Impact of Mental Math on Numerical Competency and Cognitive Abilities in Public Elementary Schools: A Randomized Controlled Trial
How to cite
Khan, Rida; MARCHETTA, Francesca; FERNANDEZ, Eric ROCA. 2026. "Impact of Mental Math on Numerical Competency and Cognitive Abilities in Public Elementary Schools: A Randomized Controlled Trial." AEA RCT Registry. AEARCTR-0018349. https://doi.org/10.1257/rct.18349-1.0
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
Background and Motivation:
Mathematical competency in early primary education is a strong predictor of long-term academic achievement and economic outcomes. However, numeracy in Pakistan's public primary schools remains critically low. A key driver of weak mathematics performance is an over-reliance on rote memorisation and algorithmic procedures, at the expense of flexible numerical reasoning. Mental math — the ability to perform calculations efficiently and accurately without written aids — develops this flexibility by training students to apply strategic thinking to numerical problems. Despite a growing body of evidence from high-income contexts, there is limited rigorous evidence on whether structured mental math instruction, delivered through teacher training, can improve mathematical fluency and broader cognitive outcomes in a low-resource South Asian setting.
Intervention:
This study evaluates a mental math training programme delivered to mathematics teachers in public primary schools in District Attock, Punjab, Pakistan. The intervention trains teachers in a set of evidence-based mental computation strategies — including partitioning and recombining, sequencing, use of near-multiples, and doubling and halving — which they then integrate into regular classroom instruction for Grade 3 and Grade 4 students. The training is designed to be low-cost, replicable, and compatible with the existing national curriculum.
Study Design:
This is a parallel two-arm, school-level randomized controlled trial. Sixty public primary schools across Tehsil Attock and Tehsil Hassanabdal were randomly assigned to treatment (30 schools) or control (30 schools), stratified by tehsil and school gender type. Randomisation was conducted at the school level to avoid contamination between treatment and control classrooms. Control schools continue with standard instruction throughout the study period.
Data:
Data will be collected at baseline and endline. Student assessment measuring direct calculation, applied word problems, mathematics fluency, non-verbal reasoning through matrix pattern completion, cognitive inhibitory control through the Stroop colour-word task, and a standardized oral mental math sequence.
Primary Outcomes:
The primary outcomes are student scores on the mental math sequence, the direct calculation section, and the mathematics fluency section of the assessment. Secondary outcomes include applied word problem performance, non-verbal reasoning scores, and inhibitory control as measured by the Stroop task.
Heterogeneous Effects:
The analysis will examine heterogeneous treatment effects by student gender, baseline mathematics performance, school type (mixed versus single-sex), and household socioeconomic status.
Sample and Timeline:
The study covers approximately 2,800 Grade 3 and Grade 4 students across 60 schools.
Intervention
This study evaluates a mental math teacher training programme implemented in government primary schools in District Attock, Punjab, Pakistan. Mathematics teachers in treatment schools receive training in a set of evidence-based mental computation strategies, which they are then expected to integrate into their regular classroom instruction for Grade 3 and Grade 4 students.
The training covers six core mental math technique families: partitioning and recombining (breaking numbers into component parts before operating); sequencing (adding or subtracting in ordered steps); use of near-multiples and compensating (rounding to a convenient number then adjusting); doubling and halving; deriving unknown facts from known ones; and rapid recall of arithmetic facts to support fluency.
Experimental design
This is a two-arm, school-level randomized controlled trial conducted in District Attock, Punjab, Pakistan. Sixty public primary schools across two tehsils (Tehsil Attock and Tehsil Hassanabdal) were randomly assigned to treatment or control. All Grade 3 and Grade 4 students in each sampled school are included in the study population.
Baseline and endline data are collected from students in all 60 schools. Student assessments measure mathematics performance and cognitive abilities. The study follows a pre-registered analysis plan with intent-to-treat as the primary estimator.
Design details
Cited by
No citation data available for this registration yet.