Hi All
I want to perform a chow test to test for differences in regression fundtions across group. However homoskedasticity does not hold. Is there an alternative test (or a modification of Chow) that takes account of heteroskedasticity?
For completeness, I am testing the following regression relating to a wage equation:
reg lwage age potexp female pt private wales scotland ni gcse alevel degree small manager healthy
As a result, I would prefer to avoid having interactions for every explanatory variable!
The sample is large (over 3,000). Given this number of observations, would the chow test still be valid?
Best
Georgia
I want to perform a chow test to test for differences in regression fundtions across group. However homoskedasticity does not hold. Is there an alternative test (or a modification of Chow) that takes account of heteroskedasticity?
For completeness, I am testing the following regression relating to a wage equation:
reg lwage age potexp female pt private wales scotland ni gcse alevel degree small manager healthy
As a result, I would prefer to avoid having interactions for every explanatory variable!
The sample is large (over 3,000). Given this number of observations, would the chow test still be valid?
Best
Georgia
