Analisis Regresi Logistik Untuk Menentukan Faktor-Faktor yang Mempengaruhi Indeks Prestasi (IPK) Mahasiswa Statistika Universitas Sumatera Utara

Abstract

This study was conducted to see the influence of gender, school origin, average length of study, the amount of pocket money and student domicile on the value of the Statistics student achievement index of the University of North Sumatra. This research was conducted using the Binary Logistic Regression method. To determine the most suitable logistic regression model, a logistic regression analysis was carried out by testing the entire parameter using the G test and wald test to test each parameter coefficient and explain its descriptive. So 143 students were obtained with a cumulative achievement index of ≥ 3.5 and 12 students with a cumulative achievement index of < 3.5 out of 155 statistics students. The logistic regression model that has the smallest statistical value of G is the most feasible model to use. Based on the results of this study, the best model was obtained, namely: ln with a G test statistic of 13,647. Based on the parameter significance test using the wald test, that the sex variable has a significant effect on the level of cumulative achievement index value of statistics students.