Download Chapter

Chapter 3 Tools and approaches

Monitoring and evaluation

Photo:

In nearly all humanitarian aid programmes, the tasks of ongoing monitoring, programme evaluation and learning are vital to success. Monitoring and evaluation activities are particularly important in complex urban settings, where initial programme design is likely to require modifications and adaptations in response to a rapidly changing operating environment.

The subject of monitoring and evaluation in humanitarian action is vast, and a wealth       of tools and approaches are available (several are given in this section). This section,+This section benefited in particular from inputs from Paul Knox Clarke, Amelie Sundberg, Neil Dillon and Leah Campbell of the ALNAP Secretariat.  like others, does not intend to review the entirety of monitoring and evaluation, but aims instead to highlight key principles, challenges and opportunities relating to urban areas. The section introduces some of the challenges in urban monitoring and evaluation, identifies emerging lessons and approaches for urban monitoring and evaluation and discusses remote monitoring in conflict situations. As with some other sections of this GPR, this is an area of emerging rather than established good practice.

Monitoring and evaluation ties in closely to a number of other sections discussed in this GPR, especially those within the project management cycle, in particular assessments (Section 3.6) and design and management (Section 3.9).

3.10.1 Challenges in urban monitoring and evaluation

The complexities of designing and enacting urban programmes are discussed throughout this GPR. The process of monitoring and evaluating programme activities and outcomes needs to deal with this complexity and the ways in which humanitarians respond to it (including by adapting to rapid changes in the environment, engaging with a multitude of actors and in some instances undertaking multi-sectoral programming).

Two reports from ALNAP+See A. T. Warner, What is Monitoring in Humanitarian Action? Describing Practice and Identifying Challenges (London: ALNAP/ODI, 2017) (https://www.alnap.org/help-library/what-is-monitoring-in-humanitarian-action-describing-practice-and-identifying); and ALNAP, Evaluation of Humanitarian Action Guide (London: ALNAP/ODI, 2016) (www.alnap.org/system/files/content/resource/files/main/alnap-evaluation-humanitarian-action-2016.pdf). describe some general challenges to monitoring and evaluation in humanitarian interventions, which are particularly relevant for urban emergencies:

  • Identifying what needs to be known in the balance between robustness and speed is particularly difficult in cities due to the interconnectedness and density of people and services, which often cannot be untangled in a clear The monitoring information needed is also likely to change as an emergency evolves, but not all organisations are accepting of an iterative approach (for a further discussion on tools relating to this, see Section 3.9 concerning design and management).
  • Considering how to use secondary data rather than resorting to primary data collection (which can be time-consuming and costly to collect) is important in urban contexts.
  • Humanitarian evaluations are often undertaken in periods of severe disruption, which can limit access to information and key informants
  • In situations of conflict, it may be difficult for evaluators to access affected communities, limiting the amount of information available from the users of humanitarian goods and services (in urban environments, access may also be a problem even where there is no overt conflict – in certain particularly violent parts of a city, for example).
  • Conflicts can also generate sharply differing points of view, which may make it difficult to arrive at an objective assessment of a particular issue. To reduce the risk of bias, triangulation (basing findings on more than one source) is always necessary.

Several other challenges to evaluation are intensified in urban environments:

  • Given that cities are dynamic, the humanitarian situation, and humanitarian needs, can change very This means that the objectives and activities of urban programmes often change. Monitoring and evaluation systems need to be flexible enough to capture these changes, and broad enough to identify unexpected results.
  • The density and interconnectedness of services in urban areas means that humanitarian needs are generally multi-sectoral and intertwined. For example, disruption in electricity generation can lead to water and sanitation systems closing down, with knock-on effects on health and education services. Closure of markets can prevent people from accessing food, health services or employment – which in turn can make it even harder to access To understand the humanitarian situation, and to capture the effects of humanitarian programmes, monitoring systems and evaluations will generally need to be multi-sectoral, and will often need to be multi-agency.
  • Any specific humanitarian intervention will very often have a number of important but indirect effects, which will tend to occur in combination with the effects of other interventions. For example, the provision of child-friendly spaces may help improve the psychological and physical well-being of children, but it is possible that this effect was also a result of better provision of water (so that children or their parents did not have to go longer distances in an unsafe environment to collect water). Where the chain between cause and effect is long, and where many elements are involved, evaluators need to think about how they consider causality: they cannot assume that an intervention will lead, in a straight line, to an outcome.
  • The sheer numbers of people in urban environments mean that deciding who to include – and who to leave out – in any form of consultation may be challenging. It is also important to remember that communities may not be defined by physical boundaries. Communities of practice and interest may also be relevant (see Section 1.1 on ways of seeing the city for a discussion of communities).
  • Given the complexity of cities, the learning functions of evaluations – understanding what works and why – are particularly important. This challenges evaluations to rigorously consider the reasons for success or failure.
  • The diversity of urban populations means that there will often be multiple interests involved in any intervention, making humanitarian activities the focus of much political Evaluators should be prepared to navigate the politics of the city.
  • Potential key informants may be from a wide range of sources, including local and national government, journalists, business people and academics. Gatekeepers may or may not be apparent, but will almost always be there. Challenges in negotiating with gangs, for instance, and seeking access to particular neighbourhoods may be exacerbated in urban environments (see Section 1.5 on urban actors).

