Article Per Year (5 Year)
p-Index From 2021 - 2026
0.23
P-Index
This Author published in this journals
All Journal Indonesian Journal of Science and Mathematics Education
Punch, Masebe Tshidiso
Unknown Affiliation
Education Mathematics

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
The impact of selection bias in the analysis of 2018 grade 12 exam data Sibanda, Emmanuel; Peter, Mathye; Punch, Masebe Tshidiso; Rankhumise, Mmushetji Petrus
Indonesian Journal of Science and Mathematics Education Vol. 9 No. 2 (2026): Indonesian Journal of Science and Mathematics Education
Publisher : Universitas Islam Negeri Raden Intan Lampung

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24042/ijsme.v9i2.23962

Abstract

Many studies comparing the performance of students taking 12th-grade exams assume comparable study groups. In practice, selection bias affects comparisons because these studies rely on observational data. This study compares findings from analyses of matched and unmatched datasets to demonstrate the impact of selection bias. The propensity score technique matched learners who took the 2018 Grade 12 examinations in the Western Cape and Gauteng Provinces. Based on the unpaired dataset, the probability of a student graduating in the Western Cape Province is the same as in Gauteng Province. However, analysis of the paired dataset reveals that the probability of a student graduating in the Western Cape Province is 96% lower than that of their counterparts in Gauteng Province. The paired dataset suggests that Gauteng Province provides a more favorable environment or conditions for students to complete Grade 12 than the Western Cape Province. It is concluded that pre-existing differences influence the validity and reliability of the results. Differences in findings between the unpaired and paired datasets highlight the importance of adjusting for measured baseline factors before comparing the two provinces. The implication is that baseline adjustment is essential for valid educational comparisons.
Co-Authors Peter, Mathye Rankhumise, Mmushetji Petrus Sibanda, Emmanuel