000			06540ntm a2200337 i 4500
003			MY-KuUP
005			20261006152520.0
006			t||||fr|||| 000 0
007			ta
008			261006t20262026my a|||fr|||| 000 0 eng d
020			_aTHE0010819 (Local) _qHardback
040			_aUMPSA _beng _cUMPSA _erda
090			_aPSM .S53 2026 r Thesis
100	0		_aSharainie Sahrin, _eauthor.
245	1	0	_aModelling secondary students’ mathematics performance using a hybrid model of structural equation modelling and artificial neural network / _cSharainie Sahrin
264		1	_aKuantan, Pahang : _bUMPSA _c2026
264		4	_c© 2025
300			_axvii, 309 pages : _bIllustrations ; _c30 cm.+ _e1 CD ROM.
336			_2rdacontent _atext
337			_2rdamedia _aunmediated
338			_2rdacarrier _avolume
347			_2rda _atext file _bPDF
500			_aCenter for Mathematical Sciences
502			_aThesis (Doctor of Philosophy) -- Universiti Malaysia Pahang - 2026
504			_aIncludes bibliographical reference
520	3		_aThe persistent underachievement of Malaysian students in mathematics has become a critical issue, as reflected in various international assessments. This ongoing issue raises concerns about its potential impact on national examination result, Malaysian Certificate of Education (SPM) and the future enrolment in science-related fields. Recent international assessments such as Programme for International Student Assessment (PISA) and Trends in International Mathematics and Science Study (TIMSS) have highlighted a concerning decline in Malaysia’s mathematics performance. Reports indicate that Malaysia continues to lag behind the ASEAN average. Additionally, the Malaysian Education Development Plan 2013-2025 highlighted that only 35% of Malaysian students demonstrate a minimal understanding of fundamental mathematical concepts. This shows underscore the urgency to identify key factors. Although prior studies have identified various determinants of mathematics achievement, limited research has modelled secondary students’ mathematics performance using an integrated approach that captures both linear and nonlinear relationships. Addressing this gap, the present study aims to examine the key factors influencing lower secondary students’ mathematics performance using structural equation modelling (SEM), to develop a hybrid structural equation modelling and artificial neural network (SEM-ANN) model for predicting students’ mathematics performance and to evaluate and compare the predictive performance of the hybrid SEM-ANN model with the individual SEM and ANN models. The instrument used is online questionnaire that was validated by using Exploratory Factor Analysis (EFA) with a pilot sample of 102 students showing that the Cronbach’s alpha values were significant. The instrument was analysed using IBM SPSS 26.0. A sample comprised 302 lower secondary students (Form 1, Form 2 and Form 3) from selected secondary school in Malaysia. The sample is selected using the non-probability sampling approach (voluntary response sampling). The questionnaire measured four main constructs which are students’ interest across five mathematics domains (algebra, statistics, number and operations, geometry and discrete mathematics), teachers’ personality, peer influence and perceived technology. Reliability analysis indicated satisfactory Cronbach’s alpha values for all constructs. The dependent variable which is students’ mathematics performance was measured using their final semester 2 mathematics examination score. Subsequently, Confirmatory Composite Analysis (CCA) was conducted using a separate sample of 302 students to assess the reliability and validity of the measurement model. The reliability was evaluated using Cronbach’s alpha and composite reliability (CR) to ensure internal consistency of the measurement items. Convergent validity (CV) was assessed using average variance extracted (AVE) ensuring the measurement model is free from redundant items. Discriminant validity (DV) was examined using Heterotrait-Monotrait ratio (HTMT) to confirm that constructs are not highly correlated with each other. The causal relationships among constructs were then validated using PLS-SEM. Latent variable scores obtained from PLS-SEM were then used as inputs for the ANN model to capture nonlinear relationships and determine the relative importance of predictors through sensitivity analysis. The PLS-SEM model revealed that all four factors which were students’ interest, perceived technology, peer influence and teachers’ personality were significantly influence students’ performance in mathematics at 0.05 level of significance The sensitivity analysis revealed that peer influence, with 100% normalized importance, was the most influential factor, indicating students’ strong consensus on the importance of their peers in enhancing mathematical understanding. This was followed by students’ interest at 90.1%, perceived technology at 71.9% and teachers’ personality at 64.1%. To further assess robustness, simulated datasets of 500 and 1000 cases were generated. Model comparisons were performed among three approaches which are SEM, ANN using mean scores and hybrid SEM-ANN using latent variable scores from SEM. The model accuracy was evaluated using root mean squared error (RMSE). The results indicated that the hybrid model surpassed the individual SEM and ANN models, as it achieved the lowest RMSE value among the three. This study contributes methodologically by integrating SEM and ANN which provides a comprehensive and robust modelling approach considering the linear and nonlinear patterns in examining mathematics performance among lower secondary students. The findings provide empirical evidence to guide policymakers and educators in refining learning strategies, strengthening peer-supported learning environments, enhancing technology integration and fostering students’ interest in mathematics. Such initiatives are critical to improve mathematics achievement and sustaining Malaysia’s long-term development in science and technology. Without such efforts, fewer students pursue science streams due to fear or disinterest in mathematics. Consequently, the decline in interest could hinder Malaysia’s progress in technological advancement.
610	2	0	_aCenter for Mathematical Sciences _xDissertations
650		0	_aUniversities and colleges _xDissertations
650		0	_aTheses _xDissertations
942			_2lcc _cTHESIS
999			_c105680 _d105686