Survey for Adolescent Development Programme - 2009-2011

Baseline and Follow-up Impact Evaluation Surveys

Implementation of the field activities of ELA in Tanzania started in 2009 with 100 adolescent centers in 10 branches located in Irigna and Dodoma, with a mix of urban and rural locations. We over-selected the number of communities to construct a control group. Prior to launching the interventions, the program implementation team identified 15 communities in each branch as potential sites for setting up the clubs. Stratified at the branch level, these 150 communities were randomly divided into two groups- 100 treatment communities and 50 control communities. All the treatment communities received the ELA intervention (i.e. a club, life-skills training, livelihood training and community meetings). For simplicity, we refer to this set of interventions as ‘Club’. The treatment communities were then further randomized into two groups –‘Club only’ and ‘Club + microfinance’. Therefore, there are three types of communities:

1. Group A: ‘Club only’ communities offering adolescent development clubs and associated interventions
2. Group B: ‘Club + MF’ offering microcredit in addition to Group A interventions
3. Group C: Control communities with no interventions
   In every community, a census of all eligible girls was done by the program team and used as our sampling frame. A baseline survey of 30-40 girls from each village was conducted during January-July, 2009. Table A1 shows the balance checks of the three groups in terms of their baseline characteristics. The normalized difference between the two treatment groups and the control group is less than 0.1 for all variables, which is lower than the rule of thumb value suggested by Imbens and Wooldridge (2009), and significant in only 1 of 40 cases. The clubs were established in 2009 after the baseline survey. Microfinance rollout started in early 2010 in the second treatment group after all the clubs were set up. In order to establish a panel data set, the follow-up survey was conducted during June- November of 2011 on the same adolescent girls surveyed at baseline.