33
Technical Paper Vulnerability Analysis: Concepts and Case Studies in emergency, recovery and development settings January 2004 WFP - VAM Rome

Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

  • Upload
    others

  • View
    3

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

Technical Paper

Vulnerability Analysis:

Concepts and Case Studies

in emergency, recovery and development settings

January 2004

WFP - VAM Rome

Page 2: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

2

ACKNOWLEDGEMENTS Under the general guidance of Annalisa Conte (Chief, VAM Unit) this paper has been prepared by the VAM Unit at WFP Headquarters by:

• Marco Knowles (Consultant)

• Eric Kenefick (Emergency VAM officer)

• Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions of the WFP staff, Government counter parts, NGO partners and funding agencies in Dominican Republic, Ivory Coast, Malawi, Mali, Sierra Leone and the Western Sahel as well as WFP staff in Regional Bureaux in Dakar, Johannesburg and Panama City. Lastly, we would like to thank Sheila Grudem and Robert Opp for their helpful editorial comments.

Page 3: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

3

TABLE OF CONTENTS

SECTION I: INTRODUCTION ...........................................................................4

Background and Objectives......................................................................... 4

Structure of the report ............................................................................... 4

Methodology............................................................................................. 4

SECTION II: KEY CONCEPTS OF FOOD SECURITY AND VULNERABILITY ....................................................................................................5

Food security ............................................................................................ 5

Primary Data Needs for Community, Household and Individual Levels ............... 6

Livelihoods, Risk and Vulnerability ............................................................... 6

Data Collection Methods ............................................................................. 7

Data analysis ............................................................................................ 8

Links to Programming ................................................................................ 8

SECTION III: CASE STUDIES ........................................................................10

Malawi ....................................................................................................11

Ivory Coast..............................................................................................14

Western Sahel .........................................................................................17

Sierra Leone ............................................................................................20

Mali ........................................................................................................24

Dominican Republic ..................................................................................27

SECTION IV: RESOURCES NECESSARY FOR VULNERABILITY ASSESSMENT AND ANALYSIS ........................................................................31

Financial inputs and sources of funding........................................................31

Staff requirements for the Vulnerability Assessments.....................................31

Temporal requirements .............................................................................32

Logistical Requirements.............................................................................33

Page 4: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

4

SECTION I: INTRODUCTION

Background and Objectives

Over the past two years, the Vulnerability Analysis and Mapping (VAM) unit in WFP-HQ has provided technical support to several food security and vulnerability studies around the world under a variety of circumstances in order to provide vital information necessary for appropriate emergency responses or programme development. This paper is intended to illustrate the many contexts under which vulnerability analysis has been conducted. Therefore, selected examples are presented here to document the processes followed and the methods utilized. The paper highlights the need to approach each scenario as a unique situation and to develop an analytical methodology which provides information to answer the following questions:

Who are the food insecure?

How many are they?

Where are they located?

Why are they vulnerable to food insecurity?

Is there a role for food aid?

Structure of the paper

The paper consists of four sections: I) Introduction; II) Key Concepts; III) Case Studies and IV) Resources Necessary to undertake Vulnerability studies. Section II highlights key concepts of food security, livelihoods, risk and risk management and their relation to food security and vulnerability analysis. Appropriate uses of quantitative and qualitative methods are also presented in this Section with the basics on data analysis and interpretation and their links to programming. Section III provides an extensive step-by-step review of the approaches and methods used for data collection and data analysis in six vulnerability studies: Malawi, Ivory Coast and Western Sahel (Emergency); Sierra Leone (Recovery) Mali and Dominican Republic (Development). These case studies were chosen as good examples of VAM analytical work jointly conducted by VAM/HQ and VAM/Field in different settings and for different purposes. Nonetheless, we recognize that there are some aspects of these activities that can be improved. Section IV covers the various resources necessary to conduct vulnerability studies, in terms of financial, human, temporal and logistical requirements.

Methodology

These six examples of vulnerability analysis were selected to present a wide range of situations under which data collection and analysis are required. Three were undertaken in both slow-onset and complex emergencies, while others were undertaken in the context of recovery and rehabilitation or development. Some studies were part of a regional collaboration with many partners; others were national in scope and carried out with WFP national partners, while others were WFP alone. Partnerships played a key role in the choice of the method(s) used for data collection and analysis. In order to collect information on the approaches employed, a framework for documenting the six processes was developed and submitted to the VAM-HQ staff who had served as key technical experts for the selected studies over the past two years. Sections III and IV of this report are the result of the synthesis of the process information for the six vulnerability studies.

Page 5: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

5

SECTION II: KEY CONCEPTS OF FOOD SECURITY and VULNERABILITY The concepts of food security are multidimensional and need to be analysed at different levels - community, household and individual - in any vulnerability analysis. Over and above the analysis of the three dimensions of food security, some additional elements must be taken into account: (a) livelihoods, (b) risk exposure and risk management strategies and (c) vulnerability. Analysis of each element, and the multi-directional linkages among elements, assists in creating a better understanding of food security in terms of availability, access and utilization. Information for the analysis of food security and vulnerability at different levels often requires both quantitative and qualitative data. Quantitative methods help in systematically measuring the food security situation, thus providing a “representative” snapshot at a particular point in time across a number of communities, households or individuals. Qualitative methods, on the other hand, attempt to capture the processes and interactions between social, political, and economic risks which contribute to greater vulnerability to food and livelihood insecurity—as perceived by men and women within specific communities. A wide range of factors emerge from the use of qualitative methods, especially insights into the possible causal explanations of particular problems. When combined, both methods allow for a more comprehensive picture of food security and vulnerability including the analysis of population’s livelihoods. However, analysis of food and livelihood security data requires specialized skills, not only in terms of computing skills but also in synthesis and interpretation of the data within the context in which it was collected – be it emergency, recovery or development.

Food security

At the 1996 World Food Summit (WFS) it was agreed that, “Food security exists when all people, at all times, have physical and economic access to sufficient, safe and nutritious food to meet their dietary needs and food preferences for an active and healthy life.” (CFS, 1996) This definition of food security integrates food availability (at household level, regional and national levels), food access and the biological utilisation of food. These three factors are interrelated. For instance, food availability in local markets is useless unless people have access to food. Also, access is useless unless people are free from poor health that reduces biological utilisation (Lieldholm and Weber 1994; Webb and Rogers 2003). This demonstrates that, to be effective, interventions intended to increase food security need to be developed on the basis of knowledge on the status of each of these factors and of the processes influencing availability, access and utilization. Food availability usually is measured at the national and household levels. National food availability from domestic production is enhanced by net food imports. Household food availability is primarily determined by household food production activities. This may include cultivating food crops, fishing, livestock rearing and hunting and gathering. Food access is determined by people’s purchasing power which can vary according to market location, price policies, seasonal food availability and people’s sources of income. Food is also accessed through exchange, in-kind income, gifts from family or friends and food aid. Food utilization incorporates food safety and quality, sufficient intake and absorption of food by the body, along with health and care factors that result in sound nutritional status and growth. As regards the absorption efficiency of the body, lack of nutrition education and the disruption of health infrastructure have detrimental impacts (Webb and Rogers, 2003). The vicious cycle of disease and malnutrition are important in individual food

Page 6: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

6

security analysis. For example, diarrhoeal diseases can affect nutritional status by reducing appetite, the body’s effectiveness in absorbing nutrients and increasing consumption needs. Indicators to capture morbidity, such as recent occurrence of diarrhoea or incidence of malaria are relevant. Interpreted with care, anthropometric measurements can also be used to assess food utilization. The “at all times” dimension of food security implies that an understanding of the current status should be coupled with an understanding of the seasonal components of their current status. Frameworks for analysing food security must, therefore, capture the dynamics of food security over time—or the vulnerability of households to shortfalls in food availability, access or utilisation. Households are considered vulnerable to food insecurity if a negative event is likely to constrain future food access, availability or utilisation options. Vulnerability is determined not only by exposure to risks, but also by the characteristics of households exposed to those risks. Depending on the type of livelihoods pursued by households, some are better able to manage the negative effects of risks, thus decreasing their vulnerability to a set of adverse outcomes - harvest failures (i.e., availability); reduced household income (i.e., access); increased incidence of illnesses (i.e., utilization).

Primary Data Needs for Community, Household and Individual Levels

Primary data collection, from a vulnerability analysis perspective, is necessary when existing secondary data do not sufficiently explain the food security situation. In this context, the goal of any primary data collection and analysis exercise is to fill in these data gaps with relevant information at community, household and individual levels. Community level information is used to help understand the local context including demographics, economic issues, the risk context, community assets, and services. This allows for development of community profiles and a basic understanding of the livelihood strategies found within particular communities as well as for triangulation with household and individual data. The information collected enumerates physical structures/assets, social networks, availability and access to health and education services as well as natural resources such as rivers, forests or grazing land. Household level information that is obtained from structured interviews with household members is used for assessing household availability and access to food through purchase, production, exchange/barter, gifts or food aid. In addition, valuable information on household level expenditure and consumption as well as more detailed information on main livelihood activities can also be enumerated. Individual food security is usually assessed for young children and their mothers through anthropometric measurements (weight and height) as well as through structured interviews to collect information on access to and utilization of health services, caring and feeding practices, hygiene and recent morbidity.

