledge when recording findings. This became evident during observation periods, though over time they became accustomed to my presence.

When training the ten interviewers to conduct the household surveys I ensured that they were also made aware of the ethical procedures I expected them to abide by. I wanted to give something back to the community as a token of thanks for allowing me their time to partake in my research, so I hired residents from the slum to be the interviewers, which brought with it unique challenges. I made sure that each interviewer operated in a different zone from the one they reside in, to reduce any bias in respondent answers and make the respondents feel more comfortable disclosing sensitive data. Many of the questions asked during both the semi-structured interviews and the surveys involved personal information about stress and social circumstances being shared. It was therefore important to be aware of cultural customs and acknowledge when and when not to probe for further information if the respondent began to feel uncomfortable.

4.8 Research limitations

Given that my research was conducted in locations already established by T-Group, this played a considerable role in the ways that research activities played out. Many residents within the area

(33)

25

were familiar with seeing researchers and participating in interviews and surveys. However, this propagated both advantages and disadvantages. I found that many respondents already had a good knowledge of water acquisition issues which made for some interesting discussions and the field guides also understood how I wanted to direct my questions and study. However, a significant proportion of respondents expressed their dismay of seeing yet another study in the area and had become disheartened with what they viewed as seeing very little change relative to the amount of research conducted over the years. Because of this, I believe that some respondents did not reveal their true sufferings and water security issues as they had become tired of answering questions. There is also the danger that some participants may have provided information during the interviews and surveys which they subjectively feel is in T-Groups interests, rather than being an accurate reflection of their water security experiences.

Another interesting limitation that affected the research was mistrust. The household interviews and more so from the surveys, found that many residents in the area feared government presence. They voiced that they were afraid the government would try and raise their taxes if they thought they were earning more money than they claimed. Some even equated the chemicals used to treat the metered water as poison used by the government to try and kill their community off. These fears sometimes made it difficult for participants to fully trust the research and several of the survey questions were often met with reluctance to answer them.

The language barrier also affected the flow of research, particularly during interviews with households where the field guide would have to translate for me in real time from Luganda to English. This made asking a series of open-ended questions difficult and I had to be conscious of asking leading questions. Ideally the surveys would have been done prior to the household interviews, as this would have helped to better inform the questions asked in the interviews. However the position of Site PI for the HWISE project only came to light in the latter half of the fieldwork period. More time in the field would have allowed for richer ethnographic data to be collected and stronger rapport to be built with the communities, although the household experiences and perceptions of water security captured during the time frame given still provide a good reflection of the issues in the area.

(34)

26 5. Historical and geopolitical context of Kampala

5.1 Informal settlements in Kampala

Much of Uganda’s current institutional, economic and societal framework can be attributed to its colonial heritage which has engrained a structural hierarchy within the country and perpetuated a distinct demarcation between the wealthy and the poor. This socio-economic divide is notably apparent in Kampala, where the poorer income classes even today are confined to informal settlements in low-lying regions of the capital (Fredby and Nilsson, 2013).

Figure 2 – Population density per acre in informal settlements (NSDFU, 2014)

The population of Kampala is increasing at an alarming rate; census data reveals that it was around 774,241 in 1991, 1,189,142 in 2002, and 1,507,114 in 2014 (Uganda Bureau of Statistics, 2016). Such an increase presents significant challenges in estimating the number of people living in, and the area of, informal settlements in Kampala. A 2014 World Bank report highlights this issue by quoting conflicting estimates from various sources: 44% of people, 25% of city (UN Habitat, 2002), 60% of

Household water (in)security in low-income settlements of Kampala

International Development Studies (MSc)

Graduate School of Social Sciences

Household Water (In)Security in

low-income settlements of

Kampala

Alex Trowell

alexander@trowell.com

Student Number: 11173300

Abstract

Key Words:

Personal motivations

Acknowledgements

Table of Contents

Acronyms and Abbreviations

Household Water (In)Security in

low-income settlements of

Kampala

1

1. Introduction

2

3

4

2. Theoretical Framework

5

6

7

8

9

10

11

12

3. Summary of Research Objective

13

Research questions

14

Physical Access to water

HOUSEHOLD WATER (IN)SECURITY

Safe Water Quality

15

4. Methodology

16

17

18

19

20

21

22

23

24

25

26

5. Historical and geopolitical context of Kampala