Article: Legal literacy assessment and prediction among students using logistic regression and latent structural modelling Journal: International Journal of Information and Communication Technology (IJICT) 2026 Vol.27 No.85 pp.1 - 19 Abstract: Traditional assessments of student legal literacy rely on static questionnaires or performance analysis, which often fail to capture structural complexity and predict highrisk behaviours accurately. To address this, we propose an integrated algorithm combining logistic regression (LR) and a latent structural model (LSM) based on the twoparameter logistic IRT model. LSM estimates latent traits across legal knowledge, concepts, and abilities, mitigating nonlinearity and measurement error, while LR incorporates demographic, behavioural, and latent trait data to predict future highrisk behaviour. A joint optimisation mechanism quantifies literacy and enables early warning of risks. Experiments show the fusion model achieves 88.5% accuracy in highrisk behaviour prediction, outperforming standalone LR (78.2%) and structureddataonly models (75.5%). It also yields an F1 score of 0.875 and recall of 86.0%, demonstrating robust performance. The approach offers a refined framework for legal literacy assessment and supports targeted prevention strategies in educational management. Inderscience Publishers - linking academia, business and industry through research

Title: Legal literacy assessment and prediction among students using logistic regression and latent structural modelling

Authors: Bing Liu

Addresses: School of Marxism, Shaanxi Technical College of Finance and Economics, Xianyang, 712000, China

Abstract: Traditional assessments of student legal literacy rely on static questionnaires or performance analysis, which often fail to capture structural complexity and predict highrisk behaviours accurately. To address this, we propose an integrated algorithm combining logistic regression (LR) and a latent structural model (LSM) based on the twoparameter logistic IRT model. LSM estimates latent traits across legal knowledge, concepts, and abilities, mitigating nonlinearity and measurement error, while LR incorporates demographic, behavioural, and latent trait data to predict future highrisk behaviour. A joint optimisation mechanism quantifies literacy and enables early warning of risks. Experiments show the fusion model achieves 88.5% accuracy in highrisk behaviour prediction, outperforming standalone LR (78.2%) and structureddataonly models (75.5%). It also yields an F1 score of 0.875 and recall of 86.0%, demonstrating robust performance. The approach offers a refined framework for legal literacy assessment and supports targeted prevention strategies in educational management.

Keywords: legal literacy assessment; logistic regression; latent structural model; LSM; latent traits; high-risk prediction.

DOI: 10.1504/IJICT.2026.155486

International Journal of Information and Communication Technology, 2026 Vol.27 No.85, pp.1 - 19

Received: 25 Feb 2026
Accepted: 29 Mar 2026
Published online: 03 Aug 2026 *

Title: Legal literacy assessment and prediction among students using logistic regression and latent structural modelling

Keep up-to-date