Comprehensive Food Security and Vulnerability Analysis 2009
Malawi, 2009
Get Microdata
Reference ID
MWI_2009_CFSVA_v01_M
Producer(s)
World Food Programme
Created on
Feb 19, 2014
Last modified
Mar 29, 2019
Page views
12673
Downloads
423
Identification
Survey ID Number
MWI_2009_CFSVA_v01_M
Title
Comprehensive Food Security and Vulnerability Analysis 2009
Country
| Name | Country code |
|---|---|
| Malawi | MWI |
Study type
Comprehensive Food Security & Vulnerability Analysis [hh/cfsva]
Series Information
The CFSVA process generates a document that describes the food security status of various segments of a population over various parts of a country or region, analyses the underlying causes of vulnerability, and recommends appropriate interventions to deal with the problems. CFSVAs are undertaken in all crisis-prone food-insecure countries. The shelf life of CFSVAs is determined by the indicators being collected and reported. In most situations, CFSVA findings are valid for three to five years, unless there are drastic food security changes in the meantime.
Abstract
The overall objective of the Comprehensive Food Security and Vulnerability Analysis (CFSVA) was to analyze the food security and vulnerability of the rural population, and to provide baseline information to actors regarding food insecurity. The study sought to answer five questions:
• Who are the people at risk of food insecurity?
• How many are they?
• Where do they live?
• Why are they food insecure?
• How can food assistance and other interventions make a difference in reducing poverty, hunger and supporting livelihoods?
The specific objectives of the rural Malawi CFSVA were to:
• identify geographic and socio-economic groups that are food insecure or vulnerable to food insecurity;
• analyze underlying causes of food insecurity and malnutrition, and explore the links between food security and nutrition;
• identify the major constraints to improving food security and review coping mechanisms used by vulnerable groups;
• support the design of livelihood group-specific poverty and food insecurity reduction programmes; and
• provide baseline data against which poverty-reduction and food security programmes will be measured.
• Who are the people at risk of food insecurity?
• How many are they?
• Where do they live?
• Why are they food insecure?
• How can food assistance and other interventions make a difference in reducing poverty, hunger and supporting livelihoods?
The specific objectives of the rural Malawi CFSVA were to:
• identify geographic and socio-economic groups that are food insecure or vulnerable to food insecurity;
• analyze underlying causes of food insecurity and malnutrition, and explore the links between food security and nutrition;
• identify the major constraints to improving food security and review coping mechanisms used by vulnerable groups;
• support the design of livelihood group-specific poverty and food insecurity reduction programmes; and
• provide baseline data against which poverty-reduction and food security programmes will be measured.
Kind of Data
Sample survey data [ssd]
Unit of Analysis
- Household
- Community
- Anthropometric measures for women and children
- Community
- Anthropometric measures for women and children
Version
Version Date
2012-09-27
Scope
Notes
- HOUSEHOLD: (1) demographics; (2) housing and facilities; (3) productive/non-productive assets and access to credit; (4) agriculture; (5) livelihood activities; (6) expenditures; (7) food sources and consumption; (8) shocks and coping strategies; (9) maternal health and nutrition; (10) child health and nutrition. Section 9 and 10 included measurement of weight, height and MUAC for women and children.
- FOCUS GROUPS focused on 6 areas: (1) demographics; (2) livestock ownership, crop production and income activities; (3) food security (including food utilization and shocks); (4) commodity markets and credit/loan accessibility; (5) health and nutrition; (6) seasonality. Section 6 included the utilization of a seasonal calendar to explore the seasonality of crop production; livestock (flock migration, pests and diseases, sells); hunger periods; diseases incidence; household expenditures; wage labor; and income activities.
- FOCUS GROUPS focused on 6 areas: (1) demographics; (2) livestock ownership, crop production and income activities; (3) food security (including food utilization and shocks); (4) commodity markets and credit/loan accessibility; (5) health and nutrition; (6) seasonality. Section 6 included the utilization of a seasonal calendar to explore the seasonality of crop production; livestock (flock migration, pests and diseases, sells); hunger periods; diseases incidence; household expenditures; wage labor; and income activities.
Coverage
Geographic Coverage
National coverage of rural Malawi.
Universe
The sample universe for this study was households of rural Malawi with children aged 6-59 months. Anthropometric data was taken for children and women in the household.
A household is defined as a group of people currently eating from the same pot “under the same roof” (or in same compound if the household has more than one structure).
A household is defined as a group of people currently eating from the same pot “under the same roof” (or in same compound if the household has more than one structure).
