COVID-19 Panel Phone Survey of Households, 2020 - Mali

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

In the WAEMU countries, COVID-19 is expected to affect households in many ways. First, governments might reduce social transfers to households due to the decline in revenue arising from the potential COVID-19 economic recession. Second households deriving income from vulnerable sectors such as tourism and related activities will likely face risk of unemployment or loss of income. Third an increase in prices of imported goods can also negatively impact household welfare, as a direct consequence of the increase of these imported items or as indirect increase of prices of local good manufactured using imported inputs. In this context, there is a need to produce high frequency data to help policy makers in monitoring the channels by which the pandemic affects households and assessing its distributional impact. To do so, the sample of the longitudinal survey will be a sub-sample of the 2018/19 household survey in each country.

For Mali, the survey which is implemented by the National Statistical Office (INSTAT), is conducted using cell phone numbers of household members collected during the 2018/19 survey. This has the advantage of conducting cost effectively welfare analysis without collecting new consumption data. The 35 minutes questionnaires covered 10 modules (knowledge, behaviour, access to services, food security, employment, safety nets, shocks, etc…). Data collection is planned for six months (six rounds) and the questionnaire is designed with core modules and rotating modules. Survey data collection started on May 11th, 2020 and households are expected to be called back every three to four weeks.

The main objectives of the survey are to:
• Identify type of households directly or indirectly affected by the pandemic;
• Identify the main channels by which the pandemic affects households;
• Provide relevant data on income and socioeconomic indicators to assess the welfare impact of the pandemic.

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

National coverage

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

Households

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

The survey covered only households of the 2018/19 survey which excluded populations in prisons, hospitals, military barracks, and school dormitories.

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

Sample survey data [ssd]

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

SAMPLING PROCEDURE
The Mali COVID-19 impact monitoring survey is a high frequency Computer Assisted Telephone Interview (CATI). The survey's sample was drawn from the population of the 2018/19 - Enquête Harmonisée des Conditions de Vie des Ménages (EHCVM) -, which was conducted between October 2018 and July 2019. EHCVM is itself a sample survey representative at national, regional and by urban/rural. For the 7,000 HHs in EHCVM, phone numbers were collected for about 90 percent of them. Each HH has between 1-4 phone numbers. The sampling, which was similar across WAEMU, aimed at having representative estimates by three zones: the capital city of Bamako, other urban areas and the rural area. The minimum sample size was 1,908 for which 1,766 were successfully interviewed, that is about 98 % of the expected minimal sample size at the national level. Given that Mali is conducting a phone survey for the first time, a total of 2,270 were drawn (25% increase) to take into account unknown non-response rates or presence of invalid numbers in the database.

The total number of completed interviews in round one is 1,766.
The total number of completed interviews in round two is 1,935.
The total number of completed interviews in round three is 1,901.
The total number of completed interviews in round four is 1,797.
The total number of completed interviews in round five is 1,766.

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

Computer Assisted Telephone Interview [cati]

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

All the interview materials were translated in french for the NSO. The questionnaire was administered in local languages with about varying length (30-35 minutes) and covered the following topics:

1- Household Roster
2- Knowledge of COVID-19
3- Behaviour and Social Distancing
4- Access to Basic Services
5- Employment and Income
6- Prices and Food Security
7- Other Impacts of COVID-19
8- Income Loss
9- Coping/Shocks
10- Social Safety Nets
11- Fragility
12- Governance and socio-political crisis

Cleaning operations
---------------------------

At the end of data collection, the raw dateset was cleaned by the NSO. This included formatting, and correcting results based on monitoring issues, enumerator feedback and survey changes.

Response rate
---------------------------

The minimum sample expected is 1,809 households (with 603 households per domain). This sample was therefore 99% covered for Bamako, about 100% for other urban areas and 91% for rural areas. Overall, the minimum sample is 98% covered. This level of coverage provides reliable data at national level and for each domain.

Round one response rate was 77.8%.
Round two response rate was 85.2%.
Round three response rate was 83.7%.
Round four response rate was 79.2%.
Round five response rate was 79.7%.