3.10.2 Emerging lessons and approaches for urban monitoring and evaluation

Lessons and approaches to urban monitoring and evaluation include:

Build monitoring and evaluation into the programme from the beginning

In urban programmes, it is especially important to include monitoring and evaluation as part of the original programme design. Urban programmes tend to be particularly ‘information heavy’, as a result of the number and diversity of people and elements involved, and the need to capture changes in the context. For programmes to be successful, resources for information collection and – particularly – analysis need to be made available at the planning stage. Many urban programmes are also fairly small compared to the level of needs across a city. To be effective, they often rely on scaling up. This depends on good information on what worked, what didn’t and why (see Section 3.2 on area-based approaches for a further discussion of scaling up). Several of the programme design approaches outlined in Section 3.9 combine iterative programming with ongoing data collection and analysis. These approaches cannot be used successfully without a clear monitoring and evaluation plan.

ALNAP’s Evaluation of Humanitarian Action guide+ALNAP, Evaluation of Humanitarian Action Guide. outlines key issues that should be considered, from the outset, in the design of evaluation systems. These are also broadly relevant for thinking about the design of monitoring systems. They include:

  • The purpose of the evaluation. Is it to show the value of the programme to donors? To provide learning for future programmes? To allow for course correction of an existing programme? To help with scaling up an existing programme? The purpose of the evaluation will, to a degree, drive the design.
  • Key users of the Who will use it? How can they be engaged in the design to ensure that they get the outputs they need?
  • Focus questions. No evaluation can tell the user everything about an activity or programme. A good evaluation will identify the most important questions, and then use a methodology that is able to answer those questions.

Focus on specific information needs

Complex, diverse and rapidly changing urban environments create specific information needs. Monitoring and evaluation systems need to ensure that these needs are met, and that the specificities of urban response are reflected in the design of monitoring and evaluation systems.

Ideally, monitoring should consider ‘need to know versus want to know’ – in such a dense environment, where monitoring has to be timely and relevant with scarce resources, untangling information needs is critical. Monitoring needs to consider the context – are there changes in the situation? Have these led – or are they likely to lead – to changes in need? Monitoring also needs to consider the outcomes of activities – what effects (both intended and, where these can be identified, unintended) are activities having? What does this mean for the programme?

Evaluations need to focus on a number of issues:

  • Causal factors: the reasons for the relative success or failure of an intervention – either to allow for adaptive programming, or for extending or scaling up pilot programmes.
  • Tailoring to context: the degree to which approaches were appropriate to the urban context – potentially against the benchmark of best practice guidelines for working in cities
  • Interconnectedness: the extent to which the programme took account of the interconnectedness of urban response, both in terms of interactions with other actors and the potential unintended spill-over effects of programme activities in other areas
  • Dynamic programming: the rationale for programme adaptations and changes away from initial programme objectives and the quality of the decision-making that led to those changes

For further information, see:

I. Christoplos and N. Dillon with F. Bonino, Evaluation of Protection in  Humanitarian  Action (London: ALNAP/ODI, 2018) (www.alnap.org/system/files/content/resource/files/main/EPHA%20Guide%20online%20interactive.pdf).

S. Jabeen, ‘Unintended Outcomes Evaluation Approach: A Plausible Way to Evaluate Unintended Outcomes of Social Development Outcomes’, Evaluation and Program Planning 68, June 2018 (www.ncbi.nlm.nih.gov/pubmed/28965770).