Producers and sponsors
Primary investigators
| Name | Affiliation |
|---|---|
| World Food Programme | United Nations |
Producers
| Name | Affiliation | Role |
|---|---|---|
| Ministry of Agriculture and food Security | Government of Malawi | Partner |
| National Statistics Office | Government of Malawi | Partner |
| Ministry of Development, Planning and Cooperation | Government of Malawi | Partner |
| Food and Agricultural Organization | United Nations | Partner |
Funding Agency/Sponsor
| Name | Abbreviation | Role |
|---|---|---|
| The Bill and Melinda Gates Foundation | Gates | Financial support |
Sampling
Sampling Procedure
The sampling plan was developed by NSO. For the purpose of the study, the seventeen livelihood zones in which the country is divided were collapsed into twelve zone; and Enumeration Areas (EAs) were stratified according to the zones. All the cities and forests/game reserves/national parks were excluded from the sample frame. A total of 271 EAs were selected as primary sample units. The number of EAs allocated to each livelihood zone was proportional to the number of EAs in the zone. Nonetheless, in the zones with a very small number of EAs, more EAs were added to ensure the representativeness of the results. Selection of EAs across the livelihood zones was based on the 1998 National and Housing census. EA sampling was done with random sampling approach. Household selection was done using systematic sampling. Twenty households were interviewed in each EA.
Deviations from the Sample Design
1.Due to PDA related problems, in some EAs the number of interviews actually saved and imported in the dataset was less than designed 20.
2.Data collectors face a number of challenges to access the rural areas to collect data because of infrastructural issues.
3.While survey data always represent the situation at a given time, seasonality has an influence on food access and food availability. The survey took place in April 2009, during the pre-harvest period and when the lean season was just coming to an end. The overall food security situation at the time of the survey can therefore be considered more difficult than normal.
2.Data collectors face a number of challenges to access the rural areas to collect data because of infrastructural issues.
3.While survey data always represent the situation at a given time, seasonality has an influence on food access and food availability. The survey took place in April 2009, during the pre-harvest period and when the lean season was just coming to an end. The overall food security situation at the time of the survey can therefore be considered more difficult than normal.
Weighting
The method uses the total number of HHs living in each livelihood zones (LZ) and the actual number of HHs in the dataset in order to compute design and standardized weights and produce results that are representative at LZ and country level.
Standardized weights
The following steps have been made to compute the standardized (or normalized) weights.
Step 1: In excel, a table was created reporting the no. of HHs in each LZ (as per 1998 data) and the no. of HHs interviewed in each LZ (as per SPSS file). See column B and C in the excel file.
Step 2: The no. of HHs living in each LZ was divided by the total no. of HHs in the country and multiplied by 100 in order to compute the percentage of HHs living in each LZ (out of the total no. of HHs). See column D in the excel file.
Step 3: The percentage of HHs living in each LZ was multiplied by the actual sample size (4908) and divided by 100. The result is the proportional distribution of HHs across the LZs in the sample. See column E.
Step 4: The proportional sample size has been divided by the actual no. of HHs interviewed in order to get the standardized weight (or normalized weight). In SPSS the average of the standardized weight is 1 (as expected).
Design (or expansion) weights
The following steps have been made to compute the design (or expansion) weights.
Step 1: For each LZ, the probability of the HHs to be selected was computed. It is equal to the no. of HHs interviewed divided by the total no. of HHs in the LZ. See column H.
Step 2: The design weight is the inverse of the probability. See column I.
Checks: Checks show that the weighted average of the standardized weights is 1 and that the total of the design weights is equal to the total population thus confirming the correctness of the method.
Standardized weights
The following steps have been made to compute the standardized (or normalized) weights.
Step 1: In excel, a table was created reporting the no. of HHs in each LZ (as per 1998 data) and the no. of HHs interviewed in each LZ (as per SPSS file). See column B and C in the excel file.
Step 2: The no. of HHs living in each LZ was divided by the total no. of HHs in the country and multiplied by 100 in order to compute the percentage of HHs living in each LZ (out of the total no. of HHs). See column D in the excel file.
Step 3: The percentage of HHs living in each LZ was multiplied by the actual sample size (4908) and divided by 100. The result is the proportional distribution of HHs across the LZs in the sample. See column E.
Step 4: The proportional sample size has been divided by the actual no. of HHs interviewed in order to get the standardized weight (or normalized weight). In SPSS the average of the standardized weight is 1 (as expected).
Design (or expansion) weights
The following steps have been made to compute the design (or expansion) weights.
