IDELA development and validation

• In Event: 2-101 - Developing and Validating Multi-Dimensional Early Childhood Development Measures for use Across Countries and Cultures

Integrative Statement

With the inclusion of early childhood development (ECD) in the 2015 Sustainable Development Goals there is an increased need for rigorous measurement of early learning and development at local, national, and global levels. A number of tools designed to measure learning and development at the pre-primary level exist, but none have been proven to be widely applicable in low resource contexts. For example, the Early Development Index (EDI) has been shown to have strong psychometric rigor, but is administered by teachers and is therefore not appropriate for use with the large proportion of children not enrolled in pre-primary education. Other tools such as the Regional Project on Child Development Indicators (PRIDI) and the East Asia-Pacific Early Child Development Scales (EAP-ECDS) have a regional focus and have not been shown to be applicable globally.

Save the Children developed the International Development and Early Learning Assessment (IDELA) in an effort to create a tool which is appropriate and feasible in diverse low- and middle-income countries (LMIC), as well as psychometrically rigorous. This presentation will detail all IDELA validation work completed to date. Tests of inter-rater reliability, test-retest reliability, and concurrent validity with the Ages and Stages Questionnaire (ASQ) were conducted during the IDELA development period from 2011-2015 (Pisani, Borisova & Dowd, 2018). Notably, results of concurrent validity tests with the ASQ tool found all corresponding domains to be positively correlated, and revealed that IDELA scores were more normally distributed and more sensitive to intervention effects(figures 1a-d).

[Figures 1a-d]

Tests of construct validity and measurement invariance were conducted in 2016-2017. Using a sample of 682 children from rural Ethiopia researchers used exploratory and confirmatory bi-factor analyses to assess the internal structure of the assessment with respect to the four hypothesized domains of school readiness and established measurement invariance across three focal comparisons (children enrolled in center-based care versus home-based care; girls versus boys; and treatment status in an impact evaluation). The results support the conclusion that the IDELA is useful for making inferences about children's school readiness (Wolf et al., 2017).

Recently, similar analyses were conducted using data from Bhutan (N = 1326) and India (N = 900), and cross-country invariance was tested across five countries (N = 4,970; Afghanistan, Bolivia, Ethiopia, Uganda, Vietnam). Within-country bi-factor analyses, multi-domain/hierarchical models consistently fit the data better than unidimensional models, suggesting that the four hypothesized domains (early language, numeracy, social-emotional, gross/fine motor) are indeed distinct, albeit highly correlated. Results of the measurement invariance analyses find evidence for configural invariance, but not metric or scalar invariance across countries. Implications for current directions in international ECD measurement will be discussed.

Authors

• Lauren Pisani, UCL
  Presenting Author

• Iva Borisova, UNICEF
  Non-Presenting Author

• Amy Jo Dowd, Save the Children
  Non-Presenting Author

• Sharon Wolf, University of Pennsylvania
  Non-Presenting Author

• Peter Halpin, New York University
  Non-Presenting Author

• Hiro Yoshikawa, New York University
  Non-Presenting Author

• Alice Wuermli, New York University
  Non-Presenting Author