Question 2 (40 pts.) The following question is motivated by Carpenter and Dobkin (2009) who estimate the effect of alcohol consumption C, measured as the number of alcohol drinks consumed per day, on mortality M, an indicator variable taking on a value of 1 if deceased and 0 otherwise. They apply a regression discontinuity design (RDD) by exploiting the minimum drinking age set at the age of 21, Le. A0 = 21. The population regression function (PRF) is given by: M,- = a + pL, + m. For this question, use the following sample: a) (2 pts.) Briefly explain whether L is a valid indicator variable for legal drinking age that is suitable for implementing a sharp regression discontinuity design. b) (2 pts.) Briefly explain whether A is a valid treatment assignment variable (or forcing variable). c) (2 pts.) If some of the observations in the control group had fake IDs with an incorrect birthday written on them, briefly describe the nature of the statistical problem it would create