J. Puri et al., What Methods May Be Used in Impact Evaluations of Humanitarian Assistance?, Working Paper 22 (New Delhi: International Initiative for Impact Evaluation (3ie), 2015) (www.alnap.org/system/files/content/resource/files/main/wp-22-humanitarian-methods-working-paper-top.pdf).

Engage in complexity and systems thinking

Overall, urban monitoring and evaluation means engaging in complexity and systems thinking (see Section 1.1, on ways of seeing the city). Guides that outline steps for undertaking evaluations in complex settings or using a systems-based approach to evaluation are:

M. Bamberger, J. L. Vaessen and E. R. Raimondo (eds), Dealing with Complexity in Development Evaluation (Thousand Oaks, CA: Sage, 2015).

B. Hargreaves, Evaluating System Change: A Planning Guide (Princeton, NJ: Mathematica Policy Research, 2010) (www.mathematica-mpr.com/our-publications-and-findings/publications/evaluating-system-change-a-planning-guide).

Williams and R. Hummelbrunner, Systems Concepts in Action: A Practitioner’s Toolkit (Stanford, CA: Stanford University Press, 2010) (www.sup.org/books/title/?id=18331).

J. Rogers, ‘Using Program  Theory  to  Evaluate  Complicated  and  Complex  Aspects of Interventions’, Evaluation 14, 2008 (https://journals.sagepub.com/doi/pdf/10.1177/1356389007084674).

This USAID paper considers issues specific to monitoring in situations of complexity: Discussion Note: Complexity-Aware Monitoring, July 2018 (https://usaidlearninglab.org/sites/default/files/resource/files/cleared_dn_complexity-aware_monitoring.pdf).

Box 3.15 Using indicators in monitoring a multi-sectoral programme in Guatemala City

In Barrio Mio, an area-based disaster risk reduction and response project in Guatemala City, Project Concern International (PCI) implemented a multi-sectoral programme with activities ranging from women’s savings groups to the installation of retaining walls and water tanks and building the GIS capacities of municipal actors. Initially, the project relied on a set of indicators from its funding proposal. These were separated by sector, and covered issues such as the number of shelters incorporating hazard mitigation measures and the number of people demonstrating good handwashing practices.

A case study on the project found that, while these indicators helped to demonstrate achieved deliverables, they failed to ‘capture the richness of the Barrio Mio project and what it’s been able to achieve – which is far beyond the level of ambition that these indicators suggest’. The Barrio Mio team developed a number of complementary indicators in addition to the list from the donor, though tensions remained ‘between what some describe as a “myopic” focus on the list of indicators and the overall impact the project has had’.

Source: Adapted from L. Campbell, Barrio Mio and Katye: PCI’s Neighbourhood Approach in Cities (London: ALNAP, 2019) (https://www.alnap.org/help-library/barrio-mio-and-katye-pcis-neighbourhood-approach-in-cities).

Be aware of the multiple information sources available in the city

Cities are information-rich. Local government, service providers, chambers of commerce, journalists and many others may collect the information that monitoring or evaluation systems require. It may also be possible to make use of geospatial approaches (see Section 3.4 on mapping and geospatial analysis) to identify, or triangulate, changes in context or the effects of programmes. When using secondary data, however, it is important to ensure that the data adequately represents the populations of greatest concern. Official data often ignores certain parts of cities (such as informal settlements), or is aggregated at a high level, and so effectively hides the reality of life for the poorest or most marginalised groups.

When investigating the degree to which outcomes are achieved, or the constraints to achieving them, qualitative data is also important. One literature review on urban crises advises that ‘qualitative data may be required to capture impacts and outcomes that are more difficult to quantify (e.g. impacts on local power structures and urban socio-economic realities)’.+D. Brown et al., Urban Crises and Humanitarian Responses: A Literature Review (London: UCL, 2015) (www.urban-response.org/system/files/content/resource/files/main/bartlett.pdf). Qualitative approaches are critical to explaining how and why something occurs.+M. Skovdal and F. Cornish, Qualitative Research for Development: A Guide for Practitioners (Rugby: Practical Action, 2015) (http://eprints.lse.ac.uk/64207/). A 2016 learning workshop of urban disaster risk reduction practitioners emphasised the need for monitoring and evaluation tools to recognise the complex social dynamics in urban neighbourhoods, and the use of qualitative indicators to understand the context.+J. P. Sarmiento et al. (eds), Urban Disaster Risk: Systematization of Neighborhood Practices (Miami, FL: Florida International University Extreme Events Institute, 2016).

