Question

1. What is heteroskedasticity? Why is it problematic? Explain.

Answer 1

Heteroskedasticity is phenomena wherein variability ofvariable is unequal and across entire range of values. If we plot scateerplot then it represents cone shaped diagram because of huge variance.

For example, age and income represent heteroskedasticity as age rises income rises and hence scatterplot shows up like cone.

In OLS regression we assume that equal weight is given to all observations, however due to presence of heteroskedasticity the cases with larger disturbances have more pull than other observations. Even the standard errors arebiased which causes another problem due to this. For example people with higher age jn above example will get better preference as they will have higher income and thus causes biasedness.