Livelihoods, Risk and Vulnerability

Livelihoods and risk are important components of food security analyses. In simple terms, livelihood strategies are the combination of activities that people pursue in order to improve their welfare in some way: stabilising the fluctuation of income; ensuring children are able to go to school; being able to afford access to health services; or better managing natural resources. Such activities are based on a set of tangible and intangible assets available and accessible to households. These include:

• Human assets: health, level of education, skills • Financial assets: credit, savings, remittances, income • Social assets: extended family, local organisations, community networks • Physical assets: schools, health clinics, markets, roads, tools • Natural assets: land, water, forests, livestock, common property resources

Livelihood analysis is helpful for two reasons. First, it allows for a better understanding of household priorities, as livelihood activities have different goals—better access to health

Page 7: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

7

care, accumulation of savings, or diversifying sources of food and income. Second, analysis of the types of livelihood activities adopted by food-secure households is useful in determining how best to assist food-insecure households—especially with food assistance that can leverage productive livelihood activities. Given that livelihood strategies change over time and from one season to another, determining the risks facing households naturally complements any livelihood analysis. As noted in the previous section, the analytical goals are to determine:

• The type of risks affecting households; • The effects of said risks on household welfare; and • The ability of livelihood strategies to manage and mitigate adverse effects of risks,

thus minimising welfare losses. Much like asset endowments, risks, too, can be categorised in broad terms such as:

• Health and Life Cycle Risks: illness, injury, occupational accidents, disability, epidemics (i.e., HIV/AIDS), famines, old age, death in the family;

• Social and Political Risks: crime, domestic violence, civil war, exclusion based on ethnicity, class or gender, political unrest, coup d’états;

• Natural Risks: heavy rainfall, landslides, floods, droughts; • Economic Risks: unemployment, harvest failure, business failure, resettlement,

financial crises, personal debt, macroeconomic policy failures; and • Environmental Risks: pollution, poor sanitation systems, deforestation and land

degradation. Livelihoods analysis does not have a particular order or sequence in terms of organising data from quantitative or qualitative surveys. However, analysts do need to decide on areas of foci which are likely to provide insights on livelihood realities. Data collection for risk and livelihoods can benefit from household questionnaires containing modules on:

• Household demographics; • Asset ownership; • Sources of income; and • Past risk exposure and management.

Community and focus group interviews can also be used to complement these data, in particular modules on community economy and infrastructure, health and education facilities, agricultural patterns and specific modules related to risk occurrence and response. When organising information for analysis, livelihood strategies can be prioritised according to their relative importance in:

• Being steady sources of income; and • Being used as a real or potential response to a particular risk factor.

The same can be undertaken for prioritising risks based on the:

• Level of exposure by households to a particular risk (i.e., how many households face this particular risk); and

• Effect on household welfare (i.e., results in welfare losses). Again, some level of clustering is possible to facilitate future analysis by grouping households who have similar sources of livelihoods (i.e., farming, wage labour, migration) and/or clustering according to responses to risk.

Data Collection Methods

Probability surveys using structured household questionnaires are the most common form of quantitative data collection. In order to have valid, reliable and generalizable conclusions from a household survey, the sample must be drawn using appropriate and methodologically rigorous sampling techniques. In turn, questionnaires must be developed and tested by those familiar with quantitative data collection and analysis. Enumerators are trained to primarily collect the information rather than

Page 8: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

8

undertake both data collection and analysis. Anthropometric measurements are also considered to be quantitative because they provide a measurable outcome. The most commonly used qualitative data collection methods are focus groups and participatory appraisal (PA) techniques. These methods are flexible and are used to collect information for understanding local factors related to food security and vulnerability. Qualitative information is usually collected at the community or group (i.e., men, women) level to complement quantitative household and individual information. Designing appropriate methods for data collection requires a proper understanding of the specific information needs. Additionally, those collecting qualitative data need to have strong skills in facilitation and interpretation of “narratives”—piecing together different bits of non-numerical information. It is important to reiterate that qualitative methods, alone, cannot quantify levels of food insecurity. This can only be achieved through quantitative data. However, the two methods, when combined, provide the best overall approach to understanding food security and vulnerability.

Data analysis

Much of the analysis of quantitative data in these case studies was conducted by persons with expertise in using specialised software such as ADDATI, SPSS or S-Plus. In many analyses, it is important to provide simply descriptive information such as percentage of female headed households by district, or percentage of children suffering from chronic malnutrition—by age group. Quantitative data analysis allows the user to look not only for numeric differences between groups/regions but also to see if these differences are large enough not to have occurred by chance (i.e., statistically significant). Other techniques used in quantitative analyses have been linear regression, multivariate analysis and cluster analysis. Qualitative data analysis also requires a high level of expertise, especially in isolating and understanding commonalities and differences among study populations—based on narratives not numbers. In contrast with quantitative methods, analysis often begins during the data collection phase, with initial assumptions being generated by those collecting the data with participants in focus or community group interviews. Initial assumptions are then revisited when reviewing focus group transcripts. Linking qualitative and quantitative findings, however, takes time. Much like quantitative data analysis, specific variables are analysed across focus groups and community discussions—leading to higher level clustering of issues which are common and highlighting important divergences that can have important implications for programming. Both quantitative and qualitative analyses are important in food security and vulnerability analyses. Quantitative data are mostly used to describe and quantify the situation, while qualitative data describe the dynamic nature of community and household food security.

Links to Programming

Creating household food security profiles for sectoral and beneficiary targeting

Irrespective of some of the contextual commonalities shared by the population under study, findings are not monolithic. There are likely to be significant differences in household food security explained by differences in asset holding, livelihood activities and vulnerability. It is, therefore, necessary to account for this heterogeneity. Categorizing the population under study into household food security groups allows the development of programming responses that are suited to the characteristics of the different groups, thus improving sectoral targeting.

Page 9: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

9

Moreover, profiling of food secure as well as food insecure groupings is useful in programming responses since it allows understanding of why in the same area some households are more successful than others in achieving food security. Creating community food security profiles for geographic targeting

Sub-national or village profiles are created by calculating the percentage of each household food security typology in each geographic unit (e.g. District). These profiles assist in targeting assistance geographically by identifying areas and villages with the greatest prevalence of food insecure households. The description of the livelihood/food security characteristics associated with the different profiles also supports sectoral targeting. Estimating emergency food needs

Baseline food security and vulnerability studies allow the estimation of the total number of people affected by chronic and transitory food insecurity (i.e. vulnerable) as well as their geographic location. In an emergency situation, this reference information along with the assessed current food availability and accessibility allow for a more thorough estimation of people in need of emergency food assistance and the tonnage of food aid required at different administrative levels. In the case of slow-onset emergency situations, such as the southern Africa food crisis, the analysis required use of sampling and statistical techniques to make estimates of population percentages at district level and also to provide a description of the vulnerable populations for targeting purposes. Targeting of particular ‘safety nets’ programmes

Since WFP’s food aid response programmes are limited in scope, it is important to link the results of vulnerability analysis to the options for programming, based upon the severity vulnerability severity among the population in general and among particular social groups. The methods used to collect the data should then yield information on maternal and child health and nutrition for MCH programmes, availability and access to education and enrolment for School Feeding programmes and on household food and livelihood security for Food for Work programme consideration. Recommending appropriate actions, food aid vs. non food-aid

Recently, policy and institutional analyses have been included in vulnerability studies together with the analysis of traditional food security indicators. This multi-dimensional approach assists in more precise identification of food security and well-being constraints and the appropriate type of interventions, which may be food based or non-food based, inviting opportunities for new and creative partnerships with other UN agencies or NGO partners. For example, if the government’s education policies are found to be a constraint to food security, there may be the need for advocacy to reform education policies rather than provision of food aid. The possible negative implications of using food aid can also be understood. For instance, if the food security problem is related to food access, food aid may be useful in overcoming short-term hunger but in the long-run may create dependency and will not be effective in addressing the underlying causes of food security.

Page 10: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

10

SECTION III: CASE STUDIES The processes and methodologies for each of the six vulnerability analysis case studies presented in this paper are documented in three stages: • Preparation: Organizational and practical activities, necessary to be carried out prior

to the field work such as identification of field enumerators, development of field instruments for data collection, logistics;

• Data Collection: Practical activities based on the purpose and objectives of the study.

Includes decisions on data collection methods, trainings to ensure the quality of the data collected, and survey teams management;

• Data Analysis, Report Writing and Presentation: Desk activities aimed at meeting

the purpose and objectives of the vulnerability study (i.e., data entry and analysis), report writing, mapping and presentation of findings to stakeholders.

The first step in carrying out vulnerability studies is to achieve clear consensus with all stakeholders involved in the process on the purpose and objectives of the study and to identify the data collection area and the unit(s) of analysis. Generally, once the survey area has been well defined, the second step is to select a number of villages to represent the area as a whole. Villages are randomly selected - usually from available lists. The sampled villages are generally informed in advance of the arrival of the enumerator teams. Focus and community group interviews are normally carried out with relatively small groups of villagers to discuss issues related to their food security, risk and livelihoods. Whenever the unit of study is the household, a list of all households in the village is created and a number of eligible households (usually 10-15) are selected using probability sampling techniques. Household interviews are then carried out and often young children (and their mothers) are weighed and measured at the end of the interview. Data analysis involves analysing qualitative data collected through the focus groups and quantitative data gathered using the household questionnaires. Using multivariate statistics, household food security profiles are created from a relatively small set of indicators: (a) food diversity; (b) food frequency; (c) food sources; (d) shares of food and food expenditures; (e) sources of income; and (f) assets ownership. Percentages of the different household food security profiles are then calculated at village and district level. The prevalence and the combination of the food security profiles within a village and a district are then used to create geographic food security typologies. These are then mapped to present the geographic distribution of food security and vulnerability. A description of the processes and methodologies used in the different vulnerability studies is provided in the following pages. They are grouped according to the type of WFP response that they were intended to inform: emergency, protracted relief and recovery, or development.

