Human Development Cash Transfer Program Impact Evaluation - Baseline Survey 2017
ideas42 is partnering with the World Bank’s Madagascar Social Protection Team to support the Government of Madagascar in designing and implementing a rigorous impact evaluation that meets its policy needs by clarifying the effectiveness of the HDCT program. An important focus of the evaluation is in piloting certain community participation, behavioral science, and motivational interventions that may improve the effectiveness of the program and help guide its ultimate scalability. Data sources for the evaluation include:
1. A Proxy Means Test (PMT) in all the intervention areas that will identify only households in the bottom 30th percentile of income that have children ages 0-10 as eligible
2. Administrative data, including school attendance, collected by the Madagascar Ministry of Education
3. Baseline, Midline and Endline Household Surveys measuring household consumption, assets, parenting behaviors, child educational development, and other key outcome measures before, during, and after the program
4. Child Development Assessments using the Malawi Developmental Assessment Tool (MDAT), which was designed for an African context and further adapted to the Malagasy context.
The purpose of the baseline survey is to collect pre-HDCT measures on outcomes of interest (e.g., household consumption, parenting behaviors, household assets, child cognitive development) and to verify balanced randomization of subjects across all intervention groups. Results from the baseline survey will be compared to the findings from the midline and endline surveys (which will be conducted after households begin receiving the transfers) to determine how the outcomes of interest change by virtue of the transfer.
Household Consumption and Education Baseline
The household consumption and education baseline was designed by pulling key questions used in other national-level surveys carried out in Madagascar by INSTAT (Institut National de la Statistique de Madagascar), as well as in other relevant surveys conducted in Africa (e.g., World Bank surveys). The data collection began in August 2016 and ended in December 2016. 4,484 households were selected for the baseline survey using a random generator from the eligible households based on the PMT.
The data was collected by an independent survey firm (CAETIC Développement) based in Madagascar. The survey was done by direct interview of the household head and / or his spouse on the basis of a standard and pre-coded questionnaire. The survey used a tablet computer (electronic questionnaire) instead of a traditional paper questionnaire. On average the duration of the interview at a household was 52 minutes. The enumerator asked the questions on the survey, prompted oral responses from the participant, and collected those responses on the tablet. The modules in the survey are as follows:
- Household Demographic Information
- Food-related Household Consumption
- Non-food-related Household Consumption
- General Household Expenses (Schooling, Transportation, etc)
- Agricultural Production
- Sources of Household Revenue
- Women’s Empowerment
- Parenting Practices
- Food Security
Producers and sponsors
The Household Consumption and Education baseline survey was conducted in the six HDCT districts of Madagascar (Tomasina II, Mahanoro, Vohipeno, Ambohimahasoa, Betioky Sud, and Faratsiho): 4,484 randomly sampled households composed of 27,437 individuals were covered by the survey, representing
individuals from 378 fokontany across 52 Communes.
For the Malagasy Adaptation of the Malawi Developmental Assessment Tool (MDAT), 3366 households were eligible (have at least one eligible child) and surveyed using MDAT. The endline will be the full 8222 households, but the baseline enabled us do basic randomization checks and provides some information on the vulnerable population for the government before the 3-year experiment ends. Given these constraints, our approach to baseline sampling was as follows:
1. Sampled from all 309 treatment villages (surveyed 7-8 randomly selected eligible households per village)
2. For pure control, randomly sampled 65 villages total. This meant sampling 5 villages from each of the 13 control communes, with 9-10 people from each village. This is because cash is evaluated at the commune level – there are 110 control villages, and we did not necessarily need to go into all of them. Since there are 13 control communes, we needed to sample from about 5 villages from each commune, with 9-10 people from each village. That is roughly 600 people from 65 total villages. Other important notes are that four of the 13 control communes had only five eligible villages. One of the communes only had three villages, so five villages per commune was the only realistic sampling ceiling.
3. We randomly selected 3000 eligible households and only did the MDAT (the early childhood cognitive development measure instrument) for households that had children ages 2-6. That said, once we had the household list for the baseline, we could pre-select with which of the children in a given household we could conduct the MDAT.
Absent respondents/refusals/not found:
For absentee respondents/refusals/not found, field enumerators had a backup list of randomly selected households from which could replace the absentee household. Field followed order of the randomly generated backup list systematically to prevent the introduction of any bias into the sampling process.
Variables in the dataset used to identify the various levels of stratification:
- i0_2 - District
- i0_3 - Commune
- i0_4 - Fokontany
More details on sampling are provided in the HDCT Sampling Approach Brief available for download.
Dates of Data Collection
Data Collection Mode
Computer Assisted Personal Interview [capi]
Malagasy-adapted MDAT Baseline
While multiple assessment tools exist, most have been created for a developed country context. One tool designed for use in developing countries, the Malawi Developmental Assessment Tool (MDAT) (Gladstone et al 2010), was specifically designed for a rural African context and is publically available for adaptation to multiple countries. We
(1) adapted this tool to a Malagasy context, creating a Madagascar-specific tool that can be used by future researchers and early childhood development specialists to assess developmental status and
(2) used this tool to conduct a baseline assessment of child developmental status for the HDCT program. More specifically, the MDAT assesses child status across four domains: gross motor, fine motor, language, and social abilities.
Baseline data collection was conducted from October 2016 to December 2016. The Malagasy-adapted version of the MDAT was used to assess the developmental status of a subset of children who reside in households that are part of the HDCT program. An Early Childhood Development Expert (ECDE) oversaw the baseline survey process. The ECDE worked in partnership with a Malagasy Local Coordinator (LC) and CAETIC to conduct the baseline survey. Survey enumerators were trained and supervised by the ECDE, the LC, and CAETIC. Enumerators could not administer the MDAT to subjects unless they have been certified by the LC.
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 country, acronym and year of implementation)
- the survey reference number
- the source and date of download
World Bank. Human Development Cash Transfer Program Impact Evaluation - Baseline Survey 2017 (HDCTIE-BL 2017). Ref: MDG_2016_HDCT-BL_v01_M. Downloaded from [uri] on [date].
Disclaimer and copyrights
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.