Research and Application of Graduate Satisfaction Based on SEM -- Take USTB as An Example

Based on the structural equation model and the characteristics of University Of Science and Technology Beijing(USTB), a graduate satisfaction evaluation model is constructed by using PLS estimation algorithm, and the fitting degree of the model is tested. It is found that the perceived quality has a great influence on the graduate satisfaction. It has a certain reference value for the improvement of school teaching and the application of structural equation model in the field of teaching. Introduction Structural equation model, also known as latent variable model, is the combination of factor analysis and path analysis [1]. It is an important statistical method for quantitative research in contemporary behavioral and social fields. It is an commonly used statistical modeling technology, which can introduce the latent variables to study the causal link between observed variables and find out the main factors which affecting teaching evaluation. Chin [2] believes that the choice of two modeling methods depends mainly on the research purpose; Joreskog & Word [3] believes that when the purpose of the research is theoretical test and the theoretical knowledge is sufficient, LISREL should be used. From a theoretical point of view, China's college satisfaction model is mainly based on foreign models, and there is no mature and official framework. So the graduate satisfaction index model constructed in this paper is not yet mature, and the relationship between variables in the model needs to be verified in empirical research. Therefore, based on the above analysis, this paper should adopt PLS modeling technology. Empirical Research on Graduate Satisfaction Model Graduate Satisfaction Assessment Model Framework Setting In order to understand the satisfaction of the USTB undergraduate students, from the perspective of students to find out the problems that the school may have in related work, the Academic Affairs Office conducted a questionnaire survey on the USTB 2018 Senior. This article constructed its index system according to the items involved in the questionnaire. The survey in this study used the 5-point Likert scale. The higher the score, the higher the satisfaction. The scale including 3 main indicators: Teaching perception, Service perception and Student satisfaction. Reliability and Validity Analysis In May 2018, the Academic Affairs Office conducted a questionnaire survey of USTB 2018 seniors, distributed 3,249 questionnaires and collected 2,841 questionnaires. After the deletion of outliers and missing values, the total number of valid questionnaires was 2,650. The effective recovery rate of the questionnaire was 81.6%. International Conference on Modeling, Analysis, Simulation Technologies and Applications (MASTA 2019) Copyright © 2019, the Authors. Published by Atlantis Press. This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/). Advances in Intelligent Systems Research, volume 168