For further information, see:

M. Skovdal and F. Cornish, Qualitative Research for Development: A Guide for Practitioners (Rugby: Practical Action, 2015) (http://eprints.lse.ac.uk/64207/).

M. Quinn Patton and M. Cochran, ‘A Guide to Using Qualitative Research Methodology’, MSF, 2012 (https://evaluation.msf.org/sites/evaluation/files/a_guide_to_using_qualitative_research_methodology.pdf).

ICRC, Acquiring and Analysing Data in Support of Evidence-based Decision Making (Geneva: ICRC, 2017) (www.icrc.org/en/publication/acquiring-and-analysing-data-support-evidence-based-decisions-guide-humanitarian-work).

M. Bamberger, J. Rugh and L. Mabry, ‘Qualitative Evaluation Approaches’ in RealWorld Evaluation: Working under Budget, Time, Data, and Political Constraints (Thousand Oaks, CA: Sage Publications, 2012).

M. Q. Patton, Qualitative Evaluation and Research Methods (Thousand Oaks, CA: Sage Publications, 2002).

Use people-centred approaches

As discussed throughout this Good Practice Review (and introduced in Section 1.1 on ways of seeing the city), people-centred approaches are key. One guide presents an approach which asks households to (retrospectively): describe their livelihoods before the disaster, immediately after the disaster and after humanitarian interventions; identify changes;  and describe the contribution interventions have made to these changes. It also lays out good practice in working with affected communities. See R. Few et al., Contribution to Change: An Approach to Evaluating the Role of Intervention in Disaster Recovery (Rugby:  Practical Action Publishing, 2013) (https://reliefweb.int/report/world/contribution-change-approach-evaluating-role-intervention-disaster-recovery).

Similarly, the Good Enough Guide: Impact Measurement and Accountability in Emergencies provides a number of simple and effective tools and principles for understanding the impact of humanitarian activities from the perspective of the people who are meant to benefit from them: see www.alnap.org/help-library/good-enough-guide-impact-measurement-and-accountability-in-emergencies.

The ‘most significant change’ (MSC) approach relies on collecting significant change stories coming out of a programme, and the systematic selection of the most significant stories  by panels of designated stakeholders or staff. It is particularly good at identifying the   more unusual or extreme effects of interventions, and for creating a shared understanding between stakeholder groups involved in a response. See R. Davies and J. Dart, The ‘Most Significant Change’(MSC) Technique: A Guide to Its Use, 2005 (www.mande.co.uk/wp-content/uploads/2005/MSCGuide.pdf).

In addition, a number of toolkits and guidance notes are available to support participatory approaches to information collection and analysis. They include:

Consider working ‘backwards’, from outcomes to interventions

An important element of many people-centred approaches to evaluation is that they invert the ‘normal’ sequence of evaluation. Rather than starting with an intervention and working forwards to try to identify the results of the intervention, they begin with the changes that people have seen and work backwards to see how these changes link to the intervention. As a result, they are generally better adapted to evaluating urban programmes, where there are often long and complicated causal chains between the humanitarian response and the effects on people’s lives, and where multiple interventions may have contributed to the final outcome. That said, they may be less good at fulfilling donor requirements to show how a single, specific intervention worked. Some of these approaches may also be too demanding for certain types of humanitarian crises, such as rapid-onset emergencies.

The Good Enough Guide, MSC and Contribution to Change discussed above all take this approach, as does the increasingly popular ‘outcome harvesting’ method, which uses a six- step process to identify positive and negative, intended and unintended outcomes, and then articulates verifiable connections between these outcomes and initiatives of interest: see www.outcomemapping.ca/download/wilsongrau_en_Outome%20Harvesting%20Brief_revised%20Nov%202013.pdf.

Traditional approaches for evaluating urban interventions can still be relevant and appropriate. The ALNAP evaluation guide presents different evaluative options, such as project evaluation, process evaluation and impact evaluation and accompanying methods, including case studies, process reviews, outcome reviews, before and after comparisons, interrupted time series and comparison groups.+ALNAP, Evaluation of Humanitarian Action Guide, pp. 193–214.

