Homoskedasticity

[Grace Kells]

I have created a "residual-vs-fitted value" plot in order to check for the assumption of homosckedasticity of my residuals. However I don't know how to interpret the results since my dependent variable is discrete, could the result of the graph be read as a random cloud?

[Giovanni Abronzino]

.

[Grace Kells]

I have created a "residual-vs-fitted value" plot in order to check for the assumption of homosckedasticity of my residuals. However I don't know how to interpret the results since my dependent variable is discrete, could the result of the graph be read as a random cloud?

[Carlo Lazzaro]

Grace:
it seems that when the fitted values increase from 5 onwards, residual dispersion increses as well up to 6 and then stays about the same.
Did you perform -estat hettest- after -regress-?

[Grace Kells]

yes I did white test and estat hettest, but with white test there is Heteroskedasticity while with hettest there is Homoskedasticity

[Carlo Lazzaro]

Grace:
in doubt go -robust- standard error.