Question

In which of the following situations would one be most comfortable that omitted variable bias is not a concern?
a. When all of the variables in a model are statistically significant
b. When an estimated model has a high R-squared.
c. When an estimated model has a high adjusted R-squared.
d. When no one in a room full of experts can think of a variable that was not included in the model that could be causing omitted variable bias

Answer 1

Regression gives biased results when an important variable is omitted. It happens due to misspecification of a linear specification model. The effect may be an overestimation, or, underestimation. The omitted variable must be correlated with a dependent variable, and, one or more of independent variables.

High R-Squared, or, adjusted R squared does not mean that there is no omitted variable bias.

a) It is not a concern when all variables are statistically significant.