09 Validity

Question 1

threats to internal validity

Answer 1

omitted variable bias
functional form misspecification
measurement error
sample selection
simultaneous causality
heteroskedasticity and/or correlated error terms

Apart from the last one, all discussed threats to internal validity lead to a violation of E [ui | Xi] = 0

Question 2

threats to external validity

Answer 2

differences in populations

differences in settings

Question 3

internal validity

Answer 3

in what degree the statistical inferences about causal effects are valid for the population and setting being studies

Question 4

external validity

Answer 4

in what degree the statistical inferences can be generalized to other populations and settings

Question 5

what kind of measurement error is the most problematic?

Answer 5

measurement error in the independent variables

Question 6

sample selection bias

Answer 6

bias from missing data based on Y

Question 7

solution to omitted variable bias

Answer 7

If observed: include them in the regression

If not: use panel data or instrumental variables

Question 8

What threats to internal validity may be reduced by instrumental variables?

Answer 8

omitted variable bias
measurement errors
simultaneous causality