Clustered robust standard errors

[Amelia Smith]

I am working with panel data and have performed a clustered robust standard error test on my data however I am unsure how to interrupt the output from stata. I have attached the stata output from the regression without the clustered robust standard errors and the output with them to compare.

Without:

with:

[Dimitriy V. Masterov]

This question is not really answerable as is. Which parameter do you need help interpreting? It might also help if you explain what your panel structure a bit. Are your panels cities over time?

In general, the estimates/coefficients will be the same, but the standard errors will be different when you cluster. Since your SEs shrink when you cluster and if Uber entry is constant within each panel (Uber is either in or out in an area at a point in time), the within cluster correlation of residuals is likely to be negative. There's a decent explanation here.