Page 11: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

11

Emergency

Malawi

Background and Objectives

The Malawi Vulnerability Assessments were carried out in response to regional food crisis characterized by sequential poor harvests as well as general poverty and an economic downturn. In this context, the assessment and analysis activities were intended to provide information on both food and non-food needs. Regional and National Vulnerability Assessment Committees (VACs) were responsible for the overall coordination and implementation of the assessment activities and were headed by the Southern African Development Community – Food, Agriculture and Natural Resources division (SADC-FANR) and national governments, respectively.

In addition to national and regional government representation, VACs included participation by international NGOs and UN agencies. The regional versus national hierarchy within the co-ordination mechanisms—along with the multiple members and objectives of the assessment—hampered the ability to respond to country specific information needs and raised new issues related to the assessment objectives and methodology. Preparation

The WFP Regional and Country Offices played a key role in negotiating and coordinating with the Government and partners, by appointing a full-time national VAM officer to participate in the VAC meetings and assessments and provided vehicles for the fieldwork. Given the limited capacity of the National VAC, VAM-HQ offered its technical expertise and support particularly for the data collection, analysis and reporting, participating in the first two assessment rounds carried out within a six-month timeframe.

Context: Emergency - Slow onset drought/crop failure

Objectives: Sub-national food needs estimates and targeting criteria for regional EMOP.

Coverage: National as part of 6 country Regional Vulnerability Assessment Committee activity.

Partners: At regional and national levels, all NGOs, Government, and UN agencies with funding from WFP and coordination from Regional Vulnerability Assessment Committee (RVAC) which had WFP representation.

Duration: Round 1 – three months; Round 2 – four months

Successes: • Information was timely and usable by WFP and partners to distribute general food

rations. • Process helped to develop government capacity and strengthen coordination of

partners within the country. • Funding for vulnerability analyses was adequate and readily available. • Two sampled household and community surveys were carried out and results

presented and published.

Difficulties: • RVAC/EMOP wanted universal approach for 6 countries although each country was

experiencing a unique complex emergency with various combinations of political, structural, economic and food security emergencies.

• In-house technical capacity of RVAC and Malawi VAC was limited with RVAC focused mostly on coordination and MVAC lacking any resident technical ability.

• Too many partners with both national and regional reporting obligations as well as a top-down approach by RVAC with little national level consultation.

• Hidden agendas by certain individuals and agencies distracted from the main focus of the assessment and analysis activities.

Page 12: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

12

In response to the Regional VAC’s desire for a uniform approach for all six countries in its constituency, household and community questionnaires were centrally developed by Save the Children Fund-UK (SCF-UK) staff in Harare at the Regional VAC. The questionnaires were inadequate for the scope of the assessment and were substantially modified by WFP-VAM, drawing from its own experiences as well as from the Malawi DHS (2000), World Bank Socio-Economic Survey (1998) and IFPRI. Data Collection

Data collection was carried out by Malawian enumerators. A two-day enumerator training was designed and imparted by Save the Children-UK, WFP/VAM and FEWS-Net. The training was meant to ensure that the questionnaire modules were well understood and translatable, and also provided information on the use of wealth ranking tools and GPS units. In addition, the training addressed issues related to village and household sampling, logistics and administration (finances, codes of conduct, etc.). Too little time was dedicated to the field-testing (half a day only) and during the first week of data collection many issues on the use of the questionnaire had to be clarified. For the second round assessment, in order to minimize these issues, about eighty-five percent of the enumerators hired had participated in the first round. The work plan for data collection was developed through discussions in the National VAC with substantial inputs from WFP-VAM. The WFP-Malawi national VAM officer worked closely with the National Economic Council (NEC) and VAC members to organise vehicles and logistics. Since the necessary financial resources were available, the tight time constraint was less of an issue.

Since one of the main objectives of the vulnerability assessment and analysis activity was to identify food and non-food needs throughout the country, the data collection area was country wide. Stratified random sampling, using livelihood zones within the Extension Planning Areas (EPAs) as the stratum, was applied in order to select villages. This ensured that the sample contained villages from each livelihood zone within an EPA. In the second assessment, villages were randomly sampled from different livelihood zones with probability proportional to size (PPS). However, the sample frame used for this random sampling had been drawn for another purpose for which the total number of households in the country had been overestimated and thus the number of households in some villages was overestimated. In most of the selected villages an official from local government or an enumerator informed the village in advance about the arrival of the assessment team. Advance notification can be crucial when conducting household interviews to ensure that villagers are at home. The first assessment round was conducted during a season when people had more free time (post-harvest/pre-planting) so advance notification was less of an issue. The enumerator teams were accompanied to the villages by an agricultural extension worker who helped to introduce the enumerators to villages. This person received a small monetary compensation for his/her time – this compensation was a part of the survey budget. In each village key informant interviews were used to discuss various food security issues including crop production, access to inputs and water, the main sources of food and income, market access and prices, and HIV/AIDS awareness. Key informants were also used to classify households using a wealth ranking exercise. Household sampling was based upon the results of the village wealth ranking where households were selected by socio-economic group based upon their representation in the village. It was found that by selecting households from various wealth groups, enumerators were able to probe more when locally identified well-off households claimed to have nothing during the crisis. Male and female household members were interviewed together. The household questionnaire involved a considerable amount of estimating on behalf of the interviewees

Sampling: if the wealth ranking exercise indicated that 60% of the households were ‘poor’, 30% were ‘middle’ and 10% were ‘better off’ then the team would randomly select 6 poor households, 3 middle and 1 well-off household.

Page 13: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

13

(Household Food Economy Approach). As a result, the reliability and validity of the data were often only as good as the ability of the particular enumerator. For example households would purposely underestimate their stores and ability to purchase food in hopes of obtaining assistance. However for the households, there were also genuine difficulties in estimating past income and food production and in making projections for these. Nutritional data were not collected directly as part of the rolling assessments because UNICEF coordinated a series of district-level anthropometric data collection exercises with various NGOs in the country. UNICEF endorsed a 30 by 30 cluster methodology within an entire district to obtain prevalence of wasting and stunting for most districts. Since the nutritional data were collected to estimate district level prevalence only there were no opportunities to make direct linkages between household food security and child (< 5 years) malnutrition. No data were collected on adult malnutrition although it had been recommended since child wasting was very low (< 3%), even in the middle of the crisis. Data Analysis

Enumerator team leaders were responsible for reviewing all household questionnaires before leaving the villages. Completed questionnaires were continually sent to Lilongwe from the field to facilitate the data entry process. The questionnaires were reviewed and coded in Lilongwe before data entry began. In the first round, the data were entered by the VAM-HQ staff person using Epi-Info while in the second round two data entry clerks were hired to enter data into an Access file. In both instances data entry and cleaning took about 4 weeks of intense work. In the first round, more than 1100 questionnaires were entered while more than 1600 questionnaires were entered in the second round. The assessment methodology included the use of both qualitative and quantitative data with more reliance on quantitative information. One of the main objectives was to estimate the percentage of the population in need of food aid at the district level and to investigate potential causes of the differences between districts. All analysis was conducted using SPSS software. In the first round of the vulnerability assessment the percentage of the population in need of food aid from September till the end of the year was calculated using the household data and involved two stages of data analysis (see box). Changes were made to the household questionnaire for the second round which facilitated a more direct calculation of the food gap for each household, eliminating the need to use the more sophisticated statistical techniques to estimate the percentage of households in need of food aid for each district. Descriptive analyses were used to build profiles of the most food insecure, with particular attention paid to households headed by women or the elderly, those keeping orphans, those with chronically ill productive members and those with no productive members. The results of both rounds of rolling assessments, including methods and vulnerability profiles are found in two publications: Malawi Emergency Food Security Assessment Report – September 2002 and Malawi Emergency Food Security Assessment Report – February 2003. Both are found on the Southern Africa Development Community (SADC) Food, Agriculture and Natural Resources (FANR) website.

In Round One, the first stage of analysis involved the estimation of the household ‘food gap’. However, as the absolute food gap measure for most households was thought to be inaccurate due to problems in estimating past and future food availability and access, linear regression was used to build a model to predict the food gap using eleven predictor variables and taking into account co-linearity between variables.

The second stage involved estimating district-level food needs. This was determined by creating a weighted index measuring household vulnerability to food insecurity using the beta coefficients (their absolute contribution) from the eleven-predictor variables in the regression model for each household.

Although viable results were obtained from the analysis, this involved substantial statistical skills to overcome the problems in estimating a food gap (Household Food Economy Approach).