Step 1: For each LZ, the probability of the HHs to be selected was computed. It is equal to the no. of HHs interviewed divided by the total no. of HHs in the LZ. See column H.
Step 2: The design weight is the inverse of the probability. See column I.
Checks: Checks show that the weighted average of the standardized weights is 1 and that the total of the design weights is equal to the total population thus confirming the correctness of the method.
Data Collection
Dates of Data Collection
| Start | End |
|---|---|
| 2009-04 | 2009-05 |
Data Collection Mode
Face-to-face [f2f]
Supervision
10 team leaders, 8 supervisors and 2 principles supervisors are involved in this exercise.
Data Collection Notes
Data collection was done with PDAs.
Questionnaires
Questionnaires
Two instruments were used for primary data collection: a household questionnaire administered to randomly selected households and focus group discussions.
Household questionnaire: The instrument is a structured questionnaire composed mainly by close-ended questions with response options provided by the enumerators. The tool was reviewed by the National CFSVA Taskforce chaired by MVAC. Members of the Taskforce comprised WFP, NSO, MVAC, the Ministry of Agriculture and Food Security, FAO, the Ministry of Finance, and FEWSNET. The survey instrument focused on 10 components: (1) demographics; (2) housing and facilities; (3) productive/non-productive assets and access to credit; (4) agriculture; (5) livelihood activities; (6) expenditures; (7) food sources and consumption; (8) shocks and coping strategies; (9) maternal health and nutrition; (10) child health and nutrition. Section 9 and 10 included measurement of weight, height and MUAC for women and children. The agricultural production section thoroughly collected information on production, inputs and fertilizers, utilization of harvest and seasonal market dependency.
Focus Groups: Subsequent to the administration of household questionnaires, the teams returned to the field for conducting focus group discussions. In each selected village, the community leader was asked to identify between 6 and 12 persons representative of the larger population and able to provide the most meaningful information. Discussions took place following the FG protocols and were done without the presence of the local authorities. The focus groups focused on 6 areas: (1) demographics; (2) livestock ownership, crop production and income activities; (3) food security (including food utilization and shocks); (4) commodity markets and credit/loan accessibility; (5) health and nutrition; (6) seasonality. Section 6 included the utilization of a seasonal calendar to explore the seasonality of crop production; livestock (flock migration, pests and diseases, sells); hunger periods; diseases incidence; household expenditures; wage labor; and income activities.
The questionnaire was developed in English and then translated in Chichewa.
Household questionnaire: The instrument is a structured questionnaire composed mainly by close-ended questions with response options provided by the enumerators. The tool was reviewed by the National CFSVA Taskforce chaired by MVAC. Members of the Taskforce comprised WFP, NSO, MVAC, the Ministry of Agriculture and Food Security, FAO, the Ministry of Finance, and FEWSNET. The survey instrument focused on 10 components: (1) demographics; (2) housing and facilities; (3) productive/non-productive assets and access to credit; (4) agriculture; (5) livelihood activities; (6) expenditures; (7) food sources and consumption; (8) shocks and coping strategies; (9) maternal health and nutrition; (10) child health and nutrition. Section 9 and 10 included measurement of weight, height and MUAC for women and children. The agricultural production section thoroughly collected information on production, inputs and fertilizers, utilization of harvest and seasonal market dependency.
Focus Groups: Subsequent to the administration of household questionnaires, the teams returned to the field for conducting focus group discussions. In each selected village, the community leader was asked to identify between 6 and 12 persons representative of the larger population and able to provide the most meaningful information. Discussions took place following the FG protocols and were done without the presence of the local authorities. The focus groups focused on 6 areas: (1) demographics; (2) livestock ownership, crop production and income activities; (3) food security (including food utilization and shocks); (4) commodity markets and credit/loan accessibility; (5) health and nutrition; (6) seasonality. Section 6 included the utilization of a seasonal calendar to explore the seasonality of crop production; livestock (flock migration, pests and diseases, sells); hunger periods; diseases incidence; household expenditures; wage labor; and income activities.
The questionnaire was developed in English and then translated in Chichewa.
Data Processing
Data Editing
Data collection was done with PDAs. The files from the PDAs were imported in SPSS for the analysis. SPSS and ADDAWIN were used to conduct a Principle Component Analysis (PCA) and a cluster analysis. Z-scores for wasting, stunting and underweight were calculated using WHO Anthro. All other analyses were done using SPSS.
