Partial least squares path modelling of UMP students’ attitudes on statistics achievements

• Main Author: Hassan Rahnaward, Ghulami
• Format: Thesis
• Language: English
• Published: 2015
• Subjects: TA Engineering (General). Civil engineering (General)
• Online Access: http://umpir.ump.edu.my/id/eprint/11244/19/Partial%20least%20squares%20path%20modelling%20of%20UMP%20students%E2%80%99%20attitudes%20on%20statistics%20achievements.pdf; http://umpir.ump.edu.my/id/eprint/11244/

The objective of this study is to examine the structural relationships in the hypothesised model that consisted of antecedent constructs of students‘ attitudes i.e. Affect, Cognitive Competence, Value, Difficulty, Interest, and Effort and their impacts on the statistics achievement. This study employed an experimental survey design which involved 718 undergraduate engineering students who enrolled in applied statistics courses in Universiti Malaysia Pahang. The samples were from different faculties such as Faculty of Industrial Sciences and Technology, Faculty of Electrical and Electronic Engineering, Faculty of Civil Engineering and Earth Resources, Faculty of Chemical Engineering and Natural Resources, Faculty of Industrial Management, Faculty of Computer Systems and Software Engineering, Faculty of Manufacturing Engineering, and Faculty of Mechanical Engineering. The Survey of Attitudes towards Statistics, with a seven point Likert scale was adopted and adapted from reliable instrument. The structural equation model through partial least squares estimation (SEM-PLS) using SmartPLS software was employed to evaluate the hypothesised relationships in the ―Students Attitudes – Achievement towards Statistics Model‖. From the modelling results, it revealed that all the hypothesised relationships are significant at p < 0.05 (t-value > 1.645), one-tailed test by using 5000 bootstrapped samples. In other words, the antecedent constructs of students‘ attitudes towards statistics influenced the Statistics Achievement both directly and indirectly. Overall, the hypothesised structural equation model predicted 43.3% of the total variance in Statistics Achievement with a statistical power analysis of G*Power at 95% confidence level. In addition, the f2 effect size for exogenous latent constructs and Q2 predictive relevance for endogenous latent constructs are also positive, while q2 effect sizes through blindfolding procedures are also acceptable. In conclusion, the empirical value showed that the model has the predictive capability in explaining the variances in the Statistics Achievement that are predicted by the antecedent constructs of attitude among the samples under study.