ECON 104 Lecture 22: 22
Document Summary
Interpretation changed from x1 by delta x1 while holding x2 constant. After the change, x1 changes by delta x1 and y changes by delta y. Comparison of regression before and after accommodating for new omitted variable with 2 regressors. Intercept changed, makes sense because the definition changes. The coefficient and slope for str is the impact of the omitted variable bias. New regressor that is correlated with str and test scores. Should expect coefficient in front of str to change as well, because bias was towards -ve infinity. So accounting for the new regressor/ removing omitted variable bias -> estimate becomes less negative for str. Most of the time, we don"t know if omitted variable bias is important or not. There is probably an omitted variable bias that is generated by pctel. Working with a random sample even though pctel not correlated with str in population.
