MRC, Department of Management, FEB UI Holds Training on Quantitative Data Analysis Using AMOS


MRC, Department of Management, FEB UI Holds Training on Quantitative Data Analysis Using AMOS

Nino Eka Putra ~ FEB UI Public Relations Officer

DEPOK, 22/7/2020 – The Management Research Center (MRC) at the Department of Management, Faculty of Economics and Business, Universitas Indonesia, held an online training program for quantitative data analysis using AMOS on Wednesday (22/7/2020).

The training presentation was delivered by M. Riyadh Rizky Adam, S.T., M.S.M., lecturer at the Swiss German University, and was moderated by Gita Gayatri, Ph.D., lecturer at FEB UI. The training was conducted via Zoom and YouTube and was joined by lecturers, researchers and students.

Arief Wibisono Lubis, Ph.D., Head of the Department of Management, FEB UI, said in his opening remark that the department pays serious attention to the effort to increase the number of publication of papers written by lecturers and students. The training on the use of AMOS is expected to help participants improve their technical capability in preparing quality research papers for publication and increase the number of published papers.

Riyadh Rizky Adam explained that the Analysis of Moment Structures (AMOS) developed by James L. Arbuckle is a computer program that can be used to model structural equation modeling (SEM). SEM is a statistical model (multivariate analysis technique) that seeks to explain the relationship between several variables in the individual or object being studied. SEM has three variables, namely univariate analysis involving one variable, bivariate analysis involving two variables, and multivariate analysis involving more than two variables.

According to Hair et al., 2019, SEM is better than general regression and other multivariate techniques, because as long as the estimation of all interdependence relationships is multiple and concurrent, they can represent unobserved concepts, including their relationship with latent variables; can estimate errors in the estimation process; defines a theoretical model to explain the entire set of relationships, and provides a modification index with which the model can be revised to obtain statistically better research findings.

“Basically, SEM involves two popular multivariate techniques, namely confirmatory factor analysis and multiple regression, which aim to test a series of relationships between one or more independent variables (usually considered predictor/causal variables because they can make predictions), and one or more dependent variables (response or result variables). The modification index gives direction for the revision of the model and better research findings with model match,” said Rizky.

Rizky went on to say that there are four characteristics of SEM. First, variables (latent and observed), second, models (structural models and measurements), third, errors (structural errors and measurements) and fourth, the use of a path diagram to describe/determine the SEM model to make it more accurate and easy compared to the mathematical model.

In addition, SEM adheres to the philosophy of reflective thinking, which is to reflect exogenous variables without giving form. On the contrary, giving form is a formative process. To make SEM data processing easier, of course we need statistical software. Currently, there are various kinds of software available for SEM data processing, including Lisrel, AMOS and Smart PLS.

“There are opinions that if you conduct formative research, you should use smart Partial Least Square (PLS) SEM with free characteristics/distribution, not SEM, which is Lisrel AMOS. If your data is disturbed/its distribution is not normal and the sample is relatively small, it is better to use smart PLS. Factor analysis that applies in SEM, namely confirmatory factor analysis, determines first the model and then the indicators for each latent variable,” Rizky explained by way of ending his presentation. (hjtp)