摘要
---------------------------
在瓦加杜古货币联盟（WAEMU）国家，COVID-19预计将以多种方式影响家庭。首先，由于COVID-19经济衰退可能导致的收入下降，政府可能会减少对家庭的福利转移。其次，从旅游业及其相关活动等脆弱部门获得收入的家庭可能面临失业或收入损失的风险。第三，进口商品价格的上涨也可能对家庭福利产生负面影响，这不仅是因为这些进口商品价格上涨的直接后果，还可能是因为使用进口投入品生产的本地商品价格的间接上涨。在此背景下，有必要生成高频数据，以帮助政策制定者监控疫情影响家庭的渠道，并评估其分配影响。为此，纵向调查样本将来自每个国家的2018/19年家庭调查的子样本。

对于马里，由国家统计局（INSTAT）实施的调查，使用的是在2018/19年调查期间收集的家庭成员的手机号码。这有利于在不收集新的消费数据的情况下进行成本效益高的福利分析。时长为35分钟的问卷涵盖了10个模块（知识、行为、服务获取、粮食安全、就业、安全网、冲击等……）。数据收集计划为六个月（六轮），问卷设计了核心模块和轮换模块。调查数据收集始于2020年5月11日，预计每三到四周对家庭进行一次回访。

调查的主要目标是：
• 确定直接或间接受到疫情影响的家庭类型；
• 确定疫情影响家庭的主要渠道；
• 提供有关收入和社会经济指标的相关数据，以评估疫情的社会福利影响。

地理覆盖范围
---------------------------
全国覆盖

分析单元
---------------------------
家庭

总体
---------------------------
本调查仅涵盖2018/19年调查的家庭，不包括监狱、医院、军营和学校宿舍等人口。

数据类型
---------------------------
样本调查数据 [ssd]

抽样程序
---------------------------
抽样程序
马里COVID-19影响监测调查是一种高频计算机辅助电话访谈（CATI）。调查样本来自2018/19年-《家庭生活条件综合调查》（EHCVM）-的人口，该调查于2018年10月至2019年7月进行。EHCVM本身是一项在全国、地区以及城市/农村具有代表性的样本调查。在EHCVM的7,000户家庭中，约90%的家庭收集了电话号码。每户家庭有1-4个电话号码。与WAEMU其他国家相似的抽样旨在通过三个区域：巴马科首都、其他城市地区和农村地区，获得具有代表性的估计。最小样本量为1,908户，其中1,766户成功接受访谈，即国家层面的预期最小样本量的约98%。鉴于马里首次进行电话调查，总共抽取了2,270户（增加25%），以考虑未知的非响应率或数据库中无效号码的存在。

第一轮完成的访谈总数为1,766。
第二轮完成的访谈总数为1,935。
第三轮完成的访谈总数为1,901。
第四轮完成的访谈总数为1,797。
第五轮完成的访谈总数为1,766。

数据收集方式
---------------------------
计算机辅助电话访谈 [cati]

研究工具
---------------------------
所有访谈材料均被翻译成法语供国家统计局使用。问卷以当地语言进行，时长约为30-35分钟，涵盖了以下主题：
1- 家庭成员名单
2- COVID-19知识
3- 行为和社会距离
4- 基本服务获取
5- 就业和收入
6- 价格和粮食安全
7- COVID-19的其他影响
8- 收入损失
9- 应对/冲击
10- 社会安全网
11- 脆弱性
12- 治理和社会政治危机

数据清洗操作
---------------------------
数据收集结束后，国家统计局对原始数据集进行了清洗。这包括格式化，并根据监控问题、调查员反馈和调查变更，纠正结果。

响应率
---------------------------
预期的最小样本量为1,809户家庭（每个领域603户）。因此，样本量因此覆盖了99%的巴马科，约100%的其他城市地区和91%的农村地区。总体而言，最小样本量覆盖率为98%。这一覆盖率提供了全国和每个领域的可靠数据。第一轮的响应率为77.8%。第二轮的响应率为85.2%。第三轮的响应率为83.7%。第四轮的响应率为79.2%。第五轮的响应率为79.7%。