Page 14: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

14

Emergency

Ivory Coast

Background and Objectives

A Rapid Risk Assessment in the Ivory Coast was undertaken within the context of an unfolding civil conflict. Given the security situation, the assessment was subject to last-minute changes in the planned itinerary as well as areas to be visited. The objective of the assessment and analysis was to identify people’s perceptions of the evolution of the crisis in terms of their current capacity to deal with the change and also to mitigate or cope with future risks. The findings of the assessment, particularly those related to risk management were meant to be used to establish a food security and risk monitoring system. Preparation

WFP initiated and led this assessment with the support and involvement of the Government, the National Statistics Office, the Office of Coordination of Humanitarian Affairs (OCHA) and the US Office of Disaster Assistance (OFDA). VAM-HQ provided the expertise for the survey design, training of enumerators and survey implementation. The experience of WFP CO staff proved invaluable in ensuring that the logistical aspects of the study were taken care of efficiently and effectively. The Government identified 12 enumerators - six men and six women with backgrounds in sociology and statistics - who were temporarily hired by WFP. They all spoke French, but had varied proficiency in the local languages. Data Collection

Context: Emergency – Rapid onset civil conflict/complex emergency

Objectives: Identify people’s perceptions of their capacity to cope with the current crisis as well as to mitigate or cope with future risks. Findings were to be used to establish a food security and risk monitoring system.

Coverage: Purposive sampling in areas around the four WFP sub-offices which was to be representative of most of the rural areas affected by the conflict.

Partners: WFP initiated and led assessment with support and involvement of the Government, the National Statistics Office, OCHA and OFDA.

Duration: Approximately 6 weeks in total.

Successes: • The results of the study were available within 6 weeks. • Sampling strategy and combined use of qualitative and spatial techniques was

innovative. • Additional output included a proposal for a food security and risk monitoring

system. • Data collection was successfully carried out during a civil conflict without incident.

Difficulties: • Study design and coverage was limited due to the conflict, requiring many last

minute changes to original plan. • Despite the objective of using study to design and implement a monitoring system,

there was no monitoring and evaluation officer participating in the vulnerability analysis activity.

• The proposal for the food security and risk monitoring system was never implemented due to difficulties of finding and hiring a qualified and willing person to manage the process.

Page 15: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

15

Tools were selected on their potential for providing indicators needed for a monitoring system based on people’s priorities. Key questions related to different risks and risk management strategies and their effects on mobility were explored. Thus a structured qualitative instrument was designed for the exercise.

Enumerators participated in a three day training course in preparation for data collection. The first section of the training focused on facilitation techniques and the remaining sessions were devoted to discussing the risk assessment framework and components of food security. A major contribution of the facilitators in this period was the evaluation and finalisation of the assessment instruments. Considering the context of civil unrest, safety and security issues were two other important elements of the training. The WFP Participatory Toolkit was a useful source from which to draw suitable assessment tools to assist enumerators in conducting focus groups discussions. These tools along with guidance sheets and reporting forms were prepared with the support of VAM-HQ. The work plan for the entire assessment including preparation, data collection and analysis was jointly prepared by VAM and the WFP CO. The most important consideration in developing and implementing the work plan was that of security. The time provided for the assessment was only two weeks, but the support provided by the Country Office helped ensure smooth implementation of the work plan. Because of the security situation, communities included in the assessment were chosen along three regional transects surrounding the four WFP sub-offices. This coincided with government controlled areas (which had seen the highest influx of IDPs) and the two rebel controlled areas where populations were cut-off from public services. This purposive sampling approach was the most appropriate under the particular security conditions. The risk assessment did not include a household survey questionnaire because the objective of the vulnerability analysis was to identify preliminary indicators for a future food security and risk monitoring system. Moreover, the idiosyncratic effects of a civil strife shock seem to be better captured by using socio-demographic focus groups rather than households. There was no need to assess the nutritional status of children at this point in the crisis. It was planned to be undertaken at a later stage. More than fifty focus groups were held in the 14 selected villages along the three transects. The groups consisted of women aged 15-35 years; men aged 15-35 years; community leaders; and families hosting IDPs. Focus group discussions proved to be the most appropriate way to: (a) select indicators for monitoring; (b) capture the dynamic aspects of the situation; and (c) understand the wide range and depth of issues related to people’s food security (i.e. conflict, displacement). Data Analysis

Qualitative data were analysed by VAM-HQ staff in Rome over a one week period—a relatively short time for a risk and livelihoods analysis framework. Variables were organised around the following analytical categories:

• Priority risks; • Responses by men and women; • Risk management strategies; and • Assets on which these livelihood/management strategies were based.

Food consumption and changes in food habits provided the basis for a secondary layer of analysis for selecting the appropriate indicators. Some difficulties were encountered in ensuring that suggestions for monitoring met WFP evaluation standards since, contrary to what had been planned, a monitoring and evaluation officer did not take part in the assessment and analysis.

Discussions with each focus group lasted about two hours and covered issues of food security, risk and risk management. Each enumerator used four participatory tools to facilitate the discussions: four boxes, force-field analysis, ranking and food frequency. Limiting the size of young men’s and women’s focus groups, that were originally meant to be of about 6-10 people, turned out to be difficult in areas with many IDPs.

Page 16: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

16

Information collected through focus groups was analysed along easily identifiable socio-demographic characteristics. Since displacement was the major issue affecting the people during the conflict, information from young men and women, being the primary sources of labour within the country, was analysed together. Similarly, data collected from host families were also analysed together since these households provide the basic needs to IDPs: (a) food; (b) shelter; and (c) protection. Simply meeting the basic needs was placing considerable strain on hosting households, thus highlighting the importance of monitoring the effects of increasing numbers of household members. Information collected from local leaders were analysed in light of the fact that they are recognised to be the authorities and community members expect them to take action to solve the ongoing problems. The set of variables described above were subsequently analysed to assess:

• food consumption and their changes; • food sources and their changes; • resources/assets to support livelihood/risk management strategies; • risk and risk exposure; and • strengths, weaknesses, opportunities, and constraints of existing livelihood

strategies. The spatial distribution of the groups was mapped to look for similarities and differences across groups and between geographic locations. Indicators for monitoring were then selected on the basis of which risks posed the greatest threat to food security (i.e., changes in consumption patterns) by analysing the effectiveness of risk management/livelihood strategies with respect to these risks. The results of the risk assessment were presented in three interrelated documents:

• Situation Analysis and Overview; • Risk Assessment; and • Proposal for a Food Security and Risk Monitoring System.

These documents are available on the WFP/VAM Web site (http://home.wfp.org/vam) or can be requested at the VAM Unit in WFP-HQs.

It was felt that traditional livelihood profiles place too much emphasis on types of employment rather than on the capacities of households and communities to manage risk. This results in a focus on sectors rather than people (i.e., agriculture, fishing, pastoralists).

Page 17: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

17

Emergency

Western Sahel

Background and Objectives

The Vulnerability Assessments in the Western Sahel which included Cape Verde, Mauritania, Senegal and the Gambia, were carried out to assess the impact of the drought upon the rural populations and to determine whether an external food aid intervention was required. Preparation

The planned assessments were presented by WFP to a forum of interested NGOs, donors and UN agencies—all of whom have ongoing collaboration with WFP within the region. Partner inputs varied across countries but in all, the participating Ministries provided personnel and in some cases also vehicles. WFP Country Offices managed budgets and organised logistics, training facilities and forums for discussion. Approximately eight enumerators were nominated by government partners for each country study and were paid for by WFP.1 Assessment teams were composed of two community facilitators, four enumerators for household interviews, and two nutritionists. All team members were professionals working in their respective fields of expertise in the Ministries of Health, Education, Agriculture and Planning. Previous arrangements between WFP and counterpart governments meant that it was not possible to provide enumerators with adequate cash incentives, creating some initial problems in ensuring continuous participation.

1 Financial resources were provided through the DFID ISP.

Context: Emergency - Slow onset drought in Cape Verde, Mauritania, Senegal, and the Gambia.

Objectives: To assess the impact of the drought upon the rural populations and to determine whether an external food aid intervention was required.

Coverage: Only drought-affected regions of each country.

Partners: Primarily WFP along with government ministries of Health, Education, Agriculture and Planning. Some NGO partner participation.

Duration: Entire process took 6-8 weeks.

Successes: • Good government participation and rapid turn-around of results to be used for

EMOP preparation. • Successful combination of quantitative and qualitative techniques used in the field.

Community-level incorporation of nutritional information was useful. • Regional approach was innovative. • Analysis successfully identified communities that were affected by drought but did

not require food assistance.

Difficulties: • Not enough time to train enumerator teams. • Linkages between household food security and child malnutrition could not be

made with the approach used. • Mostly led by regional and HQ VAM staff with little opportunity to build capacity

among country office staff and partners.

