Analysis of the Multicollinearity Problem and Its Impact on the Multiple Linear Regression Model

Abstract

This study aimed to identify the problem of multicollinearity, the most important tests used to detect it, and the causes underlying this problem. The study utilized data from a previous research project, which were collected through a questionnaire. After collecting and analyzing the data, it was found that the dataset suffers from a multicollinearity problem. Accordingly, the data were analyzed using several diagnostic tests for detecting multicollinearity, and the results of these tests were compared. The statistical analysis of the model revealed the presence of multicollinearity among the explanatory variables. The three diagnostic tests-the Farrar-Glauber test, the Frisch test, and the Variance Inflation Factor (VIF)-indicated that variables X5 and X3 were the main contributors to this problem. Consequently, these variables were gradually removed from the model, which helped reduce intercorrelation among the explanatory variables and improved the stability of the regression coefficients. The results of the three tests used in the study further showed that X5 had the greatest impact on the multicollinearity problem, while X3 was identified as an insignificant variable and a primary source of the problem.