Identification of the decline in learning outcomes in statistics courses using the chi-squared automatic interaction detection method

Abstract

The decline in learning outcomes in Statistics is a problem that must be immediately solved. Many factors affect learning outcomes such as; intelligence, interest, talent, motivation, learning patterns, learning resources, and so on. The data used in this study are secondary data from TU-FIPPS and BAAK which consists of data on final grades of Statistics students of the Economic Education Study Program, Indraprasta PGRI University. In addition, researchers also use primary data collected through surveys. Data analysis using the Chi-squared Automatic Interaction Detection (CHAID) method. Based on the results of this study, explanatory variables that affect the decline in student learning outcomes Statistics are variables of motivation and group learning frequency. The explanatory variable which has the strongest influence with the decrease in student learning outcomes Statistics is motivation. The value of the accuracy of student classifications produced using CHAID analysis of eighty-five point one percent. The classification tree model of CHAID analysis in this study is better and more accurate for estimating the decline in learning outcomes Statistics in the downward category than in the non-down or very down categories.