Dataset on the Efficiency of Indigenous School in Malaysia using DEA

Published: 4 September 2025| Version 1 | DOI: 10.17632/5zt9d63v9x.1
Contributor:
Mohd Nur Ruzainy Alwi

Description

This dataset offers an in-depth look at how 41 Indigenous (Orang Asli) schools perform across seven states in Malaysia: Johor, Kedah, Kelantan, Pahang, Perak, Selangor, and Terengganu. These schools make up 44% of the total nationwide. We collected the data using a stratified random sampling method and a questionnaire based on the OECD PISA 2017 tools. The dataset includes five main educational inputs: framework, pedagogy, culture, space, and connection. It also covers four educational outcomes: attendance, passing rates, completion rates, and transition to secondary school. We recorded responses on a 5-point Likert scale and applied percentage categories to specific outcome variables. The dataset features both raw survey responses and processed Data Envelopment Analysis (DEA) efficiency scores, along with school rankings. This resource is useful for researchers, educators, and policymakers. Researchers can replicate DEA analysis, explore different DEA models, or conduct comparative studies in indigenous and marginalized education settings. Educators can gain insights into effective teaching and learning methods. Policymakers can use evidence-based information to create interventions that aim to improve educational equity and efficiency.

Files

Steps to reproduce

Research Design & Sampling Cross-sectional survey with stratified random sampling from 93 Orang Asli schools across seven states (Johor, Pahang, Perak, Kelantan, Negeri Sembilan, Selangor, Terengganu). Sampled 41 schools (44%), highest from Perak (14), Pahang/Selangor (6 each), with one administrator per school. Instrument Development Adapted OECD PISA 2017 questionnaire for indigenous context. Five inputs: Framework, Pedagogy, Culture, Space, Connection (5-point Likert scale). Four outputs: Attendance, Passing Rates, Completion Rates, Transition to Secondary School (percentage or Likert scale). Data Collection In-person questionnaire administration with immediate collection. Secured JAKOA/MOE permissions, informed consent, and data anonymization. Data Processing Aggregated Likert responses into composite scores. Handled missing data via mean substitution or listwise deletion. Analysis Used input-oriented CCR DEA model in Excel Solver for efficiency scores (0–1) and rankings. Reproducibility Access PISA 2017 tools from oecd.org; update school list via JAKOA/MOE. Replicate sampling, survey, and DEA in Excel or R (Benchmarking package). Dataset available in Excel files; no reagents used.

Institutions

  • Universiti Tenaga Nasional

Categories

Management, Indigenous Knowledge, Indigenous as Vulnerable Population, Marginalised Population, Community Education

Funders

Licence