Uji Validitas Dan Relibilitas Instrumen Pengukuran Implementasi E-Learning Terhadap Kepuasan Pengguna Pada Masa Pandemi COVID 19 Di Lingkungan Politeknik Negeri Banjarmasin

Abstract

Implementation of online learning requires adequate preparation and support of resources, first is the software or tools used in online lectures, where Poliban has e-learning that can be used or can use other tools. Both hardware or computer equipment with the required specifications, one of which is a camera used for video conferencing, and what is very important is a stable internet network connection. Therefore, if at least these three things can be fulfilled, then online lectures can be carried out well. However, in reality, there are several problems faced by both lecturers and students, especially related to the unstable internet network signal, so that the implementation of online lectures cannot run optimally. To find out the problems faced by users, a measurement instrument for the implementation of E-learning on user satisfaction is needed. In making a good instrument model, a test is needed to ensure that the indicators used in the instrument used are valid and reliable and have a good level of reliability.
 Based on the problems mentioned above, this research will conduct instrument testing using validity tests and reliability tests where in the analysis process a Partial Least Square (PLS) application framework is used in SEM which is a statistical model that provides an approximate calculation of strength. hypothetical relationships among variables in a theoretical model
 The results of the research that have been carried out show that the results of the validity test of the instrument for measuring the implementation of e-learning on user satisfaction at the Banjarmasin state polytechnic get a loading factor value above 0.5 on the intended construct, these results indicate that the indicators used in this study are valid. or has met convergent validity, then based on the reliability test it shows that the composite reliability value for all constructs is above 0.7 which indicates that all constructs in the estimated model meet the criteria for discriminant validity. The lowest composite reliability value is 0.837 in the System Quality construct, this shows that the test instrument shows a good level of reliability, or in other words, the questionnaire results can be trusted.