Implementation of maximum likelihood estimation (MLE) in the assessment of pro-environmental tools measurement models for engineering students

• Main Authors: Ali, Muh. Ichsan, Idkhan, A. Muh., Kamin, Yusri, Hasim, Abdul Hafid
• Format: Article
• Published: Insight Society 2022
• Subjects: TJ Mechanical engineering and machinery
• Online Access: http://eprints.utm.my/id/eprint/102713/; http://dx.doi.org/10.18517/ijaseit.12.4.16724

The criterion validity test is used to compare the research construct with other tools that have been declared valid and reliable by correlating. In this study, data was obtained by distributing questionnaires and drawing conclusions by concluding the answers from the research subjects. The samples in this study were students of the Faculty of Engineering, Universitas Negeri Makassar, amounting to 200 research subjects. The researcher will collect the data by distributing questionnaires directly to respondents and through a google form with three research constructs. This is done to meet the requirements of the number of respondents who recommend conducting research whose data analysis uses the Maximum Likelihood Estimation (MLE). The IBM AMOS program was used to analyze Confirmatory Factor Analysis (CFA). The Goodness of Fit test (CMIN/DF, GFI, RMSEA, CFI, PNFI) shows that the proposed model is fit and can be continued for further analysis. Convergent validity (loading factor) constructs on the get a value of > 0.7 with a probability value of (p < 0.05), which means that the validity of the indicator has good reliability. Convergent validity gets a value > 0.7 with a probability value (p < 0.05) which means the indicator validity has good reliability. The average variance extract (AVE) value obtained is > 0.5, which means that the indicators in the developed model are proven to measure variable constructs. From the results of the research conducted, the reliability and validation values of the tools are consistent and reliable; from these results, the tools can be used repeatedly in research. A test is said to have high reliability if it provides data with consistent (fixed) results even though it is given at different times to the same subjects.