Page 18: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

18

Data Collection

Although team members were professionals, all required training on data collection. The training was provided by the VAM-HQ staff and it focused on how to carry out quantitative and qualitative data collection. Instruments for data collection were prepared in the countries by VAM (Regional VAM officer, and VAM-HQ staff), whereas the work plans for the country assessments were prepared with the participation of the Country Offices, taking into account the need to yield reliable findings within a short timeframe. Data collection instruments became de facto training materials, since each component of the questionnaires was discussed so as to clarify what information needed to be collected. The assumption that training could be undertaken in one day was inappropriate since several errors were found on completed questionnaire forms during the data analysis. Simple random sampling was used to select survey villages within the drought-affected regions of each country. Once the survey villages had been selected, provincial administrators were informed of the up-coming activities. Enumerator teams, however, had the responsibility of contacting selected villages three days prior to the survey visit. On arrival into selected villages, a community meeting was held to introduce the team members and explain the purpose of the visit and their respective tasks. Community discussions on the current food security strategies and how they have changed during the current year were introduced in the first stage. The community also developed a shock/risk calendar of recent events and their effects on food security. Strategies to manage these shocks were discussed and the most successful ones were determined by the community. For the household questionnaire, households were selected by enumerators using random selection methods (i.e. drawing of numbered pieces of paper, each representing a household in the village). To assess nutritional status, in each village thirty children between 6 and 59 months of age were randomly selected. The selected children were weighed and measured by nutritionists in the assessment teams. For this activity, the support of qualified partners such as UNICEF was sought in each country. Data Analysis

Qualitative data analysis was undertaken by the Regional VAM officer in Dakar using the risk and livelihoods analysis framework, disaggregating by gender. This involved linking together in a causal chain: risk exposure, risk effects, risk management and livelihood activities, and response capacities. Narratives emerging from qualitative data collection were checked for internal consistency, frequency of mention and extensiveness. This was used to contextualise the quantitative household data which was integrated in an overall report as a stand-alone section. This added value to the assessment since it provided information on people’s priorities. The analysis of the data was, however, partly constrained since the lead analyst lacked some of the necessary qualitative data analysis skills. Household questionnaires were checked, cleaned and coded by hand and then entered into SPSS by a team of 20 data entry staff over a period of 10 days. The large number of data entry staff made it possible to complete this process quickly. Data entry was supervised by the Regional VAM assistant who had also written the programme for data entry. Since the purpose of the vulnerability analysis was to identify whether or not there was a need for a food aid emergency response and if that was the case where and for whom, quantitative data analysis largely focused on household food consumption, income sources and expenditure levels. To determine household consumption welfare, data on food frequency and source of different foods (proxies of food access and availability) were used to establish a

Page 19: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

19

consumption threshold (based on kcal intake). Households that fell below a pre-determined caloric threshold were considered currently food insecure. Qualitative data on food consumption, food sources, expenditures and risk exposure and management were used as variables around which to describe the livelihood/risk management profiles of the household food security groups. The inclusion of risk exposure and management variables allowed to understand whether and how to use food-aid to strengthen people’s risk management strategies. The vulnerability level was calculated on the basis of total estimated resources for the year and whether these were sufficient to meet food needs for the year. If the former fell below the latter then the household was considered currently vulnerable to food insecurity. Household agricultural production, food received in-kind, and reported annual income were used to calculate household resources. Combined household resources were compared against the cash value of the household calorie requirement for that year.2 The cash value of the food requirement was calculated in the market value of the most commonly eaten cereal, in this case, maize. If household resources for the year exceed the cash value of the households calorie requirement for the year, the household could meet a minimum level of consumption and was not considered vulnerable that year. In order to quantify the number of people needing food aid, the three components of the analysis (food consumption, risk management, and vulnerability) were examined together. Again, using multivariate analysis tools, the following steps were undertaken: • Using the prevalence and distribution of household food security typologies within all

sampled communities, community vulnerability profiles were created which contained information on the proportion of food insecure people found in each typology.

• This information was used to determine the number of people in each sample that may benefit from WFP food assistance.

The results of the Vulnerability Assessments are found in Country Reports. The publications are available on WFP/VAM Web Site (http://home.wfp.org/vam) or can be requested at the VAM Unit in WFP-HQs.

2 See FAO/WHO/UNU Guidelines on energy requirements based on sex and age.

Using both multivariate and qualitative analysis techniques, indicators of current and future vulnerability were used to classify households based on their food security status. Five household categories emerged as a result of this analytical process:

• Currently Food Insecure – not meeting minimal consumption levels, nor have resources to meet minimum consumption level this year;

• Vulnerable to Food Insecurity – just meet minimum food consumption levels, but do not have resources to maintain level until next harvest;

• Potentially Vulnerable to Food Insecurity – more diverse consumption patterns, have resources to maintain current consumption, but will not be able to manage a shock/risk to current consumption;

• Less Vulnerable to Food Insecurity – diverse food sources and types, have resources to maintain current consumption and manage future consumption risks/shocks.; and

• Food Secure – diverse sources and types of food; have resources to maintain current and future consumption.

Page 20: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

20

Protracted Relief and Recovery

Sierra Leone

Background and Objectives

The Sierra Leone food security, livelihoods, health and nutrition survey was carried out in the aftermath of a protracted civil conflict. The overall objective was to collect baseline information to inform policy and guide the formulation of food and non-food based safety net programs. The scope of the study was nation-wide and meant to comprehensively analyse food security and vulnerability to be used as a reference for decision-making, programme monitoring and reporting. Preparation

Preparations for the data collection activity were jointly undertaken by WFP, CARE, Catholic Relief Services (CRS), World Vision International (WVI) and Action Contre la Faim (ACF). UNHCR, UNICEF and OCHA and the Government of Sierra Leone were also key partners. Meetings among these partners were held to define and organise the study and began almost 12 months before the actual data collection activity began. The Regional VAM officer played a key role in facilitating discussions on the scope of the study but also on the methodology for data collection and analysis. WFP provided technical backstopping in planning and implementing the survey, through the Regional VAM officer and staff from

Context: Relief and Recovery after Protracted Civil Conflict

Objectives: To collect baseline information to inform policy and guide the formulation of food and non-food based safety net programs

Coverage: Countrywide except for Freetown.

Partners: WFP as lead, along with ACF, CARE, CRS, World Vision, UNICEF, UNHCR and the Government of Sierra Leone. Funding from Canadian Government and USAID/OFDA.

Duration: Planning and partnership building took one year while the actual data collection, analysis and reporting took 4-5 months.

Successes: • Good collaboration between all players with WFP as lead agency. • Funding was timely and adequate while each agency donated staff time and

vehicles for data collection. • More than 90 enumerators on 13 teams to collect data during a 2 week period with

enough vehicles. • Successful integration of child and maternal anthropometry to household food

security. • Good coverage with at least 11 villages in each of the 13 rural districts surveyed. • First comprehensive food security and vulnerability study conducted since the end

of the war.

Difficulties: • No use of qualitative tools for data collection and analysis. • Capacity of government staff was limited. • Lack of in-country technical capacity and thus heavy reliance on regional and HQ

VAM staff. • Lack of up-to-date census data made the sampling a challenge. Poor infrastructure

(roads and bridges) added to travel times and often required selection of alternative villages for the sample.

Page 21: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

21

VAM-HQ. The Country Office provided overall coordination and logistical support for the data collection exercise. Data Collection

In total, more than 90 enumerators from numerous agencies and ministries were involved in the study. Each enumeration team had a leader, generally an international staff from one of the partner agencies, to coordinate logistics and to implement the survey protocol in the sampled villages. Each team also had a nutritionist to measure children under five and their mothers. WFP supported the cost of the Government staff deployed for the field work3. Otherwise, the various agencies paid the costs for their staff. Three three-day trainings for the enumerators were conducted over a 10 day period. The training covered the following topics: (a) overview of survey objectives; (b) definition of food security; (c) village sampling; (d) household sampling; (e) community questionnaire review; (f) household questionnaire review (to learn and to make revisions with inputs from enumerators); (g) overview of anthropometry – interpretation of z-scores and Body Mass Index (BMI); (h) age estimation for children less than 5 years of age; (i) use of UNICEF SECA scales for weighing children and adults; (j) planning and logistics; (k) village protocol; (l) administrative issues; and (m) field testing of methods and questionnaires. UNICEF and the Ministry of Health provided critical support to the training on the use of scales and measuring boards and age estimation of young children without birth records. In order to test the survey questionnaires, teams went to a nearby community just outside of Freetown on the second half of the second day of training. This was felt to be too short-a-time to adequately test the instruments and build field capacity for the enumerators. The community questionnaire was patterned after one used for the 1999 Ghana Living Standards Measurement Study (LSMS) conducted by the World Bank. The household questionnaire was designed using examples from LSMS and Demographic and Health Surveys (DHS) along with valuable inputs from the multi-agency working group and the enumerators. A salient feature of the questionnaire and sampling methodology was that they made a direct linkage between household livelihood and food security information and maternal/child health and nutritional outcomes. Given that the assessment and analysis activity was country-wide, data collection included all 12 rural districts and the rural part of Western Area (outside of Freetown). Since the teams only had 2 weeks for data collection, it was decided that they could, at best, survey 11 villages within that time period. To maximize district coverage, sampling was done by Section rather than Chiefdom4. Eleven Sections were randomly selected in each district using a random selection function in SPSS. In addition, 2 replacement Sections were also randomly chosen in case all of the villages in the original Sections were not reasonably accessible. It was decided that only villages with more than 12 households and less than 3,000 persons would be eligible for sampling and that within each Section. Therefore 1 village and 2 replacement villages were selected using probability sampling. Since there had not been a census since 1985 an accurate sample frame was not available. Hence the need to sample additional villages to serve as replacements if the originals were inaccessible or no longer existed. It was each team’s responsibility to notify villages at least a day ahead of time about their selection for the survey and the team’s arrival. This was achieved with the help of local authorities. Local officials were given a small financial compensation for the time dedicated to accompanying the teams of enumerators in the villages and introducing them. Following local custom, upon arrival in the villages, the team leader introduced the enumerators to the village through the village head. The Sierra Leone assessment did not include community focus group discussions but rather key informant community interviews which were mostly carried out by the team

3 Funds received by WFP from the Canadian Government and USAID/OFDA. 4 Administrative divisions include: Country, District, Chiefdom, Section and Village

