Comprehensive Food Security, Nutrition and Vulnerability Survey 2010 - Madagascar

Abstract
---------------------------

Efforts to improve Madagascar’s food security and nutrition over the last two years have been thwarted by political instability which has disrupted data collection. With up-to-date critical information needed to contribute to evidence based decision making, UNICEF and WFP agreed to carry out a joint Comprehensive Food and Nutrition Security and Vulnerability Analysis (CFSVA+N) in 2010. The survey’s primary objectives are to:
• Provide an accurate and detailed assessment of the current food and nutrition security situation
• Assess the causes and risk factors for food and nutrition insecurity
• Identify potential ways to mitigate food and nutrition insecurity
• Reveal pockets of vulnerability where special assistance may be required.

Geographic coverage
---------------------------

Rural areas of Madagascar

Analysis unit
---------------------------

- Household
- Community

Universe
---------------------------

The survey covered household heads and women of reproductive age (15-49 years) in each sampled household. Household is defined as one (or more) people living under the same roof or different roofs, but sharing the main meals and respecting the authority of the same person (the head of the household).

Kind of data
---------------------------

Sample survey data [ssd]

Sampling procedure
---------------------------

In order to have a representative sample by region and livelihood zone, a two-stage cluster sampling was applied.

First stage, the number of clusters per region was calculated based on the required sample size to determine malnutrition and food security by region (details on sample size calculation are reported below). A minimum of 25 clusters per region was set. Since the Fokontany is the smallest administrative unit with population data available, this was established as the administrative unit from which clusters were selected. A list of all Fokontany and their estimated population was obtained from INSTAT.Urban communities as defined by the latest DHS-IV were not included in the sampling frame. Using this list and the livelihood zones, 176 segments were defined corresponding to both the regions and the livelihood zones. The number of clusters in each of the 176 segments was determined based on the population per segment weighted from the total regional population. For each segment, the Fokontany and their population were introduced into ENA and the required number of cluster per segment was selected using "probability proportional to population size" (PPS) sampling technique. This process was repeated for each of the 176 segments. In total, 606 clusters were selected across the 22 regions and distributed in the 8 Livelihood zones.

Second stage: the secondary sample unit was the household as defined by INSTAT. Within the selected clusters, households were selected from an exhaustive of households using systematic random sampling.

For the anthropometric data, the households' number (as defined by the sample size calculation) was selected using a calculated interval sampling (i.e. from a Fokontany with 200 households, 20 households were selected using a sample interval of 10). The households for the food security survey were sampled from the larger list of households already selected for the anthropometric survey. This was done using an independently calculated interval sampling (i.e. from the 20 households selected for anthropometric data, 5 households were selected for food security and health data using a sample interval of 4).

Sampling deviation
---------------------------

Thirty-three clusters out of the 606 selected were not visited due to bad weather or insecurity.

Mode of data collection
---------------------------

Face-to-face [f2f]

Research instrument
---------------------------

Household questionnaires, community questionnaires and nutrition questionnaires (which included anthropometric measurements) were used to collect the data.

摘要
---------------------------
过去两年中，马达加斯加在提升食品安全和营养方面的努力因政治不稳定而受到阻碍，这导致了数据收集的中断。鉴于需要最新的关键信息以支持基于证据的决策制定，联合国儿童基金会（UNICEF）和世界粮食计划署（WFP）于2010年同意共同开展全面的食物和营养安全及脆弱性分析（CFSVA+N）。该调查的主要目标是：
• 提供对当前食物和营养安全状况的准确且详尽的评估
• 评估食物和营养不安全的成因和风险因素
• 确定缓解食物和营养不安全的潜在途径
• 揭示需要特别援助的脆弱区域。
地理覆盖范围
---------------------------
马达加斯加的农村地区
分析单元
---------------------------
- 家庭
- 社区
总体
---------------------------
调查涵盖了每个抽样家庭中的户主和育龄妇女（15-49岁）。家庭被定义为居住在同一屋顶或不同屋顶但共享主餐并尊重同一人（户主）权威的一个（或多个）人。
数据类型
---------------------------
样本调查数据 [ssd]
抽样程序
---------------------------
为了获得按地区和生计区具有代表性的样本，采用了两阶段聚类抽样。
第一阶段：根据区域内的所需样本量计算每个区域的聚类数量，以确定区域性的营养不良和食品安全（样本量计算的详细信息见下文）。每个区域至少设定25个聚类。由于Fokontany是最小的具有人口数据的行政单位，因此将其确定为从其选择聚类的行政单位。从INSTAT获得了所有Fokontany及其估计人口的一份清单。根据最新的DHS-IV定义的城市社区未纳入抽样框架。使用此清单和生计区，定义了176个片段，这些片段对应于地区和生计区。每个176个片段中的聚类数量是根据每个片段的人口加权总区域人口确定的。对于每个片段，将Fokontany及其人口引入ENA，并使用“按人口比例抽样”（PPS）技术选择每个片段所需的聚类数量。此过程对每个176个片段重复进行。总共在22个地区和8个生计区中选择了606个聚类。
第二阶段：次级抽样单元是INSTAT定义的家庭。在选定的聚类中，使用系统随机抽样从详尽的户名单中选择了家庭。
对于人体测量数据，使用计算间隔抽样（即，从有200户家庭的Fokontany中，选择20户，抽样间隔为10）选择了家庭的数量。食物安全调查的家庭是从已经为人体测量数据选择的较大户名单中抽样的。这是通过独立计算的间隔抽样完成的（即，从为人体测量数据选择的20户中，选择5户用于食物安全和健康数据，抽样间隔为4）。
抽样偏差
---------------------------
由于恶劣天气或不安全，606个选定的聚类中有33个未能访问。
数据收集方式
---------------------------
面对面 [f2f]
研究工具
---------------------------
使用家庭问卷、社区问卷和营养问卷（包括人体测量）收集数据。