Evaluation is not the only way to promote organisational learning, nor is it necessarily    the most cost-effective. Formal evaluation of humanitarian action sits alongside a range   of additional learning and accountability tools, from beneficiary tracking to monitoring systems and After-Action Reviews. Other learning processes to consider in humanitarian action are also presented in the ALNAP guide.

Consider using iterative approaches

As noted throughout this GPR, cities are dynamic environments, where needs often change quickly. For this reason, many approaches to monitoring and evaluation emphasise an ongoing, iterative process that relies less on establishing whether pre-defined indicators are being achieved and more on understanding what is changing, and how (and whether) the humanitarian response is achieving these changes.

Examples of this type of iterative approach include:

The Good Enough Guide and MSC approaches outlined above are both intended to be used iteratively.

It is worth noting that simply having an iterative approach is not enough: effective monitoring and evaluation of urban humanitarian action requires not only that monitoring and evaluation systems are in place, but also that they are linked to the relevant decision- making procedures and systems. Early engagement with key users (see above) can go some way to addressing this problem. For more suggestions on ensuring that evaluative (and, by extension, monitoring) information is used, see A. Hallam and F. Bonino, Using Evaluation for a Change: Insights from Humanitarian Practitioners (London: ALNAP/ODI, 2013) (www.alnap.org/system/files/content/resource/files/main/alnap-study-using-evaluation-for-a-change.pdf).

Box 3.16 Iterative monitoring and evaluation of the 2010 Haiti earthquake

Following the 2010 earthquake in Haiti, Groupe URD conducted a number of iterative+See https://www.urd.org/en/. and real-time evaluations. Iterative evaluation aims to ‘analyse how a programme is being implemented in relation to changes in context and needs, and to ensure that the programme remains relevant and that there is effective coordination between the actors involved’. Between 2012 and 2015, Groupe URD maintained a ‘Haiti observatory’ to conduct iterative monitoring, promoting learning and good practice.

More information about the work of the Haiti observatory can be found at https://www.urd.org/en/research-page/?zone_geo=crise-central-america-caribbean;haiti

Work collaboratively with other stakeholders

Another common feature of many monitoring and evaluation approaches  suited  to  urban environments is that they are intended to be used by multiple stakeholders, including humanitarian agencies, local government and civil society. In fact, many aim to simultaneously produce information and build shared understanding of the context and a shared commitment to response activities.

Shared or joint approaches have important advantages in urban contexts, where, in the wake of a humanitarian crisis, many organisations will be working on response activities that will influence (and hopefully support) each other. Joint monitoring and evaluation could facilitate an understanding broader than any one project, and ‘may provide better opportunities to document challenges, shortcomings, failures and successes’, as well as potentially revealing systemic issues, rather than individual cases.+Brown et al., Urban Crises and Humanitarian Responses.

It should be noted that collaboration can often be difficult, particularly where the agencies involved are competing for funding or where they have very different organisational structures and cultures. See Section 2.1 on coordination for further discussion on this. See also T. Beck and M. Buchanan-Smith,  Joint Evaluations Coming of Age? The Quality and Future  Scope of Joint Evaluations (London: ALNAP/ODI, 2008) (www.alnap.org/help-library/joint-evaluations-coming-of-age-the-quality-and-future-scope-of-joint-evaluations). The UN Evaluation Group’s Resource Pack on Joint Evaluations (2014) may also be helpful: www.unevaluation.org/document/detail/1620.

3.10.3 Remote monitoring in conflict situations

Monitoring in conflict situations is inevitably governed by access and security concerns. In cases where access by international humanitarian actors is restricted, remote monitoring may take place. In contexts where access is severely limited, remote monitoring, as a wider part of remote management, may be the only option. Remote management can be defined as:

A reactive stance in response to insecurity that involves some delegation of authority and decision-making responsibility to national implementers. There is commonly a moderate investment in capacity building for nationals and procedures in place that enable better communication, monitoring, and quality. Assumes that decision-making and authority will revert back to international [staff] following the restoration of security.+S. Choudhri, K. Cordes and N. Miller, Humanitarian Programming and Monitoring in Inaccessible Conflict Situations: A Literature Review, Health Cluster, 2017 (www.who.int/health-cluster/resources/publications/remote-lit-review.pdf).

