Poster #43 - Impacts of tablet-based curriculum intervention on children’s literacy and numeracy learning trajectories in Malawi

• In Event: Poster Session 17

Abstract

Tablet-based curriculum programs have emerged as promising interventions for addressing poor learning outcomes in developing contexts, where schools are challenged to provide quality learning amidst expanding enrollment (Escueta, Quan, Nickown & Oreopoulos 2017, Pitchford et.al. 2015, World Bank 2021). Rigorous research conducted in Africa since 2015 has established that onebillion’s software, onecourse, an award-winning tablet-based curriculum, produces meaningful impacts in literacy and numeracy (Levesque, Bardack, and Chigeda 2020; Levesque, Bardack, Chigeda, Bahlibi, and Winiko 2022; Pitchford, Hubber, and Chigeda 2017). Yet, no analysis to date has examined the impact of these programs on children’s learning rates over time.

The current study addresses this gap by examining children’s learning trajectories in literacy and numeracy for a subsample of 441 learners who participated in all three assessments during a recent 2-year RCT conducted in Malawi. The RCT used a non-clustered, blocked individual random assignment design. Independently within each of two schools, children were randomly assigned to two treatment groups (literacy and math) and one control group within four gender (male and female) and age category (6-7 years and 8-10 years) strata. Standardized mean differences met group design standards for baseline equivalence of the final analytic sample, ranging from 0.00 to 0.08 standard deviations in absolute value. Over two years, children were assessed at baseline (October 2019), midline (October 2020) and endline (November 2021). Primary outcomes for the study were based on the Malawi adaptations of the Early Grade Reading Assessment (EGRA) and Early Grade Mathematics Assessment (EGMA) in the Chichewa language: average EGRA percent correct (literacy intervention) and average EGMA percent correct (math intervention). Due to the COVID-19 pandemic, children experienced multiple school closures, resulting in 5 months of intervention from baseline to midline and 8 months of intervention from midline to endline (totaling 13 months of intervention).
We employ growth curve modeling to investigate whether differences in children’s rates of learning in literacy and numeracy vary over time depending on exposure to tablet-based curricula. Each model includes the following time-invariant covariates: gender, age category, strata fixed effects, treatment status, and school (urban versus peri-urban). We additionally include a cross-level interaction between time and treatment status to examine whether associations between achievement and time depend on assignment to treatment or control groups.
Tables 1 and 2 present nested multilevel models for literacy (Model 1) and numeracy (Model 2) outcomes. Results from likelihood ratio tests indicate that a random coefficient model, which includes a random slope for time in addition to a random intercept, significantly improves model fit for Models 1 and 2. Significant cross-level interactions between treatment status and time indicate that, compared to the control group, the literacy group showed high rates of learning in literacy (Model 1b) whereas the numeracy group showed even higher rates of learning in math over time (Model 2b). The broader implication of these findings, which shed new light on the program’s impact on trajectories in math and literacy achievement over two interrupted school years during the COVID-19 pandemic, will be discussed.

Authors

• Sarah Bardack, Imagine Worldwide
  Presenting Author

• Karen Levesque, Imagine Worldwide

• Abraham Bahlibi

• Antonie Chigeda, Imagine Worldwide

• Symon Winiko, University of Malawi