Improving Adolescent Lives in India: An Impact Evaluation

In 2015,UNICEF and IKEA Foundation launched the "Improving Adolescents' Lives in South Asia" program, with interventions in India, Pakistan and Afghanistan. The overall aim of the program implementation is to contribute to developing a model for scale-up in South Asia.
In India, the program operates in four states: In Andhra Pradesh (southeast India), the neighboring states of West Bengal and Jharkhand (east India) and in Assam (northeast India). As part of the program, context-specific pilot activities were launched in each state, with the goal of improving adolescents' lives by impacting the adolescents, parents and communities as a whole. The interventions include information dissemination through media outlets such as radios, print media, and television, as well as capacity building activities geared toward adolescents, their parents, and other members of the community. Therefore, outputs and outcomes are defined around three pillars: Pillar I on adolescents, Pillar II on stakeholders (parents and communities) and Pillar III on service providers. Therefore, the strategies of the program aim at enabling adolescents to protect themselves and enhancing supportive community and law enforcement structures through awareness raising, capacity building and mobilization of all stakeholders. Through these various channels, the interventions aim to shape attitudes towards marriage, gender roles, education, and adolescent empowerment, to thereby have an impact in reducing child marriage, increasing secondary education enrollment and decreasing adolescent pregnancy. This is the "Basic Package" that was provided across all states.
The evaluation design followed a two-stage randomization, where in the first stage a Basic Package including the aforementioned activities was applied, while the second-stage randomization included additional activities under the "Plus Package". The Plus Package activities were not uniformly implemented throughout the assigned program areas, but were rather randomly assigned to receive a package of interventions.
While the Basic Package was designed to include outreach, awareness raising, and capacity building activities across the three pillars and adolescent girls' groups, the Plus Package activities additionally comprised adolescent boys' groups, parents' groups, or a combination of both. Therefore, villages that received these additional efforts may experience greater improvements in the outcomes of interest through an increased engagement of additional target groups, as more comprehensive awareness and capacity building is expected to be linked with greater effectiveness.

2017-01-01

2019-12-01

Adolescent empowerment (awareness on child rights, confidence, corporal punishment, self-efficacy, general well-being), child marriage rates and practices, engagement rates and practices, child pregnancy rates, as well as education attendance, and attitudes towards education.

Information on program delivery (awareness, implementation, take-up); intermediate outcomes, such as on communication (intergenerational and intragenerational dialogue as well as interpersonal communication competence), service provision (access to services over time, adolescent-friendly services, marriage service provision), social norms and practices (on child marriage, child rights, equal opportunities)

(Clustered) Randomized Control Trial (RCT)

1. T0 120 villages where only the basic package treatment (incl. adolescent girls groups, mass media, or other activities) is implemented. (I.e. this is treatment arm (a).) In the evaluation of the additional effects of the plus activities this treatment arm T0 serves as “control” villages
2. T1 120 villages with boys’ groups: “treatment arm 1” / Adolescent Boys Group (BG)
3. T2 120 villages with parents’ groups: “treatment arm 2” / Parents Group (PG)
4. T3 120 villages with boys and parents’ groups: “treatment arm 3”/ BG + PG

Random allocation with matching at the block level. A re-randomization procedure was used to allocate the treatment arms.

Two stage randomization at block and village level

Yes

700 villages in 82 blocks

7000 adolescents, parents and community leaders; due to the COVID-19 outbreak we are in the process of amending our plan. The overall target is to collect information from 7 people per village, asking the adolescents and parents as well as the community leaders

Between 63 and 73 treatment and 40 or 80 for control villages (clusters) per treatment arms, for the second stage randomization. For the first stage, 42 treated blocks versus 40 control blocks.

Basic power calculations show that the school enrollment indicator has close to 100% power with the given sample size (for effect sizes between 3% and 5%). Adolescent pregnancy has power of 70% for a change in effect size of about 3%. Other indicators that appear to have sufficient power (for changes in effect sizes in the range of 9% and 12%) of at least 80% are improvement in awareness of harms of child marriage and menstrual hygiene.