Page 22: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

22

leader. Random selection of households was carried out in two different ways based on the size of the community and in order to maximize the number of young children in the sample. In communities with more than 30 households, only those households with children less than five years of age were eligible for selection (normally about 60-70% of households in the rural villages) while in smaller communities any household was eligible for selection. Where the interviewers did not speak the local language, local interpreters were used for household interviews. Ideally it would have been best to have questionnaires translated into the main languages spoken in the country and to have recruited more enumerators who were fluent in those languages. If there were two or more women with young children (< 5 years) in the household (multigenerational or polygamous households) only one was selected to be interviewed for the modules on maternal and child health and nutrition. Then she and all of her children under five years were also weighed and measured. Measurements of young children and their mothers were taken after the completion of the household interviews. Those responsible for weighing and measuring were previously trained on the methodology to be used. Age was estimated using information from yellow immunization cards when available and then using a local events and agriculture-based seasonal calendar for references. Children were weighed using the UNICEF Mother-child (SECA) scales and the length/height was measured using newly constructed height boards. The equipment was on loan from UNICEF Sierra Leone with support from the Ministry of Health. Adult women of reproductive age (15-49 years old) with at least one child less than five years of age were also weighed and measured. This proved useful for assessing their nutritional outcomes; for triangulating child anthropometric data; and for assessing the relationship between maternal and child malnutrition. Data Analysis

The time required for data entry was reduced by undertaking manual data cleaning and coding before entering the data. For this a WFP National staff member supervised a team of six data entry staff, provided by the various agencies, that read through all of the completed questionnaires and made corrections and coding changes on the forms. A master list of codes was provided by VAM-HQ while additional codes were created during the pre-entry cleaning process. The data collection activity yielded 4 different data bases for analysis: Village database; Individual household member database; household level database; and child level database. Initial data analysis was conducted in Freetown by the VAM-HQ staff member using SPSS. The working group was interested in differences between districts for key variables since there had not been a comprehensive household food security study done in the country since the war. The report yielded descriptive findings from the community questionnaire on demographics, economic and agricultural activities, as well as access to and utilization of markets, education and health facilities. From the household questionnaire, results were presented for: household demographics; effects of the war; housing; lighting, water and sanitation; asset ownership; agriculture; household income and expenditure; food consumption; and livelihood shocks and coping strategies. More in-depth analyses were performed to link maternal and child health and nutrition outcomes to household characteristics and food security status.

Data from the questionnaires was entered using data entry programmes prepared in Epi-info by VAM-HQ. Epi-Info was found to be a convenient software package since it was easy to use, free, could be installed on any computer and has a function to calculate z-scores that is particularly useful in the analysis of anthropometric data. Cleaning and entry of the household and community data took about four weeks.

Page 23: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

23

Household food security typologies and levels were created using multivariate statistical analysis on four main variables, representing food availability and access: a) consumption frequency of staple and non-staple foods such as groundnuts and legumes, meat, vegetables and fruits; b) the sources of each food item; c) share of household expenditure for food and other basic needs (e.g. health, education, etc.); and d) share of expenditure for individual food items. The percentage of households in each of the five typologies was calculated and compiled in a matrix by district. Added to this matrix were the prevalence of child stunting and wasting and prevalence of maternal malnutrition (low BMI) as indicators of food utilization. From this table, six district food security typologies were identified, based on the prevalence and combination of the different food security components. This descriptive information on the level of food security by district was then mapped for visual interpretation of the findings. The findings of the baseline food security study are available on the WFP/VAM Web site (http://home.wfp.org/vam) or can be requested at the VAM Unit in WFP-HQs.

Five household food security typologies emerged as a result of the analytical process:

• Chronically food insecure – not meet minimal consumption needs and have very high shares of food expenditure - 28% of all households

• At risk to become chronically food insecure - despite partial reliance on own production, shares of food expenditure quite high - (19%)

• Potentially Vulnerable to food insecurity - with greater access to own production – (25%)

• Better food security – reliance on purchases (20%)

• Food secure households (8%)

Page 24: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

Development

Mali

Background and Objectives

The objective of this food security and vulnerability analysis was to provide a sound problem analysis for the preparation of the Country Strategy Outline and the Country Programme, as well as roll-out the newly designed VAM Standard Analytical Framework (SAF). Preparation

Since a main objective of the assessment and analysis activity was to apply and test the newly designed SAF while addressing the need for providing necessary information to decision-makers for the preparation of the Country Programme, the Government, UN agencies and NGOs were approached and informed only at a later stage when WFP had already planned its field work. These organizations did not demonstrate an interest in taking part in the study, most likely due to the fact that WFP did not dedicate enough time to explain the purposes and advantages of the survey and hence WFP undertook the assessment alone. The assessment was planned, organized and managed by the Regional VAM officer and VAM-HQ. Questionnaires were prepared on the basis of previous VAM data collection in the region (Sahel). Similarly, the work plan was prepared in the country taking into consideration the poor road infrastructure and the difficulties in reaching certain areas. Despite this there were considerable delays in the implementation of the work plan because drivers did not know the area well enough or were not experienced in driving on dirt roads.

Context: Development

Objectives: To provide a sound problem analysis for the preparation of the Country Strategy Outline and the Country Programme, as well as roll-out the newly designed VAM Standard Analytical Framework (SAF).

Coverage: Previously identified food insecure districts.

Partners: WFP and government only.

Duration: Planning to final reporting took about 6 months.

Successes: • Study design utilized findings from previous VAM work done in the country. • The analysis utilized spatial information along with household and community food

security data. • Design attempted to determine differences in food security between households

with malnourished children and those where children were healthy. • Analysis resulted in the creation of village food security profiles which could then

be mapped for spatial analysis using agro-ecological indicators.

Difficulties: • No participation from partner agencies. • Logistics were problematic – the drivers were not familiar with the areas under

study and time was lost in locating sampled villages. • Villages were not notified in advance of the arrival of the survey teams. • Lack of adequate supervision for all data entry activities.

Page 25: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

25

Data Collection

Twenty-five enumerators and five team leaders were hired for the study and were paid for through WFP funds. To ensure that enumerators possessed the adequate skills for data collection the National Census Bureau provided the list of those that had worked for the census data collection. Language skills were also taken into account for the selection of enumerators since different local languages were required. A three-day training, including the questionnaire testing in a nearby rural community, was provided to the enumerators and team leaders. The training was meant to explain the purposes of the field work including: (a) food security and vulnerability concepts; (b) instruments for data collection; and (c) discussion of work plans. The geographical area surveyed was selected using the results of district level secondary data analysis as the first part of the SAF process. Based on the findings from that study, VAM had identified food insecure districts where food aid could play important role in addressing some of the underlying causes of food insecurity. The study also identified districts where the potential role of food aid was less clear or where food insecurity did not appear to have a direct link with food consumption. It was thus decided to collect data only where food insecurity was likely linked to acute food availability and accessibility and such, where food aid could have a comparative advantage. Issues of security and physical access were also taken into account in defining the study area. Villages with a population between 200 and 800 people were selected for the field work as this population size represented 80% of the rural settlements in the country. Moreover, a detailed spatial analysis using remote sensing data allowed for the identification of homogeneous agro-ecological zones from which villages were randomly selected. Villages selected for the survey were not informed in advance about the arrival of the enumerator teams. Despite this, people in the villages were available to participate in the survey since this was carried out in the lean season for agricultural work. The team leader was responsible for presenting to the community the purpose of the visit and for explaining the work they would like to undertake within the community. A community questionnaire including general questions on available economic activities, job opportunities, access to primary facilities, risks and food security etc. was used to lead the discussions with the different representatives of the community, i.e. elderly, authority, women, men and youth. While part of the team was leading these discussions, the two team members responsible for the nutrition assessment were collecting anthropometric data for all children between 6 and 59 months of age. Anthropometric data was collected for all children in each of the sampled villages, before sampling households. Based on this data, children were categorised as malnourished or not malnourished. The nutrition assessment was carried out to collect food security outcome data but also to apply a stratified household sampling strategy. Stratified sampling was used to select households for inclusion in the household survey and to ensure that the selected households represented certain characteristics such as female head of household, presence of chronic malnutrition and presence of children older than five years of age. These strata were used on the assumption that they were correlated with household food accessibility and utilization. The main objective was to determine whether those households with malnourished children would benefit from food assistance. In this context, four households with malnourished children and four without malnourished children were randomly selected. Moreover, secondary data analysis had showed that female headed households tended to be relatively disadvantaged. Therefore, in order to have a significantly representative sample, two female headed households and two households with children above 5 years of age were also randomly selected from a list provided by the community. Later, the analysis of the data confirmed the larger hardship faced by female headed households but disproved the assumption that the households with malnourished children would in all circumstances require some form of food assistance.

Page 26: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

26

Data Analysis

Team leaders were responsible for reviewing questionnaires before leaving each village, to ensure that they were completed correctly. Data entry was carried out in the different WFP sub-offices given the long distances for gathering all questionnaires in Bamako at the Country Office. This important but time consuming activity was particularly lengthy because where there was no supervisor the staff did not complete their task. When data entry was supervised, it was carried out effectively. In order to create household food security profiles, five key variables were analysed: a) availability of own cereal production; b) share of household expenditure for food and other basic needs; c) share of expenditure for individual food items; d) asset ownership; e) sources of income. Analysis was carried out using multivariate statistical techniques in order to capture the many dimensions of food security from the intra- and inter-variable relationships, that would otherwise be lost if each variable were analyzed as an independent piece of information. Five different household food security typologies were identified through the analytical process and further characterized using additional information on household demographics, literacy and nutrition outcomes. These household typologies were then broken down at village level for all sampled villages. When a village contained 80% or more households from one food-security typology, it was described with that predominant profile. Where there was not such a dominant profile represented but rather a combination of two or more household typologies, villages were described accordingly, in order to represent the household food security diversity present in the village. Villages with similar food security profiles were then grouped together forming a total of four village food security profiles. The strong correlation between these village typologies and other agro-ecological indicators available for the surveyed areas from high-resolution remote sensing images, allowed for the extrapolation of the village classification to all villages5 included in the surveyed area. Finally, maps presenting the geographical distribution of the four village food security profiles were created. The findings of the Mali food security and vulnerability analysis are available on the WFP/VAM Web site (http://home.wfp.org/vam) or can be requested at the VAM Unit in WFP-HQs.

