Analysis of 2022 and 2023 Osun State basic education certificate examination mathematics items using item response theory
Implications for large scale assessment
Bibliographic Data
| ID | 22203408 |
|---|---|
| Authors | Oluwaseyi Aina Gbolade Opesemowo (0000-0003-0242-7027, University of Johannesburg, corresponding author), Kehinde Olufunke Opatunji (Obafemi Awolowo University), Temitope Babatimehin (Obafemi Awolowo University), Titilope Rachael Opesemowo (0000-0002-0553-7355, University of Johannesburg) |
| Year | 2026 |
| Volume | 13 |
| Pages | 102381 |
| Publication date | 2026-06-01 |
| Open Access | Yes |
| Type | ARTICLE |
| Venue | Social Sciences & Humanities Open (JOURNAL) |
| Journal identifiers | ISSN: 2590-2911 • E-ISSN: 2590-2911 |
| Publisher | Elsevier BV (PUBLISHER) |
| DOI | 10.1016/j.ssaho.2025.102381 |
| OpenAlex | W7118247027 |
| Language | EN |
| References cited | 23 |
This study examined the item parameters of 2022 and 2023 Osun State Basic Education Certificate Examination (BECE) Mathematics items using Item Response Theory (IRT), with a focus on discrimination, difficulty, and differences in the item parameters. While most Nigerian studies focus on national assessments, limited attention has been given to state-level examinations, leaving the quality and fairness of items underexplored. An ex-post facto design was adopted, with a population of 34,899 (2022) and 35,857 (2023) candidates. A systematic random sample of 1052 responses was analysed using a 3-parameter logistic model. The instrument consisted of adopted 2022 and 2023 BECE Mathematics items, comprising 60 multiple-choice questions. We analyse data using the Jmetrik software. Results showed that 32 items in 2022 and 37 items in 2023 met acceptable psychometric standards, while nine and one items, respectively, exhibited poor difficulty indices. Difficulty analysis indicated that nine items in 2022 and one item in 2023 demonstrated poor indices. No significant difference was observed in the psychometric properties of the two examination years (t = 0.75, df = 1050, p > .05). These results highlight inconsistencies in item quality across years, raising concerns about fairness and reliability in high-stakes assessments. The study contributes to discourse on large-scale assessments by situating Nigerian state-level examinations within global debates on fairness, accountability, and psychometric rigour. It further underscores the value of IRT in aligning Nigeria's assessment practices with international best practices. These findings underscore the importance of enhancing the psychometric quality of mathematics items and aligning evaluation practices with international standards of educational equity
Certificate · Item analysis · Item response theory · Logistic regression · Population · E-Learning and COVID-19 · Psychometric Methodologies and Testing · Student Assessment and Feedback
Exploring undue advantage of differential item functioning in high-stakes assessments
Perception and production of L2 Arabic emphatic consonants
Fitting the Rasch Model to Account for Variation in Item Discrimination
Reimagining our futures together
Large-scale assessments of students’ learning and education policy
Influencing public school policy in the United States
Coefficient alpha and the internal structure of tests
| Highly cited | No |
|---|