The Adolescent Girls Initiative (AGI) pilot was implemented by the Government of Rwanda as part of an eight-country initiative led by the World Bank aimed at promoting the economic empowerment of adolescent girls. The development objective of the Rwanda AGI was to improve employment, incomes and empowerment of disadvantaged adolescent girls and young women (aged 16-24), and to test two integrated models for promoting these goals.
The Rwanda AGI had three components: Component I: Skills Development and Entrepreneurship Support, Component II: Scholarships to Resume Formal Education, Component III: Project Implementation Support
This evaluation focused exclusively on Component I, which was carried out by the Workforce Development Authority (WDA), under the supervision of the Ministry of Gender and Family Promotion (MIGEPROF). It was delivered sequentially to roughly 2,000 vulnerable girls and young women in three equal-sized cohorts between 2012 and 2014. The project was targeted geographically in four districts (Gasabo, Kicukiro, Gicumbi, and Rulindo), where nine vocational training centers (VTCs) provided the training.
The three objectives of the evaluation were:
- To examine how well the AGI project delivered the planned activities
- To assess the usefulness of the training provided
- To measure the change in beneficiary outcomes before and after the AGI project.
The evaluation was conducted on the second cohort of beneficiaries, from which 160 girls were randomly selected to participate in baseline and endline surveys.
Kind of data
Sample survey data [ssd]
v01, edited anonymous dataset for public distribution
The project targeted geographically to four districts that already had training centers: Gasabo, Kicukiro, Gicumbi and Rulindo.
Unit of analysis
Producers and sponsors
The World Bank - Africa Gender Innovation Lab
The World Bank
Adolescent Girls Initiative Multi-Donor Trust Fund
After the initial pre-screening for eligibility, the sample was stratified by the sector of participants' residence and selected through a public lottery conducted by Workforce Development Authority and the Ministry of Gender and Family Promotion in each of the 11 recruitment sectors. The girls were invited to attend, and directly after the lottery, Laterite Limited - an independently contracted research firm - conducted uniform random sampling (in Excel) to select a subset of admitted applicants for the baseline survey. However, the baseline survey was administered only to those who were physically present at the lottery. In 6 of the 11sectors of recruitment, girls who did not appear for the lottery were excluded from the project, so the evaluation sample reflects the project sample. In the other 5 sectors, absent applicants who were randomly selected for project admission were still allowed to join, but they were still excluded from the baseline survey. Specifically, cohort 2 had 1,364 applicants who passed the screening committee and 712 were randomly selected for project admission. Further, unsuccessful but eligible applicants were allowed to enter the lottery for the third cohort, which started just one month after the second cohort. Hence, there was no feasible way to use the rejected applicants as a control group for an evaluation.
A follow-up survey was administered to 160 of the 182 randomly sampled beneficiaries that responded to the baseline survey. Though special effort was made to follow up with the 43 individuals from the baseline survey who did not complete the project, the team was only able to interview 21 of them.
Dates of collection
Mode of data collection
Computer Assisted Personal Interview [capi]
After the collection of survey data, Laterite Limited prepared the data for analysis by correcting duplicate identification numbers, renaming endline variable names in order to match baseline variable names, dropping confidential personal identification variables (e.g. name, mobile phone number), GPS coordinates, device numbers, codifying variables stored as names of income-generating activities (IGAs), and merging baseline and endline datasets.
A number of additional changes to the data were made during the quantitative analysis:
- Values of specific variables (e.g. business type, first or second income-generating activity) recorded as "other" that fit existing answer options were re-codified;
- To address inconsistencies between different sections of the survey, values entered for the IGA screening sections (whether respondents was engaged in any household agricultural activities, wage employment, non-farm business or internship) were corrected based on information provided in subsequent, more detailed, questions on the two main income-generating activities and/or business. No changes were made in the absence of supporting information. Where both wage employment and non-farm businesses were indicated for the same IGA, answers to screening questions were reconciled based on whether the respondent reported working for herself (business) or for a non-relative (paid job).
- Because 86 out of 160 values for age at baseline were missing in the merged dataset provided by Laterite Limited, data on age was extracted from the baseline dataset;
- Outliers - 3 income values (extra 0 at the end, or amount entered as in-kind daily payment instead of monthly income) and 4 in-kind amount values (divided by 10 to fit in ranges of reported in-kind amounts for same occupation) were considered typos; for the remaining outliers, values above the 99th quintile were dropped from the estimations.
After cleaning the data to resolve inconsistencies and drop outliers, basic descriptive statistics were computed to give an overall picture of the beneficiaries' status and statistical analyses were conducted to compare the average values of each indicator using Student's t-test.
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
Shubha Chakravarty, The World Bank - Africa Gender Innovation Lab; Sarah Haddock, The World Bank. Rwanda Adolescent Girls Initiative (AGI) Evaluation 2012-2014. Ref. RWA_2012-2014_AGIE_v01_M. Dataset downloaded from [URL] 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.