5 More than 4000 villages were classified

Five household food security typologies emerged as a result of the analytical process:

• Chronically food insecure – never meet their minimum food needs nor other basic needs - 16% of all households

• Unsuccessfully coping with a regular food gap - (20%)

• Coping with a regular food gap – but unable to meet other basic needs, health and education in particular (25%)

• Successfully performing – they meet food and non-food needs (23%)

• Surplus agriculture farmers - (15%)

Page 27: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

27

Development

Dominican Republic

Background and Objectives

At the request of the Government of the Dominican Republic and the Inter-American Development Bank (IDB), WFP undertook a food security and vulnerability study in the poorest areas of Dominican Republic. The objective was to identify the link between poverty and food insecurity, and to analyse whether food aid is needed. Preparation

Given the background of the study, the Technical Secretariat of the Office of the President (STP) and the Ministries of Planning and Statistics, Environment, Disaster Preparedness, Education and Health were identified as partner institutions. These partners defined the scope of work by participating in the conceptual and planning phases. VAM staff from both the Regional Bureau and HQ were given the task of coordinating the preparation and implementation and to provide technical support in terms of methodology development, training, data collection and analysis and report writing. Moreover, a lead consultant was recruited from WFP-Cuba to coordinate the data collection and analysis process. The WFP Country Office was responsible for overall coordination and management of financial resources. The IDB provided the bulk of funds while the STP provided additional technical assistance. Constraints related to this partnership arrangement included delays in government inputs – especially financial and human resource commitments – and the limited experience of the WFP Country Office in field assessments. Data Collection

A total of 24 enumerators, provided by the WFP’s major partners in the assessment and analysis activity, were recruited to undertake the data collection in the field. Six teams of four people each were put in place. Five teams were responsible for the household

Context: Development

Objectives: To identify the link between poverty and food insecurity, and to analyse whether food aid was needed.

Coverage: Poorest areas of the country as identified by the Government.

Partners: Technical Secretariat of the Office of the President (STP) and the Ministries of Planning and Statistics, Environment, Disaster Preparedness, Education and Health. Funding from Inter-American Development Bank.

Duration: More than 8 months.

Successes: • Good use of both qualitative and quantitative methods. • Available secondary data was abundant and useful in planning and in interpretation

of findings. • Multivariate techniques found only 6% of households with chronic food insecurity,

demonstrating the sensitivity of this technique to a wide range of development settings.

Difficulties: • Not enough time and money available for adequate training, field testing and travel

during data collection. • Sampling only allowed results to be generalized to the macro-level zones rather

than provinces. • Coordination between qualitative and quantitative analysts was problematic due to

distance and additional work obligations for both.

Page 28: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

28

surveys whereas the remaining team was responsible for community and focus group surveys throughout the sampled communities. Enumerators had backgrounds in sociology, health/nutrition and statistics. The high skill level of the enumerators facilitated data collection, despite some weaknesses in the training received. Training needs for the qualitative data collection team were determined though discussions that assessed their understanding of the analytical framework and their facilitation skills. Training modules included food security and risk analysis frameworks, questionnaire design, and facilitation skills. However the training was conducted by VAM-HQ staff who were not fluent in Spanish. This problem was partially offset as participants had good knowledge of English, enabling major messages to be understood. The instruments for field data collection (i.e., focus groups tools and household questionnaires) were prepared by the Regional VAM officer and VAM-HQ and were built on experiences from other VAM studies conducted in Western Africa and Haiti. The household questionnaire was comprehensive but was found to be too long and the design and layout were difficult for enumerators to complete. Changes were made to the instrument over the course of the training—paring down to the most essential modules. Participatory tools were selected by VAM-HQ on the basis of their adeptness in assessing risk exposure and management and livelihoods and adapted to the local context. Based on the previous experience in Western Sahel, write-up sheets were simplified in order to make them more user-friendly. Modules were easily understood by enumerators but the qualitative questionnaire was also too long. Field testing of both qualitative and quantitative questionnaires took place in two peri-urban communities close to Santo Domingo. Afterwards some elements in both questionnaires were clarified and the design simplified as necessary. The work plan was jointly developed by the WFP Country Office, Regional VAM officer, Lead Consultant and IDB. There was considerable delay in delivering the final product compared to what had been scheduled in the work plan because activities such as data cleaning, analysis and reporting were not included for unknown reasons. Given the scope and the objective of the study, the data collection areas were identified using the World Bank data on poverty. Initially it had been established that the sample had to contain 150 villages, as used in the World Bank Living Standards Measurement Survey, in order to be statistically representative. This was considered unfeasible and reduced to 100 villages. Despite the reduction in number, it was still too large a sample—creating financial and logistical problems through the course of the exercise. Stratified random sampling based upon income and expenditure variables was used to ensure that only the poorest communities would be selected. Five geographical zones were identified by the Government as being the poorest and, thus, were selected as the study areas. In a country with high levels of income inequality, it was found to be useful to focus on the lowest income quintiles—which could be found within these zones. Once communities were identified, the WFP Country Office informed the local administration about the study and asked local authorities to inform villages about the arrival of the teams. The large sample size was the most important consideration in defining the individual enumerator teams’ work-plans since it imposed a very tight schedule. Focus group discussions with separate groups of women; men; and youth lasted about one full day (i.e. three focus groups lasting, on average, two hours each), including travel time. These focus groups provided qualitative information on livelihood assets, perceptions of food security risk factors and risk management strategies. Consultations with larger community groups were also held, with the intention of gaining an overview of community issues and then discussing these in greater detail with each focus group. Since many of the youth were not present when the qualitative enumerator team arrived in the villages and since the work load was already heavy for the focus group team the youth focus group was dropped for some of the last villages visited. Issues related to different

Page 29: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

29

age groups (elderly and youth) were, nevertheless, included in risk and livelihood analysis discussions with the focus groups of men and women. Household data collection was carried out by different teams. Between 12 and 15 households were selected in each village using probability sampling techniques. Household interviews took approximately 45 minutes and were conducted in Spanish. The questionnaires included modules on household demographics, income sources and expenditures, levels of education/literacy, recent morbidity, food frequency, diversity and sources, participation in community organizations and risk perception and exposure. Data Analysis

Qualitative data were analysed – disaggregated by gender- by VAM-HQ using the risk management and livelihoods analytical framework. This involved linking together, in a causal chain, risk exposure, risk effects and risk/livelihood management activities. Considerable time was required for analysing the qualitative data and for consultations between quantitative and qualitative data analysts in order to make effective and complimentary use of the two types of data — all data analysis discussions took place through email or telephone. Household questionnaires were coded by hand and then entered into Access. However, the program had not been designed carefully enough to facilitate data entry and to minimize key punch errors, due to time and funding constraints. A team from the Ministry of Environment was responsible for the data entry which took about three weeks and was under the direct supervision of the GIS-technical expert of the Ministry of Environment in collaboration with the WFP Consultant who coordinated the field work. Data cleaning was performed by the Regional VAM officer, the WFP consultant and the technical person from the Ministry of Environment. Household data were analysed using multivariate statistical techniques in order to build household food security typologies. Five key variables, together representing food availability and access, were used for creating the household food security typologies. These were a) dietary diversity; b) frequency of food items consumed; c) share of household expenditure for food and other basic needs (e.g. health, education, etc.); d) share of expenditure for individual food items; and e) sources of household income. Clustering analysis was then used to create household food security typologies ensuring a strong similarity among the households within each typology and solid heterogeneity between typologies. Once the food security typologies were created, it was found there was limited inter-household differentiation in terms of expenditure patterns. All households tend to have very high shares of food expenditure but a greater food diversity and frequency. To ensure the accuracy of the typologies, these were cross-tabulated with other variables, which included assets ownership (stove, fridge, television, radio, bike, motorbike, car and mobile phone), lighting devices, water access and supply, sex of head of household and his/her education level, household age distribution, household size and age dependency ratio. The sampling technique allowed the results to be generalizable only to macro-level zones, rather than provinces. Identifying the food security level of each zone involved categorising zones according to the dominant household food security typology. A map was used to display the location of each zone with an insert concisely describing the combination of the different household food security profiles in each zone. Due to the homogeneity of the zones, spatial analysis was not used. However, the map was found to be a useful communication tool.

Eight household food security typologies emerged as a result of the analytical process:

• Chronic food insecure – do not have enough food for each day of the week – 6%;

• Vulnerable – Only cereals are consumed everyday – 11%;

• Potentially vulnerable – more diversified diet but only one commodity per day – 5%

• Relatively good food security – Better food diversity and frequency – 18%;

• Good food security – Preference for meat – 8%

• Good food security – Good food diversity – 30%

• Very good food security – 23%

Page 30: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

30

The findings of the Dominican Republic food security and vulnerability analysis are available at the WFP/VAM Web site (http://home.wfp.org/vam) or can be requested at the VAM Unit in WFP-HQs.

