Impact Evaluation of Community Preschool Expansion: Endline Survey


Cambodia. Location within country: 13 provinces (Kampong Chhnang, Kampong Speu, Kampot, Kandal, Koh Kong, Kratie, Mondulkiri, Preah Sihanouk, Prey Veng, Ratanakiri, Steung Treng, Svay Rieng, Takeo)

Project Description

The World Bank, in collaboration with the Ministry of Education, Youth and Sports (MOEYS), supports the Second Education Sector Support Project (SESSP) funded by Global Partnership for Education (GPE). One component of SESSP seeks to expand the availability and quality of Early Childhood Care and Development (ECCD) services. The objectives of the ECCD interventions are to assist MOEYS to expand access to quality Early Childhood Education (ECE) for 3-5 year old’s through construction of facilities, provision of materials and training of staff, as well as to build the demand for ECCD services among families from disadvantaged backgrounds. The program was also designed with a rigorous evaluation to measure the impact of community ECCD services on child cognitive and psychosocial development—with a focus on school readiness. To this end, the World Bank in collaboration with MOEYS is conducting the impact evaluation over three years, from 2016 to 2018.

After having conducted the baseline, Angkor Research was awarded the midline through a rigorous MOEYS procurement process. ARC involvement in this large-scale randomized control trial includes sample selection in isolated communities with insufficient baseline sample, tracking and re-interview of households interviewed at the baseline, instrument review/translation, primary CAPI (tablet-based) data collection and data management, reconciliation and linkages with the baseline and midline for approximately 7,300 households and 305 preschools and villages (divided into treatment and control communities) in 13 provinces, to evaluate the service delivery and effects of preschools.

A range of household and child indicators were tested this round, including anthropometrics and cognitive testing for children. Household and individual data is longitudinal, and linked across survey rounds. Data collection also included interviews with village authorities and community preschool teachers, as well as community preschool facility assessments, photos and GPS locations in all 305 sample villages. Additional endline activities included school mapping of all preschools and primary schools within 3km of target villages, teacher interviews and classroom observations (both in-person and video recorded) were collected in around 327 preschools in target and surrounding villages.