<![CDATA[Regression analysis is widely used in research. However, often in using this analysis the assumption of non-multicollinearity is not fulfilled. Handling of these problems can be done using Ridge Regression (RR) and Adjusted Ridge Regression (AR) methods. This study aims to compare the performance of RR and AR in handling multicollinearity among explanatory variables in multiple regression analysis using data simulation. The simulated data contain various multicollinearity level (ρ = 0.6, 0.8, 0.9) with of each different sample size (n = 20, 50, 100). The performance of the two methods are compared using Mean Square Errors (MSE). The result shows that the AR method and the RR method produce a smaller MSE value when the sample size used is larger. The MSE value generated by the AR method tends to be smaller than the RR method which can be seen from each data repetition used. It shows that the AR method is relatively more effective than the RR method for dealing with multicollinearity problems.]]>
Copyrights © 2023