Page 31: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

31

SECTION IV: RESOURCES NECESSARY FOR VULNERABILITY ASSESSMENT AND ANALYSIS

This section highlights the range of inputs used to conduct various types of studies with focus on financial, human, temporal and logistical resource areas. It is important to consider these in the planning and implementation processes for data collection, analysis and reporting.

Financial inputs and sources of funding

There were several main sources of funding for the various vulnerability assessments highlighted in this report with the most likely source being WFP (EMOP, PRRO, ISP, Canadian, etc.) for Ivory Coast, Malawi, Mali and Western Sahel. The Inter-American Development Bank also contributed funds towards the Dominican Republic assessment, through the Government. The main sources of funding for Sierra Leone were the Canadian Government and USAID/OFDA. The NGO/UN partners covered the costs of their staff in the field as well as providing data entry personnel and computers for that process. The most expensive component of all the assessments was the hiring of enumerators and international staff. In many cases, these costs represented nearly 70% of the total. Other expense line items are training, fuel, vehicle hire, administrative/stationary and contingency. However, net expenditures can be reduced if partner agencies (NGOs, UN) donate their staff and vehicles for the assessment activity. Additional savings were added in all cases by enlisting the technical support of VAM-HQ staff where the survey activities covered only the airfare and DSA while in the field. Honoraria were paid by WFP-HQ. The total cost of the Ivory Coast assessment was $17,500, with $10,000 for VAM HQ staff travel and DSA, $2,500 for six enumerators in the field and $5,000 for fuel. In the Western Sahel, the costs of the assessments ranged from $35,000 to $50,000 each, depending on the coverage and time for data collection. In Mali the total cost was of about $60,000, of which more than 50% was for car rental. All enumerators were hired and although their monthly cost was about $600 including DSA, the large number of enumerators needed to carry out the data collection within a period of 30 to 40 days increased the overall cost of the assessment. In Malawi, Round One cost $41,000 including the expenses of the VAM-HQ staff inputs while Round Two was $35,000 plus an additional $5,000 for VAM-HQ technical staff expenses. The total cost of the Sierra Leone assessment was about $75,000, which included travel and DSA cost of VAM-HQ technical staff. The Dominican Republic assessment was the most expensive – around $90,000 in total.

Staff requirements

When planning and implementing a vulnerability assessment, there are some key staff roles that should be carefully filled to ensure that the exercise can be carried out successfully. Firstly, an overall coordinator or manager is vital so all inputs, queries and requests can be managed from a single central person. About half of the assessments presented in this paper were multi-agency activities and required an enormous amount of coordination between agencies, both at the national and regional level, in some cases. For WFP, it is important that not only country office staff are involved and informed, but also that there is participation from the appropriate Regional Bureau and VAM-HQ. Other key human resource needs to consider are technical people to assist with methodology, sampling, questionnaire design, data entry systems, data analysis and interpretation of findings. In addition, it is very important to have staff skilled in training the enumerators in participatory methods, quantitative data collection and, in some cases, child and adult anthropometry. In Malawi, the assessments were carried out through a National Vulnerability Assessment Committee (VAC), which was guided by the Regional Vulnerability Assessment Committee (RVAC). The VAC was seated in the National Economic Council with a government

Page 32: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

32

Chairperson who presided over meetings. The actual coordinator of the assessment activities was a consultant provided by Save-UK who served as the central point for all assessment information. She was also responsible for helping to build government capacity to organize and manage the activities but was not based in the Ministry. FEWS-Net (2 Malawian staff) provided substantial technical and practical guidance in planning the assessments, were team leaders in the field, and played an important role in training the enumerators for both rounds of assessments. The lead person from the National Economic Council was instrumental in recruiting government staff and vehicles for data collection in the field. From the WFP Country office, the National VAM officer was the main representative of WFP in all MVAC meetings and played a key role in organizing logistics and finances for the data collection activities. He was also a team leader for both assessments. A VAM-HQ staff member who was specialized in quantitative methods was the main technical person for both assessments. He provided substantial input for survey design, sampling and questionnaire development and was fully responsible for data entry, cleaning, analysis and report writing. He also did the majority of the secondary data research and analysis for both VAC reports and assisted in training of the enumerators in Round One. The Sierra Leone country-wide assessment was coordinated by a WFP international staff member based in the country office. The WFP CO also provided 2 national staff members who assisted in training, sampling and supervising data entry, in addition to numerous field staff who served as enumerators during the data collection. International and National staff from partner agencies (ACF, CARE, CRS, WVI & UNHCR) all provided key inputs into the core planning, coordinating, and training activities as well as the data collection exercise. WFP and UNICEF nutritionists were responsible for training survey members on collection of anthropometric data and also supervised activities in some of the enumerator areas. ODD-VAM was instrumental in planning and in bringing together the various stakeholders in-country. A technical staff from VAM-HQ travelled to Sierra Leone and provided almost all of the technical assistance including sampling, questionnaire development, training, data entry programs, data cleaning and analysis and report writing. Additional VAM-HQ staff conducted the multivariate and cluster analysis for food security profiling and HQ cartographers prepared thematic maps with some of the results. The Mali study was coordinated by the Regional VAM officer and VAM HQs (Socio-economist). It included an international consultant with expertise in spatial analysis and sampling and a regional consultant with expertise in the Sahel food security issues. These two main consultants were responsible for the management and supervision of the team leaders whereas the team leaders were responsible for the quality control of the enumerators’ work. The Western Sahel and Ivory Coast the assessments were managed and coordinated by the Regional VAM officer with additional support from VAM-HQ for both activities. National consultants and/or national WFP staff also actively participated in the field work. In Dominican Republic the coordination of the assessment was ensured by the Regional VAM officer. A Regional consultant was hired to undertake a situation analysis prior to the lunch of the data collection. Two technical experts were sent from VAM HQ to assist with the sampling, the design of qualitative and quantitative instruments and enumerators’ training. The analysis and reporting was done by the Regional VAM officer (quantitative) and VAM-HQ (qualitative and household food security profiling).

Temporal requirements

The time required to conduct vulnerability assessments depends greatly upon the level and complexity of partnerships (regional vs. local; multiple partners/stakeholders vs. WFP alone), the urgency of the situation and the immediate use of the resulting information. Other factors that influence the amount of time needed to plan, conduct, analyze, report and disseminate information are financial and human resources. In most of the case

Page 33: Vulnerability Analysis - Some Examples - Final AND · 2008-09-23 · • Samir Wanmali (Livelihoods specialist) The WFP VAM- Unit would also like to recognize the important contributions

33

studies, there was not enough time dedicated to training of enumerators and then field testing of the instruments and methodology. In Dominican Republic, the entire process took over eight months to complete, with a substantial part of the time dedicated to planning and data collection (5 months) with about 2-3 months allocated to data analysis and reporting. In Malawi, the vulnerability assessments were a part of a series of rolling assessments in six countries and were coordinated by a regional body. In the context of a slow-onset emergency and the need to estimate food needs over three month periods, it was necessary to move quickly in data collection, analysis and dissemination. In both rounds, this took approximately 2 months to complete. However, the finalization of the report in Round One took an additional month while in Round Two; it took 2 more months, due to the extensive partnerships and hierarchy of coordination structures for the assessment. For Ivory Coast, the assessment was undertaken during a complex emergency and thus final reporting was completed within 4-6 weeks – one week for planning, a few days for training, and 10 days for data collection. Analysis and reporting was completed in the remaining weeks. The assessments in the Western Sahel were planned, implemented, completed and reported upon in a 6-8 week period. Expediency was required since the results were used to prepare a regional EMOP. The vulnerability assessment in Sierra Leone was intended to serve as a baseline for monitoring food security in a recovery and rehabilitation setting and thus the entire process from initial stakeholder planning meetings to launch of the final report took about 12 months. The planning process took about 5-6 months while questionnaire development and sampling took about 3 weeks, training was done in 10 days, data collection lasted 2 weeks and data cleaning and entry took about 3-4 weeks. Analysis was conducted in stages and lasted about 6 weeks while report writing was done over a 3-4 week period. One person did the bulk of analysis and report writing. In Mali the assessment from the planning stage to the final report took about six months. Data entry and data cleaning were the most time consuming activities due to the poor human capacity available. Bad road conditions and drivers’ inexperience with the geographical area also helped to increase the number of days necessary to complete the field work. These activities used almost the 2/3 of the entire period necessary to carry out the assessment. Once the data was entered and cleaned, the data analysis at household and village level was completed in less than four weeks.

Logistical Requirements

The main factors in this particular category are related to transport. In all countries, ensuring adequate transportation for enumerator teams proved to be challenging. For the most part, WFP and its partners managed to find the necessary vehicles through use of own fleet, those donated by partner agencies (UN, NGO, Government, etc.) or through private hire. It is imperative that vehicles are in good working condition and travel with spare parts, especially in remote and rough terrain. Other issues related to logistics are drivers’ knowledge of the enumerating area and the possibility to use motorcycles and/or boats to reach sampled communities. For example, in Malawi, there is Likoma Island district in the middle of Lake Malawi. In order to include them in the assessment, the team had to hire a boat to reach the island, interview households and return to the mainland in one day. In Sierra Leone, the teams often had to cross rivers and streams on foot to reach villages, due to poor and war-damaged infrastructure. In Mali, the drivers were unfamiliar with the areas where they were travelling and also were inexperienced in driving on rough terrain.