* Variances of are not the same.
* Conditional variances of are no longer constant.
* Sources of heteroscedasticity:
* Following the error-learning models, as people learn, their errors of behavior become smaller over time or the number of errors becomes more consistent.
* As incomes grow, people have more discretionary income and hence more scope for choice about the disposition of their income.
* As data collecting technique improve, is likely to decrease.
* Heteroscedasticity can also arise as a result of the presence of outliers.
* Skewness in the distribution of one or more regressors included in the model.
* Incorrect data transformation and incorrect functional form.
* By graphical method
Do not have constant variance
* By formal method
* Park test
* Functional form
* Since is generally not known, Park suggest using as a proxy and running the following regression:
* If β turn out to be statistically significant, it would suggest that heteroscedasticity is present in the data.
* Glejser test
* Similar in spirit to the Park test. After obtaining the residuals from the OLS regression, Glejser suggest the absolute values of on the variable that is thought to be closely associated with . In his experiments, Glejser uses the following functional term:
* The error term has some problem in that its expected value is nonzero, it is serially correlated and ironically, it is heteroscedasticity.
* Spearman’s Rank Correlation test
* Step 1: Fit the regression to the data on Y and X and obtain the residuals .
* Step 2: Ignoring the sign of , that is, taking their absolute value , rank both and (or ) according to an ascending order and compute the Spearman’s rank correlation coefficient given previously.
* Step 3:...