We have collected the following variables out of a random sample.
- Final Exam Score
- Percent of lectures attended
- Level of course Dummys
- Required Course Dummy
- Cumulative GPA
- Gender Dummy
A) What kind of model should be used to estimate final exam scores in terms of the other variables? Write down the mathematical model to estimate.
B) Now rewrite the model -- using c and s subscripts to indicate class and student respectively. Indicate using the subscripts which variables vary by class only and student only and by both class and student.
C) Suppose we estimate this model with OLS, and there is a correlation between unobserved student traits (in the error term) with attendance. How does this bias the coefficient of attendance?