A study is internally valid if:

a) the sample is representative of the population to which you wish to generalize.

b) the error terms are homoscedastic.

c) the coefficient estimator is unbiased.

d) none of the above are sufficient for internal validity

Suppose you want to estimate a regression in Stata in which one of the explanatory variables, state,

is a categorical variable. Which of the following regression commands will properly use the state

variable?

a) reg y x state

b) reg y x i.state

c) reg y x, cluster(state)

d) none of the above will work correctly