Other Processing
In particular, a household Wealth Index (WI) was computed as a proxy measure of wealth. In particular, a principal component analysis (PCA) was conducted using wealth-related variables. After a careful analysis of the frequency distribution of non-productive assets and housing facilities, the following variables were used for the computation of the final WI:
1.material of wall (cement/burnt bricks vs unburnt bricks/mud/ wood/straw)
2.material of roof (tiles/iron/asbestos vs wood/plastic/grass/thatched)
3.material of floor (cement/concrete/tiles vs mud/sand/wood)
plus ownership of: 4) at least one bed; 5) at least one table; 6) at least one chair; 7) at least one mobile; 8) at least one watch; 9) at least one mosquito net; 10) at least one bicycle; 11) at least one radio; 12) at least one battery; 13) at least one pressing iron; 14) a banking account
The first component was selected to represent a proxy measure of wealth. It conserved 33 percent of the total variance. Wealth quintiles were computed, ranging from the poorest to the wealthiest.
Nutrition
Z-scores for wasting (WHZ), stunting (HAZ) and underweight (WAZ) were computed using WHO Anthro, and were imported in SPSS for the analysis. Z-scores are based on the new child growth standards released by the World Health Organization (WHO) in 2006.
1.material of wall (cement/burnt bricks vs unburnt bricks/mud/ wood/straw)
2.material of roof (tiles/iron/asbestos vs wood/plastic/grass/thatched)
3.material of floor (cement/concrete/tiles vs mud/sand/wood)
plus ownership of: 4) at least one bed; 5) at least one table; 6) at least one chair; 7) at least one mobile; 8) at least one watch; 9) at least one mosquito net; 10) at least one bicycle; 11) at least one radio; 12) at least one battery; 13) at least one pressing iron; 14) a banking account
The first component was selected to represent a proxy measure of wealth. It conserved 33 percent of the total variance. Wealth quintiles were computed, ranging from the poorest to the wealthiest.
Nutrition
Z-scores for wasting (WHZ), stunting (HAZ) and underweight (WAZ) were computed using WHO Anthro, and were imported in SPSS for the analysis. Z-scores are based on the new child growth standards released by the World Health Organization (WHO) in 2006.
Data Appraisal
Data Appraisal
Plausibility checks were conducted on the data to reduce errors. Age and sex distribution of measured children was compared to the expected distribution, standard deviation, skewness and kurtosis of the z-scores were calculated; heaping of age and weight were examined to understand the magnitude and distribution of bias (e.g. in particular areas or teams). Children whose ages were not properly recorded or flagged for invalid entries (epi-flags) were excluded from the analysis after checking for data entry errors.
Access policy
Contacts
| Name | Affiliation | URL | |
|---|---|---|---|
| Vulnerability Analysis and Mapping | World Food Programme | wfp.vaminfo@wfp.org | Link |
Citation requirements
Use of the dataset must be acknowledged using a citation which would include:
- the Identification of the Primary Investigator
- the title of the survey (including acronym and year of implementation)
- the survey reference number
- the source and date of download
Exemple:
World Food Programme. Malawi Comprehensive Food Security and Vulnerability Analysis 2009. Ref. MWI_2009_CFSVA_v01_M. Dataset downloaded from http://nada.vam.wfp.org/index.php/catalog on [date].
- the Identification of the Primary Investigator
- the title of the survey (including acronym and year of implementation)
- the survey reference number
- the source and date of download
Exemple:
World Food Programme. Malawi Comprehensive Food Security and Vulnerability Analysis 2009. Ref. MWI_2009_CFSVA_v01_M. Dataset downloaded from http://nada.vam.wfp.org/index.php/catalog on [date].
Disclaimer and copyrights
Disclaimer
The user of the data acknowledges that the original collector of the data, the authorized distributor of the data, and the relevant funding agency bear no responsibility for use of the data or for interpretations or inferences based upon such uses.
Metadata production
DDI Document ID
DDI_MWI_2009_CFSVA_v01_M
Producers
| Name | Abbreviation | Affiliation | Role |
|---|---|---|---|
| Souleika Abdillahi | SA | WFP | Data Archivist |
| World Bank, Development Data Group | DECDG | The World Bank | Reviewed the DDI |
Date of Metadata Production
2012-09-27
DDI Document version
Version 02 (February 2014). Edited version, the initial version (Version 01 - September 2012, DDI-MWI-WFP-CFSVA-2009-v1.0) DDI was done by Souleika Abdillahi (WFP).
Following DDI elements are edited, DDI ID, Study ID, Abbreviation, and Abstract. External resources (questionnaire and report) are attached to the DDI.
Following DDI elements are edited, DDI ID, Study ID, Abbreviation, and Abstract. External resources (questionnaire and report) are attached to the DDI.