COMPARISON OF DIFFERENT TESTS FOR DETECTING HETEROSCEDASTICITY IN DATASETS.

Heteroscedasticity occurs mostly because of beneath mistakes in variable, incorrect data transformation, incorrect functional form, omission of important variables, non-detailed model, outliers and skewness in the distribution of one or more independent variables in the model. All analysis were carried out in R statistical package using Imtest, zoo and package.base. Five heteroscedasticity tests were selected, which are Park test, Glejser test, Breusch-Pagan test, White test and Goldfeld test, and used on simulated datasets ranging from 20,30,40,50,60,70,80,90 and 100 sample sizes respectively at different level of heteroscedasticity (that is at low level when sigma = 0.5, mild level when sigma = 1.0 and high level when sigma = 2.0). Also, the significance criterion alpha = 0.05. However, each test was repeated 1000 times and the percentage of rejection was computed over 1000 trials. For Glejser test, the average empirical type I error rate for the test reject more than expected while Goldfeld has the least power value. Therefore, Glejser test has the highest capacity to detect heteroscedasticity in most especially on simulated datasets. [ABSTRACT FROM AUTHOR]