A 2017 literature review of humanitarian programming and monitoring in inaccessible conflict settings+Ibid. noted that ‘remote operations require increased monitoring and reporting requirements than traditional programming due to the lack of field presence and direct oversight by international organizations, but often have fewer resources to meet these increased demands’. Challenges include ‘limited opportunities for data collection, poor quality data and inaccurate information, and lack of monitoring skills and capacity of local staff, among others’.

An important issue here concerns risk transfer to local staff and other local actors. Concerning this, the same literature review concluded that ‘Remote operations involve the transfer of risk from international to local actors, who are assumed to be at lower risk for targeting and therefore safer when implementing. This is often a false assumption as they face unique threats that are often not acknowledged in security assessments. Additionally, local actors are infrequently present at trainings on security, and are often left with minimal security- related equipment when expatriates evacuate’.+Ibid., p. 8.

Where access is impossible, such as in conflict situations, existing data and resources are sometimes used. For example, Food and Agriculture Organisation (FAO) and World Food Programme (WFP) used government data for their 2018 monitoring report on food security in 16 conflict-affected countries.+FAO and WFP, Monitoring Food Security in Countries with Conflict Situations: A Joint FAO/WFP Update for the United Nations Security Council, January 2018, Issue 3 (www.fao.org/3/I8386EN/i8386en.pdf), p. iii.

Box 3.17 Libya Joint Market Monitoring Initiative (JMMI)

The Joint Market Monitoring Initiative (JMMI) was established in 2017 by REACH and the Libya Cash and Markets Working Group (CWG) to monitor market dynamics in order to improve cash programming. REACH describes the methodology for data collection as follows: ‘The methodology for the JMMI is based on purposive sampling. In each assessed market, at least four prices per item need to be collected from different shops to ensure the quality and consistency of collected data.

‘Partner field teams, in coordination with the CWG, identify shops to assess based on the following criteria: 1. Shops need to be large enough to sell all or most assessed items. 2. Prices in these shops need to be good indicators of the general price levels in the assessed area. 3. Shops should be located in different areas within the assessed city or baladiya.

‘In locations where it is not possible to identify four large markets that fulfil criterion (1), smaller shops, such as grocery shops, vegetable vendors, butchers and bakeries, are added to the shop list, as long as they fit criteria (2) and (3), in order to guarantee at least four prices per item of interest. Each month, price data is collected from the same shops whenever possible to ensure comparability across months.

‘The CWG primarily targets urban areas throughout Libya, aiming to ensure coverage of markets that serve as commercial hubs for surrounding regions. Data is collected via the KoBo mobile data collection application. The CWG maintains a joint KoBo account for the JMMI. The data collection tool is published alongside the dataset every month and disseminated to the humanitarian community’.

Monitoring takes place monthly of a ‘minimum expenditure basket’, comprising both food and non-food items. REACH reports that ‘By following the price developments of products such as bread, beans, soap and fuel, REACH and the CWG have been able to provide humanitarian actors [with] information on the financial burdens faced by households dependent on market priced goods in their respective localities’. The information indicates variations between cities and regions, wherein ‘the assessment noted clear spatial patterns both in May and June [2018] with basket costs generally lowest in coastal port cities and highest in southern Libya’.

Source: REACH, Libya: What Does It Take to Make Ends Meet: Understanding Financial Burdens with the Aid of the Minimum Expenditure Basket (Geneva: REACH, 2018) (https://reliefweb.int/report/libya/libya-joint-market-monitoring-initiative-jmmi-1-10-october-2018).

For more discussion on remote monitoring, see:

Workshop Summary: Remote Monitoring, Evaluation and Accountability in the Syria Response, ALNAP and DEC, 27 June 2014 (www.alnap.org/system/files/content/resource/files/main/alnap-dec-syria-workshop-summary-final.pdf).

E. Sagmeister and J. Steets, The Use of Third-party Monitoring in Insecure Contexts: Lessons from Afghanistan, Somalia and Syria, SAVE Resource Paper, October 2016 (https://www.gppi.net/media/SAVE__2016__The_use_of_third-party_monitoring_in_insecure_contexts.pdf).

B. Norman, Monitoring and Accountability Practices for Remotely Managed Projects Implemented in Volatile Operating Environments, Tearfund, 2012 (https://www.elrha.org/wp-content/uploads/2015/01/Remote20Monitoring20and20Accountability20Practice20_web2028229.pdf).